{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9126,
  "global_step": 22815,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 2e-06,
      "loss": 2.3586,
      "theoretical_loss": 3.484799908140819,
      "tokens_seen": 1661665280
    },
    {
      "epoch": 0.0,
      "learning_rate": 4e-06,
      "loss": 2.2339,
      "theoretical_loss": 3.4847880270596066,
      "tokens_seen": 1661730816
    },
    {
      "epoch": 0.0,
      "learning_rate": 6e-06,
      "loss": 2.2016,
      "theoretical_loss": 3.4847761465781497,
      "tokens_seen": 1661796352
    },
    {
      "epoch": 0.0,
      "learning_rate": 8e-06,
      "loss": 2.2196,
      "theoretical_loss": 3.484764266696394,
      "tokens_seen": 1661861888
    },
    {
      "epoch": 0.0,
      "learning_rate": 1e-05,
      "loss": 2.1452,
      "theoretical_loss": 3.4847523874142867,
      "tokens_seen": 1661927424
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.2e-05,
      "loss": 2.1924,
      "theoretical_loss": 3.4847405087317727,
      "tokens_seen": 1661992960
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.4e-05,
      "loss": 2.3474,
      "theoretical_loss": 3.4847286306487986,
      "tokens_seen": 1662058496
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.6e-05,
      "loss": 2.3439,
      "theoretical_loss": 3.48471675316531,
      "tokens_seen": 1662124032
    },
    {
      "epoch": 0.0,
      "learning_rate": 1.8e-05,
      "loss": 2.1693,
      "theoretical_loss": 3.4847048762812536,
      "tokens_seen": 1662189568
    },
    {
      "epoch": 0.0,
      "learning_rate": 2e-05,
      "loss": 2.0245,
      "theoretical_loss": 3.4846929999965757,
      "tokens_seen": 1662255104
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.2e-05,
      "loss": 2.071,
      "theoretical_loss": 3.484681124311222,
      "tokens_seen": 1662320640
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.4e-05,
      "loss": 1.9012,
      "theoretical_loss": 3.4846692492251385,
      "tokens_seen": 1662386176
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.6e-05,
      "loss": 1.8901,
      "theoretical_loss": 3.484657374738272,
      "tokens_seen": 1662451712
    },
    {
      "epoch": 0.0,
      "learning_rate": 2.8e-05,
      "loss": 1.9415,
      "theoretical_loss": 3.4846455008505677,
      "tokens_seen": 1662517248
    },
    {
      "epoch": 0.0,
      "learning_rate": 3e-05,
      "loss": 1.9999,
      "theoretical_loss": 3.484633627561972,
      "tokens_seen": 1662582784
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.2e-05,
      "loss": 1.9898,
      "theoretical_loss": 3.484621754872432,
      "tokens_seen": 1662648320
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.4000000000000007e-05,
      "loss": 1.6089,
      "theoretical_loss": 3.4846098827818928,
      "tokens_seen": 1662713856
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.6e-05,
      "loss": 1.7149,
      "theoretical_loss": 3.4845980112903003,
      "tokens_seen": 1662779392
    },
    {
      "epoch": 0.0,
      "learning_rate": 3.8e-05,
      "loss": 1.6944,
      "theoretical_loss": 3.484586140397602,
      "tokens_seen": 1662844928
    },
    {
      "epoch": 0.0,
      "learning_rate": 4e-05,
      "loss": 1.4261,
      "theoretical_loss": 3.484574270103743,
      "tokens_seen": 1662910464
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.2000000000000004e-05,
      "loss": 1.7173,
      "theoretical_loss": 3.484562400408671,
      "tokens_seen": 1662976000
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.4e-05,
      "loss": 1.5526,
      "theoretical_loss": 3.4845505313123297,
      "tokens_seen": 1663041536
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.6e-05,
      "loss": 1.7101,
      "theoretical_loss": 3.4845386628146677,
      "tokens_seen": 1663107072
    },
    {
      "epoch": 0.0,
      "learning_rate": 4.8e-05,
      "loss": 1.5717,
      "theoretical_loss": 3.4845267949156296,
      "tokens_seen": 1663172608
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.4392717480659485,
      "objective/train/avg_token_score": 0.027796979993581772,
      "objective/train/docs_used": 941634,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5729044675827026,
      "objective/train/original_loss": 3.6452524662017822,
      "objective/train/theoretical_loss": 3.4845149276151623,
      "objective/train/tokens_used": 22097376,
      "objective/train/unlikelihood_loss": 0.09146079421043396,
      "theoretical_loss": 3.4845149276151623,
      "tokens_seen": 1663238144
    },
    {
      "epoch": 0.0,
      "learning_rate": 5e-05,
      "loss": 1.6885,
      "theoretical_loss": 3.4845149276151623,
      "tokens_seen": 1663238144
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.2e-05,
      "loss": 1.4219,
      "theoretical_loss": 3.484503060913212,
      "tokens_seen": 1663303680
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.4e-05,
      "loss": 1.5592,
      "theoretical_loss": 3.484491194809725,
      "tokens_seen": 1663369216
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.6e-05,
      "loss": 1.1071,
      "theoretical_loss": 3.484479329304647,
      "tokens_seen": 1663434752
    },
    {
      "epoch": 0.0,
      "learning_rate": 5.800000000000001e-05,
      "loss": 1.4801,
      "theoretical_loss": 3.484467464397925,
      "tokens_seen": 1663500288
    },
    {
      "epoch": 0.0,
      "learning_rate": 6e-05,
      "loss": 1.6332,
      "theoretical_loss": 3.484455600089505,
      "tokens_seen": 1663565824
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.2e-05,
      "loss": 1.4538,
      "theoretical_loss": 3.4844437363793332,
      "tokens_seen": 1663631360
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.4e-05,
      "loss": 1.5288,
      "theoretical_loss": 3.4844318732673556,
      "tokens_seen": 1663696896
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.6e-05,
      "loss": 1.4459,
      "theoretical_loss": 3.484420010753519,
      "tokens_seen": 1663762432
    },
    {
      "epoch": 0.0,
      "learning_rate": 6.800000000000001e-05,
      "loss": 1.4075,
      "theoretical_loss": 3.484408148837769,
      "tokens_seen": 1663827968
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.000000000000001e-05,
      "loss": 1.2731,
      "theoretical_loss": 3.4843962875200525,
      "tokens_seen": 1663893504
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.2e-05,
      "loss": 1.2299,
      "theoretical_loss": 3.484384426800316,
      "tokens_seen": 1663959040
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.4e-05,
      "loss": 1.6132,
      "theoretical_loss": 3.4843725666785046,
      "tokens_seen": 1664024576
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.6e-05,
      "loss": 1.639,
      "theoretical_loss": 3.484360707154566,
      "tokens_seen": 1664090112
    },
    {
      "epoch": 0.0,
      "learning_rate": 7.8e-05,
      "loss": 1.3998,
      "theoretical_loss": 3.484348848228446,
      "tokens_seen": 1664155648
    },
    {
      "epoch": 0.0,
      "learning_rate": 8e-05,
      "loss": 1.6466,
      "theoretical_loss": 3.4843369899000907,
      "tokens_seen": 1664221184
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.2e-05,
      "loss": 1.4191,
      "theoretical_loss": 3.484325132169446,
      "tokens_seen": 1664286720
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.400000000000001e-05,
      "loss": 1.3433,
      "theoretical_loss": 3.4843132750364596,
      "tokens_seen": 1664352256
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.599999999999999e-05,
      "loss": 1.6113,
      "theoretical_loss": 3.4843014185010768,
      "tokens_seen": 1664417792
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.8e-05,
      "loss": 1.5214,
      "theoretical_loss": 3.4842895625632444,
      "tokens_seen": 1664483328
    },
    {
      "epoch": 0.0,
      "learning_rate": 8.999999999999999e-05,
      "loss": 1.5353,
      "theoretical_loss": 3.4842777072229083,
      "tokens_seen": 1664548864
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.2e-05,
      "loss": 1.4214,
      "theoretical_loss": 3.484265852480015,
      "tokens_seen": 1664614400
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.400000000000001e-05,
      "loss": 1.4677,
      "theoretical_loss": 3.4842539983345118,
      "tokens_seen": 1664679936
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.6e-05,
      "loss": 1.4979,
      "theoretical_loss": 3.484242144786344,
      "tokens_seen": 1664745472
    },
    {
      "epoch": 0.0,
      "learning_rate": 9.800000000000001e-05,
      "loss": 1.5655,
      "theoretical_loss": 3.4842302918354577,
      "tokens_seen": 1664811008
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.4052419364452362,
      "objective/train/avg_token_score": 0.023708028718829155,
      "objective/train/docs_used": 943199,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3439230918884277,
      "objective/train/original_loss": 3.491624593734741,
      "objective/train/theoretical_loss": 3.4842184394818,
      "objective/train/tokens_used": 23735776,
      "objective/train/unlikelihood_loss": 0.11052291095256805,
      "theoretical_loss": 3.4842184394818,
      "tokens_seen": 1664876544
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0001,
      "loss": 1.2824,
      "theoretical_loss": 3.4842184394818,
      "tokens_seen": 1664876544
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000102,
      "loss": 1.3515,
      "theoretical_loss": 3.4842065877253177,
      "tokens_seen": 1664942080
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000104,
      "loss": 1.327,
      "theoretical_loss": 3.4841947365659562,
      "tokens_seen": 1665007616
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000106,
      "loss": 1.4169,
      "theoretical_loss": 3.4841828860036625,
      "tokens_seen": 1665073152
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000108,
      "loss": 1.4514,
      "theoretical_loss": 3.484171036038383,
      "tokens_seen": 1665138688
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011,
      "loss": 1.374,
      "theoretical_loss": 3.4841591866700643,
      "tokens_seen": 1665204224
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000112,
      "loss": 1.5037,
      "theoretical_loss": 3.4841473378986523,
      "tokens_seen": 1665269760
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000114,
      "loss": 1.5909,
      "theoretical_loss": 3.4841354897240935,
      "tokens_seen": 1665335296
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00011600000000000001,
      "loss": 1.3882,
      "theoretical_loss": 3.4841236421463346,
      "tokens_seen": 1665400832
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000118,
      "loss": 1.4278,
      "theoretical_loss": 3.4841117951653224,
      "tokens_seen": 1665466368
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00012,
      "loss": 1.2305,
      "theoretical_loss": 3.4840999487810027,
      "tokens_seen": 1665531904
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000122,
      "loss": 1.3588,
      "theoretical_loss": 3.4840881029933226,
      "tokens_seen": 1665597440
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000124,
      "loss": 1.1805,
      "theoretical_loss": 3.484076257802228,
      "tokens_seen": 1665662976
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000126,
      "loss": 1.4335,
      "theoretical_loss": 3.4840644132076655,
      "tokens_seen": 1665728512
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000128,
      "loss": 1.5079,
      "theoretical_loss": 3.484052569209582,
      "tokens_seen": 1665794048
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013000000000000002,
      "loss": 1.319,
      "theoretical_loss": 3.4840407258079233,
      "tokens_seen": 1665859584
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000132,
      "loss": 1.4067,
      "theoretical_loss": 3.4840288830026367,
      "tokens_seen": 1665925120
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000134,
      "loss": 1.2983,
      "theoretical_loss": 3.484017040793668,
      "tokens_seen": 1665990656
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013600000000000003,
      "loss": 1.2986,
      "theoretical_loss": 3.4840051991809644,
      "tokens_seen": 1666056192
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00013800000000000002,
      "loss": 1.2998,
      "theoretical_loss": 3.4839933581644718,
      "tokens_seen": 1666121728
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014000000000000001,
      "loss": 1.5612,
      "theoretical_loss": 3.4839815177441373,
      "tokens_seen": 1666187264
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00014199999999999998,
      "loss": 1.3289,
      "theoretical_loss": 3.4839696779199065,
      "tokens_seen": 1666252800
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000144,
      "loss": 1.4071,
      "theoretical_loss": 3.483957838691727,
      "tokens_seen": 1666318336
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000146,
      "loss": 1.1692,
      "theoretical_loss": 3.483946000059545,
      "tokens_seen": 1666383872
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000148,
      "loss": 1.4058,
      "theoretical_loss": 3.483934162023307,
      "tokens_seen": 1666449408
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.3277125954627991,
      "objective/train/avg_token_score": 0.01894918829202652,
      "objective/train/docs_used": 943856,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0503660440444946,
      "objective/train/original_loss": 3.4445393085479736,
      "objective/train/theoretical_loss": 3.4839223245829594,
      "objective/train/tokens_used": 25374176,
      "objective/train/unlikelihood_loss": 0.0962013378739357,
      "theoretical_loss": 3.4839223245829594,
      "tokens_seen": 1666514944
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00015,
      "loss": 1.1905,
      "theoretical_loss": 3.4839223245829594,
      "tokens_seen": 1666514944
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000152,
      "loss": 1.5246,
      "theoretical_loss": 3.4839104877384486,
      "tokens_seen": 1666580480
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000154,
      "loss": 1.3905,
      "theoretical_loss": 3.483898651489722,
      "tokens_seen": 1666646016
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000156,
      "loss": 1.4391,
      "theoretical_loss": 3.4838868158367253,
      "tokens_seen": 1666711552
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000158,
      "loss": 1.2598,
      "theoretical_loss": 3.4838749807794063,
      "tokens_seen": 1666777088
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016,
      "loss": 1.4413,
      "theoretical_loss": 3.4838631463177103,
      "tokens_seen": 1666842624
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000162,
      "loss": 1.5131,
      "theoretical_loss": 3.483851312451584,
      "tokens_seen": 1666908160
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000164,
      "loss": 1.333,
      "theoretical_loss": 3.4838394791809746,
      "tokens_seen": 1666973696
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016600000000000002,
      "loss": 1.6326,
      "theoretical_loss": 3.4838276465058287,
      "tokens_seen": 1667039232
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00016800000000000002,
      "loss": 1.5681,
      "theoretical_loss": 3.4838158144260927,
      "tokens_seen": 1667104768
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017,
      "loss": 1.1809,
      "theoretical_loss": 3.483803982941714,
      "tokens_seen": 1667170304
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017199999999999998,
      "loss": 1.2542,
      "theoretical_loss": 3.4837921520526374,
      "tokens_seen": 1667235840
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000174,
      "loss": 1.3605,
      "theoretical_loss": 3.483780321758811,
      "tokens_seen": 1667301376
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000176,
      "loss": 1.6095,
      "theoretical_loss": 3.4837684920601815,
      "tokens_seen": 1667366912
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000178,
      "loss": 1.4505,
      "theoretical_loss": 3.483756662956695,
      "tokens_seen": 1667432448
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00017999999999999998,
      "loss": 1.2638,
      "theoretical_loss": 3.483744834448298,
      "tokens_seen": 1667497984
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000182,
      "loss": 1.4185,
      "theoretical_loss": 3.483733006534938,
      "tokens_seen": 1667563520
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000184,
      "loss": 1.2676,
      "theoretical_loss": 3.483721179216561,
      "tokens_seen": 1667629056
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000186,
      "loss": 1.7296,
      "theoretical_loss": 3.4837093524931144,
      "tokens_seen": 1667694592
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00018800000000000002,
      "loss": 1.6747,
      "theoretical_loss": 3.4836975263645438,
      "tokens_seen": 1667760128
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019,
      "loss": 1.463,
      "theoretical_loss": 3.483685700830797,
      "tokens_seen": 1667825664
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000192,
      "loss": 1.2903,
      "theoretical_loss": 3.48367387589182,
      "tokens_seen": 1667891200
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000194,
      "loss": 1.3391,
      "theoretical_loss": 3.48366205154756,
      "tokens_seen": 1667956736
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019600000000000002,
      "loss": 1.6197,
      "theoretical_loss": 3.4836502277979626,
      "tokens_seen": 1668022272
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00019800000000000002,
      "loss": 1.3482,
      "theoretical_loss": 3.4836384046429765,
      "tokens_seen": 1668087808
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.3819648027420044,
      "objective/train/avg_token_score": 0.015858473256230354,
      "objective/train/docs_used": 945126,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2602568864822388,
      "objective/train/original_loss": 3.6047937870025635,
      "objective/train/theoretical_loss": 3.4836265820825467,
      "objective/train/tokens_used": 27012576,
      "objective/train/unlikelihood_loss": 0.07870878279209137,
      "theoretical_loss": 3.4836265820825467,
      "tokens_seen": 1668153344
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 1.3153,
      "theoretical_loss": 3.4836265820825467,
      "tokens_seen": 1668153344
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000202,
      "loss": 1.313,
      "theoretical_loss": 3.483614760116621,
      "tokens_seen": 1668218880
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000204,
      "loss": 1.4786,
      "theoretical_loss": 3.4836029387451455,
      "tokens_seen": 1668284416
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000206,
      "loss": 1.4841,
      "theoretical_loss": 3.483591117968067,
      "tokens_seen": 1668349952
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000208,
      "loss": 1.4115,
      "theoretical_loss": 3.483579297785333,
      "tokens_seen": 1668415488
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00021,
      "loss": 1.5871,
      "theoretical_loss": 3.483567478196889,
      "tokens_seen": 1668481024
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000212,
      "loss": 1.5155,
      "theoretical_loss": 3.483555659202683,
      "tokens_seen": 1668546560
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000214,
      "loss": 1.6009,
      "theoretical_loss": 3.4835438408026613,
      "tokens_seen": 1668612096
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000216,
      "loss": 1.1647,
      "theoretical_loss": 3.483532022996771,
      "tokens_seen": 1668677632
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000218,
      "loss": 1.3228,
      "theoretical_loss": 3.483520205784958,
      "tokens_seen": 1668743168
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022,
      "loss": 1.4557,
      "theoretical_loss": 3.48350838916717,
      "tokens_seen": 1668808704
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000222,
      "loss": 1.3735,
      "theoretical_loss": 3.4834965731433534,
      "tokens_seen": 1668874240
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000224,
      "loss": 1.3874,
      "theoretical_loss": 3.4834847577134553,
      "tokens_seen": 1668939776
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00022600000000000002,
      "loss": 1.3996,
      "theoretical_loss": 3.4834729428774223,
      "tokens_seen": 1669005312
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000228,
      "loss": 1.2483,
      "theoretical_loss": 3.4834611286352013,
      "tokens_seen": 1669070848
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023,
      "loss": 1.5426,
      "theoretical_loss": 3.4834493149867387,
      "tokens_seen": 1669136384
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023200000000000003,
      "loss": 1.1371,
      "theoretical_loss": 3.483437501931982,
      "tokens_seen": 1669201920
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023400000000000002,
      "loss": 1.5659,
      "theoretical_loss": 3.483425689470878,
      "tokens_seen": 1669267456
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000236,
      "loss": 1.5264,
      "theoretical_loss": 3.4834138776033736,
      "tokens_seen": 1669332992
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00023799999999999998,
      "loss": 1.3228,
      "theoretical_loss": 3.483402066329415,
      "tokens_seen": 1669398528
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.00024,
      "loss": 1.4733,
      "theoretical_loss": 3.48339025564895,
      "tokens_seen": 1669464064
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000242,
      "loss": 1.329,
      "theoretical_loss": 3.4833784455619248,
      "tokens_seen": 1669529600
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000244,
      "loss": 1.4319,
      "theoretical_loss": 3.4833666360682867,
      "tokens_seen": 1669595136
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000246,
      "loss": 1.4501,
      "theoretical_loss": 3.4833548271679824,
      "tokens_seen": 1669660672
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.000248,
      "loss": 1.3319,
      "theoretical_loss": 3.4833430188609587,
      "tokens_seen": 1669726208
    },
    {
      "epoch": 0.0,
      "objective/train/aligned_token_count": 0.45992180705070496,
      "objective/train/avg_token_score": 0.009118808433413506,
      "objective/train/docs_used": 945699,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6401410102844238,
      "objective/train/original_loss": 4.021496772766113,
      "objective/train/theoretical_loss": 3.4833312111471626,
      "objective/train/tokens_used": 28650976,
      "objective/train/unlikelihood_loss": 0.06683876365423203,
      "theoretical_loss": 3.4833312111471626,
      "tokens_seen": 1669791744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025,
      "loss": 1.6621,
      "theoretical_loss": 3.4833312111471626,
      "tokens_seen": 1669791744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000252,
      "loss": 1.361,
      "theoretical_loss": 3.483319404026541,
      "tokens_seen": 1669857280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000254,
      "loss": 1.1887,
      "theoretical_loss": 3.4833075974990404,
      "tokens_seen": 1669922816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000256,
      "loss": 1.4238,
      "theoretical_loss": 3.4832957915646086,
      "tokens_seen": 1669988352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00025800000000000004,
      "loss": 1.4582,
      "theoretical_loss": 3.4832839862231926,
      "tokens_seen": 1670053888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00026000000000000003,
      "loss": 1.509,
      "theoretical_loss": 3.4832721814747383,
      "tokens_seen": 1670119424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000262,
      "loss": 1.6246,
      "theoretical_loss": 3.4832603773191932,
      "tokens_seen": 1670184960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000264,
      "loss": 1.3161,
      "theoretical_loss": 3.4832485737565047,
      "tokens_seen": 1670250496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000266,
      "loss": 1.6053,
      "theoretical_loss": 3.483236770786619,
      "tokens_seen": 1670316032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000268,
      "loss": 1.2865,
      "theoretical_loss": 3.4832249684094836,
      "tokens_seen": 1670381568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027,
      "loss": 1.3504,
      "theoretical_loss": 3.4832131666250454,
      "tokens_seen": 1670447104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027200000000000005,
      "loss": 1.4324,
      "theoretical_loss": 3.483201365433251,
      "tokens_seen": 1670512640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027400000000000005,
      "loss": 1.478,
      "theoretical_loss": 3.4831895648340474,
      "tokens_seen": 1670578176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027600000000000004,
      "loss": 1.3201,
      "theoretical_loss": 3.4831777648273823,
      "tokens_seen": 1670643712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00027800000000000004,
      "loss": 1.3004,
      "theoretical_loss": 3.4831659654132023,
      "tokens_seen": 1670709248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028000000000000003,
      "loss": 1.2886,
      "theoretical_loss": 3.4831541665914543,
      "tokens_seen": 1670774784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028199999999999997,
      "loss": 1.4002,
      "theoretical_loss": 3.4831423683620852,
      "tokens_seen": 1670840320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028399999999999996,
      "loss": 1.0925,
      "theoretical_loss": 3.4831305707250424,
      "tokens_seen": 1670905856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00028599999999999996,
      "loss": 1.594,
      "theoretical_loss": 3.483118773680273,
      "tokens_seen": 1670971392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000288,
      "loss": 1.5114,
      "theoretical_loss": 3.4831069772277234,
      "tokens_seen": 1671036928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00029,
      "loss": 1.4329,
      "theoretical_loss": 3.483095181367341,
      "tokens_seen": 1671102464
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000292,
      "loss": 1.5268,
      "theoretical_loss": 3.4830833860990733,
      "tokens_seen": 1671168000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000294,
      "loss": 1.3822,
      "theoretical_loss": 3.4830715914228665,
      "tokens_seen": 1671233536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000296,
      "loss": 1.5873,
      "theoretical_loss": 3.4830597973386683,
      "tokens_seen": 1671299072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000298,
      "loss": 1.5036,
      "theoretical_loss": 3.4830480038464255,
      "tokens_seen": 1671364608
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.3960166275501251,
      "objective/train/avg_token_score": 0.020477375015616417,
      "objective/train/docs_used": 946982,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2867282629013062,
      "objective/train/original_loss": 4.248522758483887,
      "objective/train/theoretical_loss": 3.4830362109460857,
      "objective/train/tokens_used": 30289376,
      "objective/train/unlikelihood_loss": 0.06127317622303963,
      "theoretical_loss": 3.4830362109460857,
      "tokens_seen": 1671430144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0003,
      "loss": 1.2874,
      "theoretical_loss": 3.4830362109460857,
      "tokens_seen": 1671430144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000302,
      "loss": 1.5552,
      "theoretical_loss": 3.483024418637595,
      "tokens_seen": 1671495680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000304,
      "loss": 1.4532,
      "theoretical_loss": 3.4830126269209014,
      "tokens_seen": 1671561216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000306,
      "loss": 1.3209,
      "theoretical_loss": 3.4830008357959517,
      "tokens_seen": 1671626752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000308,
      "loss": 1.5412,
      "theoretical_loss": 3.4829890452626926,
      "tokens_seen": 1671692288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031,
      "loss": 1.0823,
      "theoretical_loss": 3.482977255321072,
      "tokens_seen": 1671757824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000312,
      "loss": 1.4433,
      "theoretical_loss": 3.482965465971036,
      "tokens_seen": 1671823360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000314,
      "loss": 1.3343,
      "theoretical_loss": 3.4829536772125325,
      "tokens_seen": 1671888896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000316,
      "loss": 1.4329,
      "theoretical_loss": 3.482941889045509,
      "tokens_seen": 1671954432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00031800000000000003,
      "loss": 1.2974,
      "theoretical_loss": 3.4829301014699117,
      "tokens_seen": 1672019968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00032,
      "loss": 1.5543,
      "theoretical_loss": 3.482918314485688,
      "tokens_seen": 1672085504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000322,
      "loss": 1.6948,
      "theoretical_loss": 3.4829065280927853,
      "tokens_seen": 1672151040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000324,
      "loss": 1.358,
      "theoretical_loss": 3.482894742291151,
      "tokens_seen": 1672216576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000326,
      "loss": 1.5256,
      "theoretical_loss": 3.4828829570807316,
      "tokens_seen": 1672282112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000328,
      "loss": 1.2557,
      "theoretical_loss": 3.4828711724614747,
      "tokens_seen": 1672347648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033,
      "loss": 1.4626,
      "theoretical_loss": 3.4828593884333277,
      "tokens_seen": 1672413184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033200000000000005,
      "loss": 1.4653,
      "theoretical_loss": 3.4828476049962367,
      "tokens_seen": 1672478720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033400000000000004,
      "loss": 1.4492,
      "theoretical_loss": 3.4828358221501503,
      "tokens_seen": 1672544256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033600000000000004,
      "loss": 1.3397,
      "theoretical_loss": 3.4828240398950143,
      "tokens_seen": 1672609792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00033800000000000003,
      "loss": 1.2303,
      "theoretical_loss": 3.4828122582307777,
      "tokens_seen": 1672675328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034,
      "loss": 1.3492,
      "theoretical_loss": 3.482800477157386,
      "tokens_seen": 1672740864
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000342,
      "loss": 1.4296,
      "theoretical_loss": 3.4827886966747874,
      "tokens_seen": 1672806400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00034399999999999996,
      "loss": 1.6639,
      "theoretical_loss": 3.482776916782929,
      "tokens_seen": 1672871936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000346,
      "loss": 1.3111,
      "theoretical_loss": 3.482765137481757,
      "tokens_seen": 1672937472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000348,
      "loss": 1.4424,
      "theoretical_loss": 3.4827533587712205,
      "tokens_seen": 1673003008
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.42277616262435913,
      "objective/train/avg_token_score": 0.034132879227399826,
      "objective/train/docs_used": 947668,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4029995203018188,
      "objective/train/original_loss": 4.207383155822754,
      "objective/train/theoretical_loss": 3.4827415806512656,
      "objective/train/tokens_used": 31927776,
      "objective/train/unlikelihood_loss": 0.05666952580213547,
      "theoretical_loss": 3.4827415806512656,
      "tokens_seen": 1673068544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035,
      "loss": 1.7249,
      "theoretical_loss": 3.4827415806512656,
      "tokens_seen": 1673068544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000352,
      "loss": 1.4411,
      "theoretical_loss": 3.4827298031218397,
      "tokens_seen": 1673134080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000354,
      "loss": 1.4599,
      "theoretical_loss": 3.48271802618289,
      "tokens_seen": 1673199616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000356,
      "loss": 1.4334,
      "theoretical_loss": 3.4827062498343637,
      "tokens_seen": 1673265152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000358,
      "loss": 1.4262,
      "theoretical_loss": 3.4826944740762085,
      "tokens_seen": 1673330688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00035999999999999997,
      "loss": 1.5422,
      "theoretical_loss": 3.4826826989083712,
      "tokens_seen": 1673396224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000362,
      "loss": 1.3631,
      "theoretical_loss": 3.4826709243307996,
      "tokens_seen": 1673461760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000364,
      "loss": 1.5299,
      "theoretical_loss": 3.4826591503434408,
      "tokens_seen": 1673527296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000366,
      "loss": 1.4358,
      "theoretical_loss": 3.4826473769462414,
      "tokens_seen": 1673592832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000368,
      "loss": 1.3031,
      "theoretical_loss": 3.4826356041391495,
      "tokens_seen": 1673658368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037,
      "loss": 1.372,
      "theoretical_loss": 3.482623831922113,
      "tokens_seen": 1673723904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000372,
      "loss": 1.4469,
      "theoretical_loss": 3.4826120602950774,
      "tokens_seen": 1673789440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000374,
      "loss": 1.3485,
      "theoretical_loss": 3.4826002892579915,
      "tokens_seen": 1673854976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00037600000000000003,
      "loss": 1.1995,
      "theoretical_loss": 3.482588518810802,
      "tokens_seen": 1673920512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000378,
      "loss": 1.4059,
      "theoretical_loss": 3.4825767489534574,
      "tokens_seen": 1673986048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00038,
      "loss": 1.4645,
      "theoretical_loss": 3.482564979685903,
      "tokens_seen": 1674051584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000382,
      "loss": 1.6177,
      "theoretical_loss": 3.482553211008088,
      "tokens_seen": 1674117120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000384,
      "loss": 1.5993,
      "theoretical_loss": 3.482541442919959,
      "tokens_seen": 1674182656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000386,
      "loss": 1.5792,
      "theoretical_loss": 3.482529675421463,
      "tokens_seen": 1674248192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000388,
      "loss": 1.2894,
      "theoretical_loss": 3.4825179085125484,
      "tokens_seen": 1674313728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039000000000000005,
      "loss": 1.3151,
      "theoretical_loss": 3.4825061421931616,
      "tokens_seen": 1674379264
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039200000000000004,
      "loss": 1.484,
      "theoretical_loss": 3.4824943764632503,
      "tokens_seen": 1674444800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039400000000000004,
      "loss": 1.4304,
      "theoretical_loss": 3.4824826113227623,
      "tokens_seen": 1674510336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00039600000000000003,
      "loss": 1.3797,
      "theoretical_loss": 3.4824708467716445,
      "tokens_seen": 1674575872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000398,
      "loss": 1.1721,
      "theoretical_loss": 3.4824590828098447,
      "tokens_seen": 1674641408
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.47403469681739807,
      "objective/train/avg_token_score": 0.020986413583159447,
      "objective/train/docs_used": 948408,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4957624673843384,
      "objective/train/original_loss": 4.087368488311768,
      "objective/train/theoretical_loss": 3.48244731943731,
      "objective/train/tokens_used": 33566176,
      "objective/train/unlikelihood_loss": 0.058626897633075714,
      "theoretical_loss": 3.48244731943731,
      "tokens_seen": 1674706944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004,
      "loss": 1.6842,
      "theoretical_loss": 3.48244731943731,
      "tokens_seen": 1674706944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000402,
      "loss": 1.3746,
      "theoretical_loss": 3.4824355566539875,
      "tokens_seen": 1674772480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000404,
      "loss": 1.4025,
      "theoretical_loss": 3.4824237944598257,
      "tokens_seen": 1674838016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00040600000000000006,
      "loss": 1.4322,
      "theoretical_loss": 3.482412032854771,
      "tokens_seen": 1674903552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000408,
      "loss": 1.5092,
      "theoretical_loss": 3.482400271838771,
      "tokens_seen": 1674969088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00041,
      "loss": 1.3048,
      "theoretical_loss": 3.482388511411774,
      "tokens_seen": 1675034624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000412,
      "loss": 1.1327,
      "theoretical_loss": 3.4823767515737267,
      "tokens_seen": 1675100160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000414,
      "loss": 1.3768,
      "theoretical_loss": 3.4823649923245767,
      "tokens_seen": 1675165696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000416,
      "loss": 1.3944,
      "theoretical_loss": 3.4823532336642717,
      "tokens_seen": 1675231232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00041799999999999997,
      "loss": 1.5612,
      "theoretical_loss": 3.482341475592759,
      "tokens_seen": 1675296768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00042,
      "loss": 1.6033,
      "theoretical_loss": 3.4823297181099857,
      "tokens_seen": 1675362304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000422,
      "loss": 1.2907,
      "theoretical_loss": 3.4823179612159,
      "tokens_seen": 1675427840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000424,
      "loss": 1.4496,
      "theoretical_loss": 3.4823062049104485,
      "tokens_seen": 1675493376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000426,
      "loss": 1.3219,
      "theoretical_loss": 3.4822944491935797,
      "tokens_seen": 1675558912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000428,
      "loss": 1.6165,
      "theoretical_loss": 3.4822826940652405,
      "tokens_seen": 1675624448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043,
      "loss": 1.46,
      "theoretical_loss": 3.482270939525379,
      "tokens_seen": 1675689984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000432,
      "loss": 1.3409,
      "theoretical_loss": 3.482259185573942,
      "tokens_seen": 1675755520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00043400000000000003,
      "loss": 1.5918,
      "theoretical_loss": 3.4822474322108774,
      "tokens_seen": 1675821056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000436,
      "loss": 1.3411,
      "theoretical_loss": 3.4822356794361324,
      "tokens_seen": 1675886592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000438,
      "loss": 1.4671,
      "theoretical_loss": 3.482223927249655,
      "tokens_seen": 1675952128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00044,
      "loss": 1.407,
      "theoretical_loss": 3.482212175651393,
      "tokens_seen": 1676017664
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000442,
      "loss": 1.4158,
      "theoretical_loss": 3.4822004246412934,
      "tokens_seen": 1676083200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000444,
      "loss": 1.2442,
      "theoretical_loss": 3.4821886742193042,
      "tokens_seen": 1676148736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000446,
      "loss": 1.6039,
      "theoretical_loss": 3.4821769243853717,
      "tokens_seen": 1676214272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000448,
      "loss": 1.3757,
      "theoretical_loss": 3.4821651751394453,
      "tokens_seen": 1676279808
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.43438416719436646,
      "objective/train/avg_token_score": 0.02250559814274311,
      "objective/train/docs_used": 949481,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3553886413574219,
      "objective/train/original_loss": 4.167667388916016,
      "objective/train/theoretical_loss": 3.4821534264814717,
      "objective/train/tokens_used": 35204576,
      "objective/train/unlikelihood_loss": 0.05399244651198387,
      "theoretical_loss": 3.4821534264814717,
      "tokens_seen": 1676345344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045000000000000004,
      "loss": 1.5774,
      "theoretical_loss": 3.4821534264814717,
      "tokens_seen": 1676345344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045200000000000004,
      "loss": 1.1835,
      "theoretical_loss": 3.4821416784113985,
      "tokens_seen": 1676410880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00045400000000000003,
      "loss": 1.4729,
      "theoretical_loss": 3.4821299309291733,
      "tokens_seen": 1676476416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000456,
      "loss": 1.4368,
      "theoretical_loss": 3.4821181840347437,
      "tokens_seen": 1676541952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000458,
      "loss": 1.5042,
      "theoretical_loss": 3.4821064377280577,
      "tokens_seen": 1676607488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046,
      "loss": 1.4255,
      "theoretical_loss": 3.4820946920090625,
      "tokens_seen": 1676673024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000462,
      "loss": 1.3812,
      "theoretical_loss": 3.4820829468777057,
      "tokens_seen": 1676738560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046400000000000006,
      "loss": 1.384,
      "theoretical_loss": 3.482071202333935,
      "tokens_seen": 1676804096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046600000000000005,
      "loss": 1.3564,
      "theoretical_loss": 3.482059458377698,
      "tokens_seen": 1676869632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00046800000000000005,
      "loss": 1.4955,
      "theoretical_loss": 3.4820477150089433,
      "tokens_seen": 1676935168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047,
      "loss": 1.3263,
      "theoretical_loss": 3.4820359722276164,
      "tokens_seen": 1677000704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000472,
      "loss": 1.6779,
      "theoretical_loss": 3.4820242300336677,
      "tokens_seen": 1677066240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000474,
      "loss": 1.4524,
      "theoretical_loss": 3.4820124884270425,
      "tokens_seen": 1677131776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047599999999999997,
      "loss": 1.5956,
      "theoretical_loss": 3.4820007474076893,
      "tokens_seen": 1677197312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00047799999999999996,
      "loss": 1.5913,
      "theoretical_loss": 3.481989006975556,
      "tokens_seen": 1677262848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00048,
      "loss": 1.4133,
      "theoretical_loss": 3.4819772671305906,
      "tokens_seen": 1677328384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000482,
      "loss": 1.2181,
      "theoretical_loss": 3.48196552787274,
      "tokens_seen": 1677393920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000484,
      "loss": 1.337,
      "theoretical_loss": 3.4819537892019525,
      "tokens_seen": 1677459456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000486,
      "loss": 1.2865,
      "theoretical_loss": 3.4819420511181756,
      "tokens_seen": 1677524992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000488,
      "loss": 1.6925,
      "theoretical_loss": 3.481930313621357,
      "tokens_seen": 1677590528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00049,
      "loss": 1.2952,
      "theoretical_loss": 3.4819185767114442,
      "tokens_seen": 1677656064
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000492,
      "loss": 1.3687,
      "theoretical_loss": 3.4819068403883855,
      "tokens_seen": 1677721600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000494,
      "loss": 1.4566,
      "theoretical_loss": 3.4818951046521276,
      "tokens_seen": 1677787136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000496,
      "loss": 1.3783,
      "theoretical_loss": 3.4818833695026195,
      "tokens_seen": 1677852672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498,
      "loss": 1.4967,
      "theoretical_loss": 3.4818716349398082,
      "tokens_seen": 1677918208
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.42913001775741577,
      "objective/train/avg_token_score": 0.028014685958623886,
      "objective/train/docs_used": 950694,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2823351621627808,
      "objective/train/original_loss": 4.018346309661865,
      "objective/train/theoretical_loss": 3.481859900963642,
      "objective/train/tokens_used": 36842976,
      "objective/train/unlikelihood_loss": 0.06534267216920853,
      "theoretical_loss": 3.481859900963642,
      "tokens_seen": 1677983744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0005,
      "loss": 1.4589,
      "theoretical_loss": 3.481859900963642,
      "tokens_seen": 1677983744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499979797979798,
      "loss": 1.7066,
      "theoretical_loss": 3.481848167574068,
      "tokens_seen": 1678049280
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999595959595959,
      "loss": 1.4743,
      "theoretical_loss": 3.481836434771034,
      "tokens_seen": 1678114816
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004999393939393939,
      "loss": 1.4814,
      "theoretical_loss": 3.4818247025544884,
      "tokens_seen": 1678180352
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499919191919192,
      "loss": 1.5105,
      "theoretical_loss": 3.481812970924379,
      "tokens_seen": 1678245888
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00049989898989899,
      "loss": 1.3163,
      "theoretical_loss": 3.481801239880652,
      "tokens_seen": 1678311424
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998787878787879,
      "loss": 1.5049,
      "theoretical_loss": 3.4817895094232574,
      "tokens_seen": 1678376960
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998585858585859,
      "loss": 1.7115,
      "theoretical_loss": 3.481777779552141,
      "tokens_seen": 1678442496
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998383838383839,
      "loss": 1.3624,
      "theoretical_loss": 3.481766050267253,
      "tokens_seen": 1678508032
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004998181818181818,
      "loss": 1.4109,
      "theoretical_loss": 3.481754321568539,
      "tokens_seen": 1678573568
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997979797979798,
      "loss": 1.6151,
      "theoretical_loss": 3.4817425934559476,
      "tokens_seen": 1678639104
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997777777777778,
      "loss": 1.4839,
      "theoretical_loss": 3.4817308659294266,
      "tokens_seen": 1678704640
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997575757575757,
      "loss": 1.3913,
      "theoretical_loss": 3.4817191389889244,
      "tokens_seen": 1678770176
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997373737373737,
      "loss": 1.4744,
      "theoretical_loss": 3.481707412634388,
      "tokens_seen": 1678835712
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004997171717171717,
      "loss": 1.2652,
      "theoretical_loss": 3.481695686865766,
      "tokens_seen": 1678901248
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996969696969697,
      "loss": 1.5878,
      "theoretical_loss": 3.481683961683005,
      "tokens_seen": 1678966784
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996767676767676,
      "loss": 1.2625,
      "theoretical_loss": 3.4816722370860544,
      "tokens_seen": 1679032320
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996565656565657,
      "loss": 1.3929,
      "theoretical_loss": 3.4816605130748615,
      "tokens_seen": 1679097856
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996363636363637,
      "loss": 1.4169,
      "theoretical_loss": 3.4816487896493733,
      "tokens_seen": 1679163392
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004996161616161617,
      "loss": 1.4403,
      "theoretical_loss": 3.4816370668095393,
      "tokens_seen": 1679228928
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995959595959596,
      "loss": 1.5393,
      "theoretical_loss": 3.4816253445553063,
      "tokens_seen": 1679294464
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995757575757576,
      "loss": 1.6212,
      "theoretical_loss": 3.4816136228866217,
      "tokens_seen": 1679360000
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995555555555555,
      "loss": 1.4312,
      "theoretical_loss": 3.481601901803435,
      "tokens_seen": 1679425536
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995353535353536,
      "loss": 1.4936,
      "theoretical_loss": 3.481590181305693,
      "tokens_seen": 1679491072
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004995151515151515,
      "loss": 1.29,
      "theoretical_loss": 3.4815784613933434,
      "tokens_seen": 1679556608
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.3876466155052185,
      "objective/train/avg_token_score": 0.019771317020058632,
      "objective/train/docs_used": 951339,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2079968452453613,
      "objective/train/original_loss": 3.654224395751953,
      "objective/train/theoretical_loss": 3.481566742066335,
      "objective/train/tokens_used": 38481376,
      "objective/train/unlikelihood_loss": 0.10271399468183517,
      "theoretical_loss": 3.481566742066335,
      "tokens_seen": 1679622144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994949494949495,
      "loss": 1.1997,
      "theoretical_loss": 3.481566742066335,
      "tokens_seen": 1679622144
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994747474747474,
      "loss": 1.3423,
      "theoretical_loss": 3.4815550233246153,
      "tokens_seen": 1679687680
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994545454545454,
      "loss": 1.416,
      "theoretical_loss": 3.481543305168132,
      "tokens_seen": 1679753216
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994343434343435,
      "loss": 1.4918,
      "theoretical_loss": 3.4815315875968333,
      "tokens_seen": 1679818752
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004994141414141415,
      "loss": 1.6278,
      "theoretical_loss": 3.4815198706106676,
      "tokens_seen": 1679884288
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993939393939394,
      "loss": 1.4463,
      "theoretical_loss": 3.481508154209582,
      "tokens_seen": 1679949824
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993737373737374,
      "loss": 1.4316,
      "theoretical_loss": 3.4814964383935245,
      "tokens_seen": 1680015360
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993535353535354,
      "loss": 1.5408,
      "theoretical_loss": 3.4814847231624437,
      "tokens_seen": 1680080896
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993333333333334,
      "loss": 1.4201,
      "theoretical_loss": 3.4814730085162875,
      "tokens_seen": 1680146432
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004993131313131313,
      "loss": 1.3325,
      "theoretical_loss": 3.4814612944550034,
      "tokens_seen": 1680211968
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992929292929293,
      "loss": 1.4038,
      "theoretical_loss": 3.4814495809785395,
      "tokens_seen": 1680277504
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992727272727272,
      "loss": 1.4769,
      "theoretical_loss": 3.4814378680868447,
      "tokens_seen": 1680343040
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992525252525252,
      "loss": 1.5646,
      "theoretical_loss": 3.4814261557798654,
      "tokens_seen": 1680408576
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992323232323232,
      "loss": 1.3709,
      "theoretical_loss": 3.4814144440575507,
      "tokens_seen": 1680474112
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004992121212121212,
      "loss": 1.3969,
      "theoretical_loss": 3.4814027329198485,
      "tokens_seen": 1680539648
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991919191919192,
      "loss": 1.4675,
      "theoretical_loss": 3.481391022366707,
      "tokens_seen": 1680605184
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991717171717172,
      "loss": 1.5211,
      "theoretical_loss": 3.481379312398073,
      "tokens_seen": 1680670720
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991515151515152,
      "loss": 1.424,
      "theoretical_loss": 3.4813676030138963,
      "tokens_seen": 1680736256
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991313131313132,
      "loss": 1.285,
      "theoretical_loss": 3.481355894214124,
      "tokens_seen": 1680801792
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004991111111111111,
      "loss": 1.5255,
      "theoretical_loss": 3.481344185998704,
      "tokens_seen": 1680867328
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990909090909091,
      "loss": 1.3467,
      "theoretical_loss": 3.4813324783675847,
      "tokens_seen": 1680932864
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990707070707071,
      "loss": 1.414,
      "theoretical_loss": 3.481320771320714,
      "tokens_seen": 1680998400
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004990505050505051,
      "loss": 1.4323,
      "theoretical_loss": 3.4813090648580403,
      "tokens_seen": 1681063936
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499030303030303,
      "loss": 1.2216,
      "theoretical_loss": 3.4812973589795106,
      "tokens_seen": 1681129472
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000499010101010101,
      "loss": 1.6606,
      "theoretical_loss": 3.4812856536850747,
      "tokens_seen": 1681195008
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.57746821641922,
      "objective/train/avg_token_score": 0.003087435383349657,
      "objective/train/docs_used": 952727,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.965943694114685,
      "objective/train/original_loss": 3.664557933807373,
      "objective/train/theoretical_loss": 3.4812739489746796,
      "objective/train/tokens_used": 40119776,
      "objective/train/unlikelihood_loss": 0.07379762828350067,
      "theoretical_loss": 3.4812739489746796,
      "tokens_seen": 1681260544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498989898989899,
      "loss": 1.3937,
      "theoretical_loss": 3.4812739489746796,
      "tokens_seen": 1681260544
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498969696969697,
      "loss": 1.4396,
      "theoretical_loss": 3.4812622448482733,
      "tokens_seen": 1681326080
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498949494949495,
      "loss": 1.5074,
      "theoretical_loss": 3.481250541305804,
      "tokens_seen": 1681391616
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498929292929293,
      "loss": 1.3075,
      "theoretical_loss": 3.4812388383472204,
      "tokens_seen": 1681457152
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004989090909090909,
      "loss": 1.4787,
      "theoretical_loss": 3.4812271359724702,
      "tokens_seen": 1681522688
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988888888888889,
      "loss": 1.4177,
      "theoretical_loss": 3.481215434181501,
      "tokens_seen": 1681588224
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988686868686869,
      "loss": 1.3858,
      "theoretical_loss": 3.481203732974262,
      "tokens_seen": 1681653760
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988484848484849,
      "loss": 1.5262,
      "theoretical_loss": 3.481192032350701,
      "tokens_seen": 1681719296
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988282828282828,
      "loss": 1.5809,
      "theoretical_loss": 3.481180332310765,
      "tokens_seen": 1681784832
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004988080808080808,
      "loss": 1.505,
      "theoretical_loss": 3.481168632854404,
      "tokens_seen": 1681850368
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987878787878788,
      "loss": 1.3779,
      "theoretical_loss": 3.4811569339815644,
      "tokens_seen": 1681915904
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987676767676767,
      "loss": 1.3716,
      "theoretical_loss": 3.4811452356921953,
      "tokens_seen": 1681981440
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987474747474747,
      "loss": 1.572,
      "theoretical_loss": 3.4811335379862456,
      "tokens_seen": 1682046976
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987272727272728,
      "loss": 1.4112,
      "theoretical_loss": 3.4811218408636617,
      "tokens_seen": 1682112512
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004987070707070708,
      "loss": 1.4665,
      "theoretical_loss": 3.481110144324393,
      "tokens_seen": 1682178048
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986868686868687,
      "loss": 1.4609,
      "theoretical_loss": 3.481098448368387,
      "tokens_seen": 1682243584
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986666666666667,
      "loss": 1.6894,
      "theoretical_loss": 3.4810867529955933,
      "tokens_seen": 1682309120
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986464646464647,
      "loss": 1.3648,
      "theoretical_loss": 3.481075058205958,
      "tokens_seen": 1682374656
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986262626262627,
      "loss": 1.5816,
      "theoretical_loss": 3.4810633639994313,
      "tokens_seen": 1682440192
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004986060606060606,
      "loss": 1.6544,
      "theoretical_loss": 3.48105167037596,
      "tokens_seen": 1682505728
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985858585858586,
      "loss": 1.4945,
      "theoretical_loss": 3.4810399773354925,
      "tokens_seen": 1682571264
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985656565656565,
      "loss": 1.3454,
      "theoretical_loss": 3.4810282848779783,
      "tokens_seen": 1682636800
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985454545454545,
      "loss": 1.2652,
      "theoretical_loss": 3.481016593003364,
      "tokens_seen": 1682702336
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985252525252525,
      "loss": 1.3969,
      "theoretical_loss": 3.4810049017115983,
      "tokens_seen": 1682767872
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004985050505050506,
      "loss": 1.5008,
      "theoretical_loss": 3.4809932110026303,
      "tokens_seen": 1682833408
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.4591275751590729,
      "objective/train/avg_token_score": 0.02467002533376217,
      "objective/train/docs_used": 953335,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5312504768371582,
      "objective/train/original_loss": 4.14010763168335,
      "objective/train/theoretical_loss": 3.4809815208764077,
      "objective/train/tokens_used": 41758176,
      "objective/train/unlikelihood_loss": 0.059290263801813126,
      "theoretical_loss": 3.4809815208764077,
      "tokens_seen": 1682898944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984848484848485,
      "loss": 1.423,
      "theoretical_loss": 3.4809815208764077,
      "tokens_seen": 1682898944
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984646464646465,
      "loss": 1.6025,
      "theoretical_loss": 3.480969831332878,
      "tokens_seen": 1682964480
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984444444444445,
      "loss": 1.6665,
      "theoretical_loss": 3.4809581423719904,
      "tokens_seen": 1683030016
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984242424242425,
      "loss": 1.472,
      "theoretical_loss": 3.480946453993693,
      "tokens_seen": 1683095552
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004984040404040404,
      "loss": 1.4715,
      "theoretical_loss": 3.480934766197934,
      "tokens_seen": 1683161088
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983838383838384,
      "loss": 1.4211,
      "theoretical_loss": 3.480923078984662,
      "tokens_seen": 1683226624
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983636363636363,
      "loss": 1.495,
      "theoretical_loss": 3.480911392353825,
      "tokens_seen": 1683292160
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983434343434344,
      "loss": 1.2509,
      "theoretical_loss": 3.4808997063053706,
      "tokens_seen": 1683357696
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983232323232323,
      "loss": 1.3817,
      "theoretical_loss": 3.480888020839248,
      "tokens_seen": 1683423232
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004983030303030303,
      "loss": 1.2817,
      "theoretical_loss": 3.480876335955406,
      "tokens_seen": 1683488768
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982828282828282,
      "loss": 1.4437,
      "theoretical_loss": 3.4808646516537918,
      "tokens_seen": 1683554304
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982626262626263,
      "loss": 1.5573,
      "theoretical_loss": 3.4808529679343536,
      "tokens_seen": 1683619840
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982424242424243,
      "loss": 1.6952,
      "theoretical_loss": 3.480841284797041,
      "tokens_seen": 1683685376
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982222222222223,
      "loss": 1.2028,
      "theoretical_loss": 3.4808296022418013,
      "tokens_seen": 1683750912
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004982020202020202,
      "loss": 1.3336,
      "theoretical_loss": 3.480817920268583,
      "tokens_seen": 1683816448
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981818181818182,
      "loss": 1.3224,
      "theoretical_loss": 3.480806238877335,
      "tokens_seen": 1683881984
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981616161616162,
      "loss": 1.392,
      "theoretical_loss": 3.4807945580680055,
      "tokens_seen": 1683947520
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981414141414142,
      "loss": 1.3183,
      "theoretical_loss": 3.480782877840542,
      "tokens_seen": 1684013056
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981212121212121,
      "loss": 1.5275,
      "theoretical_loss": 3.480771198194894,
      "tokens_seen": 1684078592
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004981010101010101,
      "loss": 1.3843,
      "theoretical_loss": 3.4807595191310092,
      "tokens_seen": 1684144128
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498080808080808,
      "loss": 1.4248,
      "theoretical_loss": 3.4807478406488364,
      "tokens_seen": 1684209664
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498060606060606,
      "loss": 1.5212,
      "theoretical_loss": 3.4807361627483235,
      "tokens_seen": 1684275200
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498040404040404,
      "loss": 1.5897,
      "theoretical_loss": 3.4807244854294197,
      "tokens_seen": 1684340736
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004980202020202021,
      "loss": 1.4325,
      "theoretical_loss": 3.480712808692072,
      "tokens_seen": 1684406272
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000498,
      "loss": 1.4759,
      "theoretical_loss": 3.4807011325362307,
      "tokens_seen": 1684471808
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.46138808131217957,
      "objective/train/avg_token_score": 0.026906196027994156,
      "objective/train/docs_used": 954676,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2592566013336182,
      "objective/train/original_loss": 3.659281015396118,
      "objective/train/theoretical_loss": 3.4806894569618425,
      "objective/train/tokens_used": 43396576,
      "objective/train/unlikelihood_loss": 0.058182328939437866,
      "theoretical_loss": 3.4806894569618425,
      "tokens_seen": 1684537344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497979797979798,
      "loss": 1.2782,
      "theoretical_loss": 3.4806894569618425,
      "tokens_seen": 1684537344
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497959595959596,
      "loss": 1.3906,
      "theoretical_loss": 3.480677781968857,
      "tokens_seen": 1684602880
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.000497939393939394,
      "loss": 1.5195,
      "theoretical_loss": 3.480666107557221,
      "tokens_seen": 1684668416
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004979191919191919,
      "loss": 1.494,
      "theoretical_loss": 3.4806544337268854,
      "tokens_seen": 1684733952
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978989898989899,
      "loss": 1.5243,
      "theoretical_loss": 3.4806427604777967,
      "tokens_seen": 1684799488
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978787878787879,
      "loss": 1.3817,
      "theoretical_loss": 3.4806310878099045,
      "tokens_seen": 1684865024
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978585858585858,
      "loss": 1.3818,
      "theoretical_loss": 3.4806194157231567,
      "tokens_seen": 1684930560
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978383838383838,
      "loss": 1.3979,
      "theoretical_loss": 3.480607744217502,
      "tokens_seen": 1684996096
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004978181818181818,
      "loss": 1.5915,
      "theoretical_loss": 3.4805960732928876,
      "tokens_seen": 1685061632
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977979797979799,
      "loss": 1.6397,
      "theoretical_loss": 3.480584402949264,
      "tokens_seen": 1685127168
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977777777777778,
      "loss": 1.7836,
      "theoretical_loss": 3.4805727331865786,
      "tokens_seen": 1685192704
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977575757575758,
      "loss": 1.4281,
      "theoretical_loss": 3.48056106400478,
      "tokens_seen": 1685258240
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977373737373738,
      "loss": 1.5813,
      "theoretical_loss": 3.480549395403817,
      "tokens_seen": 1685323776
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004977171717171717,
      "loss": 1.5063,
      "theoretical_loss": 3.4805377273836378,
      "tokens_seen": 1685389312
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976969696969697,
      "loss": 1.3745,
      "theoretical_loss": 3.4805260599441903,
      "tokens_seen": 1685454848
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976767676767677,
      "loss": 1.5523,
      "theoretical_loss": 3.480514393085424,
      "tokens_seen": 1685520384
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976565656565656,
      "loss": 1.3951,
      "theoretical_loss": 3.4805027268072877,
      "tokens_seen": 1685585920
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976363636363636,
      "loss": 1.3416,
      "theoretical_loss": 3.4804910611097286,
      "tokens_seen": 1685651456
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004976161616161616,
      "loss": 1.5686,
      "theoretical_loss": 3.4804793959926963,
      "tokens_seen": 1685716992
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975959595959596,
      "loss": 1.489,
      "theoretical_loss": 3.480467731456139,
      "tokens_seen": 1685782528
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975757575757575,
      "loss": 1.341,
      "theoretical_loss": 3.4804560675000054,
      "tokens_seen": 1685848064
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975555555555556,
      "loss": 1.2872,
      "theoretical_loss": 3.4804444041242437,
      "tokens_seen": 1685913600
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975353535353536,
      "loss": 1.5454,
      "theoretical_loss": 3.480432741328803,
      "tokens_seen": 1685979136
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004975151515151516,
      "loss": 1.5041,
      "theoretical_loss": 3.4804210791136314,
      "tokens_seen": 1686044672
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004974949494949495,
      "loss": 1.5567,
      "theoretical_loss": 3.4804094174786773,
      "tokens_seen": 1686110208
    },
    {
      "epoch": 0.01,
      "objective/train/aligned_token_count": 0.5329912304878235,
      "objective/train/avg_token_score": 0.01913226768374443,
      "objective/train/docs_used": 955358,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6620068550109863,
      "objective/train/original_loss": 3.886705160140991,
      "objective/train/theoretical_loss": 3.4803977564238897,
      "objective/train/tokens_used": 45034976,
      "objective/train/unlikelihood_loss": 0.04157567396759987,
      "theoretical_loss": 3.4803977564238897,
      "tokens_seen": 1686175744
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0004974747474747475,
      "loss": 1.5445,
      "theoretical_loss": 3.4803977564238897,
      "tokens_seen": 1686175744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974545454545454,
      "loss": 1.4179,
      "theoretical_loss": 3.4803860959492177,
      "tokens_seen": 1686241280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974343434343435,
      "loss": 1.4412,
      "theoretical_loss": 3.4803744360546087,
      "tokens_seen": 1686306816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004974141414141414,
      "loss": 1.5498,
      "theoretical_loss": 3.4803627767400123,
      "tokens_seen": 1686372352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973939393939394,
      "loss": 1.4011,
      "theoretical_loss": 3.480351118005376,
      "tokens_seen": 1686437888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973737373737373,
      "loss": 1.4917,
      "theoretical_loss": 3.48033945985065,
      "tokens_seen": 1686503424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973535353535353,
      "loss": 1.392,
      "theoretical_loss": 3.4803278022757818,
      "tokens_seen": 1686568960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973333333333334,
      "loss": 1.4304,
      "theoretical_loss": 3.4803161452807196,
      "tokens_seen": 1686634496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004973131313131314,
      "loss": 1.4172,
      "theoretical_loss": 3.4803044888654133,
      "tokens_seen": 1686700032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972929292929293,
      "loss": 1.2836,
      "theoretical_loss": 3.480292833029811,
      "tokens_seen": 1686765568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972727272727273,
      "loss": 1.5495,
      "theoretical_loss": 3.480281177773861,
      "tokens_seen": 1686831104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972525252525253,
      "loss": 1.3092,
      "theoretical_loss": 3.4802695230975127,
      "tokens_seen": 1686896640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972323232323233,
      "loss": 1.3053,
      "theoretical_loss": 3.480257869000714,
      "tokens_seen": 1686962176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004972121212121212,
      "loss": 1.3317,
      "theoretical_loss": 3.4802462154834135,
      "tokens_seen": 1687027712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971919191919192,
      "loss": 1.484,
      "theoretical_loss": 3.4802345625455606,
      "tokens_seen": 1687093248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971717171717171,
      "loss": 1.4059,
      "theoretical_loss": 3.480222910187104,
      "tokens_seen": 1687158784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971515151515152,
      "loss": 1.5274,
      "theoretical_loss": 3.4802112584079916,
      "tokens_seen": 1687224320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971313131313131,
      "loss": 1.1674,
      "theoretical_loss": 3.4801996072081725,
      "tokens_seen": 1687289856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004971111111111111,
      "loss": 1.721,
      "theoretical_loss": 3.4801879565875957,
      "tokens_seen": 1687355392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497090909090909,
      "loss": 1.2672,
      "theoretical_loss": 3.4801763065462095,
      "tokens_seen": 1687420928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970707070707071,
      "loss": 1.6776,
      "theoretical_loss": 3.4801646570839626,
      "tokens_seen": 1687486464
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970505050505051,
      "loss": 1.3374,
      "theoretical_loss": 3.480153008200804,
      "tokens_seen": 1687552000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004970303030303031,
      "loss": 1.2859,
      "theoretical_loss": 3.480141359896683,
      "tokens_seen": 1687617536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000497010101010101,
      "loss": 1.5063,
      "theoretical_loss": 3.4801297121715464,
      "tokens_seen": 1687683072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496989898989899,
      "loss": 1.3917,
      "theoretical_loss": 3.4801180650253447,
      "tokens_seen": 1687748608
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.4268084168434143,
      "objective/train/avg_token_score": 0.010676675476133823,
      "objective/train/docs_used": 956671,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4835927486419678,
      "objective/train/original_loss": 4.212736129760742,
      "objective/train/theoretical_loss": 3.480106418458026,
      "objective/train/tokens_used": 46673376,
      "objective/train/unlikelihood_loss": 0.08470705896615982,
      "theoretical_loss": 3.480106418458026,
      "tokens_seen": 1687814144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496969696969697,
      "loss": 1.4357,
      "theoretical_loss": 3.480106418458026,
      "tokens_seen": 1687814144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496949494949495,
      "loss": 1.3479,
      "theoretical_loss": 3.4800947724695392,
      "tokens_seen": 1687879680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969292929292929,
      "loss": 1.2908,
      "theoretical_loss": 3.4800831270598334,
      "tokens_seen": 1687945216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004969090909090909,
      "loss": 1.506,
      "theoretical_loss": 3.4800714822288565,
      "tokens_seen": 1688010752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968888888888889,
      "loss": 1.5604,
      "theoretical_loss": 3.4800598379765577,
      "tokens_seen": 1688076288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968686868686868,
      "loss": 1.435,
      "theoretical_loss": 3.4800481943028863,
      "tokens_seen": 1688141824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968484848484849,
      "loss": 1.3855,
      "theoretical_loss": 3.48003655120779,
      "tokens_seen": 1688207360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968282828282829,
      "loss": 1.579,
      "theoretical_loss": 3.4800249086912185,
      "tokens_seen": 1688272896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004968080808080808,
      "loss": 1.5069,
      "theoretical_loss": 3.4800132667531205,
      "tokens_seen": 1688338432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967878787878788,
      "loss": 1.5546,
      "theoretical_loss": 3.4800016253934447,
      "tokens_seen": 1688403968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967676767676768,
      "loss": 1.2675,
      "theoretical_loss": 3.4799899846121396,
      "tokens_seen": 1688469504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967474747474748,
      "loss": 1.5451,
      "theoretical_loss": 3.4799783444091545,
      "tokens_seen": 1688535040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967272727272727,
      "loss": 1.4007,
      "theoretical_loss": 3.4799667047844376,
      "tokens_seen": 1688600576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004967070707070707,
      "loss": 1.4922,
      "theoretical_loss": 3.479955065737938,
      "tokens_seen": 1688666112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966868686868687,
      "loss": 1.6596,
      "theoretical_loss": 3.479943427269605,
      "tokens_seen": 1688731648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966666666666666,
      "loss": 1.7516,
      "theoretical_loss": 3.4799317893793873,
      "tokens_seen": 1688797184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966464646464646,
      "loss": 1.437,
      "theoretical_loss": 3.479920152067233,
      "tokens_seen": 1688862720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966262626262627,
      "loss": 1.3793,
      "theoretical_loss": 3.4799085153330918,
      "tokens_seen": 1688928256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004966060606060607,
      "loss": 1.6005,
      "theoretical_loss": 3.479896879176912,
      "tokens_seen": 1688993792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965858585858586,
      "loss": 1.5023,
      "theoretical_loss": 3.479885243598643,
      "tokens_seen": 1689059328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965656565656566,
      "loss": 1.4687,
      "theoretical_loss": 3.4798736085982336,
      "tokens_seen": 1689124864
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965454545454546,
      "loss": 1.6716,
      "theoretical_loss": 3.479861974175632,
      "tokens_seen": 1689190400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965252525252525,
      "loss": 1.6124,
      "theoretical_loss": 3.4798503403307874,
      "tokens_seen": 1689255936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004965050505050505,
      "loss": 1.2172,
      "theoretical_loss": 3.4798387070636494,
      "tokens_seen": 1689321472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964848484848485,
      "loss": 1.7067,
      "theoretical_loss": 3.479827074374166,
      "tokens_seen": 1689387008
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.4199046790599823,
      "objective/train/avg_token_score": 0.011589493602514267,
      "objective/train/docs_used": 957370,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3580185174942017,
      "objective/train/original_loss": 4.241112232208252,
      "objective/train/theoretical_loss": 3.4798154422622867,
      "objective/train/tokens_used": 48311776,
      "objective/train/unlikelihood_loss": 0.06464245170354843,
      "theoretical_loss": 3.4798154422622867,
      "tokens_seen": 1689452544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964646464646464,
      "loss": 1.4656,
      "theoretical_loss": 3.4798154422622867,
      "tokens_seen": 1689452544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964444444444444,
      "loss": 1.4522,
      "theoretical_loss": 3.4798038107279603,
      "tokens_seen": 1689518080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964242424242424,
      "loss": 1.5003,
      "theoretical_loss": 3.4797921797711355,
      "tokens_seen": 1689583616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004964040404040405,
      "loss": 1.4859,
      "theoretical_loss": 3.4797805493917613,
      "tokens_seen": 1689649152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963838383838384,
      "loss": 1.5211,
      "theoretical_loss": 3.4797689195897865,
      "tokens_seen": 1689714688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963636363636364,
      "loss": 1.5367,
      "theoretical_loss": 3.4797572903651606,
      "tokens_seen": 1689780224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963434343434344,
      "loss": 1.4631,
      "theoretical_loss": 3.479745661717832,
      "tokens_seen": 1689845760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963232323232324,
      "loss": 1.4319,
      "theoretical_loss": 3.47973403364775,
      "tokens_seen": 1689911296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004963030303030303,
      "loss": 1.2408,
      "theoretical_loss": 3.4797224061548633,
      "tokens_seen": 1689976832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962828282828283,
      "loss": 1.3776,
      "theoretical_loss": 3.4797107792391206,
      "tokens_seen": 1690042368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962626262626262,
      "loss": 1.4295,
      "theoretical_loss": 3.4796991529004715,
      "tokens_seen": 1690107904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962424242424243,
      "loss": 1.5602,
      "theoretical_loss": 3.4796875271388648,
      "tokens_seen": 1690173440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962222222222222,
      "loss": 1.3299,
      "theoretical_loss": 3.4796759019542494,
      "tokens_seen": 1690238976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004962020202020202,
      "loss": 1.4424,
      "theoretical_loss": 3.479664277346574,
      "tokens_seen": 1690304512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961818181818181,
      "loss": 1.6119,
      "theoretical_loss": 3.479652653315788,
      "tokens_seen": 1690370048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961616161616161,
      "loss": 1.5685,
      "theoretical_loss": 3.4796410298618405,
      "tokens_seen": 1690435584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961414141414142,
      "loss": 1.4242,
      "theoretical_loss": 3.4796294069846803,
      "tokens_seen": 1690501120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961212121212122,
      "loss": 1.4835,
      "theoretical_loss": 3.4796177846842564,
      "tokens_seen": 1690566656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004961010101010101,
      "loss": 1.3436,
      "theoretical_loss": 3.479606162960518,
      "tokens_seen": 1690632192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960808080808081,
      "loss": 1.6164,
      "theoretical_loss": 3.4795945418134133,
      "tokens_seen": 1690697728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960606060606061,
      "loss": 1.6758,
      "theoretical_loss": 3.479582921242893,
      "tokens_seen": 1690763264
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004960404040404041,
      "loss": 1.298,
      "theoretical_loss": 3.4795713012489045,
      "tokens_seen": 1690828800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496020202020202,
      "loss": 1.4461,
      "theoretical_loss": 3.4795596818313976,
      "tokens_seen": 1690894336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000496,
      "loss": 1.4362,
      "theoretical_loss": 3.4795480629903217,
      "tokens_seen": 1690959872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959797979797979,
      "loss": 1.3436,
      "theoretical_loss": 3.479536444725625,
      "tokens_seen": 1691025408
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.38599705696105957,
      "objective/train/avg_token_score": 0.048682406544685364,
      "objective/train/docs_used": 958661,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2436673641204834,
      "objective/train/original_loss": 4.153236389160156,
      "objective/train/theoretical_loss": 3.4795248270372574,
      "objective/train/tokens_used": 49950176,
      "objective/train/unlikelihood_loss": 0.07140035182237625,
      "theoretical_loss": 3.4795248270372574,
      "tokens_seen": 1691090944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004959595959595959,
      "loss": 1.2475,
      "theoretical_loss": 3.4795248270372574,
      "tokens_seen": 1691090944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495939393939394,
      "loss": 1.6063,
      "theoretical_loss": 3.479513209925167,
      "tokens_seen": 1691156480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495919191919192,
      "loss": 1.4383,
      "theoretical_loss": 3.479501593389304,
      "tokens_seen": 1691222016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958989898989899,
      "loss": 1.6435,
      "theoretical_loss": 3.479489977429617,
      "tokens_seen": 1691287552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958787878787879,
      "loss": 1.5405,
      "theoretical_loss": 3.4794783620460548,
      "tokens_seen": 1691353088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958585858585859,
      "loss": 1.5615,
      "theoretical_loss": 3.479466747238567,
      "tokens_seen": 1691418624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958383838383839,
      "loss": 1.51,
      "theoretical_loss": 3.4794551330071024,
      "tokens_seen": 1691484160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004958181818181818,
      "loss": 1.4175,
      "theoretical_loss": 3.47944351935161,
      "tokens_seen": 1691549696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957979797979798,
      "loss": 1.5496,
      "theoretical_loss": 3.4794319062720396,
      "tokens_seen": 1691615232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957777777777778,
      "loss": 1.4502,
      "theoretical_loss": 3.479420293768339,
      "tokens_seen": 1691680768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957575757575757,
      "loss": 1.4327,
      "theoretical_loss": 3.479408681840459,
      "tokens_seen": 1691746304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957373737373737,
      "loss": 1.4366,
      "theoretical_loss": 3.4793970704883477,
      "tokens_seen": 1691811840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004957171717171717,
      "loss": 1.2897,
      "theoretical_loss": 3.4793854597119545,
      "tokens_seen": 1691877376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956969696969698,
      "loss": 1.6798,
      "theoretical_loss": 3.4793738495112283,
      "tokens_seen": 1691942912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956767676767677,
      "loss": 1.2665,
      "theoretical_loss": 3.4793622398861186,
      "tokens_seen": 1692008448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956565656565657,
      "loss": 1.5467,
      "theoretical_loss": 3.4793506308365747,
      "tokens_seen": 1692073984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956363636363637,
      "loss": 1.3853,
      "theoretical_loss": 3.479339022362545,
      "tokens_seen": 1692139520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004956161616161616,
      "loss": 1.5345,
      "theoretical_loss": 3.47932741446398,
      "tokens_seen": 1692205056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955959595959596,
      "loss": 1.6217,
      "theoretical_loss": 3.4793158071408277,
      "tokens_seen": 1692270592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955757575757576,
      "loss": 1.6179,
      "theoretical_loss": 3.4793042003930377,
      "tokens_seen": 1692336128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955555555555556,
      "loss": 1.3264,
      "theoretical_loss": 3.479292594220559,
      "tokens_seen": 1692401664
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955353535353535,
      "loss": 1.6683,
      "theoretical_loss": 3.4792809886233416,
      "tokens_seen": 1692467200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004955151515151515,
      "loss": 1.6922,
      "theoretical_loss": 3.4792693836013333,
      "tokens_seen": 1692532736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954949494949495,
      "loss": 1.6503,
      "theoretical_loss": 3.4792577791544845,
      "tokens_seen": 1692598272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954747474747474,
      "loss": 1.6034,
      "theoretical_loss": 3.4792461752827446,
      "tokens_seen": 1692663808
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.4952957034111023,
      "objective/train/avg_token_score": 0.008989947848021984,
      "objective/train/docs_used": 959159,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6178122758865356,
      "objective/train/original_loss": 3.7790725231170654,
      "objective/train/theoretical_loss": 3.479234571986062,
      "objective/train/tokens_used": 51588576,
      "objective/train/unlikelihood_loss": 0.05506987124681473,
      "theoretical_loss": 3.479234571986062,
      "tokens_seen": 1692729344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954545454545455,
      "loss": 1.5392,
      "theoretical_loss": 3.479234571986062,
      "tokens_seen": 1692729344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954343434343435,
      "loss": 1.5108,
      "theoretical_loss": 3.479222969264386,
      "tokens_seen": 1692794880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004954141414141415,
      "loss": 1.404,
      "theoretical_loss": 3.4792113671176663,
      "tokens_seen": 1692860416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953939393939394,
      "loss": 1.4155,
      "theoretical_loss": 3.479199765545852,
      "tokens_seen": 1692925952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953737373737374,
      "loss": 1.4939,
      "theoretical_loss": 3.4791881645488916,
      "tokens_seen": 1692991488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953535353535354,
      "loss": 1.5546,
      "theoretical_loss": 3.479176564126736,
      "tokens_seen": 1693057024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953333333333334,
      "loss": 1.6032,
      "theoretical_loss": 3.4791649642793328,
      "tokens_seen": 1693122560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004953131313131313,
      "loss": 1.4701,
      "theoretical_loss": 3.479153365006632,
      "tokens_seen": 1693188096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952929292929293,
      "loss": 1.4542,
      "theoretical_loss": 3.479141766308584,
      "tokens_seen": 1693253632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952727272727272,
      "loss": 1.508,
      "theoretical_loss": 3.479130168185136,
      "tokens_seen": 1693319168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952525252525252,
      "loss": 1.5405,
      "theoretical_loss": 3.4791185706362384,
      "tokens_seen": 1693384704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952323232323232,
      "loss": 1.3952,
      "theoretical_loss": 3.4791069736618407,
      "tokens_seen": 1693450240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004952121212121213,
      "loss": 1.4216,
      "theoretical_loss": 3.479095377261892,
      "tokens_seen": 1693515776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951919191919192,
      "loss": 1.5664,
      "theoretical_loss": 3.4790837814363407,
      "tokens_seen": 1693581312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951717171717172,
      "loss": 1.6256,
      "theoretical_loss": 3.479072186185138,
      "tokens_seen": 1693646848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951515151515152,
      "loss": 1.3279,
      "theoretical_loss": 3.4790605915082313,
      "tokens_seen": 1693712384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951313131313132,
      "loss": 1.3493,
      "theoretical_loss": 3.479048997405571,
      "tokens_seen": 1693777920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004951111111111111,
      "loss": 1.6102,
      "theoretical_loss": 3.479037403877107,
      "tokens_seen": 1693843456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950909090909091,
      "loss": 1.2949,
      "theoretical_loss": 3.4790258109227867,
      "tokens_seen": 1693908992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495070707070707,
      "loss": 1.3058,
      "theoretical_loss": 3.479014218542561,
      "tokens_seen": 1693974528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004950505050505051,
      "loss": 1.375,
      "theoretical_loss": 3.4790026267363796,
      "tokens_seen": 1694040064
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495030303030303,
      "loss": 1.4453,
      "theoretical_loss": 3.478991035504191,
      "tokens_seen": 1694105600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000495010101010101,
      "loss": 1.55,
      "theoretical_loss": 3.4789794448459443,
      "tokens_seen": 1694171136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494989898989899,
      "loss": 1.5699,
      "theoretical_loss": 3.4789678547615894,
      "tokens_seen": 1694236672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494969696969697,
      "loss": 1.3551,
      "theoretical_loss": 3.478956265251076,
      "tokens_seen": 1694302208
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.3170210123062134,
      "objective/train/avg_token_score": 0.017827244475483894,
      "objective/train/docs_used": 960536,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9088751673698425,
      "objective/train/original_loss": 4.046257972717285,
      "objective/train/theoretical_loss": 3.478944676314353,
      "objective/train/tokens_used": 53226976,
      "objective/train/unlikelihood_loss": 0.10128523409366608,
      "theoretical_loss": 3.478944676314353,
      "tokens_seen": 1694367744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494949494949495,
      "loss": 1.3764,
      "theoretical_loss": 3.478944676314353,
      "tokens_seen": 1694367744
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494929292929293,
      "loss": 1.7527,
      "theoretical_loss": 3.4789330879513694,
      "tokens_seen": 1694433280
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004949090909090909,
      "loss": 1.5846,
      "theoretical_loss": 3.478921500162076,
      "tokens_seen": 1694498816
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948888888888889,
      "loss": 1.3229,
      "theoretical_loss": 3.4789099129464205,
      "tokens_seen": 1694564352
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948686868686869,
      "loss": 1.5288,
      "theoretical_loss": 3.4788983263043534,
      "tokens_seen": 1694629888
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948484848484849,
      "loss": 1.5913,
      "theoretical_loss": 3.478886740235824,
      "tokens_seen": 1694695424
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948282828282828,
      "loss": 1.4224,
      "theoretical_loss": 3.4788751547407815,
      "tokens_seen": 1694760960
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004948080808080808,
      "loss": 1.6395,
      "theoretical_loss": 3.478863569819176,
      "tokens_seen": 1694826496
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947878787878787,
      "loss": 1.4925,
      "theoretical_loss": 3.478851985470956,
      "tokens_seen": 1694892032
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947676767676767,
      "loss": 1.5345,
      "theoretical_loss": 3.478840401696071,
      "tokens_seen": 1694957568
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947474747474748,
      "loss": 1.3046,
      "theoretical_loss": 3.4788288184944713,
      "tokens_seen": 1695023104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947272727272728,
      "loss": 1.5731,
      "theoretical_loss": 3.4788172358661056,
      "tokens_seen": 1695088640
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004947070707070707,
      "loss": 1.3865,
      "theoretical_loss": 3.478805653810924,
      "tokens_seen": 1695154176
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946868686868687,
      "loss": 1.3987,
      "theoretical_loss": 3.478794072328875,
      "tokens_seen": 1695219712
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946666666666667,
      "loss": 1.4943,
      "theoretical_loss": 3.478782491419909,
      "tokens_seen": 1695285248
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946464646464647,
      "loss": 1.2217,
      "theoretical_loss": 3.4787709110839753,
      "tokens_seen": 1695350784
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946262626262626,
      "loss": 1.4646,
      "theoretical_loss": 3.478759331321023,
      "tokens_seen": 1695416320
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004946060606060606,
      "loss": 1.4741,
      "theoretical_loss": 3.478747752131002,
      "tokens_seen": 1695481856
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945858585858586,
      "loss": 1.4813,
      "theoretical_loss": 3.4787361735138624,
      "tokens_seen": 1695547392
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945656565656565,
      "loss": 1.3267,
      "theoretical_loss": 3.478724595469552,
      "tokens_seen": 1695612928
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945454545454545,
      "loss": 1.3645,
      "theoretical_loss": 3.4787130179980217,
      "tokens_seen": 1695678464
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945252525252526,
      "loss": 1.4331,
      "theoretical_loss": 3.478701441099221,
      "tokens_seen": 1695744000
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004945050505050506,
      "loss": 1.481,
      "theoretical_loss": 3.4786898647730986,
      "tokens_seen": 1695809536
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944848484848485,
      "loss": 1.4608,
      "theoretical_loss": 3.478678289019605,
      "tokens_seen": 1695875072
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944646464646465,
      "loss": 1.4611,
      "theoretical_loss": 3.478666713838689,
      "tokens_seen": 1695940608
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.3928396999835968,
      "objective/train/avg_token_score": 0.021624328568577766,
      "objective/train/docs_used": 961177,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3411294221878052,
      "objective/train/original_loss": 4.00459623336792,
      "objective/train/theoretical_loss": 3.478655139230301,
      "objective/train/tokens_used": 54865376,
      "objective/train/unlikelihood_loss": 0.05852148309350014,
      "theoretical_loss": 3.478655139230301,
      "tokens_seen": 1696006144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944444444444445,
      "loss": 1.4319,
      "theoretical_loss": 3.478655139230301,
      "tokens_seen": 1696006144
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944242424242424,
      "loss": 1.432,
      "theoretical_loss": 3.478643565194389,
      "tokens_seen": 1696071680
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004944040404040404,
      "loss": 1.4987,
      "theoretical_loss": 3.478631991730904,
      "tokens_seen": 1696137216
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943838383838384,
      "loss": 1.6415,
      "theoretical_loss": 3.478620418839795,
      "tokens_seen": 1696202752
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943636363636363,
      "loss": 1.6562,
      "theoretical_loss": 3.4786088465210123,
      "tokens_seen": 1696268288
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943434343434343,
      "loss": 1.3873,
      "theoretical_loss": 3.478597274774504,
      "tokens_seen": 1696333824
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943232323232323,
      "loss": 1.3246,
      "theoretical_loss": 3.4785857036002215,
      "tokens_seen": 1696399360
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004943030303030303,
      "loss": 1.3636,
      "theoretical_loss": 3.478574132998113,
      "tokens_seen": 1696464896
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942828282828283,
      "loss": 1.446,
      "theoretical_loss": 3.478562562968129,
      "tokens_seen": 1696530432
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942626262626263,
      "loss": 1.3962,
      "theoretical_loss": 3.4785509935102183,
      "tokens_seen": 1696595968
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942424242424243,
      "loss": 1.4457,
      "theoretical_loss": 3.4785394246243317,
      "tokens_seen": 1696661504
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942222222222223,
      "loss": 1.6089,
      "theoretical_loss": 3.478527856310417,
      "tokens_seen": 1696727040
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004942020202020202,
      "loss": 1.5574,
      "theoretical_loss": 3.4785162885684255,
      "tokens_seen": 1696792576
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941818181818182,
      "loss": 1.4105,
      "theoretical_loss": 3.478504721398306,
      "tokens_seen": 1696858112
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941616161616161,
      "loss": 1.399,
      "theoretical_loss": 3.4784931548000086,
      "tokens_seen": 1696923648
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941414141414142,
      "loss": 1.5323,
      "theoretical_loss": 3.4784815887734823,
      "tokens_seen": 1696989184
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941212121212121,
      "loss": 1.5829,
      "theoretical_loss": 3.4784700233186774,
      "tokens_seen": 1697054720
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004941010101010101,
      "loss": 1.4786,
      "theoretical_loss": 3.4784584584355436,
      "tokens_seen": 1697120256
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494080808080808,
      "loss": 1.5934,
      "theoretical_loss": 3.47844689412403,
      "tokens_seen": 1697185792
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494060606060606,
      "loss": 1.4969,
      "theoretical_loss": 3.4784353303840865,
      "tokens_seen": 1697251328
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940404040404041,
      "loss": 1.5708,
      "theoretical_loss": 3.478423767215663,
      "tokens_seen": 1697316864
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004940202020202021,
      "loss": 1.3845,
      "theoretical_loss": 3.478412204618709,
      "tokens_seen": 1697382400
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000494,
      "loss": 1.3359,
      "theoretical_loss": 3.4784006425931744,
      "tokens_seen": 1697447936
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493979797979798,
      "loss": 1.4839,
      "theoretical_loss": 3.478389081139009,
      "tokens_seen": 1697513472
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493959595959596,
      "loss": 1.5483,
      "theoretical_loss": 3.4783775202561618,
      "tokens_seen": 1697579008
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.5273093581199646,
      "objective/train/avg_token_score": 0.005273816175758839,
      "objective/train/docs_used": 961828,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.706133484840393,
      "objective/train/original_loss": 3.877432107925415,
      "objective/train/theoretical_loss": 3.4783659599445826,
      "objective/train/tokens_used": 56503776,
      "objective/train/unlikelihood_loss": 0.0712694600224495,
      "theoretical_loss": 3.4783659599445826,
      "tokens_seen": 1697644544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493939393939394,
      "loss": 1.5571,
      "theoretical_loss": 3.4783659599445826,
      "tokens_seen": 1697644544
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004939191919191919,
      "loss": 1.2926,
      "theoretical_loss": 3.478354400204222,
      "tokens_seen": 1697710080
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938989898989899,
      "loss": 1.3542,
      "theoretical_loss": 3.4783428410350297,
      "tokens_seen": 1697775616
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938787878787878,
      "loss": 1.6162,
      "theoretical_loss": 3.4783312824369537,
      "tokens_seen": 1697841152
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938585858585859,
      "loss": 1.5232,
      "theoretical_loss": 3.478319724409946,
      "tokens_seen": 1697906688
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938383838383838,
      "loss": 1.5115,
      "theoretical_loss": 3.478308166953955,
      "tokens_seen": 1697972224
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004938181818181819,
      "loss": 1.6203,
      "theoretical_loss": 3.4782966100689308,
      "tokens_seen": 1698037760
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937979797979798,
      "loss": 1.6335,
      "theoretical_loss": 3.4782850537548233,
      "tokens_seen": 1698103296
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937777777777778,
      "loss": 1.3283,
      "theoretical_loss": 3.478273498011582,
      "tokens_seen": 1698168832
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937575757575758,
      "loss": 1.5045,
      "theoretical_loss": 3.4782619428391564,
      "tokens_seen": 1698234368
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937373737373738,
      "loss": 1.5789,
      "theoretical_loss": 3.478250388237497,
      "tokens_seen": 1698299904
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004937171717171717,
      "loss": 1.4186,
      "theoretical_loss": 3.478238834206553,
      "tokens_seen": 1698365440
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936969696969697,
      "loss": 1.4197,
      "theoretical_loss": 3.4782272807462746,
      "tokens_seen": 1698430976
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936767676767677,
      "loss": 1.384,
      "theoretical_loss": 3.4782157278566115,
      "tokens_seen": 1698496512
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936565656565657,
      "loss": 1.553,
      "theoretical_loss": 3.4782041755375133,
      "tokens_seen": 1698562048
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936363636363636,
      "loss": 1.4178,
      "theoretical_loss": 3.47819262378893,
      "tokens_seen": 1698627584
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004936161616161616,
      "loss": 1.6003,
      "theoretical_loss": 3.478181072610811,
      "tokens_seen": 1698693120
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935959595959597,
      "loss": 1.0584,
      "theoretical_loss": 3.478169522003107,
      "tokens_seen": 1698758656
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935757575757576,
      "loss": 1.4205,
      "theoretical_loss": 3.478157971965767,
      "tokens_seen": 1698824192
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935555555555556,
      "loss": 1.4223,
      "theoretical_loss": 3.4781464224987415,
      "tokens_seen": 1698889728
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935353535353536,
      "loss": 1.3599,
      "theoretical_loss": 3.478134873601979,
      "tokens_seen": 1698955264
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004935151515151515,
      "loss": 1.375,
      "theoretical_loss": 3.4781233252754316,
      "tokens_seen": 1699020800
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934949494949495,
      "loss": 1.4275,
      "theoretical_loss": 3.4781117775190467,
      "tokens_seen": 1699086336
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934747474747475,
      "loss": 1.3522,
      "theoretical_loss": 3.4781002303327764,
      "tokens_seen": 1699151872
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934545454545455,
      "loss": 1.48,
      "theoretical_loss": 3.4780886837165688,
      "tokens_seen": 1699217408
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.35807672142982483,
      "objective/train/avg_token_score": 0.012293923646211624,
      "objective/train/docs_used": 963095,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1824865341186523,
      "objective/train/original_loss": 4.2231035232543945,
      "objective/train/theoretical_loss": 3.4780771376703745,
      "objective/train/tokens_used": 58142176,
      "objective/train/unlikelihood_loss": 0.0799039676785469,
      "theoretical_loss": 3.4780771376703745,
      "tokens_seen": 1699282944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934343434343434,
      "loss": 1.2993,
      "theoretical_loss": 3.4780771376703745,
      "tokens_seen": 1699282944
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004934141414141414,
      "loss": 1.4704,
      "theoretical_loss": 3.4780655921941435,
      "tokens_seen": 1699348480
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933939393939394,
      "loss": 1.3564,
      "theoretical_loss": 3.4780540472878254,
      "tokens_seen": 1699414016
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933737373737373,
      "loss": 1.487,
      "theoretical_loss": 3.47804250295137,
      "tokens_seen": 1699479552
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933535353535354,
      "loss": 1.331,
      "theoretical_loss": 3.478030959184728,
      "tokens_seen": 1699545088
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933333333333334,
      "loss": 1.6139,
      "theoretical_loss": 3.4780194159878484,
      "tokens_seen": 1699610624
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004933131313131314,
      "loss": 1.4798,
      "theoretical_loss": 3.4780078733606814,
      "tokens_seen": 1699676160
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932929292929293,
      "loss": 1.516,
      "theoretical_loss": 3.477996331303177,
      "tokens_seen": 1699741696
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932727272727273,
      "loss": 1.5999,
      "theoretical_loss": 3.4779847898152854,
      "tokens_seen": 1699807232
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932525252525253,
      "loss": 1.4377,
      "theoretical_loss": 3.477973248896956,
      "tokens_seen": 1699872768
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932323232323232,
      "loss": 1.4555,
      "theoretical_loss": 3.4779617085481385,
      "tokens_seen": 1699938304
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004932121212121212,
      "loss": 1.2517,
      "theoretical_loss": 3.4779501687687837,
      "tokens_seen": 1700003840
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931919191919192,
      "loss": 1.5707,
      "theoretical_loss": 3.477938629558841,
      "tokens_seen": 1700069376
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931717171717171,
      "loss": 1.4201,
      "theoretical_loss": 3.4779270909182607,
      "tokens_seen": 1700134912
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931515151515151,
      "loss": 1.539,
      "theoretical_loss": 3.4779155528469925,
      "tokens_seen": 1700200448
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931313131313131,
      "loss": 1.4055,
      "theoretical_loss": 3.477904015344986,
      "tokens_seen": 1700265984
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004931111111111112,
      "loss": 1.628,
      "theoretical_loss": 3.4778924784121914,
      "tokens_seen": 1700331520
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930909090909091,
      "loss": 1.4366,
      "theoretical_loss": 3.47788094204856,
      "tokens_seen": 1700397056
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930707070707071,
      "loss": 1.5349,
      "theoretical_loss": 3.4778694062540394,
      "tokens_seen": 1700462592
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930505050505051,
      "loss": 1.335,
      "theoretical_loss": 3.4778578710285815,
      "tokens_seen": 1700528128
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004930303030303031,
      "loss": 1.4479,
      "theoretical_loss": 3.477846336372136,
      "tokens_seen": 1700593664
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000493010101010101,
      "loss": 1.6763,
      "theoretical_loss": 3.4778348022846517,
      "tokens_seen": 1700659200
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000492989898989899,
      "loss": 1.5868,
      "theoretical_loss": 3.47782326876608,
      "tokens_seen": 1700724736
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929696969696969,
      "loss": 1.4506,
      "theoretical_loss": 3.4778117358163705,
      "tokens_seen": 1700790272
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.000492949494949495,
      "loss": 1.5949,
      "theoretical_loss": 3.4778002034354722,
      "tokens_seen": 1700855808
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.4352394938468933,
      "objective/train/avg_token_score": 0.022302182391285896,
      "objective/train/docs_used": 963811,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3806675672531128,
      "objective/train/original_loss": 3.8218584060668945,
      "objective/train/theoretical_loss": 3.477788671623337,
      "objective/train/tokens_used": 59780576,
      "objective/train/unlikelihood_loss": 0.0842813178896904,
      "theoretical_loss": 3.477788671623337,
      "tokens_seen": 1700921344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929292929292929,
      "loss": 1.6276,
      "theoretical_loss": 3.477788671623337,
      "tokens_seen": 1700921344
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004929090909090909,
      "loss": 1.3813,
      "theoretical_loss": 3.4777771403799136,
      "tokens_seen": 1700986880
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928888888888888,
      "loss": 1.6226,
      "theoretical_loss": 3.4777656097051524,
      "tokens_seen": 1701052416
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928686868686869,
      "loss": 1.6444,
      "theoretical_loss": 3.4777540795990034,
      "tokens_seen": 1701117952
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928484848484849,
      "loss": 1.3509,
      "theoretical_loss": 3.477742550061417,
      "tokens_seen": 1701183488
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928282828282829,
      "loss": 1.2636,
      "theoretical_loss": 3.4777310210923424,
      "tokens_seen": 1701249024
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004928080808080808,
      "loss": 1.486,
      "theoretical_loss": 3.4777194926917305,
      "tokens_seen": 1701314560
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927878787878788,
      "loss": 1.339,
      "theoretical_loss": 3.4777079648595315,
      "tokens_seen": 1701380096
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927676767676768,
      "loss": 1.5191,
      "theoretical_loss": 3.4776964375956947,
      "tokens_seen": 1701445632
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927474747474748,
      "loss": 1.4387,
      "theoretical_loss": 3.477684910900171,
      "tokens_seen": 1701511168
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927272727272727,
      "loss": 1.3134,
      "theoretical_loss": 3.4776733847729098,
      "tokens_seen": 1701576704
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004927070707070707,
      "loss": 1.4071,
      "theoretical_loss": 3.4776618592138613,
      "tokens_seen": 1701642240
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926868686868686,
      "loss": 1.3664,
      "theoretical_loss": 3.477650334222976,
      "tokens_seen": 1701707776
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926666666666666,
      "loss": 1.5606,
      "theoretical_loss": 3.477638809800204,
      "tokens_seen": 1701773312
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926464646464647,
      "loss": 1.4081,
      "theoretical_loss": 3.477627285945495,
      "tokens_seen": 1701838848
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926262626262627,
      "loss": 1.5566,
      "theoretical_loss": 3.4776157626587993,
      "tokens_seen": 1701904384
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004926060606060606,
      "loss": 1.3377,
      "theoretical_loss": 3.4776042399400673,
      "tokens_seen": 1701969920
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925858585858586,
      "loss": 1.1917,
      "theoretical_loss": 3.477592717789248,
      "tokens_seen": 1702035456
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925656565656566,
      "loss": 1.3057,
      "theoretical_loss": 3.4775811962062937,
      "tokens_seen": 1702100992
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925454545454546,
      "loss": 1.4049,
      "theoretical_loss": 3.4775696751911527,
      "tokens_seen": 1702166528
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925252525252525,
      "loss": 1.5844,
      "theoretical_loss": 3.4775581547437757,
      "tokens_seen": 1702232064
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004925050505050505,
      "loss": 1.5742,
      "theoretical_loss": 3.477546634864113,
      "tokens_seen": 1702297600
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924848484848485,
      "loss": 1.6743,
      "theoretical_loss": 3.477535115552115,
      "tokens_seen": 1702363136
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924646464646464,
      "loss": 1.7692,
      "theoretical_loss": 3.4775235968077314,
      "tokens_seen": 1702428672
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0004924444444444444,
      "loss": 1.4189,
      "theoretical_loss": 3.4775120786309124,
      "tokens_seen": 1702494208
    },
    {
      "epoch": 0.02,
      "objective/train/aligned_token_count": 0.4039589464664459,
      "objective/train/avg_token_score": 0.022879842668771744,
      "objective/train/docs_used": 965136,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2613023519515991,
      "objective/train/original_loss": 4.1900811195373535,
      "objective/train/theoretical_loss": 3.4775005610216088,
      "objective/train/tokens_used": 61418976,
      "objective/train/unlikelihood_loss": 0.07926816493272781,
      "theoretical_loss": 3.4775005610216088,
      "tokens_seen": 1702559744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004924242424242425,
      "loss": 1.3197,
      "theoretical_loss": 3.4775005610216088,
      "tokens_seen": 1702559744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004924040404040405,
      "loss": 1.4524,
      "theoretical_loss": 3.4774890439797694,
      "tokens_seen": 1702625280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923838383838384,
      "loss": 1.4616,
      "theoretical_loss": 3.4774775275053464,
      "tokens_seen": 1702690816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923636363636364,
      "loss": 1.4173,
      "theoretical_loss": 3.4774660115982883,
      "tokens_seen": 1702756352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923434343434344,
      "loss": 1.4959,
      "theoretical_loss": 3.477454496258546,
      "tokens_seen": 1702821888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923232323232323,
      "loss": 1.4659,
      "theoretical_loss": 3.47744298148607,
      "tokens_seen": 1702887424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004923030303030303,
      "loss": 1.4347,
      "theoretical_loss": 3.4774314672808098,
      "tokens_seen": 1702952960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922828282828283,
      "loss": 1.4188,
      "theoretical_loss": 3.477419953642716,
      "tokens_seen": 1703018496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922626262626262,
      "loss": 1.4657,
      "theoretical_loss": 3.477408440571739,
      "tokens_seen": 1703084032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922424242424242,
      "loss": 1.2297,
      "theoretical_loss": 3.4773969280678294,
      "tokens_seen": 1703149568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922222222222222,
      "loss": 1.2242,
      "theoretical_loss": 3.4773854161309363,
      "tokens_seen": 1703215104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004922020202020202,
      "loss": 1.3182,
      "theoretical_loss": 3.477373904761011,
      "tokens_seen": 1703280640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921818181818182,
      "loss": 1.4721,
      "theoretical_loss": 3.4773623939580034,
      "tokens_seen": 1703346176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921616161616162,
      "loss": 1.7844,
      "theoretical_loss": 3.477350883721863,
      "tokens_seen": 1703411712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921414141414142,
      "loss": 1.3457,
      "theoretical_loss": 3.477339374052542,
      "tokens_seen": 1703477248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921212121212122,
      "loss": 1.3502,
      "theoretical_loss": 3.477327864949989,
      "tokens_seen": 1703542784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004921010101010101,
      "loss": 1.6302,
      "theoretical_loss": 3.4773163564141547,
      "tokens_seen": 1703608320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920808080808081,
      "loss": 1.3801,
      "theoretical_loss": 3.4773048484449895,
      "tokens_seen": 1703673856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920606060606061,
      "loss": 1.4358,
      "theoretical_loss": 3.4772933410424436,
      "tokens_seen": 1703739392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004920404040404041,
      "loss": 1.7846,
      "theoretical_loss": 3.477281834206468,
      "tokens_seen": 1703804928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000492020202020202,
      "loss": 1.4055,
      "theoretical_loss": 3.4772703279370116,
      "tokens_seen": 1703870464
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000492,
      "loss": 1.5334,
      "theoretical_loss": 3.477258822234026,
      "tokens_seen": 1703936000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004919797979797979,
      "loss": 1.5056,
      "theoretical_loss": 3.4772473170974614,
      "tokens_seen": 1704001536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491959595959596,
      "loss": 1.1744,
      "theoretical_loss": 3.477235812527267,
      "tokens_seen": 1704067072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491939393939394,
      "loss": 1.5387,
      "theoretical_loss": 3.4772243085233945,
      "tokens_seen": 1704132608
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.3712732195854187,
      "objective/train/avg_token_score": 0.018682070076465607,
      "objective/train/docs_used": 965825,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2413780689239502,
      "objective/train/original_loss": 4.305376052856445,
      "objective/train/theoretical_loss": 3.4772128050857933,
      "objective/train/tokens_used": 63057376,
      "objective/train/unlikelihood_loss": 0.054906729608774185,
      "theoretical_loss": 3.4772128050857933,
      "tokens_seen": 1704198144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491919191919192,
      "loss": 1.4768,
      "theoretical_loss": 3.4772128050857933,
      "tokens_seen": 1704198144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918989898989899,
      "loss": 1.5636,
      "theoretical_loss": 3.4772013022144144,
      "tokens_seen": 1704263680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918787878787879,
      "loss": 1.3053,
      "theoretical_loss": 3.4771897999092083,
      "tokens_seen": 1704329216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918585858585859,
      "loss": 1.5969,
      "theoretical_loss": 3.4771782981701245,
      "tokens_seen": 1704394752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918383838383839,
      "loss": 1.5039,
      "theoretical_loss": 3.4771667969971136,
      "tokens_seen": 1704460288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004918181818181818,
      "loss": 1.2291,
      "theoretical_loss": 3.4771552963901264,
      "tokens_seen": 1704525824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917979797979798,
      "loss": 1.3308,
      "theoretical_loss": 3.477143796349113,
      "tokens_seen": 1704591360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917777777777777,
      "loss": 1.4503,
      "theoretical_loss": 3.477132296874024,
      "tokens_seen": 1704656896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917575757575758,
      "loss": 1.4605,
      "theoretical_loss": 3.47712079796481,
      "tokens_seen": 1704722432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917373737373737,
      "loss": 1.553,
      "theoretical_loss": 3.4771092996214206,
      "tokens_seen": 1704787968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004917171717171718,
      "loss": 1.5346,
      "theoretical_loss": 3.4770978018438066,
      "tokens_seen": 1704853504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916969696969697,
      "loss": 1.5756,
      "theoretical_loss": 3.4770863046319187,
      "tokens_seen": 1704919040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916767676767677,
      "loss": 1.4356,
      "theoretical_loss": 3.4770748079857072,
      "tokens_seen": 1704984576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916565656565657,
      "loss": 1.441,
      "theoretical_loss": 3.4770633119051224,
      "tokens_seen": 1705050112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916363636363637,
      "loss": 1.3771,
      "theoretical_loss": 3.4770518163901145,
      "tokens_seen": 1705115648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004916161616161616,
      "loss": 1.4827,
      "theoretical_loss": 3.4770403214406347,
      "tokens_seen": 1705181184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915959595959596,
      "loss": 1.2737,
      "theoretical_loss": 3.4770288270566327,
      "tokens_seen": 1705246720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915757575757576,
      "loss": 1.6622,
      "theoretical_loss": 3.477017333238059,
      "tokens_seen": 1705312256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915555555555556,
      "loss": 1.3899,
      "theoretical_loss": 3.4770058399848645,
      "tokens_seen": 1705377792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915353535353535,
      "loss": 1.4385,
      "theoretical_loss": 3.476994347296999,
      "tokens_seen": 1705443328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004915151515151515,
      "loss": 1.3491,
      "theoretical_loss": 3.476982855174414,
      "tokens_seen": 1705508864
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914949494949494,
      "loss": 1.4771,
      "theoretical_loss": 3.476971363617059,
      "tokens_seen": 1705574400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914747474747475,
      "loss": 1.3391,
      "theoretical_loss": 3.476959872624885,
      "tokens_seen": 1705639936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914545454545455,
      "loss": 1.4106,
      "theoretical_loss": 3.4769483821978424,
      "tokens_seen": 1705705472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914343434343435,
      "loss": 1.2723,
      "theoretical_loss": 3.4769368923358814,
      "tokens_seen": 1705771008
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.470796674489975,
      "objective/train/avg_token_score": 0.008710848167538643,
      "objective/train/docs_used": 966739,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.535874843597412,
      "objective/train/original_loss": 3.7179017066955566,
      "objective/train/theoretical_loss": 3.476925403038953,
      "objective/train/tokens_used": 64695776,
      "objective/train/unlikelihood_loss": 0.08725922554731369,
      "theoretical_loss": 3.476925403038953,
      "tokens_seen": 1705836544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004914141414141414,
      "loss": 1.7468,
      "theoretical_loss": 3.476925403038953,
      "tokens_seen": 1705836544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913939393939394,
      "loss": 1.2985,
      "theoretical_loss": 3.4769139143070076,
      "tokens_seen": 1705902080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913737373737374,
      "loss": 1.3906,
      "theoretical_loss": 3.476902426139995,
      "tokens_seen": 1705967616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913535353535354,
      "loss": 1.5775,
      "theoretical_loss": 3.4768909385378666,
      "tokens_seen": 1706033152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913333333333333,
      "loss": 1.3863,
      "theoretical_loss": 3.4768794515005723,
      "tokens_seen": 1706098688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004913131313131313,
      "loss": 1.3093,
      "theoretical_loss": 3.476867965028063,
      "tokens_seen": 1706164224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912929292929293,
      "loss": 1.4804,
      "theoretical_loss": 3.47685647912029,
      "tokens_seen": 1706229760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912727272727272,
      "loss": 1.4999,
      "theoretical_loss": 3.476844993777202,
      "tokens_seen": 1706295296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912525252525252,
      "loss": 1.4509,
      "theoretical_loss": 3.476833508998751,
      "tokens_seen": 1706360832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912323232323233,
      "loss": 1.3798,
      "theoretical_loss": 3.4768220247848873,
      "tokens_seen": 1706426368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004912121212121213,
      "loss": 1.6698,
      "theoretical_loss": 3.4768105411355608,
      "tokens_seen": 1706491904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911919191919192,
      "loss": 1.5185,
      "theoretical_loss": 3.476799058050723,
      "tokens_seen": 1706557440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911717171717172,
      "loss": 1.1568,
      "theoretical_loss": 3.4767875755303237,
      "tokens_seen": 1706622976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911515151515152,
      "loss": 1.0498,
      "theoretical_loss": 3.476776093574314,
      "tokens_seen": 1706688512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911313131313131,
      "loss": 1.3549,
      "theoretical_loss": 3.4767646121826443,
      "tokens_seen": 1706754048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004911111111111111,
      "loss": 1.4073,
      "theoretical_loss": 3.4767531313552658,
      "tokens_seen": 1706819584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910909090909091,
      "loss": 1.3959,
      "theoretical_loss": 3.4767416510921274,
      "tokens_seen": 1706885120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491070707070707,
      "loss": 1.2292,
      "theoretical_loss": 3.4767301713931813,
      "tokens_seen": 1706950656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491050505050505,
      "loss": 1.4875,
      "theoretical_loss": 3.4767186922583777,
      "tokens_seen": 1707016192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000491030303030303,
      "loss": 1.3406,
      "theoretical_loss": 3.4767072136876673,
      "tokens_seen": 1707081728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004910101010101011,
      "loss": 1.3963,
      "theoretical_loss": 3.4766957356810004,
      "tokens_seen": 1707147264
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490989898989899,
      "loss": 1.3853,
      "theoretical_loss": 3.4766842582383277,
      "tokens_seen": 1707212800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490969696969697,
      "loss": 1.2996,
      "theoretical_loss": 3.4766727813596,
      "tokens_seen": 1707278336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490949494949495,
      "loss": 1.6414,
      "theoretical_loss": 3.476661305044768,
      "tokens_seen": 1707343872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490929292929293,
      "loss": 1.3811,
      "theoretical_loss": 3.476649829293782,
      "tokens_seen": 1707409408
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.4374389052391052,
      "objective/train/avg_token_score": 0.005900775548070669,
      "objective/train/docs_used": 967356,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5339224338531494,
      "objective/train/original_loss": 4.014285087585449,
      "objective/train/theoretical_loss": 3.4766383541065933,
      "objective/train/tokens_used": 66334176,
      "objective/train/unlikelihood_loss": 0.08221994340419769,
      "theoretical_loss": 3.4766383541065933,
      "tokens_seen": 1707474944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004909090909090909,
      "loss": 1.4133,
      "theoretical_loss": 3.4766383541065933,
      "tokens_seen": 1707474944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908888888888889,
      "loss": 1.2069,
      "theoretical_loss": 3.4766268794831516,
      "tokens_seen": 1707540480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908686868686868,
      "loss": 1.2704,
      "theoretical_loss": 3.4766154054234084,
      "tokens_seen": 1707606016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908484848484849,
      "loss": 1.5368,
      "theoretical_loss": 3.4766039319273143,
      "tokens_seen": 1707671552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908282828282828,
      "loss": 1.2068,
      "theoretical_loss": 3.4765924589948196,
      "tokens_seen": 1707737088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004908080808080808,
      "loss": 1.3801,
      "theoretical_loss": 3.4765809866258754,
      "tokens_seen": 1707802624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907878787878787,
      "loss": 1.5021,
      "theoretical_loss": 3.476569514820432,
      "tokens_seen": 1707868160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907676767676768,
      "loss": 1.599,
      "theoretical_loss": 3.4765580435784402,
      "tokens_seen": 1707933696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907474747474748,
      "loss": 1.6612,
      "theoretical_loss": 3.4765465728998506,
      "tokens_seen": 1707999232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907272727272728,
      "loss": 1.7327,
      "theoretical_loss": 3.476535102784614,
      "tokens_seen": 1708064768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004907070707070707,
      "loss": 1.3258,
      "theoretical_loss": 3.4765236332326817,
      "tokens_seen": 1708130304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906868686868687,
      "loss": 1.2935,
      "theoretical_loss": 3.4765121642440038,
      "tokens_seen": 1708195840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906666666666667,
      "loss": 1.477,
      "theoretical_loss": 3.476500695818531,
      "tokens_seen": 1708261376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906464646464647,
      "loss": 1.5143,
      "theoretical_loss": 3.476489227956214,
      "tokens_seen": 1708326912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906262626262626,
      "loss": 1.5016,
      "theoretical_loss": 3.476477760657004,
      "tokens_seen": 1708392448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004906060606060606,
      "loss": 1.3412,
      "theoretical_loss": 3.4764662939208515,
      "tokens_seen": 1708457984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905858585858585,
      "loss": 1.3278,
      "theoretical_loss": 3.476454827747707,
      "tokens_seen": 1708523520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905656565656565,
      "loss": 1.4474,
      "theoretical_loss": 3.476443362137522,
      "tokens_seen": 1708589056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905454545454546,
      "loss": 1.3127,
      "theoretical_loss": 3.4764318970902464,
      "tokens_seen": 1708654592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905252525252526,
      "loss": 1.5852,
      "theoretical_loss": 3.4764204326058317,
      "tokens_seen": 1708720128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004905050505050505,
      "loss": 1.328,
      "theoretical_loss": 3.476408968684228,
      "tokens_seen": 1708785664
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904848484848485,
      "loss": 1.5521,
      "theoretical_loss": 3.4763975053253864,
      "tokens_seen": 1708851200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904646464646465,
      "loss": 1.4926,
      "theoretical_loss": 3.4763860425292576,
      "tokens_seen": 1708916736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904444444444445,
      "loss": 1.3039,
      "theoretical_loss": 3.4763745802957926,
      "tokens_seen": 1708982272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904242424242424,
      "loss": 1.5436,
      "theoretical_loss": 3.476363118624943,
      "tokens_seen": 1709047808
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.4849095940589905,
      "objective/train/avg_token_score": 0.0028560131322592497,
      "objective/train/docs_used": 968332,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6418017148971558,
      "objective/train/original_loss": 3.7804646492004395,
      "objective/train/theoretical_loss": 3.4763516575166573,
      "objective/train/tokens_used": 67972576,
      "objective/train/unlikelihood_loss": 0.1123645082116127,
      "theoretical_loss": 3.4763516575166573,
      "tokens_seen": 1709113344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004904040404040404,
      "loss": 1.516,
      "theoretical_loss": 3.4763516575166573,
      "tokens_seen": 1709113344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903838383838384,
      "loss": 1.3689,
      "theoretical_loss": 3.476340196970889,
      "tokens_seen": 1709178880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903636363636364,
      "loss": 1.6257,
      "theoretical_loss": 3.476328736987587,
      "tokens_seen": 1709244416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903434343434343,
      "loss": 1.3624,
      "theoretical_loss": 3.476317277566703,
      "tokens_seen": 1709309952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903232323232323,
      "loss": 1.4273,
      "theoretical_loss": 3.4763058187081874,
      "tokens_seen": 1709375488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004903030303030304,
      "loss": 1.3265,
      "theoretical_loss": 3.476294360411991,
      "tokens_seen": 1709441024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902828282828283,
      "loss": 1.6407,
      "theoretical_loss": 3.4762829026780655,
      "tokens_seen": 1709506560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902626262626263,
      "loss": 1.5641,
      "theoretical_loss": 3.4762714455063612,
      "tokens_seen": 1709572096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902424242424243,
      "loss": 1.4669,
      "theoretical_loss": 3.4762599888968286,
      "tokens_seen": 1709637632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902222222222222,
      "loss": 1.4743,
      "theoretical_loss": 3.476248532849419,
      "tokens_seen": 1709703168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004902020202020202,
      "loss": 1.1262,
      "theoretical_loss": 3.4762370773640834,
      "tokens_seen": 1709768704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901818181818182,
      "loss": 1.4312,
      "theoretical_loss": 3.476225622440772,
      "tokens_seen": 1709834240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901616161616162,
      "loss": 1.3713,
      "theoretical_loss": 3.4762141680794367,
      "tokens_seen": 1709899776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901414141414141,
      "loss": 1.4197,
      "theoretical_loss": 3.476202714280028,
      "tokens_seen": 1709965312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901212121212121,
      "loss": 1.3616,
      "theoretical_loss": 3.476191261042496,
      "tokens_seen": 1710030848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004901010101010101,
      "loss": 1.3839,
      "theoretical_loss": 3.4761798083667923,
      "tokens_seen": 1710096384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000490080808080808,
      "loss": 1.6839,
      "theoretical_loss": 3.476168356252868,
      "tokens_seen": 1710161920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900606060606061,
      "loss": 1.3635,
      "theoretical_loss": 3.476156904700674,
      "tokens_seen": 1710227456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900404040404041,
      "loss": 1.6225,
      "theoretical_loss": 3.476145453710161,
      "tokens_seen": 1710292992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004900202020202021,
      "loss": 1.3887,
      "theoretical_loss": 3.4761340032812793,
      "tokens_seen": 1710358528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00049,
      "loss": 1.4664,
      "theoretical_loss": 3.4761225534139806,
      "tokens_seen": 1710424064
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489979797979798,
      "loss": 1.388,
      "theoretical_loss": 3.476111104108216,
      "tokens_seen": 1710489600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489959595959596,
      "loss": 1.4894,
      "theoretical_loss": 3.476099655363936,
      "tokens_seen": 1710555136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899393939393939,
      "loss": 1.2687,
      "theoretical_loss": 3.476088207181092,
      "tokens_seen": 1710620672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004899191919191919,
      "loss": 1.298,
      "theoretical_loss": 3.4760767595596347,
      "tokens_seen": 1710686208
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.5306084752082825,
      "objective/train/avg_token_score": 0.01167221274226904,
      "objective/train/docs_used": 968886,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6835694313049316,
      "objective/train/original_loss": 4.055310249328613,
      "objective/train/theoretical_loss": 3.4760653124995144,
      "objective/train/tokens_used": 69610976,
      "objective/train/unlikelihood_loss": 0.05426410958170891,
      "theoretical_loss": 3.4760653124995144,
      "tokens_seen": 1710751744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898989898989899,
      "loss": 1.7209,
      "theoretical_loss": 3.4760653124995144,
      "tokens_seen": 1710751744
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898787878787878,
      "loss": 1.2871,
      "theoretical_loss": 3.4760538660006834,
      "tokens_seen": 1710817280
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898585858585858,
      "loss": 1.4492,
      "theoretical_loss": 3.4760424200630915,
      "tokens_seen": 1710882816
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898383838383839,
      "loss": 1.4377,
      "theoretical_loss": 3.4760309746866906,
      "tokens_seen": 1710948352
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004898181818181819,
      "loss": 1.4232,
      "theoretical_loss": 3.476019529871431,
      "tokens_seen": 1711013888
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897979797979798,
      "loss": 1.6333,
      "theoretical_loss": 3.4760080856172637,
      "tokens_seen": 1711079424
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897777777777778,
      "loss": 1.4597,
      "theoretical_loss": 3.475996641924141,
      "tokens_seen": 1711144960
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897575757575758,
      "loss": 1.5103,
      "theoretical_loss": 3.475985198792012,
      "tokens_seen": 1711210496
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897373737373738,
      "loss": 1.3766,
      "theoretical_loss": 3.4759737562208284,
      "tokens_seen": 1711276032
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004897171717171717,
      "loss": 1.532,
      "theoretical_loss": 3.4759623142105416,
      "tokens_seen": 1711341568
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896969696969697,
      "loss": 1.4063,
      "theoretical_loss": 3.4759508727611026,
      "tokens_seen": 1711407104
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896767676767676,
      "loss": 1.3932,
      "theoretical_loss": 3.475939431872462,
      "tokens_seen": 1711472640
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896565656565657,
      "loss": 1.4003,
      "theoretical_loss": 3.4759279915445713,
      "tokens_seen": 1711538176
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896363636363636,
      "loss": 1.335,
      "theoretical_loss": 3.4759165517773813,
      "tokens_seen": 1711603712
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004896161616161617,
      "loss": 1.1782,
      "theoretical_loss": 3.4759051125708433,
      "tokens_seen": 1711669248
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895959595959596,
      "loss": 1.2264,
      "theoretical_loss": 3.475893673924908,
      "tokens_seen": 1711734784
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895757575757576,
      "loss": 1.4453,
      "theoretical_loss": 3.475882235839527,
      "tokens_seen": 1711800320
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895555555555556,
      "loss": 1.4011,
      "theoretical_loss": 3.4758707983146504,
      "tokens_seen": 1711865856
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895353535353536,
      "loss": 1.4299,
      "theoretical_loss": 3.47585936135023,
      "tokens_seen": 1711931392
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004895151515151515,
      "loss": 1.5037,
      "theoretical_loss": 3.475847924946217,
      "tokens_seen": 1711996928
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894949494949495,
      "loss": 1.5703,
      "theoretical_loss": 3.475836489102562,
      "tokens_seen": 1712062464
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894747474747475,
      "loss": 1.4185,
      "theoretical_loss": 3.4758250538192166,
      "tokens_seen": 1712128000
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894545454545455,
      "loss": 1.6266,
      "theoretical_loss": 3.4758136190961313,
      "tokens_seen": 1712193536
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894343434343434,
      "loss": 1.6168,
      "theoretical_loss": 3.475802184933258,
      "tokens_seen": 1712259072
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004894141414141414,
      "loss": 1.2173,
      "theoretical_loss": 3.475790751330547,
      "tokens_seen": 1712324608
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.4318181872367859,
      "objective/train/avg_token_score": 0.056178782135248184,
      "objective/train/docs_used": 970327,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3246653079986572,
      "objective/train/original_loss": 3.9383597373962402,
      "objective/train/theoretical_loss": 3.47577931828795,
      "objective/train/tokens_used": 71249376,
      "objective/train/unlikelihood_loss": 0.07031881809234619,
      "theoretical_loss": 3.47577931828795,
      "tokens_seen": 1712390144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893939393939393,
      "loss": 1.4182,
      "theoretical_loss": 3.47577931828795,
      "tokens_seen": 1712390144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893737373737374,
      "loss": 1.1773,
      "theoretical_loss": 3.4757678858054177,
      "tokens_seen": 1712455680
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893535353535354,
      "loss": 1.624,
      "theoretical_loss": 3.4757564538829016,
      "tokens_seen": 1712521216
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893333333333334,
      "loss": 1.4873,
      "theoretical_loss": 3.4757450225203526,
      "tokens_seen": 1712586752
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004893131313131313,
      "loss": 1.3469,
      "theoretical_loss": 3.475733591717722,
      "tokens_seen": 1712652288
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892929292929293,
      "loss": 1.4407,
      "theoretical_loss": 3.4757221614749607,
      "tokens_seen": 1712717824
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892727272727273,
      "loss": 1.4212,
      "theoretical_loss": 3.47571073179202,
      "tokens_seen": 1712783360
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892525252525253,
      "loss": 1.3738,
      "theoretical_loss": 3.4756993026688514,
      "tokens_seen": 1712848896
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892323232323232,
      "loss": 1.3697,
      "theoretical_loss": 3.475687874105406,
      "tokens_seen": 1712914432
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004892121212121212,
      "loss": 1.5092,
      "theoretical_loss": 3.4756764461016347,
      "tokens_seen": 1712979968
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891919191919192,
      "loss": 1.2167,
      "theoretical_loss": 3.4756650186574882,
      "tokens_seen": 1713045504
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891717171717171,
      "loss": 1.4599,
      "theoretical_loss": 3.475653591772918,
      "tokens_seen": 1713111040
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891515151515151,
      "loss": 1.61,
      "theoretical_loss": 3.4756421654478764,
      "tokens_seen": 1713176576
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891313131313132,
      "loss": 1.679,
      "theoretical_loss": 3.475630739682313,
      "tokens_seen": 1713242112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004891111111111112,
      "loss": 1.3736,
      "theoretical_loss": 3.47561931447618,
      "tokens_seen": 1713307648
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890909090909091,
      "loss": 1.2454,
      "theoretical_loss": 3.475607889829428,
      "tokens_seen": 1713373184
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890707070707071,
      "loss": 1.5084,
      "theoretical_loss": 3.475596465742009,
      "tokens_seen": 1713438720
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004890505050505051,
      "loss": 1.5201,
      "theoretical_loss": 3.475585042213874,
      "tokens_seen": 1713504256
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489030303030303,
      "loss": 1.6822,
      "theoretical_loss": 3.4755736192449733,
      "tokens_seen": 1713569792
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000489010101010101,
      "loss": 1.4501,
      "theoretical_loss": 3.4755621968352592,
      "tokens_seen": 1713635328
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488989898989899,
      "loss": 1.4203,
      "theoretical_loss": 3.475550774984682,
      "tokens_seen": 1713700864
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889696969696969,
      "loss": 1.4855,
      "theoretical_loss": 3.4755393536931942,
      "tokens_seen": 1713766400
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889494949494949,
      "loss": 1.3609,
      "theoretical_loss": 3.4755279329607465,
      "tokens_seen": 1713831936
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004889292929292929,
      "loss": 1.5888,
      "theoretical_loss": 3.4755165127872893,
      "tokens_seen": 1713897472
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488909090909091,
      "loss": 1.5043,
      "theoretical_loss": 3.4755050931727753,
      "tokens_seen": 1713963008
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.5464931726455688,
      "objective/train/avg_token_score": 0.0072470419108867645,
      "objective/train/docs_used": 971016,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7423732280731201,
      "objective/train/original_loss": 3.5915262699127197,
      "objective/train/theoretical_loss": 3.4754936741171543,
      "objective/train/tokens_used": 72887776,
      "objective/train/unlikelihood_loss": 0.07072310149669647,
      "theoretical_loss": 3.4754936741171543,
      "tokens_seen": 1714028544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888888888888889,
      "loss": 1.4365,
      "theoretical_loss": 3.4754936741171543,
      "tokens_seen": 1714028544
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888686868686869,
      "loss": 1.434,
      "theoretical_loss": 3.475482255620379,
      "tokens_seen": 1714094080
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888484848484849,
      "loss": 1.1613,
      "theoretical_loss": 3.4754708376823995,
      "tokens_seen": 1714159616
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888282828282829,
      "loss": 1.2823,
      "theoretical_loss": 3.475459420303168,
      "tokens_seen": 1714225152
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004888080808080808,
      "loss": 1.2084,
      "theoretical_loss": 3.4754480034826356,
      "tokens_seen": 1714290688
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887878787878788,
      "loss": 1.1509,
      "theoretical_loss": 3.475436587220753,
      "tokens_seen": 1714356224
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887676767676768,
      "loss": 1.4293,
      "theoretical_loss": 3.4754251715174727,
      "tokens_seen": 1714421760
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887474747474748,
      "loss": 1.4948,
      "theoretical_loss": 3.4754137563727445,
      "tokens_seen": 1714487296
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887272727272727,
      "loss": 1.3427,
      "theoretical_loss": 3.4754023417865207,
      "tokens_seen": 1714552832
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004887070707070707,
      "loss": 1.393,
      "theoretical_loss": 3.4753909277587525,
      "tokens_seen": 1714618368
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886868686868686,
      "loss": 1.4489,
      "theoretical_loss": 3.475379514289391,
      "tokens_seen": 1714683904
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886666666666667,
      "loss": 1.3675,
      "theoretical_loss": 3.475368101378388,
      "tokens_seen": 1714749440
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886464646464647,
      "loss": 1.2169,
      "theoretical_loss": 3.475356689025694,
      "tokens_seen": 1714814976
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886262626262627,
      "loss": 1.5594,
      "theoretical_loss": 3.4753452772312614,
      "tokens_seen": 1714880512
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004886060606060606,
      "loss": 1.3374,
      "theoretical_loss": 3.475333865995041,
      "tokens_seen": 1714946048
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885858585858586,
      "loss": 1.3952,
      "theoretical_loss": 3.4753224553169835,
      "tokens_seen": 1715011584
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885656565656566,
      "loss": 1.4714,
      "theoretical_loss": 3.4753110451970417,
      "tokens_seen": 1715077120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885454545454546,
      "loss": 1.2954,
      "theoretical_loss": 3.475299635635166,
      "tokens_seen": 1715142656
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885252525252525,
      "loss": 1.5126,
      "theoretical_loss": 3.475288226631308,
      "tokens_seen": 1715208192
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004885050505050505,
      "loss": 1.307,
      "theoretical_loss": 3.475276818185419,
      "tokens_seen": 1715273728
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884848484848484,
      "loss": 1.3257,
      "theoretical_loss": 3.475265410297451,
      "tokens_seen": 1715339264
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884646464646465,
      "loss": 1.3537,
      "theoretical_loss": 3.4752540029673544,
      "tokens_seen": 1715404800
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884444444444445,
      "loss": 1.3866,
      "theoretical_loss": 3.4752425961950815,
      "tokens_seen": 1715470336
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884242424242425,
      "loss": 1.4019,
      "theoretical_loss": 3.4752311899805832,
      "tokens_seen": 1715535872
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004884040404040404,
      "loss": 1.1464,
      "theoretical_loss": 3.475219784323811,
      "tokens_seen": 1715601408
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.4002932608127594,
      "objective/train/avg_token_score": 0.017858585342764854,
      "objective/train/docs_used": 972281,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3489974737167358,
      "objective/train/original_loss": 4.905088424682617,
      "objective/train/theoretical_loss": 3.475208379224716,
      "objective/train/tokens_used": 74526176,
      "objective/train/unlikelihood_loss": 0.05417528748512268,
      "theoretical_loss": 3.475208379224716,
      "tokens_seen": 1715666944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883838383838384,
      "loss": 1.526,
      "theoretical_loss": 3.475208379224716,
      "tokens_seen": 1715666944
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883636363636364,
      "loss": 1.2685,
      "theoretical_loss": 3.475196974683251,
      "tokens_seen": 1715732480
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883434343434344,
      "loss": 1.4254,
      "theoretical_loss": 3.4751855706993657,
      "tokens_seen": 1715798016
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883232323232323,
      "loss": 1.3064,
      "theoretical_loss": 3.4751741672730123,
      "tokens_seen": 1715863552
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004883030303030303,
      "loss": 1.3617,
      "theoretical_loss": 3.4751627644041427,
      "tokens_seen": 1715929088
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882828282828283,
      "loss": 1.4178,
      "theoretical_loss": 3.4751513620927073,
      "tokens_seen": 1715994624
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882626262626263,
      "loss": 1.6226,
      "theoretical_loss": 3.475139960338659,
      "tokens_seen": 1716060160
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882424242424243,
      "loss": 1.4401,
      "theoretical_loss": 3.4751285591419476,
      "tokens_seen": 1716125696
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048822222222222225,
      "loss": 1.3443,
      "theoretical_loss": 3.475117158502526,
      "tokens_seen": 1716191232
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004882020202020202,
      "loss": 1.4155,
      "theoretical_loss": 3.4751057584203444,
      "tokens_seen": 1716256768
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048818181818181816,
      "loss": 1.2397,
      "theoretical_loss": 3.4750943588953556,
      "tokens_seen": 1716322304
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048816161616161617,
      "loss": 1.2896,
      "theoretical_loss": 3.47508295992751,
      "tokens_seen": 1716387840
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881414141414142,
      "loss": 1.4234,
      "theoretical_loss": 3.47507156151676,
      "tokens_seen": 1716453376
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048812121212121214,
      "loss": 1.6226,
      "theoretical_loss": 3.4750601636630565,
      "tokens_seen": 1716518912
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004881010101010101,
      "loss": 1.3232,
      "theoretical_loss": 3.475048766366351,
      "tokens_seen": 1716584448
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048808080808080805,
      "loss": 1.2095,
      "theoretical_loss": 3.4750373696265955,
      "tokens_seen": 1716649984
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880606060606061,
      "loss": 1.3368,
      "theoretical_loss": 3.4750259734437408,
      "tokens_seen": 1716715520
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004880404040404041,
      "loss": 1.5763,
      "theoretical_loss": 3.4750145778177393,
      "tokens_seen": 1716781056
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048802020202020203,
      "loss": 1.537,
      "theoretical_loss": 3.4750031827485417,
      "tokens_seen": 1716846592
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000488,
      "loss": 1.4221,
      "theoretical_loss": 3.4749917882360997,
      "tokens_seen": 1716912128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048797979797979795,
      "loss": 1.4776,
      "theoretical_loss": 3.4749803942803656,
      "tokens_seen": 1716977664
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.000487959595959596,
      "loss": 1.6911,
      "theoretical_loss": 3.4749690008812903,
      "tokens_seen": 1717043200
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048793939393939397,
      "loss": 1.6163,
      "theoretical_loss": 3.4749576080388254,
      "tokens_seen": 1717108736
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004879191919191919,
      "loss": 1.5933,
      "theoretical_loss": 3.4749462157529227,
      "tokens_seen": 1717174272
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004878989898989899,
      "loss": 1.4458,
      "theoretical_loss": 3.4749348240235336,
      "tokens_seen": 1717239808
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.34756842255592346,
      "objective/train/avg_token_score": 0.022121969610452652,
      "objective/train/docs_used": 972817,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2492351531982422,
      "objective/train/original_loss": 4.464137554168701,
      "objective/train/theoretical_loss": 3.47492343285061,
      "objective/train/tokens_used": 76164576,
      "objective/train/unlikelihood_loss": 0.06865523010492325,
      "theoretical_loss": 3.47492343285061,
      "tokens_seen": 1717305344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048787878787878784,
      "loss": 1.4663,
      "theoretical_loss": 3.47492343285061,
      "tokens_seen": 1717305344
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004878585858585859,
      "loss": 1.6475,
      "theoretical_loss": 3.474912042234102,
      "tokens_seen": 1717370880
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048783838383838386,
      "loss": 1.3615,
      "theoretical_loss": 3.4749006521739636,
      "tokens_seen": 1717436416
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004878181818181818,
      "loss": 1.3724,
      "theoretical_loss": 3.4748892626701444,
      "tokens_seen": 1717501952
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877979797979798,
      "loss": 1.4726,
      "theoretical_loss": 3.4748778737225976,
      "tokens_seen": 1717567488
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877777777777778,
      "loss": 1.5526,
      "theoretical_loss": 3.4748664853312734,
      "tokens_seen": 1717633024
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877575757575758,
      "loss": 1.3223,
      "theoretical_loss": 3.474855097496124,
      "tokens_seen": 1717698560
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048773737373737375,
      "loss": 1.6783,
      "theoretical_loss": 3.474843710217101,
      "tokens_seen": 1717764096
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004877171717171717,
      "loss": 1.3255,
      "theoretical_loss": 3.474832323494156,
      "tokens_seen": 1717829632
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048769696969696967,
      "loss": 1.3216,
      "theoretical_loss": 3.474820937327241,
      "tokens_seen": 1717895168
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004876767676767677,
      "loss": 1.3008,
      "theoretical_loss": 3.474809551716307,
      "tokens_seen": 1717960704
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004876565656565657,
      "loss": 1.1347,
      "theoretical_loss": 3.474798166661306,
      "tokens_seen": 1718026240
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048763636363636365,
      "loss": 1.4412,
      "theoretical_loss": 3.4747867821621896,
      "tokens_seen": 1718091776
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004876161616161616,
      "loss": 1.1156,
      "theoretical_loss": 3.4747753982189096,
      "tokens_seen": 1718157312
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875959595959596,
      "loss": 1.4876,
      "theoretical_loss": 3.474764014831417,
      "tokens_seen": 1718222848
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048757575757575757,
      "loss": 1.4029,
      "theoretical_loss": 3.4747526319996647,
      "tokens_seen": 1718288384
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875555555555556,
      "loss": 1.3172,
      "theoretical_loss": 3.474741249723603,
      "tokens_seen": 1718353920
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048753535353535354,
      "loss": 1.2698,
      "theoretical_loss": 3.4747298680031844,
      "tokens_seen": 1718419456
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004875151515151515,
      "loss": 1.5606,
      "theoretical_loss": 3.4747184868383605,
      "tokens_seen": 1718484992
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874949494949495,
      "loss": 1.4975,
      "theoretical_loss": 3.474707106229083,
      "tokens_seen": 1718550528
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048747474747474746,
      "loss": 1.5241,
      "theoretical_loss": 3.4746957261753035,
      "tokens_seen": 1718616064
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004874545454545455,
      "loss": 1.5173,
      "theoretical_loss": 3.4746843466769737,
      "tokens_seen": 1718681600
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048743434343434343,
      "loss": 1.5222,
      "theoretical_loss": 3.4746729677340453,
      "tokens_seen": 1718747136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00048741414141414144,
      "loss": 1.4428,
      "theoretical_loss": 3.47466158934647,
      "tokens_seen": 1718812672
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0004873939393939394,
      "loss": 1.2955,
      "theoretical_loss": 3.4746502115141995,
      "tokens_seen": 1718878208
    },
    {
      "epoch": 0.03,
      "objective/train/aligned_token_count": 0.3179985284805298,
      "objective/train/avg_token_score": 0.02353089302778244,
      "objective/train/docs_used": 973424,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0046734809875488,
      "objective/train/original_loss": 4.215363502502441,
      "objective/train/theoretical_loss": 3.4746388342371857,
      "objective/train/tokens_used": 77802976,
      "objective/train/unlikelihood_loss": 0.07478584349155426,
      "theoretical_loss": 3.4746388342371857,
      "tokens_seen": 1718943744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048737373737373736,
      "loss": 1.2882,
      "theoretical_loss": 3.4746388342371857,
      "tokens_seen": 1718943744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048735353535353537,
      "loss": 1.5482,
      "theoretical_loss": 3.47462745751538,
      "tokens_seen": 1719009280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004873333333333333,
      "loss": 1.1838,
      "theoretical_loss": 3.474616081348735,
      "tokens_seen": 1719074816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048731313131313134,
      "loss": 1.438,
      "theoretical_loss": 3.4746047057372014,
      "tokens_seen": 1719140352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004872929292929293,
      "loss": 1.4355,
      "theoretical_loss": 3.4745933306807313,
      "tokens_seen": 1719205888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048727272727272725,
      "loss": 1.2308,
      "theoretical_loss": 3.474581956179277,
      "tokens_seen": 1719271424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048725252525252526,
      "loss": 1.5441,
      "theoretical_loss": 3.474570582232789,
      "tokens_seen": 1719336960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048723232323232327,
      "loss": 1.4567,
      "theoretical_loss": 3.4745592088412205,
      "tokens_seen": 1719402496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048721212121212123,
      "loss": 1.5781,
      "theoretical_loss": 3.4745478360045228,
      "tokens_seen": 1719468032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871919191919192,
      "loss": 1.4201,
      "theoretical_loss": 3.4745364637226466,
      "tokens_seen": 1719533568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048717171717171714,
      "loss": 1.5531,
      "theoretical_loss": 3.4745250919955453,
      "tokens_seen": 1719599104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048715151515151515,
      "loss": 1.0908,
      "theoretical_loss": 3.47451372082317,
      "tokens_seen": 1719664640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048713131313131316,
      "loss": 1.6216,
      "theoretical_loss": 3.474502350205472,
      "tokens_seen": 1719730176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004871111111111111,
      "loss": 1.2598,
      "theoretical_loss": 3.474490980142404,
      "tokens_seen": 1719795712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870909090909091,
      "loss": 1.3527,
      "theoretical_loss": 3.4744796106339173,
      "tokens_seen": 1719861248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048707070707070704,
      "loss": 1.3626,
      "theoretical_loss": 3.474468241679964,
      "tokens_seen": 1719926784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004870505050505051,
      "loss": 1.4276,
      "theoretical_loss": 3.474456873280496,
      "tokens_seen": 1719992320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048703030303030306,
      "loss": 1.5703,
      "theoretical_loss": 3.4744455054354644,
      "tokens_seen": 1720057856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000487010101010101,
      "loss": 1.4335,
      "theoretical_loss": 3.474434138144822,
      "tokens_seen": 1720123392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048698989898989897,
      "loss": 1.4018,
      "theoretical_loss": 3.47442277140852,
      "tokens_seen": 1720188928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000486969696969697,
      "loss": 1.4003,
      "theoretical_loss": 3.47441140522651,
      "tokens_seen": 1720254464
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000486949494949495,
      "loss": 1.5975,
      "theoretical_loss": 3.474400039598745,
      "tokens_seen": 1720320000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048692929292929295,
      "loss": 1.3209,
      "theoretical_loss": 3.4743886745251755,
      "tokens_seen": 1720385536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004869090909090909,
      "loss": 1.5469,
      "theoretical_loss": 3.4743773100057544,
      "tokens_seen": 1720451072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048688888888888886,
      "loss": 1.3673,
      "theoretical_loss": 3.4743659460404333,
      "tokens_seen": 1720516608
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.5211998820304871,
      "objective/train/avg_token_score": 0.0027845173608511686,
      "objective/train/docs_used": 974650,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7277867794036865,
      "objective/train/original_loss": 3.4647514820098877,
      "objective/train/theoretical_loss": 3.4743545826291635,
      "objective/train/tokens_used": 79441376,
      "objective/train/unlikelihood_loss": 0.10759128630161285,
      "theoretical_loss": 3.4743545826291635,
      "tokens_seen": 1720582144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048686868686868693,
      "loss": 1.5171,
      "theoretical_loss": 3.4743545826291635,
      "tokens_seen": 1720582144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868484848484849,
      "loss": 1.3346,
      "theoretical_loss": 3.4743432197718978,
      "tokens_seen": 1720647680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048682828282828284,
      "loss": 1.2947,
      "theoretical_loss": 3.4743318574685875,
      "tokens_seen": 1720713216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004868080808080808,
      "loss": 1.3649,
      "theoretical_loss": 3.4743204957191844,
      "tokens_seen": 1720778752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048678787878787876,
      "loss": 1.3602,
      "theoretical_loss": 3.474309134523641,
      "tokens_seen": 1720844288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867676767676768,
      "loss": 1.3379,
      "theoretical_loss": 3.4742977738819087,
      "tokens_seen": 1720909824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867474747474748,
      "loss": 1.5094,
      "theoretical_loss": 3.4742864137939398,
      "tokens_seen": 1720975360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048672727272727274,
      "loss": 1.4478,
      "theoretical_loss": 3.474275054259685,
      "tokens_seen": 1721040896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004867070707070707,
      "loss": 1.4563,
      "theoretical_loss": 3.4742636952790984,
      "tokens_seen": 1721106432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004866868686868687,
      "loss": 1.5572,
      "theoretical_loss": 3.47425233685213,
      "tokens_seen": 1721171968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004866666666666667,
      "loss": 1.3569,
      "theoretical_loss": 3.474240978978733,
      "tokens_seen": 1721237504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048664646464646467,
      "loss": 1.2837,
      "theoretical_loss": 3.474229621658859,
      "tokens_seen": 1721303040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048662626262626263,
      "loss": 1.4573,
      "theoretical_loss": 3.474218264892459,
      "tokens_seen": 1721368576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004866060606060606,
      "loss": 1.4406,
      "theoretical_loss": 3.4742069086794864,
      "tokens_seen": 1721434112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865858585858586,
      "loss": 1.2444,
      "theoretical_loss": 3.474195553019892,
      "tokens_seen": 1721499648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865656565656566,
      "loss": 1.3948,
      "theoretical_loss": 3.4741841979136288,
      "tokens_seen": 1721565184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048654545454545456,
      "loss": 1.4,
      "theoretical_loss": 3.4741728433606482,
      "tokens_seen": 1721630720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004865252525252525,
      "loss": 1.5166,
      "theoretical_loss": 3.4741614893609016,
      "tokens_seen": 1721696256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048650505050505053,
      "loss": 1.5006,
      "theoretical_loss": 3.4741501359143423,
      "tokens_seen": 1721761792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864848484848485,
      "loss": 1.2737,
      "theoretical_loss": 3.4741387830209214,
      "tokens_seen": 1721827328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864646464646465,
      "loss": 1.3364,
      "theoretical_loss": 3.474127430680591,
      "tokens_seen": 1721892864
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048644444444444446,
      "loss": 1.4177,
      "theoretical_loss": 3.474116078893303,
      "tokens_seen": 1721958400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864242424242424,
      "loss": 1.3975,
      "theoretical_loss": 3.47410472765901,
      "tokens_seen": 1722023936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004864040404040404,
      "loss": 1.3059,
      "theoretical_loss": 3.474093376977664,
      "tokens_seen": 1722089472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863838383838384,
      "loss": 1.4123,
      "theoretical_loss": 3.474082026849216,
      "tokens_seen": 1722155008
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.3375488817691803,
      "objective/train/avg_token_score": 0.013777061365544796,
      "objective/train/docs_used": 975130,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2517693042755127,
      "objective/train/original_loss": 5.1249237060546875,
      "objective/train/theoretical_loss": 3.4740706772736187,
      "objective/train/tokens_used": 81079776,
      "objective/train/unlikelihood_loss": 0.040009669959545135,
      "theoretical_loss": 3.4740706772736187,
      "tokens_seen": 1722220544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863636363636364,
      "loss": 1.5127,
      "theoretical_loss": 3.4740706772736187,
      "tokens_seen": 1722220544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048634343434343435,
      "loss": 1.2964,
      "theoretical_loss": 3.4740593282508243,
      "tokens_seen": 1722286080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048632323232323236,
      "loss": 1.4136,
      "theoretical_loss": 3.474047979780785,
      "tokens_seen": 1722351616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004863030303030303,
      "loss": 1.4117,
      "theoretical_loss": 3.474036631863452,
      "tokens_seen": 1722417152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862828282828283,
      "loss": 1.3667,
      "theoretical_loss": 3.4740252844987785,
      "tokens_seen": 1722482688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862626262626263,
      "loss": 1.2363,
      "theoretical_loss": 3.474013937686715,
      "tokens_seen": 1722548224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048624242424242424,
      "loss": 1.4347,
      "theoretical_loss": 3.4740025914272152,
      "tokens_seen": 1722613760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048622222222222225,
      "loss": 1.3785,
      "theoretical_loss": 3.4739912457202307,
      "tokens_seen": 1722679296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004862020202020202,
      "loss": 1.4418,
      "theoretical_loss": 3.4739799005657126,
      "tokens_seen": 1722744832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048618181818181817,
      "loss": 1.4983,
      "theoretical_loss": 3.473968555963614,
      "tokens_seen": 1722810368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861616161616162,
      "loss": 1.4721,
      "theoretical_loss": 3.473957211913887,
      "tokens_seen": 1722875904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861414141414142,
      "loss": 1.4795,
      "theoretical_loss": 3.4739458684164832,
      "tokens_seen": 1722941440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048612121212121215,
      "loss": 1.3607,
      "theoretical_loss": 3.473934525471355,
      "tokens_seen": 1723006976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004861010101010101,
      "loss": 1.4223,
      "theoretical_loss": 3.4739231830784543,
      "tokens_seen": 1723072512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048608080808080806,
      "loss": 1.5619,
      "theoretical_loss": 3.473911841237733,
      "tokens_seen": 1723138048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048606060606060607,
      "loss": 1.4897,
      "theoretical_loss": 3.473900499949144,
      "tokens_seen": 1723203584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004860404040404041,
      "loss": 1.3993,
      "theoretical_loss": 3.4738891592126393,
      "tokens_seen": 1723269120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048602020202020204,
      "loss": 1.408,
      "theoretical_loss": 3.4738778190281705,
      "tokens_seen": 1723334656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000486,
      "loss": 1.3879,
      "theoretical_loss": 3.4738664793956895,
      "tokens_seen": 1723400192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048597979797979795,
      "loss": 1.4359,
      "theoretical_loss": 3.473855140315149,
      "tokens_seen": 1723465728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048595959595959596,
      "loss": 1.4494,
      "theoretical_loss": 3.473843801786501,
      "tokens_seen": 1723531264
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485939393939394,
      "loss": 1.3349,
      "theoretical_loss": 3.473832463809698,
      "tokens_seen": 1723596800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048591919191919193,
      "loss": 1.4685,
      "theoretical_loss": 3.4738211263846916,
      "tokens_seen": 1723662336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858989898989899,
      "loss": 1.3586,
      "theoretical_loss": 3.473809789511434,
      "tokens_seen": 1723727872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048587878787878784,
      "loss": 1.3079,
      "theoretical_loss": 3.473798453189878,
      "tokens_seen": 1723793408
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.4409824013710022,
      "objective/train/avg_token_score": 0.026003144681453705,
      "objective/train/docs_used": 976352,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4241589307785034,
      "objective/train/original_loss": 4.103183746337891,
      "objective/train/theoretical_loss": 3.4737871174199757,
      "objective/train/tokens_used": 82718176,
      "objective/train/unlikelihood_loss": 0.09437726438045502,
      "theoretical_loss": 3.4737871174199757,
      "tokens_seen": 1723858944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858585858585859,
      "loss": 1.4386,
      "theoretical_loss": 3.4737871174199757,
      "tokens_seen": 1723858944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048583838383838387,
      "loss": 1.3626,
      "theoretical_loss": 3.4737757822016784,
      "tokens_seen": 1723924480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004858181818181818,
      "loss": 1.4696,
      "theoretical_loss": 3.473764447534939,
      "tokens_seen": 1723990016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857979797979798,
      "loss": 1.3754,
      "theoretical_loss": 3.4737531134197095,
      "tokens_seen": 1724055552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048577777777777774,
      "loss": 1.4346,
      "theoretical_loss": 3.473741779855942,
      "tokens_seen": 1724121088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857575757575758,
      "loss": 1.4929,
      "theoretical_loss": 3.4737304468435894,
      "tokens_seen": 1724186624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048573737373737376,
      "loss": 1.4366,
      "theoretical_loss": 3.473719114382603,
      "tokens_seen": 1724252160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004857171717171717,
      "loss": 1.2766,
      "theoretical_loss": 3.4737077824729354,
      "tokens_seen": 1724317696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856969696969697,
      "loss": 1.3714,
      "theoretical_loss": 3.4736964511145385,
      "tokens_seen": 1724383232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856767676767677,
      "loss": 1.2763,
      "theoretical_loss": 3.473685120307365,
      "tokens_seen": 1724448768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856565656565657,
      "loss": 1.4836,
      "theoretical_loss": 3.4736737900513672,
      "tokens_seen": 1724514304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048563636363636365,
      "loss": 1.4698,
      "theoretical_loss": 3.473662460346497,
      "tokens_seen": 1724579840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004856161616161616,
      "loss": 1.4449,
      "theoretical_loss": 3.4736511311927067,
      "tokens_seen": 1724645376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048559595959595957,
      "loss": 1.3052,
      "theoretical_loss": 3.4736398025899486,
      "tokens_seen": 1724710912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004855757575757576,
      "loss": 1.565,
      "theoretical_loss": 3.4736284745381756,
      "tokens_seen": 1724776448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004855555555555556,
      "loss": 1.3919,
      "theoretical_loss": 3.4736171470373387,
      "tokens_seen": 1724841984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048553535353535355,
      "loss": 1.5023,
      "theoretical_loss": 3.473605820087391,
      "tokens_seen": 1724907520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004855151515151515,
      "loss": 1.2998,
      "theoretical_loss": 3.4735944936882848,
      "tokens_seen": 1724973056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854949494949495,
      "loss": 1.6207,
      "theoretical_loss": 3.473583167839972,
      "tokens_seen": 1725038592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048547474747474747,
      "loss": 1.2696,
      "theoretical_loss": 3.473571842542405,
      "tokens_seen": 1725104128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854545454545455,
      "loss": 1.6765,
      "theoretical_loss": 3.473560517795536,
      "tokens_seen": 1725169664
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048543434343434344,
      "loss": 1.5439,
      "theoretical_loss": 3.473549193599318,
      "tokens_seen": 1725235200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004854141414141414,
      "loss": 1.5656,
      "theoretical_loss": 3.4735378699537027,
      "tokens_seen": 1725300736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853939393939394,
      "loss": 1.2932,
      "theoretical_loss": 3.4735265468586425,
      "tokens_seen": 1725366272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048537373737373736,
      "loss": 1.4775,
      "theoretical_loss": 3.4735152243140894,
      "tokens_seen": 1725431808
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.4415322542190552,
      "objective/train/avg_token_score": 0.00970387365669012,
      "objective/train/docs_used": 977030,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4506078958511353,
      "objective/train/original_loss": 4.433774948120117,
      "objective/train/theoretical_loss": 3.4735039023199965,
      "objective/train/tokens_used": 84356576,
      "objective/train/unlikelihood_loss": 0.02620289847254753,
      "theoretical_loss": 3.4735039023199965,
      "tokens_seen": 1725497344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004853535353535354,
      "loss": 1.7168,
      "theoretical_loss": 3.4735039023199965,
      "tokens_seen": 1725497344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048533333333333333,
      "loss": 1.2021,
      "theoretical_loss": 3.4734925808763153,
      "tokens_seen": 1725562880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048531313131313134,
      "loss": 1.4966,
      "theoretical_loss": 3.473481259982999,
      "tokens_seen": 1725628416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852929292929293,
      "loss": 1.4124,
      "theoretical_loss": 3.4734699396399993,
      "tokens_seen": 1725693952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852727272727273,
      "loss": 1.3887,
      "theoretical_loss": 3.4734586198472686,
      "tokens_seen": 1725759488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048525252525252527,
      "loss": 1.1216,
      "theoretical_loss": 3.4734473006047595,
      "tokens_seen": 1725825024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004852323232323232,
      "loss": 1.3029,
      "theoretical_loss": 3.4734359819124245,
      "tokens_seen": 1725890560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048521212121212123,
      "loss": 1.3902,
      "theoretical_loss": 3.4734246637702153,
      "tokens_seen": 1725956096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851919191919192,
      "loss": 1.34,
      "theoretical_loss": 3.4734133461780847,
      "tokens_seen": 1726021632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851717171717172,
      "loss": 1.4358,
      "theoretical_loss": 3.4734020291359853,
      "tokens_seen": 1726087168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048515151515151516,
      "loss": 1.4642,
      "theoretical_loss": 3.473390712643869,
      "tokens_seen": 1726152704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048513131313131317,
      "loss": 1.3417,
      "theoretical_loss": 3.473379396701689,
      "tokens_seen": 1726218240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004851111111111111,
      "loss": 1.1703,
      "theoretical_loss": 3.473368081309397,
      "tokens_seen": 1726283776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850909090909091,
      "loss": 1.6721,
      "theoretical_loss": 3.473356766466945,
      "tokens_seen": 1726349312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004850707070707071,
      "loss": 1.3141,
      "theoretical_loss": 3.4733454521742866,
      "tokens_seen": 1726414848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048505050505050505,
      "loss": 1.2502,
      "theoretical_loss": 3.4733341384313734,
      "tokens_seen": 1726480384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048503030303030306,
      "loss": 1.2986,
      "theoretical_loss": 3.4733228252381574,
      "tokens_seen": 1726545920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000485010101010101,
      "loss": 1.3311,
      "theoretical_loss": 3.4733115125945924,
      "tokens_seen": 1726611456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484989898989899,
      "loss": 1.5198,
      "theoretical_loss": 3.47330020050063,
      "tokens_seen": 1726676992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484969696969697,
      "loss": 1.2449,
      "theoretical_loss": 3.4732888889562226,
      "tokens_seen": 1726742528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484949494949495,
      "loss": 1.345,
      "theoretical_loss": 3.4732775779613227,
      "tokens_seen": 1726808064
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048492929292929296,
      "loss": 1.3956,
      "theoretical_loss": 3.473266267515883,
      "tokens_seen": 1726873600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004849090909090909,
      "loss": 1.3801,
      "theoretical_loss": 3.4732549576198553,
      "tokens_seen": 1726939136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048488888888888887,
      "loss": 1.3834,
      "theoretical_loss": 3.4732436482731925,
      "tokens_seen": 1727004672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848686868686869,
      "loss": 1.4937,
      "theoretical_loss": 3.4732323394758478,
      "tokens_seen": 1727070208
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.39345064759254456,
      "objective/train/avg_token_score": 0.009139729663729668,
      "objective/train/docs_used": 978106,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4235575199127197,
      "objective/train/original_loss": 4.360210418701172,
      "objective/train/theoretical_loss": 3.473221031227773,
      "objective/train/tokens_used": 85994976,
      "objective/train/unlikelihood_loss": 0.08843415230512619,
      "theoretical_loss": 3.473221031227773,
      "tokens_seen": 1727135744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848484848484849,
      "loss": 1.5847,
      "theoretical_loss": 3.473221031227773,
      "tokens_seen": 1727135744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048482828282828285,
      "loss": 1.4794,
      "theoretical_loss": 3.47320972352892,
      "tokens_seen": 1727201280
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004848080808080808,
      "loss": 1.4167,
      "theoretical_loss": 3.473198416379242,
      "tokens_seen": 1727266816
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048478787878787876,
      "loss": 1.2351,
      "theoretical_loss": 3.4731871097786913,
      "tokens_seen": 1727332352
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048476767676767683,
      "loss": 1.3563,
      "theoretical_loss": 3.4731758037272202,
      "tokens_seen": 1727397888
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847474747474748,
      "loss": 1.291,
      "theoretical_loss": 3.473164498224782,
      "tokens_seen": 1727463424
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048472727272727274,
      "loss": 1.4047,
      "theoretical_loss": 3.4731531932713287,
      "tokens_seen": 1727528960
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004847070707070707,
      "loss": 1.3456,
      "theoretical_loss": 3.473141888866812,
      "tokens_seen": 1727594496
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048468686868686865,
      "loss": 1.3892,
      "theoretical_loss": 3.4731305850111855,
      "tokens_seen": 1727660032
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846666666666667,
      "loss": 1.3859,
      "theoretical_loss": 3.473119281704402,
      "tokens_seen": 1727725568
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846464646464647,
      "loss": 1.1975,
      "theoretical_loss": 3.473107978946413,
      "tokens_seen": 1727791104
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048462626262626263,
      "loss": 1.3792,
      "theoretical_loss": 3.4730966767371716,
      "tokens_seen": 1727856640
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004846060606060606,
      "loss": 1.3253,
      "theoretical_loss": 3.4730853750766304,
      "tokens_seen": 1727922176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048458585858585855,
      "loss": 1.5187,
      "theoretical_loss": 3.4730740739647414,
      "tokens_seen": 1727987712
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845656565656566,
      "loss": 1.4738,
      "theoretical_loss": 3.4730627734014576,
      "tokens_seen": 1728053248
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048454545454545457,
      "loss": 1.1208,
      "theoretical_loss": 3.4730514733867315,
      "tokens_seen": 1728118784
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845252525252525,
      "loss": 1.3628,
      "theoretical_loss": 3.4730401739205163,
      "tokens_seen": 1728184320
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004845050505050505,
      "loss": 1.4873,
      "theoretical_loss": 3.4730288750027634,
      "tokens_seen": 1728249856
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844848484848485,
      "loss": 1.3379,
      "theoretical_loss": 3.473017576633426,
      "tokens_seen": 1728315392
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844646464646465,
      "loss": 1.2096,
      "theoretical_loss": 3.4730062788124565,
      "tokens_seen": 1728380928
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048444444444444446,
      "loss": 1.2331,
      "theoretical_loss": 3.472994981539808,
      "tokens_seen": 1728446464
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844242424242424,
      "loss": 1.5172,
      "theoretical_loss": 3.4729836848154325,
      "tokens_seen": 1728512000
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004844040404040404,
      "loss": 1.4394,
      "theoretical_loss": 3.472972388639283,
      "tokens_seen": 1728577536
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843838383838384,
      "loss": 1.4188,
      "theoretical_loss": 3.4729610930113117,
      "tokens_seen": 1728643072
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843636363636364,
      "loss": 1.305,
      "theoretical_loss": 3.472949797931472,
      "tokens_seen": 1728708608
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.3660801649093628,
      "objective/train/avg_token_score": 0.007394853513687849,
      "objective/train/docs_used": 978835,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.256290316581726,
      "objective/train/original_loss": 4.4458208084106445,
      "objective/train/theoretical_loss": 3.472938503399715,
      "objective/train/tokens_used": 87633376,
      "objective/train/unlikelihood_loss": 0.06266473978757858,
      "theoretical_loss": 3.472938503399715,
      "tokens_seen": 1728774144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048434343434343435,
      "loss": 1.5173,
      "theoretical_loss": 3.472938503399715,
      "tokens_seen": 1728774144
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843232323232323,
      "loss": 1.6362,
      "theoretical_loss": 3.472927209415995,
      "tokens_seen": 1728839680
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004843030303030303,
      "loss": 1.2885,
      "theoretical_loss": 3.4729159159802636,
      "tokens_seen": 1728905216
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842828282828283,
      "loss": 1.5108,
      "theoretical_loss": 3.4729046230924743,
      "tokens_seen": 1728970752
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842626262626263,
      "loss": 1.5734,
      "theoretical_loss": 3.472893330752579,
      "tokens_seen": 1729036288
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048424242424242425,
      "loss": 1.2754,
      "theoretical_loss": 3.4728820389605306,
      "tokens_seen": 1729101824
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842222222222222,
      "loss": 1.3543,
      "theoretical_loss": 3.4728707477162817,
      "tokens_seen": 1729167360
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004842020202020202,
      "loss": 1.4882,
      "theoretical_loss": 3.472859457019785,
      "tokens_seen": 1729232896
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048418181818181817,
      "loss": 1.288,
      "theoretical_loss": 3.4728481668709934,
      "tokens_seen": 1729298432
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004841616161616162,
      "loss": 1.1096,
      "theoretical_loss": 3.4728368772698586,
      "tokens_seen": 1729363968
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048414141414141414,
      "loss": 1.3439,
      "theoretical_loss": 3.4728255882163346,
      "tokens_seen": 1729429504
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048412121212121215,
      "loss": 1.3121,
      "theoretical_loss": 3.4728142997103735,
      "tokens_seen": 1729495040
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004841010101010101,
      "loss": 1.3209,
      "theoretical_loss": 3.472803011751928,
      "tokens_seen": 1729560576
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048408080808080807,
      "loss": 1.3692,
      "theoretical_loss": 3.472791724340951,
      "tokens_seen": 1729626112
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004840606060606061,
      "loss": 1.5298,
      "theoretical_loss": 3.472780437477395,
      "tokens_seen": 1729691648
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048404040404040403,
      "loss": 1.4671,
      "theoretical_loss": 3.472769151161213,
      "tokens_seen": 1729757184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048402020202020204,
      "loss": 1.5067,
      "theoretical_loss": 3.4727578653923565,
      "tokens_seen": 1729822720
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000484,
      "loss": 1.5815,
      "theoretical_loss": 3.4727465801707798,
      "tokens_seen": 1729888256
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048397979797979796,
      "loss": 1.5065,
      "theoretical_loss": 3.472735295496435,
      "tokens_seen": 1729953792
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048395959595959597,
      "loss": 1.3621,
      "theoretical_loss": 3.4727240113692748,
      "tokens_seen": 1730019328
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483939393939394,
      "loss": 1.2581,
      "theoretical_loss": 3.472712727789252,
      "tokens_seen": 1730084864
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048391919191919194,
      "loss": 1.4585,
      "theoretical_loss": 3.4727014447563196,
      "tokens_seen": 1730150400
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004838989898989899,
      "loss": 1.3508,
      "theoretical_loss": 3.472690162270429,
      "tokens_seen": 1730215936
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048387878787878785,
      "loss": 1.4799,
      "theoretical_loss": 3.472678880331535,
      "tokens_seen": 1730281472
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048385858585858586,
      "loss": 1.3777,
      "theoretical_loss": 3.4726675989395894,
      "tokens_seen": 1730347008
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.46120479702949524,
      "objective/train/avg_token_score": 0.012367086485028267,
      "objective/train/docs_used": 979527,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5714759826660156,
      "objective/train/original_loss": 4.267512321472168,
      "objective/train/theoretical_loss": 3.472656318094545,
      "objective/train/tokens_used": 89271776,
      "objective/train/unlikelihood_loss": 0.061104901134967804,
      "theoretical_loss": 3.472656318094545,
      "tokens_seen": 1730412544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048383838383838387,
      "loss": 1.3963,
      "theoretical_loss": 3.472656318094545,
      "tokens_seen": 1730412544
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048381818181818183,
      "loss": 1.4758,
      "theoretical_loss": 3.4726450377963545,
      "tokens_seen": 1730478080
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837979797979798,
      "loss": 1.2657,
      "theoretical_loss": 3.47263375804497,
      "tokens_seen": 1730543616
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048377777777777774,
      "loss": 1.3777,
      "theoretical_loss": 3.472622478840346,
      "tokens_seen": 1730609152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837575757575758,
      "loss": 1.4637,
      "theoretical_loss": 3.472611200182434,
      "tokens_seen": 1730674688
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048373737373737377,
      "loss": 1.3547,
      "theoretical_loss": 3.472599922071187,
      "tokens_seen": 1730740224
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004837171717171717,
      "loss": 1.4689,
      "theoretical_loss": 3.472588644506558,
      "tokens_seen": 1730805760
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836969696969697,
      "loss": 1.6301,
      "theoretical_loss": 3.4725773674884994,
      "tokens_seen": 1730871296
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048367676767676764,
      "loss": 1.5216,
      "theoretical_loss": 3.4725660910169647,
      "tokens_seen": 1730936832
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836565656565657,
      "loss": 1.5067,
      "theoretical_loss": 3.472554815091906,
      "tokens_seen": 1731002368
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048363636363636366,
      "loss": 1.4398,
      "theoretical_loss": 3.472543539713277,
      "tokens_seen": 1731067904
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004836161616161616,
      "loss": 1.2717,
      "theoretical_loss": 3.4725322648810297,
      "tokens_seen": 1731133440
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048359595959595957,
      "loss": 1.3783,
      "theoretical_loss": 3.4725209905951173,
      "tokens_seen": 1731198976
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048357575757575764,
      "loss": 1.5397,
      "theoretical_loss": 3.472509716855493,
      "tokens_seen": 1731264512
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004835555555555556,
      "loss": 1.4858,
      "theoretical_loss": 3.4724984436621087,
      "tokens_seen": 1731330048
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048353535353535355,
      "loss": 1.3623,
      "theoretical_loss": 3.4724871710149183,
      "tokens_seen": 1731395584
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004835151515151515,
      "loss": 1.3471,
      "theoretical_loss": 3.4724758989138738,
      "tokens_seen": 1731461120
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048349494949494946,
      "loss": 1.5417,
      "theoretical_loss": 3.472464627358929,
      "tokens_seen": 1731526656
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048347474747474753,
      "loss": 1.6138,
      "theoretical_loss": 3.4724533563500355,
      "tokens_seen": 1731592192
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834545454545455,
      "loss": 1.6054,
      "theoretical_loss": 3.472442085887147,
      "tokens_seen": 1731657728
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048343434343434344,
      "loss": 1.6946,
      "theoretical_loss": 3.472430815970217,
      "tokens_seen": 1731723264
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004834141414141414,
      "loss": 1.4545,
      "theoretical_loss": 3.4724195465991974,
      "tokens_seen": 1731788800
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833939393939394,
      "loss": 1.4118,
      "theoretical_loss": 3.4724082777740413,
      "tokens_seen": 1731854336
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833737373737374,
      "loss": 1.2402,
      "theoretical_loss": 3.472397009494702,
      "tokens_seen": 1731919872
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833535353535354,
      "loss": 1.5217,
      "theoretical_loss": 3.4723857417611317,
      "tokens_seen": 1731985408
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.37738269567489624,
      "objective/train/avg_token_score": 0.007554531563073397,
      "objective/train/docs_used": 980288,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1984022855758667,
      "objective/train/original_loss": 4.563331127166748,
      "objective/train/theoretical_loss": 3.4723744745732836,
      "objective/train/tokens_used": 90910176,
      "objective/train/unlikelihood_loss": 0.06496313959360123,
      "theoretical_loss": 3.4723744745732836,
      "tokens_seen": 1732050944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048333333333333334,
      "loss": 1.1324,
      "theoretical_loss": 3.4723744745732836,
      "tokens_seen": 1732050944
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004833131313131313,
      "loss": 1.2802,
      "theoretical_loss": 3.472363207931111,
      "tokens_seen": 1732116480
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832929292929293,
      "loss": 1.3541,
      "theoretical_loss": 3.4723519418345665,
      "tokens_seen": 1732182016
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004832727272727273,
      "loss": 1.5818,
      "theoretical_loss": 3.4723406762836033,
      "tokens_seen": 1732247552
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048325252525252527,
      "loss": 1.4111,
      "theoretical_loss": 3.472329411278174,
      "tokens_seen": 1732313088
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048323232323232323,
      "loss": 1.4464,
      "theoretical_loss": 3.472318146818232,
      "tokens_seen": 1732378624
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048321212121212124,
      "loss": 1.4059,
      "theoretical_loss": 3.4723068829037294,
      "tokens_seen": 1732444160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831919191919192,
      "loss": 1.2797,
      "theoretical_loss": 3.47229561953462,
      "tokens_seen": 1732509696
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831717171717172,
      "loss": 1.3734,
      "theoretical_loss": 3.4722843567108566,
      "tokens_seen": 1732575232
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048315151515151516,
      "loss": 1.3392,
      "theoretical_loss": 3.4722730944323916,
      "tokens_seen": 1732640768
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004831313131313131,
      "loss": 1.4454,
      "theoretical_loss": 3.4722618326991785,
      "tokens_seen": 1732706304
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048311111111111113,
      "loss": 1.578,
      "theoretical_loss": 3.4722505715111707,
      "tokens_seen": 1732771840
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830909090909091,
      "loss": 1.3528,
      "theoretical_loss": 3.4722393108683205,
      "tokens_seen": 1732837376
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004830707070707071,
      "loss": 1.419,
      "theoretical_loss": 3.472228050770581,
      "tokens_seen": 1732902912
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048305050505050506,
      "loss": 1.3326,
      "theoretical_loss": 3.472216791217905,
      "tokens_seen": 1732968448
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048303030303030307,
      "loss": 1.6869,
      "theoretical_loss": 3.4722055322102463,
      "tokens_seen": 1733033984
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000483010101010101,
      "loss": 1.5042,
      "theoretical_loss": 3.472194273747557,
      "tokens_seen": 1733099520
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482989898989899,
      "loss": 1.3897,
      "theoretical_loss": 3.47218301582979,
      "tokens_seen": 1733165056
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.000482969696969697,
      "loss": 1.2143,
      "theoretical_loss": 3.4721717584569003,
      "tokens_seen": 1733230592
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048294949494949495,
      "loss": 1.5147,
      "theoretical_loss": 3.472160501628838,
      "tokens_seen": 1733296128
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048292929292929296,
      "loss": 1.3647,
      "theoretical_loss": 3.4721492453455585,
      "tokens_seen": 1733361664
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004829090909090909,
      "loss": 1.1661,
      "theoretical_loss": 3.4721379896070133,
      "tokens_seen": 1733427200
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828888888888889,
      "loss": 1.5285,
      "theoretical_loss": 3.4721267344131563,
      "tokens_seen": 1733492736
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828686868686869,
      "loss": 1.4618,
      "theoretical_loss": 3.4721154797639406,
      "tokens_seen": 1733558272
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828484848484849,
      "loss": 1.5015,
      "theoretical_loss": 3.4721042256593186,
      "tokens_seen": 1733623808
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.5816226601600647,
      "objective/train/avg_token_score": 0.014024200849235058,
      "objective/train/docs_used": 981887,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.1395461559295654,
      "objective/train/original_loss": 4.308298587799072,
      "objective/train/theoretical_loss": 3.472092972099244,
      "objective/train/tokens_used": 92548576,
      "objective/train/unlikelihood_loss": 0.05324159935116768,
      "theoretical_loss": 3.472092972099244,
      "tokens_seen": 1733689344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048282828282828285,
      "loss": 1.4546,
      "theoretical_loss": 3.472092972099244,
      "tokens_seen": 1733689344
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004828080808080808,
      "loss": 1.1628,
      "theoretical_loss": 3.4720817190836692,
      "tokens_seen": 1733754880
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048278787878787877,
      "loss": 1.2467,
      "theoretical_loss": 3.472070466612548,
      "tokens_seen": 1733820416
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827676767676768,
      "loss": 1.4641,
      "theoretical_loss": 3.4720592146858333,
      "tokens_seen": 1733885952
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827474747474748,
      "loss": 1.5125,
      "theoretical_loss": 3.4720479633034778,
      "tokens_seen": 1733951488
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048272727272727275,
      "loss": 1.3324,
      "theoretical_loss": 3.472036712465435,
      "tokens_seen": 1734017024
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004827070707070707,
      "loss": 1.3651,
      "theoretical_loss": 3.4720254621716578,
      "tokens_seen": 1734082560
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048268686868686866,
      "loss": 1.4273,
      "theoretical_loss": 3.472014212422099,
      "tokens_seen": 1734148096
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048266666666666667,
      "loss": 1.2076,
      "theoretical_loss": 3.4720029632167124,
      "tokens_seen": 1734213632
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826464646464647,
      "loss": 1.1455,
      "theoretical_loss": 3.471991714555451,
      "tokens_seen": 1734279168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048262626262626264,
      "loss": 1.5309,
      "theoretical_loss": 3.471980466438267,
      "tokens_seen": 1734344704
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004826060606060606,
      "loss": 1.5117,
      "theoretical_loss": 3.471969218865115,
      "tokens_seen": 1734410240
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048258585858585855,
      "loss": 1.5216,
      "theoretical_loss": 3.4719579718359466,
      "tokens_seen": 1734475776
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825656565656566,
      "loss": 1.52,
      "theoretical_loss": 3.4719467253507164,
      "tokens_seen": 1734541312
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825454545454546,
      "loss": 1.5335,
      "theoretical_loss": 3.471935479409376,
      "tokens_seen": 1734606848
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048252525252525253,
      "loss": 1.531,
      "theoretical_loss": 3.4719242340118797,
      "tokens_seen": 1734672384
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004825050505050505,
      "loss": 1.3912,
      "theoretical_loss": 3.471912989158181,
      "tokens_seen": 1734737920
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048248484848484845,
      "loss": 1.5328,
      "theoretical_loss": 3.471901744848231,
      "tokens_seen": 1734803456
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824646464646465,
      "loss": 1.4363,
      "theoretical_loss": 3.471890501081985,
      "tokens_seen": 1734868992
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048244444444444447,
      "loss": 1.6167,
      "theoretical_loss": 3.4718792578593956,
      "tokens_seen": 1734934528
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824242424242424,
      "loss": 1.3632,
      "theoretical_loss": 3.4718680151804158,
      "tokens_seen": 1735000064
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004824040404040404,
      "loss": 1.3851,
      "theoretical_loss": 3.4718567730449985,
      "tokens_seen": 1735065600
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823838383838384,
      "loss": 1.2801,
      "theoretical_loss": 3.4718455314530976,
      "tokens_seen": 1735131136
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823636363636364,
      "loss": 1.3273,
      "theoretical_loss": 3.471834290404665,
      "tokens_seen": 1735196672
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.00048234343434343436,
      "loss": 1.5982,
      "theoretical_loss": 3.4718230498996556,
      "tokens_seen": 1735262208
    },
    {
      "epoch": 0.04,
      "objective/train/aligned_token_count": 0.25384896993637085,
      "objective/train/avg_token_score": 0.01493894774466753,
      "objective/train/docs_used": 982466,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8242332339286804,
      "objective/train/original_loss": 4.889101505279541,
      "objective/train/theoretical_loss": 3.4718118099380213,
      "objective/train/tokens_used": 94186976,
      "objective/train/unlikelihood_loss": 0.03157869726419449,
      "theoretical_loss": 3.4718118099380213,
      "tokens_seen": 1735327744
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0004823232323232323,
      "loss": 1.1415,
      "theoretical_loss": 3.4718118099380213,
      "tokens_seen": 1735327744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004823030303030303,
      "loss": 1.321,
      "theoretical_loss": 3.4718005705197164,
      "tokens_seen": 1735393280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822828282828283,
      "loss": 1.6192,
      "theoretical_loss": 3.4717893316446933,
      "tokens_seen": 1735458816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822626262626263,
      "loss": 1.277,
      "theoretical_loss": 3.471778093312905,
      "tokens_seen": 1735524352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048224242424242425,
      "loss": 1.3975,
      "theoretical_loss": 3.4717668555243053,
      "tokens_seen": 1735589888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822222222222222,
      "loss": 1.5485,
      "theoretical_loss": 3.4717556182788476,
      "tokens_seen": 1735655424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004822020202020202,
      "loss": 1.1766,
      "theoretical_loss": 3.471744381576485,
      "tokens_seen": 1735720960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821818181818182,
      "loss": 1.3168,
      "theoretical_loss": 3.4717331454171703,
      "tokens_seen": 1735786496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821616161616162,
      "loss": 1.4985,
      "theoretical_loss": 3.471721909800857,
      "tokens_seen": 1735852032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048214141414141415,
      "loss": 1.4077,
      "theoretical_loss": 3.4717106747274986,
      "tokens_seen": 1735917568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821212121212121,
      "loss": 1.4942,
      "theoretical_loss": 3.471699440197048,
      "tokens_seen": 1735983104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004821010101010101,
      "loss": 1.5322,
      "theoretical_loss": 3.4716882062094587,
      "tokens_seen": 1736048640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048208080808080807,
      "loss": 1.4188,
      "theoretical_loss": 3.4716769727646835,
      "tokens_seen": 1736114176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004820606060606061,
      "loss": 1.5154,
      "theoretical_loss": 3.4716657398626767,
      "tokens_seen": 1736179712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048204040404040404,
      "loss": 1.2208,
      "theoretical_loss": 3.471654507503391,
      "tokens_seen": 1736245248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048202020202020205,
      "loss": 1.524,
      "theoretical_loss": 3.4716432756867794,
      "tokens_seen": 1736310784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000482,
      "loss": 1.5759,
      "theoretical_loss": 3.4716320444127957,
      "tokens_seen": 1736376320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048197979797979796,
      "loss": 1.3482,
      "theoretical_loss": 3.471620813681393,
      "tokens_seen": 1736441856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481959595959596,
      "loss": 1.4318,
      "theoretical_loss": 3.4716095834925245,
      "tokens_seen": 1736507392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048193939393939393,
      "loss": 1.423,
      "theoretical_loss": 3.4715983538461432,
      "tokens_seen": 1736572928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048191919191919194,
      "loss": 1.4259,
      "theoretical_loss": 3.4715871247422037,
      "tokens_seen": 1736638464
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818989898989899,
      "loss": 1.6113,
      "theoretical_loss": 3.4715758961806573,
      "tokens_seen": 1736704000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818787878787879,
      "loss": 1.4458,
      "theoretical_loss": 3.4715646681614594,
      "tokens_seen": 1736769536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048185858585858587,
      "loss": 1.1604,
      "theoretical_loss": 3.4715534406845627,
      "tokens_seen": 1736835072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004818383838383839,
      "loss": 1.3382,
      "theoretical_loss": 3.4715422137499194,
      "tokens_seen": 1736900608
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.43413978815078735,
      "objective/train/avg_token_score": 0.012802019715309143,
      "objective/train/docs_used": 983872,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4975614547729492,
      "objective/train/original_loss": 4.492546081542969,
      "objective/train/theoretical_loss": 3.4715309873574847,
      "objective/train/tokens_used": 95825376,
      "objective/train/unlikelihood_loss": 0.04530109092593193,
      "theoretical_loss": 3.4715309873574847,
      "tokens_seen": 1736966144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048181818181818184,
      "loss": 1.3615,
      "theoretical_loss": 3.4715309873574847,
      "tokens_seen": 1736966144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004817979797979798,
      "loss": 1.5586,
      "theoretical_loss": 3.4715197615072104,
      "tokens_seen": 1737031680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004817777777777778,
      "loss": 1.5499,
      "theoretical_loss": 3.4715085361990505,
      "tokens_seen": 1737097216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048175757575757576,
      "loss": 1.3085,
      "theoretical_loss": 3.471497311432959,
      "tokens_seen": 1737162752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048173737373737377,
      "loss": 1.4672,
      "theoretical_loss": 3.4714860872088877,
      "tokens_seen": 1737228288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048171717171717173,
      "loss": 1.3109,
      "theoretical_loss": 3.471474863526791,
      "tokens_seen": 1737293824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816969696969697,
      "loss": 1.5911,
      "theoretical_loss": 3.471463640386623,
      "tokens_seen": 1737359360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816767676767677,
      "loss": 1.4829,
      "theoretical_loss": 3.4714524177883352,
      "tokens_seen": 1737424896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816565656565657,
      "loss": 1.3043,
      "theoretical_loss": 3.4714411957318827,
      "tokens_seen": 1737490432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048163636363636366,
      "loss": 1.5117,
      "theoretical_loss": 3.4714299742172186,
      "tokens_seen": 1737555968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004816161616161616,
      "loss": 1.5361,
      "theoretical_loss": 3.4714187532442953,
      "tokens_seen": 1737621504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815959595959596,
      "loss": 1.4369,
      "theoretical_loss": 3.471407532813067,
      "tokens_seen": 1737687040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815757575757576,
      "loss": 1.3458,
      "theoretical_loss": 3.4713963129234875,
      "tokens_seen": 1737752576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815555555555556,
      "loss": 1.4904,
      "theoretical_loss": 3.4713850935755093,
      "tokens_seen": 1737818112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048153535353535356,
      "loss": 1.3929,
      "theoretical_loss": 3.4713738747690863,
      "tokens_seen": 1737883648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004815151515151515,
      "loss": 1.5227,
      "theoretical_loss": 3.471362656504172,
      "tokens_seen": 1737949184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048149494949494947,
      "loss": 1.4509,
      "theoretical_loss": 3.4713514387807196,
      "tokens_seen": 1738014720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048147474747474754,
      "loss": 1.3549,
      "theoretical_loss": 3.471340221598683,
      "tokens_seen": 1738080256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004814545454545455,
      "loss": 1.4395,
      "theoretical_loss": 3.4713290049580157,
      "tokens_seen": 1738145792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048143434343434345,
      "loss": 1.484,
      "theoretical_loss": 3.4713177888586704,
      "tokens_seen": 1738211328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004814141414141414,
      "loss": 1.3216,
      "theoretical_loss": 3.471306573300601,
      "tokens_seen": 1738276864
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048139393939393936,
      "loss": 1.6081,
      "theoretical_loss": 3.4712953582837613,
      "tokens_seen": 1738342400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048137373737373743,
      "loss": 1.5769,
      "theoretical_loss": 3.471284143808104,
      "tokens_seen": 1738407936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813535353535354,
      "loss": 1.4718,
      "theoretical_loss": 3.4712729298735834,
      "tokens_seen": 1738473472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048133333333333334,
      "loss": 1.3028,
      "theoretical_loss": 3.471261716480152,
      "tokens_seen": 1738539008
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.48277124762535095,
      "objective/train/avg_token_score": 0.009705035015940666,
      "objective/train/docs_used": 984348,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.655248999595642,
      "objective/train/original_loss": 4.150419235229492,
      "objective/train/theoretical_loss": 3.4712505036277648,
      "objective/train/tokens_used": 97463776,
      "objective/train/unlikelihood_loss": 0.052979037165641785,
      "theoretical_loss": 3.4712505036277648,
      "tokens_seen": 1738604544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004813131313131313,
      "loss": 1.4086,
      "theoretical_loss": 3.4712505036277648,
      "tokens_seen": 1738604544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048129292929292926,
      "loss": 1.3672,
      "theoretical_loss": 3.4712392913163734,
      "tokens_seen": 1738670080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812727272727273,
      "loss": 1.4242,
      "theoretical_loss": 3.4712280795459334,
      "tokens_seen": 1738735616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812525252525253,
      "loss": 1.5003,
      "theoretical_loss": 3.4712168683163966,
      "tokens_seen": 1738801152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048123232323232323,
      "loss": 1.3127,
      "theoretical_loss": 3.4712056576277175,
      "tokens_seen": 1738866688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004812121212121212,
      "loss": 1.3313,
      "theoretical_loss": 3.4711944474798493,
      "tokens_seen": 1738932224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811919191919192,
      "loss": 1.3412,
      "theoretical_loss": 3.4711832378727454,
      "tokens_seen": 1738997760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811717171717172,
      "loss": 1.3733,
      "theoretical_loss": 3.4711720288063592,
      "tokens_seen": 1739063296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048115151515151517,
      "loss": 1.4347,
      "theoretical_loss": 3.4711608202806445,
      "tokens_seen": 1739128832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048113131313131313,
      "loss": 1.1575,
      "theoretical_loss": 3.471149612295555,
      "tokens_seen": 1739194368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004811111111111111,
      "loss": 1.4967,
      "theoretical_loss": 3.4711384048510445,
      "tokens_seen": 1739259904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810909090909091,
      "loss": 1.5965,
      "theoretical_loss": 3.471127197947066,
      "tokens_seen": 1739325440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004810707070707071,
      "loss": 1.6453,
      "theoretical_loss": 3.4711159915835728,
      "tokens_seen": 1739390976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048105050505050506,
      "loss": 1.4245,
      "theoretical_loss": 3.4711047857605193,
      "tokens_seen": 1739456512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000481030303030303,
      "loss": 1.5783,
      "theoretical_loss": 3.4710935804778584,
      "tokens_seen": 1739522048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048101010101010103,
      "loss": 1.2761,
      "theoretical_loss": 3.471082375735544,
      "tokens_seen": 1739587584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480989898989899,
      "loss": 1.4248,
      "theoretical_loss": 3.47107117153353,
      "tokens_seen": 1739653120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480969696969697,
      "loss": 1.6432,
      "theoretical_loss": 3.471059967871769,
      "tokens_seen": 1739718656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048094949494949496,
      "loss": 1.394,
      "theoretical_loss": 3.4710487647502157,
      "tokens_seen": 1739784192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004809292929292929,
      "loss": 1.3901,
      "theoretical_loss": 3.4710375621688234,
      "tokens_seen": 1739849728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004809090909090909,
      "loss": 1.4413,
      "theoretical_loss": 3.471026360127545,
      "tokens_seen": 1739915264
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808888888888889,
      "loss": 1.4342,
      "theoretical_loss": 3.471015158626335,
      "tokens_seen": 1739980800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808686868686869,
      "loss": 1.2352,
      "theoretical_loss": 3.471003957665147,
      "tokens_seen": 1740046336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048084848484848485,
      "loss": 1.2348,
      "theoretical_loss": 3.470992757243934,
      "tokens_seen": 1740111872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048082828282828286,
      "loss": 1.3523,
      "theoretical_loss": 3.4709815573626495,
      "tokens_seen": 1740177408
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.33125612139701843,
      "objective/train/avg_token_score": 0.017105352133512497,
      "objective/train/docs_used": 985100,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2295209169387817,
      "objective/train/original_loss": 4.453027248382568,
      "objective/train/theoretical_loss": 3.4709703580212485,
      "objective/train/tokens_used": 99102176,
      "objective/train/unlikelihood_loss": 0.0723608136177063,
      "theoretical_loss": 3.4709703580212485,
      "tokens_seen": 1740242944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004808080808080808,
      "loss": 1.4608,
      "theoretical_loss": 3.4709703580212485,
      "tokens_seen": 1740242944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807878787878788,
      "loss": 1.4108,
      "theoretical_loss": 3.470959159219683,
      "tokens_seen": 1740308480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807676767676768,
      "loss": 1.5332,
      "theoretical_loss": 3.4709479609579077,
      "tokens_seen": 1740374016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048074747474747474,
      "loss": 1.7023,
      "theoretical_loss": 3.470936763235876,
      "tokens_seen": 1740439552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048072727272727275,
      "loss": 1.24,
      "theoretical_loss": 3.470925566053541,
      "tokens_seen": 1740505088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004807070707070707,
      "loss": 1.4697,
      "theoretical_loss": 3.4709143694108575,
      "tokens_seen": 1740570624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048068686868686867,
      "loss": 1.4578,
      "theoretical_loss": 3.4709031733077786,
      "tokens_seen": 1740636160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806666666666667,
      "loss": 1.5702,
      "theoretical_loss": 3.4708919777442575,
      "tokens_seen": 1740701696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806464646464647,
      "loss": 1.5591,
      "theoretical_loss": 3.470880782720249,
      "tokens_seen": 1740767232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048062626262626264,
      "loss": 1.2108,
      "theoretical_loss": 3.470869588235705,
      "tokens_seen": 1740832768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004806060606060606,
      "loss": 1.3671,
      "theoretical_loss": 3.4708583942905813,
      "tokens_seen": 1740898304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048058585858585856,
      "loss": 1.5984,
      "theoretical_loss": 3.4708472008848297,
      "tokens_seen": 1740963840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048056565656565657,
      "loss": 1.5255,
      "theoretical_loss": 3.4708360080184057,
      "tokens_seen": 1741029376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805454545454546,
      "loss": 1.5005,
      "theoretical_loss": 3.4708248156912616,
      "tokens_seen": 1741094912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048052525252525254,
      "loss": 1.5318,
      "theoretical_loss": 3.470813623903352,
      "tokens_seen": 1741160448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004805050505050505,
      "loss": 1.3555,
      "theoretical_loss": 3.47080243265463,
      "tokens_seen": 1741225984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048048484848484845,
      "loss": 1.324,
      "theoretical_loss": 3.47079124194505,
      "tokens_seen": 1741291520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004804646464646465,
      "loss": 1.4643,
      "theoretical_loss": 3.4707800517745646,
      "tokens_seen": 1741357056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004804444444444445,
      "loss": 1.2959,
      "theoretical_loss": 3.470768862143129,
      "tokens_seen": 1741422592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048042424242424243,
      "loss": 1.3676,
      "theoretical_loss": 3.470757673050696,
      "tokens_seen": 1741488128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004804040404040404,
      "loss": 1.4532,
      "theoretical_loss": 3.4707464844972193,
      "tokens_seen": 1741553664
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048038383838383834,
      "loss": 1.3073,
      "theoretical_loss": 3.470735296482653,
      "tokens_seen": 1741619200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004803636363636364,
      "loss": 1.3783,
      "theoretical_loss": 3.470724109006951,
      "tokens_seen": 1741684736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048034343434343437,
      "loss": 1.6342,
      "theoretical_loss": 3.470712922070067,
      "tokens_seen": 1741750272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004803232323232323,
      "loss": 1.5602,
      "theoretical_loss": 3.4707017356719545,
      "tokens_seen": 1741815808
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.4005987346172333,
      "objective/train/avg_token_score": 0.019556434825062752,
      "objective/train/docs_used": 986266,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.536611795425415,
      "objective/train/original_loss": 4.283116817474365,
      "objective/train/theoretical_loss": 3.4706905498125673,
      "objective/train/tokens_used": 100740576,
      "objective/train/unlikelihood_loss": 0.08906306326389313,
      "theoretical_loss": 3.4706905498125673,
      "tokens_seen": 1741881344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004803030303030303,
      "loss": 1.4406,
      "theoretical_loss": 3.4706905498125673,
      "tokens_seen": 1741881344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802828282828283,
      "loss": 1.4997,
      "theoretical_loss": 3.47067936449186,
      "tokens_seen": 1741946880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802626262626263,
      "loss": 1.6152,
      "theoretical_loss": 3.4706681797097847,
      "tokens_seen": 1742012416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048024242424242426,
      "loss": 1.5295,
      "theoretical_loss": 3.470656995466297,
      "tokens_seen": 1742077952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004802222222222222,
      "loss": 1.3141,
      "theoretical_loss": 3.47064581176135,
      "tokens_seen": 1742143488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048020202020202017,
      "loss": 1.3606,
      "theoretical_loss": 3.470634628594897,
      "tokens_seen": 1742209024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004801818181818182,
      "loss": 1.4034,
      "theoretical_loss": 3.4706234459668925,
      "tokens_seen": 1742274560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004801616161616162,
      "loss": 1.4074,
      "theoretical_loss": 3.47061226387729,
      "tokens_seen": 1742340096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048014141414141415,
      "loss": 1.2829,
      "theoretical_loss": 3.470601082326043,
      "tokens_seen": 1742405632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004801212121212121,
      "loss": 1.4758,
      "theoretical_loss": 3.4705899013131063,
      "tokens_seen": 1742471168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004801010101010101,
      "loss": 1.3614,
      "theoretical_loss": 3.470578720838433,
      "tokens_seen": 1742536704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048008080808080813,
      "loss": 1.3864,
      "theoretical_loss": 3.4705675409019774,
      "tokens_seen": 1742602240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004800606060606061,
      "loss": 1.4217,
      "theoretical_loss": 3.4705563615036925,
      "tokens_seen": 1742667776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048004040404040404,
      "loss": 1.1732,
      "theoretical_loss": 3.4705451826435327,
      "tokens_seen": 1742733312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000480020202020202,
      "loss": 1.4523,
      "theoretical_loss": 3.4705340043214523,
      "tokens_seen": 1742798848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00048,
      "loss": 1.4718,
      "theoretical_loss": 3.4705228265374046,
      "tokens_seen": 1742864384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479979797979798,
      "loss": 1.4891,
      "theoretical_loss": 3.470511649291344,
      "tokens_seen": 1742929920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479959595959596,
      "loss": 1.2984,
      "theoretical_loss": 3.4705004725832236,
      "tokens_seen": 1742995456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047993939393939394,
      "loss": 1.5396,
      "theoretical_loss": 3.470489296412998,
      "tokens_seen": 1743060992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047991919191919195,
      "loss": 1.6143,
      "theoretical_loss": 3.470478120780621,
      "tokens_seen": 1743126528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798989898989899,
      "loss": 1.3893,
      "theoretical_loss": 3.4704669456860455,
      "tokens_seen": 1743192064
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004798787878787879,
      "loss": 1.3921,
      "theoretical_loss": 3.4704557711292265,
      "tokens_seen": 1743257600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047985858585858587,
      "loss": 1.326,
      "theoretical_loss": 3.470444597110118,
      "tokens_seen": 1743323136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047983838383838383,
      "loss": 1.3948,
      "theoretical_loss": 3.4704334236286734,
      "tokens_seen": 1743388672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047981818181818184,
      "loss": 1.3807,
      "theoretical_loss": 3.4704222506848463,
      "tokens_seen": 1743454208
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.3908846378326416,
      "objective/train/avg_token_score": 0.00899689830839634,
      "objective/train/docs_used": 986906,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2534172534942627,
      "objective/train/original_loss": 4.556232929229736,
      "objective/train/theoretical_loss": 3.470411078278591,
      "objective/train/tokens_used": 102378976,
      "objective/train/unlikelihood_loss": 0.03879563882946968,
      "theoretical_loss": 3.470411078278591,
      "tokens_seen": 1743519744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797979797979798,
      "loss": 1.5736,
      "theoretical_loss": 3.470411078278591,
      "tokens_seen": 1743519744
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797777777777778,
      "loss": 1.304,
      "theoretical_loss": 3.4703999064098623,
      "tokens_seen": 1743585280
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047975757575757577,
      "loss": 1.4849,
      "theoretical_loss": 3.470388735078613,
      "tokens_seen": 1743650816
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004797373737373738,
      "loss": 1.3432,
      "theoretical_loss": 3.4703775642847967,
      "tokens_seen": 1743716352
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047971717171717173,
      "loss": 1.1625,
      "theoretical_loss": 3.4703663940283684,
      "tokens_seen": 1743781888
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796969696969697,
      "loss": 1.6368,
      "theoretical_loss": 3.470355224309282,
      "tokens_seen": 1743847424
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796767676767677,
      "loss": 1.4579,
      "theoretical_loss": 3.470344055127491,
      "tokens_seen": 1743912960
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047965656565656566,
      "loss": 1.4415,
      "theoretical_loss": 3.4703328864829492,
      "tokens_seen": 1743978496
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047963636363636367,
      "loss": 1.5911,
      "theoretical_loss": 3.4703217183756108,
      "tokens_seen": 1744044032
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004796161616161616,
      "loss": 1.2517,
      "theoretical_loss": 3.47031055080543,
      "tokens_seen": 1744109568
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795959595959596,
      "loss": 1.2024,
      "theoretical_loss": 3.470299383772361,
      "tokens_seen": 1744175104
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795757575757576,
      "loss": 1.3713,
      "theoretical_loss": 3.470288217276357,
      "tokens_seen": 1744240640
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795555555555556,
      "loss": 1.4253,
      "theoretical_loss": 3.4702770513173724,
      "tokens_seen": 1744306176
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047953535353535356,
      "loss": 1.406,
      "theoretical_loss": 3.470265885895361,
      "tokens_seen": 1744371712
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004795151515151515,
      "loss": 1.4899,
      "theoretical_loss": 3.4702547210102774,
      "tokens_seen": 1744437248
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794949494949495,
      "loss": 1.5386,
      "theoretical_loss": 3.470243556662075,
      "tokens_seen": 1744502784
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794747474747475,
      "loss": 1.2509,
      "theoretical_loss": 3.470232392850708,
      "tokens_seen": 1744568320
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794545454545455,
      "loss": 1.7622,
      "theoretical_loss": 3.470221229576131,
      "tokens_seen": 1744633856
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047943434343434345,
      "loss": 1.3678,
      "theoretical_loss": 3.4702100668382965,
      "tokens_seen": 1744699392
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004794141414141414,
      "loss": 1.4218,
      "theoretical_loss": 3.47019890463716,
      "tokens_seen": 1744764928
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047939393939393937,
      "loss": 1.4213,
      "theoretical_loss": 3.470187742972675,
      "tokens_seen": 1744830464
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793737373737374,
      "loss": 1.4785,
      "theoretical_loss": 3.470176581844796,
      "tokens_seen": 1744896000
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793535353535354,
      "loss": 1.5449,
      "theoretical_loss": 3.4701654212534763,
      "tokens_seen": 1744961536
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047933333333333335,
      "loss": 1.287,
      "theoretical_loss": 3.4701542611986698,
      "tokens_seen": 1745027072
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004793131313131313,
      "loss": 1.567,
      "theoretical_loss": 3.470143101680332,
      "tokens_seen": 1745092608
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.2597140669822693,
      "objective/train/avg_token_score": 0.04489205777645111,
      "objective/train/docs_used": 988319,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.912868082523346,
      "objective/train/original_loss": 4.443542003631592,
      "objective/train/theoretical_loss": 3.4701319426984156,
      "objective/train/tokens_used": 104017376,
      "objective/train/unlikelihood_loss": 0.06317847967147827,
      "theoretical_loss": 3.4701319426984156,
      "tokens_seen": 1745158144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047929292929292926,
      "loss": 1.2568,
      "theoretical_loss": 3.4701319426984156,
      "tokens_seen": 1745158144
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792727272727273,
      "loss": 1.3641,
      "theoretical_loss": 3.4701207842528747,
      "tokens_seen": 1745223680
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792525252525253,
      "loss": 1.5416,
      "theoretical_loss": 3.470109626343664,
      "tokens_seen": 1745289216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047923232323232324,
      "loss": 1.4213,
      "theoretical_loss": 3.470098468970738,
      "tokens_seen": 1745354752
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004792121212121212,
      "loss": 1.0977,
      "theoretical_loss": 3.470087312134049,
      "tokens_seen": 1745420288
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047919191919191915,
      "loss": 1.5069,
      "theoretical_loss": 3.470076155833553,
      "tokens_seen": 1745485824
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004791717171717172,
      "loss": 1.569,
      "theoretical_loss": 3.4700650000692033,
      "tokens_seen": 1745551360
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004791515151515152,
      "loss": 1.5586,
      "theoretical_loss": 3.4700538448409537,
      "tokens_seen": 1745616896
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047913131313131313,
      "loss": 1.5639,
      "theoretical_loss": 3.470042690148759,
      "tokens_seen": 1745682432
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004791111111111111,
      "loss": 1.4329,
      "theoretical_loss": 3.4700315359925726,
      "tokens_seen": 1745747968
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790909090909091,
      "loss": 1.4756,
      "theoretical_loss": 3.470020382372349,
      "tokens_seen": 1745813504
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004790707070707071,
      "loss": 1.4192,
      "theoretical_loss": 3.470009229288043,
      "tokens_seen": 1745879040
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047905050505050507,
      "loss": 1.2349,
      "theoretical_loss": 3.469998076739607,
      "tokens_seen": 1745944576
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479030303030303,
      "loss": 1.4639,
      "theoretical_loss": 3.469986924726997,
      "tokens_seen": 1746010112
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000479010101010101,
      "loss": 1.5232,
      "theoretical_loss": 3.469975773250166,
      "tokens_seen": 1746075648
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478989898989899,
      "loss": 1.1636,
      "theoretical_loss": 3.4699646223090688,
      "tokens_seen": 1746141184
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478969696969697,
      "loss": 1.4086,
      "theoretical_loss": 3.469953471903659,
      "tokens_seen": 1746206720
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047894949494949496,
      "loss": 1.3567,
      "theoretical_loss": 3.469942322033891,
      "tokens_seen": 1746272256
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004789292929292929,
      "loss": 1.6627,
      "theoretical_loss": 3.4699311726997193,
      "tokens_seen": 1746337792
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047890909090909093,
      "loss": 1.4643,
      "theoretical_loss": 3.469920023901097,
      "tokens_seen": 1746403328
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788888888888889,
      "loss": 1.393,
      "theoretical_loss": 3.46990887563798,
      "tokens_seen": 1746468864
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788686868686869,
      "loss": 1.6576,
      "theoretical_loss": 3.469897727910321,
      "tokens_seen": 1746534400
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047884848484848485,
      "loss": 1.3613,
      "theoretical_loss": 3.4698865807180748,
      "tokens_seen": 1746599936
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788282828282828,
      "loss": 1.5132,
      "theoretical_loss": 3.4698754340611955,
      "tokens_seen": 1746665472
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004788080808080808,
      "loss": 1.4484,
      "theoretical_loss": 3.4698642879396373,
      "tokens_seen": 1746731008
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.29673752188682556,
      "objective/train/avg_token_score": 0.007496143225580454,
      "objective/train/docs_used": 989102,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9727373123168945,
      "objective/train/original_loss": 4.123487949371338,
      "objective/train/theoretical_loss": 3.4698531423533545,
      "objective/train/tokens_used": 105655776,
      "objective/train/unlikelihood_loss": 0.07145555317401886,
      "theoretical_loss": 3.4698531423533545,
      "tokens_seen": 1746796544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787878787878788,
      "loss": 1.2235,
      "theoretical_loss": 3.4698531423533545,
      "tokens_seen": 1746796544
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787676767676768,
      "loss": 1.4326,
      "theoretical_loss": 3.469841997302301,
      "tokens_seen": 1746862080
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047874747474747475,
      "loss": 1.4788,
      "theoretical_loss": 3.4698308527864317,
      "tokens_seen": 1746927616
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047872727272727276,
      "loss": 1.4569,
      "theoretical_loss": 3.4698197088056997,
      "tokens_seen": 1746993152
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004787070707070707,
      "loss": 1.3571,
      "theoretical_loss": 3.4698085653600605,
      "tokens_seen": 1747058688
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047868686868686867,
      "loss": 1.3078,
      "theoretical_loss": 3.469797422449467,
      "tokens_seen": 1747124224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786666666666667,
      "loss": 1.3261,
      "theoretical_loss": 3.4697862800738752,
      "tokens_seen": 1747189760
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047864646464646464,
      "loss": 1.5378,
      "theoretical_loss": 3.4697751382332376,
      "tokens_seen": 1747255296
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047862626262626265,
      "loss": 1.3303,
      "theoretical_loss": 3.46976399692751,
      "tokens_seen": 1747320832
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004786060606060606,
      "loss": 1.513,
      "theoretical_loss": 3.469752856156645,
      "tokens_seen": 1747386368
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047858585858585856,
      "loss": 1.5127,
      "theoretical_loss": 3.469741715920598,
      "tokens_seen": 1747451904
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004785656565656566,
      "loss": 1.511,
      "theoretical_loss": 3.4697305762193227,
      "tokens_seen": 1747517440
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004785454545454546,
      "loss": 1.3687,
      "theoretical_loss": 3.469719437052774,
      "tokens_seen": 1747582976
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047852525252525254,
      "loss": 1.2167,
      "theoretical_loss": 3.4697082984209056,
      "tokens_seen": 1747648512
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004785050505050505,
      "loss": 1.3243,
      "theoretical_loss": 3.469697160323672,
      "tokens_seen": 1747714048
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047848484848484846,
      "loss": 1.7173,
      "theoretical_loss": 3.469686022761028,
      "tokens_seen": 1747779584
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047846464646464647,
      "loss": 1.2967,
      "theoretical_loss": 3.469674885732927,
      "tokens_seen": 1747845120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784444444444445,
      "loss": 1.3806,
      "theoretical_loss": 3.4696637492393236,
      "tokens_seen": 1747910656
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047842424242424244,
      "loss": 1.433,
      "theoretical_loss": 3.4696526132801724,
      "tokens_seen": 1747976192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004784040404040404,
      "loss": 1.2802,
      "theoretical_loss": 3.4696414778554274,
      "tokens_seen": 1748041728
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783838383838384,
      "loss": 1.3713,
      "theoretical_loss": 3.4696303429650435,
      "tokens_seen": 1748107264
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783636363636364,
      "loss": 1.288,
      "theoretical_loss": 3.469619208608974,
      "tokens_seen": 1748172800
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047834343434343437,
      "loss": 1.5354,
      "theoretical_loss": 3.469608074787174,
      "tokens_seen": 1748238336
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047832323232323233,
      "loss": 1.3335,
      "theoretical_loss": 3.4695969414995975,
      "tokens_seen": 1748303872
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004783030303030303,
      "loss": 1.2709,
      "theoretical_loss": 3.4695858087461993,
      "tokens_seen": 1748369408
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.356671541929245,
      "objective/train/avg_token_score": 0.015191304497420788,
      "objective/train/docs_used": 990185,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0629782676696777,
      "objective/train/original_loss": 4.2517290115356445,
      "objective/train/theoretical_loss": 3.469574676526933,
      "objective/train/tokens_used": 107294176,
      "objective/train/unlikelihood_loss": 0.06087062880396843,
      "theoretical_loss": 3.469574676526933,
      "tokens_seen": 1748434944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782828282828283,
      "loss": 1.27,
      "theoretical_loss": 3.469574676526933,
      "tokens_seen": 1748434944
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782626262626263,
      "loss": 1.3251,
      "theoretical_loss": 3.469563544841754,
      "tokens_seen": 1748500480
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047824242424242426,
      "loss": 1.374,
      "theoretical_loss": 3.4695524136906153,
      "tokens_seen": 1748566016
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782222222222222,
      "loss": 1.3274,
      "theoretical_loss": 3.4695412830734726,
      "tokens_seen": 1748631552
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004782020202020202,
      "loss": 1.5862,
      "theoretical_loss": 3.46953015299028,
      "tokens_seen": 1748697088
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047818181818181824,
      "loss": 1.3369,
      "theoretical_loss": 3.469519023440991,
      "tokens_seen": 1748762624
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781616161616162,
      "loss": 1.5887,
      "theoretical_loss": 3.4695078944255604,
      "tokens_seen": 1748828160
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047814141414141416,
      "loss": 1.4068,
      "theoretical_loss": 3.4694967659439433,
      "tokens_seen": 1748893696
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004781212121212121,
      "loss": 1.4508,
      "theoretical_loss": 3.469485637996093,
      "tokens_seen": 1748959232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047810101010101007,
      "loss": 1.3742,
      "theoretical_loss": 3.4694745105819647,
      "tokens_seen": 1749024768
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047808080808080814,
      "loss": 1.255,
      "theoretical_loss": 3.4694633837015125,
      "tokens_seen": 1749090304
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004780606060606061,
      "loss": 1.3452,
      "theoretical_loss": 3.469452257354691,
      "tokens_seen": 1749155840
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047804040404040405,
      "loss": 1.4228,
      "theoretical_loss": 3.469441131541454,
      "tokens_seen": 1749221376
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000478020202020202,
      "loss": 1.3645,
      "theoretical_loss": 3.4694300062617565,
      "tokens_seen": 1749286912
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047799999999999996,
      "loss": 1.2568,
      "theoretical_loss": 3.469418881515553,
      "tokens_seen": 1749352448
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047797979797979803,
      "loss": 1.2278,
      "theoretical_loss": 3.4694077573027977,
      "tokens_seen": 1749417984
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.000477959595959596,
      "loss": 1.3498,
      "theoretical_loss": 3.469396633623445,
      "tokens_seen": 1749483520
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047793939393939394,
      "loss": 1.201,
      "theoretical_loss": 3.4693855104774496,
      "tokens_seen": 1749549056
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004779191919191919,
      "loss": 1.4381,
      "theoretical_loss": 3.4693743878647654,
      "tokens_seen": 1749614592
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778989898989899,
      "loss": 1.3039,
      "theoretical_loss": 3.469363265785348,
      "tokens_seen": 1749680128
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778787878787879,
      "loss": 1.6264,
      "theoretical_loss": 3.46935214423915,
      "tokens_seen": 1749745664
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778585858585859,
      "loss": 1.6328,
      "theoretical_loss": 3.4693410232261277,
      "tokens_seen": 1749811200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047783838383838384,
      "loss": 1.5476,
      "theoretical_loss": 3.4693299027462343,
      "tokens_seen": 1749876736
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004778181818181818,
      "loss": 1.2902,
      "theoretical_loss": 3.469318782799425,
      "tokens_seen": 1749942272
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777979797979798,
      "loss": 1.4278,
      "theoretical_loss": 3.4693076633856545,
      "tokens_seen": 1750007808
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.4995723366737366,
      "objective/train/avg_token_score": 0.01636279560625553,
      "objective/train/docs_used": 990903,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7601079940795898,
      "objective/train/original_loss": 4.128569602966309,
      "objective/train/theoretical_loss": 3.4692965445048767,
      "objective/train/tokens_used": 108932576,
      "objective/train/unlikelihood_loss": 0.05954797938466072,
      "theoretical_loss": 3.4692965445048767,
      "tokens_seen": 1750073344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004777777777777778,
      "loss": 1.532,
      "theoretical_loss": 3.4692965445048767,
      "tokens_seen": 1750073344
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047775757575757577,
      "loss": 1.3282,
      "theoretical_loss": 3.469285426157046,
      "tokens_seen": 1750138880
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047773737373737373,
      "loss": 1.3803,
      "theoretical_loss": 3.469274308342117,
      "tokens_seen": 1750204416
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047771717171717174,
      "loss": 1.5482,
      "theoretical_loss": 3.469263191060045,
      "tokens_seen": 1750269952
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776969696969697,
      "loss": 1.6532,
      "theoretical_loss": 3.4692520743107833,
      "tokens_seen": 1750335488
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776767676767677,
      "loss": 1.4167,
      "theoretical_loss": 3.469240958094287,
      "tokens_seen": 1750401024
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047765656565656566,
      "loss": 1.193,
      "theoretical_loss": 3.469229842410511,
      "tokens_seen": 1750466560
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004776363636363636,
      "loss": 1.3584,
      "theoretical_loss": 3.4692187272594093,
      "tokens_seen": 1750532096
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047761616161616163,
      "loss": 1.439,
      "theoretical_loss": 3.469207612640936,
      "tokens_seen": 1750597632
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775959595959596,
      "loss": 1.2879,
      "theoretical_loss": 3.469196498555047,
      "tokens_seen": 1750663168
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775757575757576,
      "loss": 1.3754,
      "theoretical_loss": 3.469185385001696,
      "tokens_seen": 1750728704
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047755555555555556,
      "loss": 1.2025,
      "theoretical_loss": 3.4691742719808376,
      "tokens_seen": 1750794240
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047753535353535357,
      "loss": 1.5471,
      "theoretical_loss": 3.469163159492426,
      "tokens_seen": 1750859776
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004775151515151515,
      "loss": 1.2861,
      "theoretical_loss": 3.469152047536416,
      "tokens_seen": 1750925312
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774949494949495,
      "loss": 1.4033,
      "theoretical_loss": 3.4691409361127628,
      "tokens_seen": 1750990848
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774747474747475,
      "loss": 1.3833,
      "theoretical_loss": 3.46912982522142,
      "tokens_seen": 1751056384
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047745454545454545,
      "loss": 1.5292,
      "theoretical_loss": 3.469118714862343,
      "tokens_seen": 1751121920
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047743434343434346,
      "loss": 1.3997,
      "theoretical_loss": 3.469107605035486,
      "tokens_seen": 1751187456
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004774141414141414,
      "loss": 1.3918,
      "theoretical_loss": 3.4690964957408035,
      "tokens_seen": 1751252992
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773939393939394,
      "loss": 1.5265,
      "theoretical_loss": 3.46908538697825,
      "tokens_seen": 1751318528
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773737373737374,
      "loss": 1.2903,
      "theoretical_loss": 3.4690742787477804,
      "tokens_seen": 1751384064
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773535353535354,
      "loss": 1.4772,
      "theoretical_loss": 3.4690631710493496,
      "tokens_seen": 1751449600
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047733333333333335,
      "loss": 1.4411,
      "theoretical_loss": 3.469052063882912,
      "tokens_seen": 1751515136
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0004773131313131313,
      "loss": 1.4693,
      "theoretical_loss": 3.4690409572484215,
      "tokens_seen": 1751580672
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00047729292929292927,
      "loss": 1.6307,
      "theoretical_loss": 3.469029851145833,
      "tokens_seen": 1751646208
    },
    {
      "epoch": 0.05,
      "objective/train/aligned_token_count": 0.35160067677497864,
      "objective/train/avg_token_score": 0.0432390458881855,
      "objective/train/docs_used": 991641,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.203386664390564,
      "objective/train/original_loss": 4.336913108825684,
      "objective/train/theoretical_loss": 3.469018745575102,
      "objective/train/tokens_used": 110570976,
      "objective/train/unlikelihood_loss": 0.058832619339227676,
      "theoretical_loss": 3.469018745575102,
      "tokens_seen": 1751711744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772727272727273,
      "loss": 1.3743,
      "theoretical_loss": 3.469018745575102,
      "tokens_seen": 1751711744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772525252525253,
      "loss": 1.4335,
      "theoretical_loss": 3.469007640536182,
      "tokens_seen": 1751777280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047723232323232325,
      "loss": 1.5358,
      "theoretical_loss": 3.4689965360290285,
      "tokens_seen": 1751842816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004772121212121212,
      "loss": 1.2956,
      "theoretical_loss": 3.4689854320535956,
      "tokens_seen": 1751908352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047719191919191916,
      "loss": 1.5891,
      "theoretical_loss": 3.4689743286098382,
      "tokens_seen": 1751973888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771717171717172,
      "loss": 1.5675,
      "theoretical_loss": 3.468963225697711,
      "tokens_seen": 1752039424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771515151515152,
      "loss": 1.3356,
      "theoretical_loss": 3.4689521233171687,
      "tokens_seen": 1752104960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047713131313131314,
      "loss": 1.4772,
      "theoretical_loss": 3.468941021468166,
      "tokens_seen": 1752170496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004771111111111111,
      "loss": 1.2844,
      "theoretical_loss": 3.468929920150657,
      "tokens_seen": 1752236032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047709090909090905,
      "loss": 1.5272,
      "theoretical_loss": 3.468918819364597,
      "tokens_seen": 1752301568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770707070707071,
      "loss": 1.4227,
      "theoretical_loss": 3.4689077191099402,
      "tokens_seen": 1752367104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004770505050505051,
      "loss": 1.2629,
      "theoretical_loss": 3.4688966193866415,
      "tokens_seen": 1752432640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047703030303030303,
      "loss": 1.4331,
      "theoretical_loss": 3.468885520194656,
      "tokens_seen": 1752498176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000477010101010101,
      "loss": 1.4719,
      "theoretical_loss": 3.468874421533938,
      "tokens_seen": 1752563712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000476989898989899,
      "loss": 1.4844,
      "theoretical_loss": 3.4688633234044426,
      "tokens_seen": 1752629248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000476969696969697,
      "loss": 1.6083,
      "theoretical_loss": 3.468852225806123,
      "tokens_seen": 1752694784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047694949494949497,
      "loss": 1.3576,
      "theoretical_loss": 3.4688411287389362,
      "tokens_seen": 1752760320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004769292929292929,
      "loss": 1.5143,
      "theoretical_loss": 3.4688300322028356,
      "tokens_seen": 1752825856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004769090909090909,
      "loss": 1.4537,
      "theoretical_loss": 3.468818936197776,
      "tokens_seen": 1752891392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768888888888889,
      "loss": 1.5708,
      "theoretical_loss": 3.4688078407237124,
      "tokens_seen": 1752956928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768686868686869,
      "loss": 1.4414,
      "theoretical_loss": 3.4687967457805993,
      "tokens_seen": 1753022464
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047684848484848486,
      "loss": 1.5843,
      "theoretical_loss": 3.4687856513683917,
      "tokens_seen": 1753088000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004768282828282828,
      "loss": 1.5725,
      "theoretical_loss": 3.468774557487044,
      "tokens_seen": 1753153536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047680808080808083,
      "loss": 1.2515,
      "theoretical_loss": 3.4687634641365115,
      "tokens_seen": 1753219072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767878787878788,
      "loss": 1.7292,
      "theoretical_loss": 3.4687523713167483,
      "tokens_seen": 1753284608
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.5292032957077026,
      "objective/train/avg_token_score": 0.022525779902935028,
      "objective/train/docs_used": 992944,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7439391613006592,
      "objective/train/original_loss": 3.8954503536224365,
      "objective/train/theoretical_loss": 3.468741279027709,
      "objective/train/tokens_used": 112209376,
      "objective/train/unlikelihood_loss": 0.06822893768548965,
      "theoretical_loss": 3.468741279027709,
      "tokens_seen": 1753350144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767676767676768,
      "loss": 1.5733,
      "theoretical_loss": 3.468741279027709,
      "tokens_seen": 1753350144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047674747474747475,
      "loss": 1.3831,
      "theoretical_loss": 3.4687301872693497,
      "tokens_seen": 1753415680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767272727272727,
      "loss": 1.2289,
      "theoretical_loss": 3.4687190960416245,
      "tokens_seen": 1753481216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004767070707070707,
      "loss": 1.2174,
      "theoretical_loss": 3.468708005344487,
      "tokens_seen": 1753546752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047668686868686873,
      "loss": 1.4894,
      "theoretical_loss": 3.468696915177894,
      "tokens_seen": 1753612288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766666666666667,
      "loss": 1.7117,
      "theoretical_loss": 3.4686858255417987,
      "tokens_seen": 1753677824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047664646464646465,
      "loss": 1.5382,
      "theoretical_loss": 3.4686747364361565,
      "tokens_seen": 1753743360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047662626262626266,
      "loss": 1.6347,
      "theoretical_loss": 3.4686636478609225,
      "tokens_seen": 1753808896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004766060606060606,
      "loss": 1.3016,
      "theoretical_loss": 3.4686525598160514,
      "tokens_seen": 1753874432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765858585858586,
      "loss": 1.2402,
      "theoretical_loss": 3.4686414723014973,
      "tokens_seen": 1753939968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765656565656566,
      "loss": 1.6342,
      "theoretical_loss": 3.4686303853172165,
      "tokens_seen": 1754005504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047654545454545454,
      "loss": 1.4743,
      "theoretical_loss": 3.468619298863162,
      "tokens_seen": 1754071040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047652525252525255,
      "loss": 1.5816,
      "theoretical_loss": 3.46860821293929,
      "tokens_seen": 1754136576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004765050505050505,
      "loss": 1.3497,
      "theoretical_loss": 3.4685971275455545,
      "tokens_seen": 1754202112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764848484848485,
      "loss": 1.4686,
      "theoretical_loss": 3.4685860426819106,
      "tokens_seen": 1754267648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764646464646465,
      "loss": 1.5528,
      "theoretical_loss": 3.468574958348314,
      "tokens_seen": 1754333184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764444444444445,
      "loss": 1.4893,
      "theoretical_loss": 3.468563874544718,
      "tokens_seen": 1754398720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047642424242424244,
      "loss": 1.5118,
      "theoretical_loss": 3.4685527912710787,
      "tokens_seen": 1754464256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004764040404040404,
      "loss": 1.5134,
      "theoretical_loss": 3.4685417085273507,
      "tokens_seen": 1754529792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763838383838384,
      "loss": 1.6654,
      "theoretical_loss": 3.468530626313488,
      "tokens_seen": 1754595328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047636363636363637,
      "loss": 1.5252,
      "theoretical_loss": 3.468519544629447,
      "tokens_seen": 1754660864
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763434343434344,
      "loss": 1.2085,
      "theoretical_loss": 3.468508463475181,
      "tokens_seen": 1754726400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047632323232323233,
      "loss": 1.4196,
      "theoretical_loss": 3.4684973828506465,
      "tokens_seen": 1754791936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004763030303030303,
      "loss": 1.6456,
      "theoretical_loss": 3.468486302755797,
      "tokens_seen": 1754857472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762828282828283,
      "loss": 1.5107,
      "theoretical_loss": 3.4684752231905875,
      "tokens_seen": 1754923008
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.3735337257385254,
      "objective/train/avg_token_score": 0.013170573860406876,
      "objective/train/docs_used": 993452,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3061128854751587,
      "objective/train/original_loss": 4.716046333312988,
      "objective/train/theoretical_loss": 3.468464144154974,
      "objective/train/tokens_used": 113847776,
      "objective/train/unlikelihood_loss": 0.05805331468582153,
      "theoretical_loss": 3.468464144154974,
      "tokens_seen": 1754988544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762626262626263,
      "loss": 1.1821,
      "theoretical_loss": 3.468464144154974,
      "tokens_seen": 1754988544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047624242424242427,
      "loss": 1.3812,
      "theoretical_loss": 3.468453065648911,
      "tokens_seen": 1755054080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762222222222222,
      "loss": 1.5309,
      "theoretical_loss": 3.468441987672353,
      "tokens_seen": 1755119616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004762020202020202,
      "loss": 1.5255,
      "theoretical_loss": 3.468430910225255,
      "tokens_seen": 1755185152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761818181818182,
      "loss": 1.4808,
      "theoretical_loss": 3.468419833307572,
      "tokens_seen": 1755250688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761616161616162,
      "loss": 1.5992,
      "theoretical_loss": 3.468408756919259,
      "tokens_seen": 1755316224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047614141414141416,
      "loss": 1.4779,
      "theoretical_loss": 3.468397681060271,
      "tokens_seen": 1755381760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761212121212121,
      "loss": 1.4515,
      "theoretical_loss": 3.4683866057305623,
      "tokens_seen": 1755447296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004761010101010101,
      "loss": 1.3284,
      "theoretical_loss": 3.468375530930089,
      "tokens_seen": 1755512832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760808080808081,
      "loss": 1.4936,
      "theoretical_loss": 3.468364456658805,
      "tokens_seen": 1755578368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004760606060606061,
      "loss": 1.4289,
      "theoretical_loss": 3.4683533829166664,
      "tokens_seen": 1755643904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047604040404040406,
      "loss": 1.5279,
      "theoretical_loss": 3.468342309703627,
      "tokens_seen": 1755709440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000476020202020202,
      "loss": 1.1404,
      "theoretical_loss": 3.4683312370196426,
      "tokens_seen": 1755774976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047599999999999997,
      "loss": 1.6458,
      "theoretical_loss": 3.4683201648646675,
      "tokens_seen": 1755840512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047597979797979803,
      "loss": 1.3894,
      "theoretical_loss": 3.4683090932386573,
      "tokens_seen": 1755906048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475959595959596,
      "loss": 1.5657,
      "theoretical_loss": 3.4682980221415662,
      "tokens_seen": 1755971584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047593939393939395,
      "loss": 1.2969,
      "theoretical_loss": 3.46828695157335,
      "tokens_seen": 1756037120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004759191919191919,
      "loss": 1.4629,
      "theoretical_loss": 3.4682758815339634,
      "tokens_seen": 1756102656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047589898989898986,
      "loss": 1.4144,
      "theoretical_loss": 3.4682648120233615,
      "tokens_seen": 1756168192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047587878787878793,
      "loss": 1.2853,
      "theoretical_loss": 3.4682537430414992,
      "tokens_seen": 1756233728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004758585858585859,
      "loss": 1.1362,
      "theoretical_loss": 3.468242674588331,
      "tokens_seen": 1756299264
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047583838383838384,
      "loss": 1.4003,
      "theoretical_loss": 3.468231606663813,
      "tokens_seen": 1756364800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004758181818181818,
      "loss": 1.3655,
      "theoretical_loss": 3.4682205392678993,
      "tokens_seen": 1756430336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757979797979798,
      "loss": 1.2837,
      "theoretical_loss": 3.4682094724005457,
      "tokens_seen": 1756495872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757777777777778,
      "loss": 1.2286,
      "theoretical_loss": 3.4681984060617066,
      "tokens_seen": 1756561408
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.4603494703769684,
      "objective/train/avg_token_score": 0.007801194675266743,
      "objective/train/docs_used": 994709,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4813201427459717,
      "objective/train/original_loss": 3.937505006790161,
      "objective/train/theoretical_loss": 3.4681873402513372,
      "objective/train/tokens_used": 115486176,
      "objective/train/unlikelihood_loss": 0.07865725457668304,
      "theoretical_loss": 3.4681873402513372,
      "tokens_seen": 1756626944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757575757575758,
      "loss": 1.5432,
      "theoretical_loss": 3.4681873402513372,
      "tokens_seen": 1756626944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047573737373737373,
      "loss": 1.328,
      "theoretical_loss": 3.4681762749693927,
      "tokens_seen": 1756692480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004757171717171717,
      "loss": 1.3619,
      "theoretical_loss": 3.4681652102158282,
      "tokens_seen": 1756758016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756969696969697,
      "loss": 1.2638,
      "theoretical_loss": 3.4681541459905985,
      "tokens_seen": 1756823552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756767676767677,
      "loss": 1.5705,
      "theoretical_loss": 3.468143082293659,
      "tokens_seen": 1756889088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047565656565656567,
      "loss": 1.3297,
      "theoretical_loss": 3.4681320191249645,
      "tokens_seen": 1756954624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004756363636363636,
      "loss": 1.2614,
      "theoretical_loss": 3.46812095648447,
      "tokens_seen": 1757020160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047561616161616164,
      "loss": 1.3043,
      "theoretical_loss": 3.468109894372131,
      "tokens_seen": 1757085696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755959595959596,
      "loss": 1.4471,
      "theoretical_loss": 3.468098832787902,
      "tokens_seen": 1757151232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755757575757576,
      "loss": 1.4198,
      "theoretical_loss": 3.4680877717317387,
      "tokens_seen": 1757216768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047555555555555556,
      "loss": 1.4285,
      "theoretical_loss": 3.4680767112035955,
      "tokens_seen": 1757282304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004755353535353535,
      "loss": 1.3098,
      "theoretical_loss": 3.4680656512034282,
      "tokens_seen": 1757347840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047551515151515153,
      "loss": 1.4135,
      "theoretical_loss": 3.4680545917311916,
      "tokens_seen": 1757413376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754949494949495,
      "loss": 1.2951,
      "theoretical_loss": 3.468043532786841,
      "tokens_seen": 1757478912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754747474747475,
      "loss": 1.4952,
      "theoretical_loss": 3.4680324743703315,
      "tokens_seen": 1757544448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047545454545454545,
      "loss": 1.3478,
      "theoretical_loss": 3.4680214164816174,
      "tokens_seen": 1757609984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047543434343434347,
      "loss": 1.4044,
      "theoretical_loss": 3.468010359120655,
      "tokens_seen": 1757675520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004754141414141414,
      "loss": 1.3726,
      "theoretical_loss": 3.467999302287399,
      "tokens_seen": 1757741056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753939393939394,
      "loss": 1.5165,
      "theoretical_loss": 3.4679882459818043,
      "tokens_seen": 1757806592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753737373737374,
      "loss": 1.6779,
      "theoretical_loss": 3.4679771902038263,
      "tokens_seen": 1757872128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047535353535353535,
      "loss": 1.3127,
      "theoretical_loss": 3.4679661349534197,
      "tokens_seen": 1757937664
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047533333333333336,
      "loss": 1.4477,
      "theoretical_loss": 3.4679550802305408,
      "tokens_seen": 1758003200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004753131313131313,
      "loss": 1.3832,
      "theoretical_loss": 3.467944026035144,
      "tokens_seen": 1758068736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047529292929292927,
      "loss": 1.3038,
      "theoretical_loss": 3.4679329723671835,
      "tokens_seen": 1758134272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752727272727273,
      "loss": 1.3008,
      "theoretical_loss": 3.467921919226616,
      "tokens_seen": 1758199808
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.37341153621673584,
      "objective/train/avg_token_score": 0.01459649857133627,
      "objective/train/docs_used": 995465,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.182531476020813,
      "objective/train/original_loss": 4.433279991149902,
      "objective/train/theoretical_loss": 3.4679108666133964,
      "objective/train/tokens_used": 117124576,
      "objective/train/unlikelihood_loss": 0.05860118195414543,
      "theoretical_loss": 3.4679108666133964,
      "tokens_seen": 1758265344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752525252525253,
      "loss": 1.3201,
      "theoretical_loss": 3.4679108666133964,
      "tokens_seen": 1758265344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047523232323232325,
      "loss": 1.4253,
      "theoretical_loss": 3.4678998145274793,
      "tokens_seen": 1758330880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004752121212121212,
      "loss": 1.2983,
      "theoretical_loss": 3.467888762968821,
      "tokens_seen": 1758396416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047519191919191916,
      "loss": 1.4404,
      "theoretical_loss": 3.467877711937375,
      "tokens_seen": 1758461952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751717171717172,
      "loss": 1.4583,
      "theoretical_loss": 3.4678666614330975,
      "tokens_seen": 1758527488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751515151515152,
      "loss": 1.4141,
      "theoretical_loss": 3.467855611455944,
      "tokens_seen": 1758593024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047513131313131314,
      "loss": 1.527,
      "theoretical_loss": 3.467844562005869,
      "tokens_seen": 1758658560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004751111111111111,
      "loss": 1.3646,
      "theoretical_loss": 3.4678335130828284,
      "tokens_seen": 1758724096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047509090909090906,
      "loss": 1.2391,
      "theoretical_loss": 3.4678224646867766,
      "tokens_seen": 1758789632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750707070707071,
      "loss": 1.281,
      "theoretical_loss": 3.4678114168176695,
      "tokens_seen": 1758855168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004750505050505051,
      "loss": 1.392,
      "theoretical_loss": 3.467800369475462,
      "tokens_seen": 1758920704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047503030303030304,
      "loss": 1.2578,
      "theoretical_loss": 3.46778932266011,
      "tokens_seen": 1758986240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000475010101010101,
      "loss": 1.4468,
      "theoretical_loss": 3.467778276371568,
      "tokens_seen": 1759051776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474989898989899,
      "loss": 1.3945,
      "theoretical_loss": 3.467767230609791,
      "tokens_seen": 1759117312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474969696969697,
      "loss": 1.166,
      "theoretical_loss": 3.4677561853747356,
      "tokens_seen": 1759182848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047494949494949497,
      "loss": 1.5393,
      "theoretical_loss": 3.4677451406663558,
      "tokens_seen": 1759248384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047492929292929293,
      "loss": 1.6032,
      "theoretical_loss": 3.467734096484607,
      "tokens_seen": 1759313920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004749090909090909,
      "loss": 1.4508,
      "theoretical_loss": 3.467723052829445,
      "tokens_seen": 1759379456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047488888888888895,
      "loss": 1.5765,
      "theoretical_loss": 3.4677120097008247,
      "tokens_seen": 1759444992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748686868686869,
      "loss": 1.3183,
      "theoretical_loss": 3.4677009670987013,
      "tokens_seen": 1759510528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047484848484848487,
      "loss": 1.3187,
      "theoretical_loss": 3.4676899250230306,
      "tokens_seen": 1759576064
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748282828282828,
      "loss": 1.3925,
      "theoretical_loss": 3.4676788834737677,
      "tokens_seen": 1759641600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004748080808080808,
      "loss": 1.4587,
      "theoretical_loss": 3.467667842450868,
      "tokens_seen": 1759707136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047478787878787884,
      "loss": 1.4543,
      "theoretical_loss": 3.467656801954286,
      "tokens_seen": 1759772672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747676767676768,
      "loss": 1.3954,
      "theoretical_loss": 3.4676457619839773,
      "tokens_seen": 1759838208
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.3027859330177307,
      "objective/train/avg_token_score": 0.026031043380498886,
      "objective/train/docs_used": 996601,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0335179567337036,
      "objective/train/original_loss": 4.650774955749512,
      "objective/train/theoretical_loss": 3.4676347225398985,
      "objective/train/tokens_used": 118762976,
      "objective/train/unlikelihood_loss": 0.07695260643959045,
      "theoretical_loss": 3.4676347225398985,
      "tokens_seen": 1759903744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047474747474747476,
      "loss": 1.1674,
      "theoretical_loss": 3.4676347225398985,
      "tokens_seen": 1759903744
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004747272727272727,
      "loss": 1.283,
      "theoretical_loss": 3.467623683622003,
      "tokens_seen": 1759969280
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047470707070707067,
      "loss": 1.4078,
      "theoretical_loss": 3.467612645230248,
      "tokens_seen": 1760034816
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047468686868686874,
      "loss": 1.3631,
      "theoretical_loss": 3.4676016073645872,
      "tokens_seen": 1760100352
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746666666666667,
      "loss": 1.4091,
      "theoretical_loss": 3.467590570024977,
      "tokens_seen": 1760165888
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047464646464646465,
      "loss": 1.4418,
      "theoretical_loss": 3.4675795332113726,
      "tokens_seen": 1760231424
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746262626262626,
      "loss": 1.3409,
      "theoretical_loss": 3.467568496923729,
      "tokens_seen": 1760296960
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004746060606060606,
      "loss": 1.518,
      "theoretical_loss": 3.4675574611620013,
      "tokens_seen": 1760362496
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047458585858585863,
      "loss": 1.4428,
      "theoretical_loss": 3.4675464259261455,
      "tokens_seen": 1760428032
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745656565656566,
      "loss": 1.3315,
      "theoretical_loss": 3.467535391216116,
      "tokens_seen": 1760493568
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047454545454545454,
      "loss": 1.5314,
      "theoretical_loss": 3.46752435703187,
      "tokens_seen": 1760559104
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745252525252525,
      "loss": 1.5786,
      "theoretical_loss": 3.4675133233733613,
      "tokens_seen": 1760624640
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004745050505050505,
      "loss": 1.5052,
      "theoretical_loss": 3.467502290240546,
      "tokens_seen": 1760690176
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004744848484848485,
      "loss": 1.5072,
      "theoretical_loss": 3.4674912576333785,
      "tokens_seen": 1760755712
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004744646464646465,
      "loss": 1.3799,
      "theoretical_loss": 3.4674802255518156,
      "tokens_seen": 1760821248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047444444444444444,
      "loss": 1.359,
      "theoretical_loss": 3.467469193995812,
      "tokens_seen": 1760886784
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047442424242424245,
      "loss": 1.5932,
      "theoretical_loss": 3.4674581629653227,
      "tokens_seen": 1760952320
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004744040404040404,
      "loss": 1.2436,
      "theoretical_loss": 3.467447132460304,
      "tokens_seen": 1761017856
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743838383838384,
      "loss": 1.2935,
      "theoretical_loss": 3.467436102480711,
      "tokens_seen": 1761083392
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047436363636363637,
      "loss": 1.6615,
      "theoretical_loss": 3.4674250730264986,
      "tokens_seen": 1761148928
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047434343434343433,
      "loss": 1.4143,
      "theoretical_loss": 3.4674140440976227,
      "tokens_seen": 1761214464
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047432323232323234,
      "loss": 1.5773,
      "theoretical_loss": 3.467403015694039,
      "tokens_seen": 1761280000
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004743030303030303,
      "loss": 1.4823,
      "theoretical_loss": 3.467391987815702,
      "tokens_seen": 1761345536
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742828282828283,
      "loss": 1.282,
      "theoretical_loss": 3.4673809604625676,
      "tokens_seen": 1761411072
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047426262626262626,
      "loss": 1.5621,
      "theoretical_loss": 3.4673699336345916,
      "tokens_seen": 1761476608
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.4690249264240265,
      "objective/train/avg_token_score": 0.015736954286694527,
      "objective/train/docs_used": 997243,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6199918985366821,
      "objective/train/original_loss": 3.9470674991607666,
      "objective/train/theoretical_loss": 3.4673589073317297,
      "objective/train/tokens_used": 120401376,
      "objective/train/unlikelihood_loss": 0.07392548024654388,
      "theoretical_loss": 3.4673589073317297,
      "tokens_seen": 1761542144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742424242424243,
      "loss": 1.4835,
      "theoretical_loss": 3.4673589073317297,
      "tokens_seen": 1761542144
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047422222222222223,
      "loss": 1.4607,
      "theoretical_loss": 3.4673478815539363,
      "tokens_seen": 1761607680
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004742020202020202,
      "loss": 1.655,
      "theoretical_loss": 3.4673368563011677,
      "tokens_seen": 1761673216
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741818181818182,
      "loss": 1.4306,
      "theoretical_loss": 3.467325831573379,
      "tokens_seen": 1761738752
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047416161616161616,
      "loss": 1.4634,
      "theoretical_loss": 3.467314807370526,
      "tokens_seen": 1761804288
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047414141414141417,
      "loss": 1.4165,
      "theoretical_loss": 3.4673037836925635,
      "tokens_seen": 1761869824
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741212121212121,
      "loss": 1.4309,
      "theoretical_loss": 3.467292760539448,
      "tokens_seen": 1761935360
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004741010101010101,
      "loss": 1.3544,
      "theoretical_loss": 3.467281737911134,
      "tokens_seen": 1762000896
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004740808080808081,
      "loss": 1.4597,
      "theoretical_loss": 3.4672707158075777,
      "tokens_seen": 1762066432
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004740606060606061,
      "loss": 1.5272,
      "theoretical_loss": 3.467259694228734,
      "tokens_seen": 1762131968
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047404040404040406,
      "loss": 1.3645,
      "theoretical_loss": 3.467248673174559,
      "tokens_seen": 1762197504
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474020202020202,
      "loss": 1.2698,
      "theoretical_loss": 3.467237652645008,
      "tokens_seen": 1762263040
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000474,
      "loss": 1.1518,
      "theoretical_loss": 3.4672266326400365,
      "tokens_seen": 1762328576
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000473979797979798,
      "loss": 1.446,
      "theoretical_loss": 3.4672156131595995,
      "tokens_seen": 1762394112
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000473959595959596,
      "loss": 1.4405,
      "theoretical_loss": 3.4672045942036536,
      "tokens_seen": 1762459648
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047393939393939395,
      "loss": 1.2889,
      "theoretical_loss": 3.4671935757721535,
      "tokens_seen": 1762525184
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004739191919191919,
      "loss": 1.389,
      "theoretical_loss": 3.467182557865055,
      "tokens_seen": 1762590720
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047389898989898987,
      "loss": 1.4165,
      "theoretical_loss": 3.4671715404823136,
      "tokens_seen": 1762656256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047387878787878793,
      "loss": 1.5323,
      "theoretical_loss": 3.4671605236238854,
      "tokens_seen": 1762721792
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738585858585859,
      "loss": 1.4547,
      "theoretical_loss": 3.467149507289725,
      "tokens_seen": 1762787328
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047383838383838385,
      "loss": 1.3971,
      "theoretical_loss": 3.4671384914797887,
      "tokens_seen": 1762852864
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004738181818181818,
      "loss": 1.4061,
      "theoretical_loss": 3.4671274761940314,
      "tokens_seen": 1762918400
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047379797979797976,
      "loss": 1.2906,
      "theoretical_loss": 3.4671164614324095,
      "tokens_seen": 1762983936
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737777777777778,
      "loss": 1.4725,
      "theoretical_loss": 3.4671054471948777,
      "tokens_seen": 1763049472
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737575757575758,
      "loss": 1.2618,
      "theoretical_loss": 3.467094433481392,
      "tokens_seen": 1763115008
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.3298509418964386,
      "objective/train/avg_token_score": 0.013608359731733799,
      "objective/train/docs_used": 998400,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.140276312828064,
      "objective/train/original_loss": 4.645390033721924,
      "objective/train/theoretical_loss": 3.4670834202919085,
      "objective/train/tokens_used": 122039776,
      "objective/train/unlikelihood_loss": 0.06240946799516678,
      "theoretical_loss": 3.4670834202919085,
      "tokens_seen": 1763180544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047373737373737374,
      "loss": 1.3174,
      "theoretical_loss": 3.4670834202919085,
      "tokens_seen": 1763180544
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004737171717171717,
      "loss": 1.6851,
      "theoretical_loss": 3.4670724076263815,
      "tokens_seen": 1763246080
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736969696969697,
      "loss": 1.3992,
      "theoretical_loss": 3.4670613954847678,
      "tokens_seen": 1763311616
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736767676767677,
      "loss": 1.482,
      "theoretical_loss": 3.467050383867023,
      "tokens_seen": 1763377152
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736565656565657,
      "loss": 1.4294,
      "theoretical_loss": 3.467039372773102,
      "tokens_seen": 1763442688
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047363636363636363,
      "loss": 1.4823,
      "theoretical_loss": 3.467028362202961,
      "tokens_seen": 1763508224
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004736161616161616,
      "loss": 1.4347,
      "theoretical_loss": 3.467017352156555,
      "tokens_seen": 1763573760
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004735959595959596,
      "loss": 1.5324,
      "theoretical_loss": 3.46700634263384,
      "tokens_seen": 1763639296
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004735757575757576,
      "loss": 1.2494,
      "theoretical_loss": 3.466995333634772,
      "tokens_seen": 1763704832
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047355555555555557,
      "loss": 1.3624,
      "theoretical_loss": 3.4669843251593058,
      "tokens_seen": 1763770368
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004735353535353535,
      "loss": 1.3935,
      "theoretical_loss": 3.4669733172073975,
      "tokens_seen": 1763835904
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047351515151515154,
      "loss": 1.4901,
      "theoretical_loss": 3.466962309779003,
      "tokens_seen": 1763901440
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734949494949495,
      "loss": 1.3198,
      "theoretical_loss": 3.4669513028740777,
      "tokens_seen": 1763966976
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734747474747475,
      "loss": 1.2222,
      "theoretical_loss": 3.466940296492577,
      "tokens_seen": 1764032512
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047345454545454546,
      "loss": 1.4412,
      "theoretical_loss": 3.466929290634457,
      "tokens_seen": 1764098048
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004734343434343434,
      "loss": 1.2884,
      "theoretical_loss": 3.466918285299673,
      "tokens_seen": 1764163584
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047341414141414143,
      "loss": 1.6086,
      "theoretical_loss": 3.466907280488181,
      "tokens_seen": 1764229120
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733939393939394,
      "loss": 1.3749,
      "theoretical_loss": 3.4668962761999365,
      "tokens_seen": 1764294656
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733737373737374,
      "loss": 1.2535,
      "theoretical_loss": 3.4668852724348955,
      "tokens_seen": 1764360192
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047335353535353535,
      "loss": 1.3333,
      "theoretical_loss": 3.4668742691930126,
      "tokens_seen": 1764425728
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047333333333333336,
      "loss": 1.2971,
      "theoretical_loss": 3.4668632664742454,
      "tokens_seen": 1764491264
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004733131313131313,
      "loss": 1.3483,
      "theoretical_loss": 3.4668522642785478,
      "tokens_seen": 1764556800
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732929292929293,
      "loss": 1.2886,
      "theoretical_loss": 3.466841262605876,
      "tokens_seen": 1764622336
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732727272727273,
      "loss": 1.4083,
      "theoretical_loss": 3.466830261456187,
      "tokens_seen": 1764687872
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047325252525252525,
      "loss": 1.4522,
      "theoretical_loss": 3.4668192608294346,
      "tokens_seen": 1764753408
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.30015885829925537,
      "objective/train/avg_token_score": 0.025842411443591118,
      "objective/train/docs_used": 998902,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9383648633956909,
      "objective/train/original_loss": 4.241371154785156,
      "objective/train/theoretical_loss": 3.4668082607255757,
      "objective/train/tokens_used": 123678176,
      "objective/train/unlikelihood_loss": 0.10302804410457611,
      "theoretical_loss": 3.4668082607255757,
      "tokens_seen": 1764818944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047323232323232326,
      "loss": 1.4569,
      "theoretical_loss": 3.4668082607255757,
      "tokens_seen": 1764818944
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004732121212121212,
      "loss": 1.4089,
      "theoretical_loss": 3.4667972611445657,
      "tokens_seen": 1764884480
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731919191919192,
      "loss": 1.4655,
      "theoretical_loss": 3.46678626208636,
      "tokens_seen": 1764950016
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731717171717172,
      "loss": 1.4081,
      "theoretical_loss": 3.466775263550915,
      "tokens_seen": 1765015552
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731515151515152,
      "loss": 1.5019,
      "theoretical_loss": 3.466764265538186,
      "tokens_seen": 1765081088
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047313131313131315,
      "loss": 1.4618,
      "theoretical_loss": 3.466753268048129,
      "tokens_seen": 1765146624
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004731111111111111,
      "loss": 1.2713,
      "theoretical_loss": 3.4667422710806997,
      "tokens_seen": 1765212160
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730909090909091,
      "loss": 1.2932,
      "theoretical_loss": 3.466731274635854,
      "tokens_seen": 1765277696
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730707070707071,
      "loss": 1.6351,
      "theoretical_loss": 3.466720278713547,
      "tokens_seen": 1765343232
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004730505050505051,
      "loss": 1.5274,
      "theoretical_loss": 3.466709283313735,
      "tokens_seen": 1765408768
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047303030303030304,
      "loss": 1.4793,
      "theoretical_loss": 3.4666982884363735,
      "tokens_seen": 1765474304
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000473010101010101,
      "loss": 1.4963,
      "theoretical_loss": 3.466687294081419,
      "tokens_seen": 1765539840
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472989898989899,
      "loss": 1.3953,
      "theoretical_loss": 3.466676300248827,
      "tokens_seen": 1765605376
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472969696969697,
      "loss": 1.3848,
      "theoretical_loss": 3.4666653069385527,
      "tokens_seen": 1765670912
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.000472949494949495,
      "loss": 1.3043,
      "theoretical_loss": 3.466654314150552,
      "tokens_seen": 1765736448
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047292929292929293,
      "loss": 1.3659,
      "theoretical_loss": 3.4666433218847814,
      "tokens_seen": 1765801984
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004729090909090909,
      "loss": 1.2315,
      "theoretical_loss": 3.466632330141196,
      "tokens_seen": 1765867520
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728888888888889,
      "loss": 1.412,
      "theoretical_loss": 3.4666213389197518,
      "tokens_seen": 1765933056
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728686868686869,
      "loss": 1.527,
      "theoretical_loss": 3.4666103482204056,
      "tokens_seen": 1765998592
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047284848484848487,
      "loss": 1.1406,
      "theoretical_loss": 3.466599358043111,
      "tokens_seen": 1766064128
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047282828282828283,
      "loss": 1.4347,
      "theoretical_loss": 3.4665883683878262,
      "tokens_seen": 1766129664
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004728080808080808,
      "loss": 1.2191,
      "theoretical_loss": 3.466577379254506,
      "tokens_seen": 1766195200
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727878787878788,
      "loss": 1.3334,
      "theoretical_loss": 3.4665663906431057,
      "tokens_seen": 1766260736
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727676767676768,
      "loss": 1.3408,
      "theoretical_loss": 3.466555402553582,
      "tokens_seen": 1766326272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047274747474747476,
      "loss": 1.424,
      "theoretical_loss": 3.466544414985891,
      "tokens_seen": 1766391808
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.5058650970458984,
      "objective/train/avg_token_score": 0.020609397441148758,
      "objective/train/docs_used": 999561,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6941487789154053,
      "objective/train/original_loss": 3.954047918319702,
      "objective/train/theoretical_loss": 3.4665334279399875,
      "objective/train/tokens_used": 125316576,
      "objective/train/unlikelihood_loss": 0.04431372508406639,
      "theoretical_loss": 3.4665334279399875,
      "tokens_seen": 1766457344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727272727272727,
      "loss": 1.5081,
      "theoretical_loss": 3.4665334279399875,
      "tokens_seen": 1766457344
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004727070707070707,
      "loss": 1.2911,
      "theoretical_loss": 3.4665224414158278,
      "tokens_seen": 1766522880
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047268686868686874,
      "loss": 1.4671,
      "theoretical_loss": 3.466511455413368,
      "tokens_seen": 1766588416
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004726666666666667,
      "loss": 1.3627,
      "theoretical_loss": 3.466500469932564,
      "tokens_seen": 1766653952
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047264646464646466,
      "loss": 1.6863,
      "theoretical_loss": 3.466489484973371,
      "tokens_seen": 1766719488
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004726262626262626,
      "loss": 1.3115,
      "theoretical_loss": 3.4664785005357457,
      "tokens_seen": 1766785024
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047260606060606057,
      "loss": 1.4088,
      "theoretical_loss": 3.4664675166196437,
      "tokens_seen": 1766850560
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047258585858585864,
      "loss": 1.2904,
      "theoretical_loss": 3.4664565332250215,
      "tokens_seen": 1766916096
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725656565656566,
      "loss": 1.5834,
      "theoretical_loss": 3.4664455503518337,
      "tokens_seen": 1766981632
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047254545454545455,
      "loss": 1.5353,
      "theoretical_loss": 3.466434568000037,
      "tokens_seen": 1767047168
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725252525252525,
      "loss": 1.2129,
      "theoretical_loss": 3.4664235861695873,
      "tokens_seen": 1767112704
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004725050505050505,
      "loss": 1.4923,
      "theoretical_loss": 3.46641260486044,
      "tokens_seen": 1767178240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047248484848484853,
      "loss": 1.6744,
      "theoretical_loss": 3.466401624072552,
      "tokens_seen": 1767243776
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724646464646465,
      "loss": 1.1128,
      "theoretical_loss": 3.4663906438058785,
      "tokens_seen": 1767309312
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047244444444444444,
      "loss": 1.5299,
      "theoretical_loss": 3.4663796640603755,
      "tokens_seen": 1767374848
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724242424242424,
      "loss": 1.4135,
      "theoretical_loss": 3.466368684835999,
      "tokens_seen": 1767440384
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004724040404040404,
      "loss": 1.3394,
      "theoretical_loss": 3.466357706132705,
      "tokens_seen": 1767505920
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723838383838384,
      "loss": 1.3015,
      "theoretical_loss": 3.46634672795045,
      "tokens_seen": 1767571456
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723636363636364,
      "loss": 1.3472,
      "theoretical_loss": 3.466335750289189,
      "tokens_seen": 1767636992
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047234343434343433,
      "loss": 1.2866,
      "theoretical_loss": 3.466324773148878,
      "tokens_seen": 1767702528
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047232323232323235,
      "loss": 1.6653,
      "theoretical_loss": 3.466313796529474,
      "tokens_seen": 1767768064
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004723030303030303,
      "loss": 1.5335,
      "theoretical_loss": 3.466302820430932,
      "tokens_seen": 1767833600
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004722828282828283,
      "loss": 1.5068,
      "theoretical_loss": 3.466291844853208,
      "tokens_seen": 1767899136
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.00047226262626262627,
      "loss": 1.3592,
      "theoretical_loss": 3.4662808697962584,
      "tokens_seen": 1767964672
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0004722424242424242,
      "loss": 1.4486,
      "theoretical_loss": 3.466269895260039,
      "tokens_seen": 1768030208
    },
    {
      "epoch": 0.06,
      "objective/train/aligned_token_count": 0.3491568863391876,
      "objective/train/avg_token_score": 0.024049675092101097,
      "objective/train/docs_used": 1000589,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.332099437713623,
      "objective/train/original_loss": 4.583323001861572,
      "objective/train/theoretical_loss": 3.466258921244506,
      "objective/train/tokens_used": 126954976,
      "objective/train/unlikelihood_loss": 0.07067796587944031,
      "theoretical_loss": 3.466258921244506,
      "tokens_seen": 1768095744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047222222222222224,
      "loss": 1.5076,
      "theoretical_loss": 3.466258921244506,
      "tokens_seen": 1768095744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004722020202020202,
      "loss": 1.2966,
      "theoretical_loss": 3.466247947749615,
      "tokens_seen": 1768161280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721818181818182,
      "loss": 1.2774,
      "theoretical_loss": 3.466236974775322,
      "tokens_seen": 1768226816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047216161616161616,
      "loss": 1.267,
      "theoretical_loss": 3.4662260023215836,
      "tokens_seen": 1768292352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721414141414142,
      "loss": 1.362,
      "theoretical_loss": 3.466215030388356,
      "tokens_seen": 1768357888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047212121212121213,
      "loss": 1.427,
      "theoretical_loss": 3.4662040589755936,
      "tokens_seen": 1768423424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004721010101010101,
      "loss": 1.2644,
      "theoretical_loss": 3.466193088083254,
      "tokens_seen": 1768488960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004720808080808081,
      "loss": 1.5104,
      "theoretical_loss": 3.466182117711293,
      "tokens_seen": 1768554496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047206060606060606,
      "loss": 1.1877,
      "theoretical_loss": 3.466171147859666,
      "tokens_seen": 1768620032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047204040404040407,
      "loss": 1.3425,
      "theoretical_loss": 3.46616017852833,
      "tokens_seen": 1768685568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000472020202020202,
      "loss": 1.1985,
      "theoretical_loss": 3.4661492097172397,
      "tokens_seen": 1768751104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000472,
      "loss": 1.4138,
      "theoretical_loss": 3.466138241426352,
      "tokens_seen": 1768816640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000471979797979798,
      "loss": 1.3324,
      "theoretical_loss": 3.466127273655623,
      "tokens_seen": 1768882176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000471959595959596,
      "loss": 1.4765,
      "theoretical_loss": 3.466116306405009,
      "tokens_seen": 1768947712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047193939393939396,
      "loss": 1.2163,
      "theoretical_loss": 3.466105339674465,
      "tokens_seen": 1769013248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004719191919191919,
      "loss": 1.319,
      "theoretical_loss": 3.466094373463948,
      "tokens_seen": 1769078784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718989898989899,
      "loss": 1.3871,
      "theoretical_loss": 3.4660834077734144,
      "tokens_seen": 1769144320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718787878787879,
      "loss": 1.3804,
      "theoretical_loss": 3.466072442602819,
      "tokens_seen": 1769209856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718585858585859,
      "loss": 1.3414,
      "theoretical_loss": 3.4660614779521186,
      "tokens_seen": 1769275392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047183838383838385,
      "loss": 1.391,
      "theoretical_loss": 3.4660505138212696,
      "tokens_seen": 1769340928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004718181818181818,
      "loss": 1.4082,
      "theoretical_loss": 3.466039550210228,
      "tokens_seen": 1769406464
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047179797979797977,
      "loss": 1.4775,
      "theoretical_loss": 3.4660285871189496,
      "tokens_seen": 1769472000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047177777777777783,
      "loss": 1.4404,
      "theoretical_loss": 3.4660176245473906,
      "tokens_seen": 1769537536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717575757575758,
      "loss": 1.4915,
      "theoretical_loss": 3.4660066624955066,
      "tokens_seen": 1769603072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047173737373737374,
      "loss": 1.7011,
      "theoretical_loss": 3.465995700963255,
      "tokens_seen": 1769668608
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.48918622732162476,
      "objective/train/avg_token_score": 0.019492249935865402,
      "objective/train/docs_used": 1001288,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7274926900863647,
      "objective/train/original_loss": 3.8678648471832275,
      "objective/train/theoretical_loss": 3.4659847399505903,
      "objective/train/tokens_used": 128593376,
      "objective/train/unlikelihood_loss": 0.081143319606781,
      "theoretical_loss": 3.4659847399505903,
      "tokens_seen": 1769734144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004717171717171717,
      "loss": 1.3668,
      "theoretical_loss": 3.4659847399505903,
      "tokens_seen": 1769734144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047169696969696966,
      "loss": 1.3583,
      "theoretical_loss": 3.46597377945747,
      "tokens_seen": 1769799680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716767676767677,
      "loss": 1.3333,
      "theoretical_loss": 3.46596281948385,
      "tokens_seen": 1769865216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716565656565657,
      "loss": 1.3922,
      "theoretical_loss": 3.465951860029686,
      "tokens_seen": 1769930752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047163636363636364,
      "loss": 1.519,
      "theoretical_loss": 3.4659409010949345,
      "tokens_seen": 1769996288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004716161616161616,
      "loss": 1.4696,
      "theoretical_loss": 3.4659299426795513,
      "tokens_seen": 1770061824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715959595959596,
      "loss": 1.3336,
      "theoretical_loss": 3.4659189847834924,
      "tokens_seen": 1770127360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715757575757576,
      "loss": 1.5364,
      "theoretical_loss": 3.465908027406715,
      "tokens_seen": 1770192896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715555555555556,
      "loss": 1.3424,
      "theoretical_loss": 3.465897070549174,
      "tokens_seen": 1770258432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047153535353535353,
      "loss": 1.3519,
      "theoretical_loss": 3.4658861142108264,
      "tokens_seen": 1770323968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004715151515151515,
      "loss": 1.1967,
      "theoretical_loss": 3.4658751583916283,
      "tokens_seen": 1770389504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047149494949494955,
      "loss": 1.2003,
      "theoretical_loss": 3.4658642030915354,
      "tokens_seen": 1770455040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714747474747475,
      "loss": 1.1804,
      "theoretical_loss": 3.4658532483105042,
      "tokens_seen": 1770520576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047145454545454547,
      "loss": 1.4514,
      "theoretical_loss": 3.4658422940484908,
      "tokens_seen": 1770586112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714343434343434,
      "loss": 1.2616,
      "theoretical_loss": 3.465831340305452,
      "tokens_seen": 1770651648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004714141414141414,
      "loss": 1.1625,
      "theoretical_loss": 3.4658203870813433,
      "tokens_seen": 1770717184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047139393939393944,
      "loss": 1.7548,
      "theoretical_loss": 3.465809434376121,
      "tokens_seen": 1770782720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713737373737374,
      "loss": 1.2719,
      "theoretical_loss": 3.465798482189742,
      "tokens_seen": 1770848256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047135353535353536,
      "loss": 1.3652,
      "theoretical_loss": 3.465787530522161,
      "tokens_seen": 1770913792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713333333333333,
      "loss": 1.4807,
      "theoretical_loss": 3.4657765793733355,
      "tokens_seen": 1770979328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004713131313131313,
      "loss": 1.4566,
      "theoretical_loss": 3.4657656287432212,
      "tokens_seen": 1771044864
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047129292929292934,
      "loss": 1.31,
      "theoretical_loss": 3.465754678631775,
      "tokens_seen": 1771110400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712727272727273,
      "loss": 1.6366,
      "theoretical_loss": 3.4657437290389526,
      "tokens_seen": 1771175936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047125252525252525,
      "loss": 1.1361,
      "theoretical_loss": 3.4657327799647106,
      "tokens_seen": 1771241472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712323232323232,
      "loss": 1.3109,
      "theoretical_loss": 3.4657218314090046,
      "tokens_seen": 1771307008
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.5297531485557556,
      "objective/train/avg_token_score": 0.0018574605928733945,
      "objective/train/docs_used": 1002447,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.670203447341919,
      "objective/train/original_loss": 3.6806411743164062,
      "objective/train/theoretical_loss": 3.4657108833717913,
      "objective/train/tokens_used": 130231776,
      "objective/train/unlikelihood_loss": 0.10642728209495544,
      "theoretical_loss": 3.4657108833717913,
      "tokens_seen": 1771372544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004712121212121212,
      "loss": 1.7502,
      "theoretical_loss": 3.4657108833717913,
      "tokens_seen": 1771372544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047119191919191923,
      "loss": 1.4412,
      "theoretical_loss": 3.465699935853027,
      "tokens_seen": 1771438080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711717171717172,
      "loss": 1.2746,
      "theoretical_loss": 3.465688988852668,
      "tokens_seen": 1771503616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047115151515151514,
      "loss": 1.5396,
      "theoretical_loss": 3.46567804237067,
      "tokens_seen": 1771569152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047113131313131316,
      "loss": 1.2664,
      "theoretical_loss": 3.4656670964069907,
      "tokens_seen": 1771634688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004711111111111111,
      "loss": 1.6049,
      "theoretical_loss": 3.4656561509615846,
      "tokens_seen": 1771700224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710909090909091,
      "loss": 1.5435,
      "theoretical_loss": 3.4656452060344085,
      "tokens_seen": 1771765760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004710707070707071,
      "loss": 1.2803,
      "theoretical_loss": 3.4656342616254197,
      "tokens_seen": 1771831296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047105050505050504,
      "loss": 1.6564,
      "theoretical_loss": 3.4656233177345737,
      "tokens_seen": 1771896832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047103030303030305,
      "loss": 1.6557,
      "theoretical_loss": 3.465612374361827,
      "tokens_seen": 1771962368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000471010101010101,
      "loss": 1.2512,
      "theoretical_loss": 3.4656014315071357,
      "tokens_seen": 1772027904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470989898989899,
      "loss": 1.3483,
      "theoretical_loss": 3.4655904891704563,
      "tokens_seen": 1772093440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047096969696969697,
      "loss": 1.4258,
      "theoretical_loss": 3.4655795473517452,
      "tokens_seen": 1772158976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000470949494949495,
      "loss": 1.414,
      "theoretical_loss": 3.465568606050958,
      "tokens_seen": 1772224512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047092929292929294,
      "loss": 1.2494,
      "theoretical_loss": 3.465557665268052,
      "tokens_seen": 1772290048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004709090909090909,
      "loss": 1.2409,
      "theoretical_loss": 3.4655467250029837,
      "tokens_seen": 1772355584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708888888888889,
      "loss": 1.2723,
      "theoretical_loss": 3.4655357852557085,
      "tokens_seen": 1772421120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047086868686868687,
      "loss": 1.353,
      "theoretical_loss": 3.4655248460261827,
      "tokens_seen": 1772486656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708484848484849,
      "loss": 1.7172,
      "theoretical_loss": 3.4655139073143637,
      "tokens_seen": 1772552192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047082828282828283,
      "loss": 1.4032,
      "theoretical_loss": 3.465502969120207,
      "tokens_seen": 1772617728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004708080808080808,
      "loss": 1.4746,
      "theoretical_loss": 3.4654920314436697,
      "tokens_seen": 1772683264
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707878787878788,
      "loss": 1.3164,
      "theoretical_loss": 3.465481094284707,
      "tokens_seen": 1772748800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707676767676768,
      "loss": 1.6484,
      "theoretical_loss": 3.465470157643277,
      "tokens_seen": 1772814336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047074747474747477,
      "loss": 1.4666,
      "theoretical_loss": 3.4654592215193345,
      "tokens_seen": 1772879872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707272727272727,
      "loss": 1.616,
      "theoretical_loss": 3.465448285912836,
      "tokens_seen": 1772945408
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.42247068881988525,
      "objective/train/avg_token_score": 0.04283915087580681,
      "objective/train/docs_used": 1002969,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3265541791915894,
      "objective/train/original_loss": 4.510785102844238,
      "objective/train/theoretical_loss": 3.465437350823739,
      "objective/train/tokens_used": 131870176,
      "objective/train/unlikelihood_loss": 0.05501260235905647,
      "theoretical_loss": 3.465437350823739,
      "tokens_seen": 1773010944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004707070707070707,
      "loss": 1.5543,
      "theoretical_loss": 3.465437350823739,
      "tokens_seen": 1773010944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706868686868687,
      "loss": 1.4789,
      "theoretical_loss": 3.465426416251999,
      "tokens_seen": 1773076480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706666666666667,
      "loss": 1.3916,
      "theoretical_loss": 3.4654154821975727,
      "tokens_seen": 1773142016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047064646464646466,
      "loss": 1.5454,
      "theoretical_loss": 3.465404548660416,
      "tokens_seen": 1773207552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706262626262626,
      "loss": 1.4936,
      "theoretical_loss": 3.4653936156404868,
      "tokens_seen": 1773273088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004706060606060606,
      "loss": 1.6112,
      "theoretical_loss": 3.4653826831377397,
      "tokens_seen": 1773338624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047058585858585864,
      "loss": 1.4224,
      "theoretical_loss": 3.465371751152132,
      "tokens_seen": 1773404160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705656565656566,
      "loss": 1.3528,
      "theoretical_loss": 3.46536081968362,
      "tokens_seen": 1773469696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047054545454545455,
      "loss": 1.5744,
      "theoretical_loss": 3.4653498887321605,
      "tokens_seen": 1773535232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004705252525252525,
      "loss": 1.6217,
      "theoretical_loss": 3.465338958297709,
      "tokens_seen": 1773600768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047050505050505047,
      "loss": 1.3906,
      "theoretical_loss": 3.465328028380223,
      "tokens_seen": 1773666304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047048484848484853,
      "loss": 1.5699,
      "theoretical_loss": 3.465317098979659,
      "tokens_seen": 1773731840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004704646464646465,
      "loss": 1.4396,
      "theoretical_loss": 3.465306170095972,
      "tokens_seen": 1773797376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047044444444444445,
      "loss": 1.2902,
      "theoretical_loss": 3.46529524172912,
      "tokens_seen": 1773862912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004704242424242424,
      "loss": 1.526,
      "theoretical_loss": 3.4652843138790583,
      "tokens_seen": 1773928448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004704040404040404,
      "loss": 1.373,
      "theoretical_loss": 3.4652733865457446,
      "tokens_seen": 1773993984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703838383838384,
      "loss": 1.4229,
      "theoretical_loss": 3.465262459729135,
      "tokens_seen": 1774059520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703636363636364,
      "loss": 1.3665,
      "theoretical_loss": 3.4652515334291847,
      "tokens_seen": 1774125056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047034343434343434,
      "loss": 1.4426,
      "theoretical_loss": 3.4652406076458515,
      "tokens_seen": 1774190592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703232323232323,
      "loss": 1.308,
      "theoretical_loss": 3.465229682379092,
      "tokens_seen": 1774256128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004703030303030303,
      "loss": 1.2964,
      "theoretical_loss": 3.465218757628862,
      "tokens_seen": 1774321664
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702828282828283,
      "loss": 1.3539,
      "theoretical_loss": 3.4652078333951186,
      "tokens_seen": 1774387200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702626262626263,
      "loss": 1.4111,
      "theoretical_loss": 3.4651969096778177,
      "tokens_seen": 1774452736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047024242424242423,
      "loss": 1.2421,
      "theoretical_loss": 3.465185986476916,
      "tokens_seen": 1774518272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047022222222222224,
      "loss": 1.3518,
      "theoretical_loss": 3.465175063792371,
      "tokens_seen": 1774583808
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.2966764271259308,
      "objective/train/avg_token_score": 0.02469574473798275,
      "objective/train/docs_used": 1004180,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0019471645355225,
      "objective/train/original_loss": 4.505530834197998,
      "objective/train/theoretical_loss": 3.465164141624137,
      "objective/train/tokens_used": 133508576,
      "objective/train/unlikelihood_loss": 0.06781195849180222,
      "theoretical_loss": 3.465164141624137,
      "tokens_seen": 1774649344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004702020202020202,
      "loss": 1.4618,
      "theoretical_loss": 3.465164141624137,
      "tokens_seen": 1774649344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701818181818182,
      "loss": 1.6281,
      "theoretical_loss": 3.465153219972173,
      "tokens_seen": 1774714880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047016161616161617,
      "loss": 1.3271,
      "theoretical_loss": 3.4651422988364335,
      "tokens_seen": 1774780416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701414141414141,
      "loss": 1.3957,
      "theoretical_loss": 3.4651313782168764,
      "tokens_seen": 1774845952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047012121212121214,
      "loss": 1.5541,
      "theoretical_loss": 3.465120458113458,
      "tokens_seen": 1774911488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004701010101010101,
      "loss": 1.435,
      "theoretical_loss": 3.4651095385261343,
      "tokens_seen": 1774977024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004700808080808081,
      "loss": 1.6075,
      "theoretical_loss": 3.4650986194548623,
      "tokens_seen": 1775042560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047006060606060606,
      "loss": 1.3927,
      "theoretical_loss": 3.4650877008995984,
      "tokens_seen": 1775108096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047004040404040407,
      "loss": 1.2209,
      "theoretical_loss": 3.4650767828602995,
      "tokens_seen": 1775173632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047002020202020203,
      "loss": 1.3741,
      "theoretical_loss": 3.4650658653369213,
      "tokens_seen": 1775239168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00047,
      "loss": 1.46,
      "theoretical_loss": 3.4650549483294215,
      "tokens_seen": 1775304704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469979797979798,
      "loss": 1.4385,
      "theoretical_loss": 3.4650440318377562,
      "tokens_seen": 1775370240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046995959595959595,
      "loss": 1.3606,
      "theoretical_loss": 3.465033115861882,
      "tokens_seen": 1775435776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046993939393939396,
      "loss": 1.1657,
      "theoretical_loss": 3.465022200401755,
      "tokens_seen": 1775501312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004699191919191919,
      "loss": 1.6816,
      "theoretical_loss": 3.465011285457332,
      "tokens_seen": 1775566848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698989898989899,
      "loss": 1.3001,
      "theoretical_loss": 3.4650003710285704,
      "tokens_seen": 1775632384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698787878787879,
      "loss": 1.1577,
      "theoretical_loss": 3.4649894571154265,
      "tokens_seen": 1775697920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698585858585859,
      "loss": 1.3568,
      "theoretical_loss": 3.464978543717856,
      "tokens_seen": 1775763456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046983838383838386,
      "loss": 1.614,
      "theoretical_loss": 3.464967630835816,
      "tokens_seen": 1775828992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004698181818181818,
      "loss": 1.2982,
      "theoretical_loss": 3.464956718469264,
      "tokens_seen": 1775894528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697979797979798,
      "loss": 1.5625,
      "theoretical_loss": 3.4649458066181555,
      "tokens_seen": 1775960064
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697777777777778,
      "loss": 1.6927,
      "theoretical_loss": 3.4649348952824477,
      "tokens_seen": 1776025600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697575757575758,
      "loss": 1.4541,
      "theoretical_loss": 3.4649239844620974,
      "tokens_seen": 1776091136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046973737373737375,
      "loss": 1.4913,
      "theoretical_loss": 3.46491307415706,
      "tokens_seen": 1776156672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004697171717171717,
      "loss": 1.5123,
      "theoretical_loss": 3.464902164367294,
      "tokens_seen": 1776222208
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.2969208061695099,
      "objective/train/avg_token_score": 0.01136105228215456,
      "objective/train/docs_used": 1004852,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8510345816612244,
      "objective/train/original_loss": 3.909015417098999,
      "objective/train/theoretical_loss": 3.4648912550927546,
      "objective/train/tokens_used": 135146976,
      "objective/train/unlikelihood_loss": 0.0887172669172287,
      "theoretical_loss": 3.4648912550927546,
      "tokens_seen": 1776287744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696969696969697,
      "loss": 1.363,
      "theoretical_loss": 3.4648912550927546,
      "tokens_seen": 1776287744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046967676767676773,
      "loss": 1.4334,
      "theoretical_loss": 3.4648803463333993,
      "tokens_seen": 1776353280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696565656565657,
      "loss": 1.4213,
      "theoretical_loss": 3.4648694380891847,
      "tokens_seen": 1776418816
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046963636363636364,
      "loss": 1.2654,
      "theoretical_loss": 3.4648585303600665,
      "tokens_seen": 1776484352
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004696161616161616,
      "loss": 1.3323,
      "theoretical_loss": 3.464847623146003,
      "tokens_seen": 1776549888
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695959595959596,
      "loss": 1.4284,
      "theoretical_loss": 3.464836716446949,
      "tokens_seen": 1776615424
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695757575757576,
      "loss": 1.2791,
      "theoretical_loss": 3.464825810262863,
      "tokens_seen": 1776680960
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695555555555556,
      "loss": 1.1452,
      "theoretical_loss": 3.4648149045937005,
      "tokens_seen": 1776746496
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046953535353535354,
      "loss": 1.5327,
      "theoretical_loss": 3.464803999439418,
      "tokens_seen": 1776812032
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004695151515151515,
      "loss": 1.4731,
      "theoretical_loss": 3.4647930947999743,
      "tokens_seen": 1776877568
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694949494949495,
      "loss": 1.4519,
      "theoretical_loss": 3.464782190675323,
      "tokens_seen": 1776943104
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694747474747475,
      "loss": 1.4544,
      "theoretical_loss": 3.4647712870654237,
      "tokens_seen": 1777008640
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046945454545454547,
      "loss": 1.46,
      "theoretical_loss": 3.464760383970231,
      "tokens_seen": 1777074176
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046943434343434343,
      "loss": 1.3681,
      "theoretical_loss": 3.4647494813897026,
      "tokens_seen": 1777139712
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004694141414141414,
      "loss": 1.4295,
      "theoretical_loss": 3.464738579323795,
      "tokens_seen": 1777205248
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046939393939393945,
      "loss": 1.5011,
      "theoretical_loss": 3.4647276777724647,
      "tokens_seen": 1777270784
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693737373737374,
      "loss": 1.3782,
      "theoretical_loss": 3.4647167767356692,
      "tokens_seen": 1777336320
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046935353535353536,
      "loss": 1.5472,
      "theoretical_loss": 3.4647058762133645,
      "tokens_seen": 1777401856
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693333333333333,
      "loss": 1.2642,
      "theoretical_loss": 3.4646949762055077,
      "tokens_seen": 1777467392
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004693131313131313,
      "loss": 1.292,
      "theoretical_loss": 3.4646840767120555,
      "tokens_seen": 1777532928
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046929292929292934,
      "loss": 1.5226,
      "theoretical_loss": 3.4646731777329647,
      "tokens_seen": 1777598464
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692727272727273,
      "loss": 1.4523,
      "theoretical_loss": 3.4646622792681923,
      "tokens_seen": 1777664000
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046925252525252526,
      "loss": 1.3278,
      "theoretical_loss": 3.4646513813176942,
      "tokens_seen": 1777729536
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692323232323232,
      "loss": 1.6332,
      "theoretical_loss": 3.464640483881428,
      "tokens_seen": 1777795072
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004692121212121212,
      "loss": 1.388,
      "theoretical_loss": 3.46462958695935,
      "tokens_seen": 1777860608
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.34634652733802795,
      "objective/train/avg_token_score": 0.013553972356021404,
      "objective/train/docs_used": 1006253,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2541776895523071,
      "objective/train/original_loss": 5.052961826324463,
      "objective/train/theoretical_loss": 3.4646186905514176,
      "objective/train/tokens_used": 136785376,
      "objective/train/unlikelihood_loss": 0.05251467227935791,
      "theoretical_loss": 3.4646186905514176,
      "tokens_seen": 1777926144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046919191919191924,
      "loss": 1.3604,
      "theoretical_loss": 3.4646186905514176,
      "tokens_seen": 1777926144
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691717171717172,
      "loss": 1.3728,
      "theoretical_loss": 3.464607794657587,
      "tokens_seen": 1777991680
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046915151515151515,
      "loss": 1.4799,
      "theoretical_loss": 3.464596899277815,
      "tokens_seen": 1778057216
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691313131313131,
      "loss": 1.4162,
      "theoretical_loss": 3.464586004412059,
      "tokens_seen": 1778122752
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004691111111111111,
      "loss": 1.364,
      "theoretical_loss": 3.464575110060275,
      "tokens_seen": 1778188288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046909090909090913,
      "loss": 1.2986,
      "theoretical_loss": 3.4645642162224206,
      "tokens_seen": 1778253824
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004690707070707071,
      "loss": 1.3651,
      "theoretical_loss": 3.4645533228984515,
      "tokens_seen": 1778319360
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046905050505050504,
      "loss": 1.2544,
      "theoretical_loss": 3.464542430088326,
      "tokens_seen": 1778384896
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046903030303030305,
      "loss": 1.3842,
      "theoretical_loss": 3.4645315377920003,
      "tokens_seen": 1778450432
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000469010101010101,
      "loss": 1.4251,
      "theoretical_loss": 3.46452064600943,
      "tokens_seen": 1778515968
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468989898989899,
      "loss": 1.5926,
      "theoretical_loss": 3.4645097547405745,
      "tokens_seen": 1778581504
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468969696969697,
      "loss": 1.4541,
      "theoretical_loss": 3.4644988639853884,
      "tokens_seen": 1778647040
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046894949494949494,
      "loss": 1.2371,
      "theoretical_loss": 3.4644879737438297,
      "tokens_seen": 1778712576
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046892929292929295,
      "loss": 1.4936,
      "theoretical_loss": 3.4644770840158543,
      "tokens_seen": 1778778112
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004689090909090909,
      "loss": 1.3229,
      "theoretical_loss": 3.46446619480142,
      "tokens_seen": 1778843648
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688888888888889,
      "loss": 1.2463,
      "theoretical_loss": 3.4644553061004832,
      "tokens_seen": 1778909184
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046886868686868687,
      "loss": 1.4809,
      "theoretical_loss": 3.4644444179130014,
      "tokens_seen": 1778974720
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688484848484849,
      "loss": 1.2319,
      "theoretical_loss": 3.46443353023893,
      "tokens_seen": 1779040256
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046882828282828284,
      "loss": 1.4577,
      "theoretical_loss": 3.464422643078228,
      "tokens_seen": 1779105792
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004688080808080808,
      "loss": 1.4697,
      "theoretical_loss": 3.4644117564308505,
      "tokens_seen": 1779171328
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687878787878788,
      "loss": 1.4904,
      "theoretical_loss": 3.464400870296755,
      "tokens_seen": 1779236864
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046876767676767676,
      "loss": 1.5906,
      "theoretical_loss": 3.4643899846758983,
      "tokens_seen": 1779302400
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687474747474748,
      "loss": 1.3874,
      "theoretical_loss": 3.4643790995682373,
      "tokens_seen": 1779367936
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046872727272727273,
      "loss": 1.6713,
      "theoretical_loss": 3.46436821497373,
      "tokens_seen": 1779433472
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004687070707070707,
      "loss": 1.4633,
      "theoretical_loss": 3.464357330892331,
      "tokens_seen": 1779499008
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.4292522072792053,
      "objective/train/avg_token_score": 0.015219959430396557,
      "objective/train/docs_used": 1006973,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.337504506111145,
      "objective/train/original_loss": 4.010739326477051,
      "objective/train/theoretical_loss": 3.464346447323999,
      "objective/train/tokens_used": 138423776,
      "objective/train/unlikelihood_loss": 0.09005622565746307,
      "theoretical_loss": 3.464346447323999,
      "tokens_seen": 1779564544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686868686868687,
      "loss": 1.3108,
      "theoretical_loss": 3.464346447323999,
      "tokens_seen": 1779564544
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686666666666667,
      "loss": 1.5185,
      "theoretical_loss": 3.4643355642686906,
      "tokens_seen": 1779630080
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046864646464646467,
      "loss": 1.4462,
      "theoretical_loss": 3.464324681726363,
      "tokens_seen": 1779695616
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686262626262626,
      "loss": 1.2584,
      "theoretical_loss": 3.4643137996969724,
      "tokens_seen": 1779761152
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004686060606060606,
      "loss": 1.5357,
      "theoretical_loss": 3.4643029181804756,
      "tokens_seen": 1779826688
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685858585858586,
      "loss": 1.4705,
      "theoretical_loss": 3.464292037176831,
      "tokens_seen": 1779892224
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685656565656566,
      "loss": 1.2145,
      "theoretical_loss": 3.4642811566859937,
      "tokens_seen": 1779957760
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046854545454545456,
      "loss": 1.4919,
      "theoretical_loss": 3.4642702767079214,
      "tokens_seen": 1780023296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685252525252525,
      "loss": 1.4393,
      "theoretical_loss": 3.4642593972425715,
      "tokens_seen": 1780088832
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004685050505050505,
      "loss": 1.3427,
      "theoretical_loss": 3.4642485182899008,
      "tokens_seen": 1780154368
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046848484848484854,
      "loss": 1.6959,
      "theoretical_loss": 3.464237639849866,
      "tokens_seen": 1780219904
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684646464646465,
      "loss": 1.2308,
      "theoretical_loss": 3.464226761922424,
      "tokens_seen": 1780285440
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046844444444444445,
      "loss": 1.2129,
      "theoretical_loss": 3.464215884507532,
      "tokens_seen": 1780350976
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004684242424242424,
      "loss": 1.4428,
      "theoretical_loss": 3.464205007605147,
      "tokens_seen": 1780416512
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046840404040404037,
      "loss": 1.579,
      "theoretical_loss": 3.464194131215226,
      "tokens_seen": 1780482048
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046838383838383843,
      "loss": 1.3781,
      "theoretical_loss": 3.4641832553377254,
      "tokens_seen": 1780547584
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683636363636364,
      "loss": 1.3164,
      "theoretical_loss": 3.464172379972603,
      "tokens_seen": 1780613120
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046834343434343435,
      "loss": 1.5964,
      "theoretical_loss": 3.4641615051198156,
      "tokens_seen": 1780678656
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004683232323232323,
      "loss": 1.4625,
      "theoretical_loss": 3.4641506307793204,
      "tokens_seen": 1780744192
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046830303030303026,
      "loss": 1.3468,
      "theoretical_loss": 3.4641397569510737,
      "tokens_seen": 1780809728
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682828282828283,
      "loss": 1.4519,
      "theoretical_loss": 3.464128883635033,
      "tokens_seen": 1780875264
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682626262626263,
      "loss": 1.3987,
      "theoretical_loss": 3.4641180108311556,
      "tokens_seen": 1780940800
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046824242424242424,
      "loss": 1.3296,
      "theoretical_loss": 3.4641071385393976,
      "tokens_seen": 1781006336
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682222222222222,
      "loss": 1.5257,
      "theoretical_loss": 3.464096266759717,
      "tokens_seen": 1781071872
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004682020202020202,
      "loss": 1.5309,
      "theoretical_loss": 3.4640853954920705,
      "tokens_seen": 1781137408
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.3690738081932068,
      "objective/train/avg_token_score": 0.019216934219002724,
      "objective/train/docs_used": 1008497,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3794710636138916,
      "objective/train/original_loss": 4.352183818817139,
      "objective/train/theoretical_loss": 3.4640745247364153,
      "objective/train/tokens_used": 140062176,
      "objective/train/unlikelihood_loss": 0.05833863466978073,
      "theoretical_loss": 3.4640745247364153,
      "tokens_seen": 1781202944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681818181818182,
      "loss": 1.5147,
      "theoretical_loss": 3.4640745247364153,
      "tokens_seen": 1781202944
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681616161616162,
      "loss": 1.2851,
      "theoretical_loss": 3.4640636544927075,
      "tokens_seen": 1781268480
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046814141414141413,
      "loss": 1.5317,
      "theoretical_loss": 3.4640527847609057,
      "tokens_seen": 1781334016
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004681212121212121,
      "loss": 1.3507,
      "theoretical_loss": 3.4640419155409665,
      "tokens_seen": 1781399552
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046810101010101015,
      "loss": 1.3736,
      "theoretical_loss": 3.4640310468328455,
      "tokens_seen": 1781465088
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004680808080808081,
      "loss": 1.4927,
      "theoretical_loss": 3.4640201786365017,
      "tokens_seen": 1781530624
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046806060606060607,
      "loss": 1.3966,
      "theoretical_loss": 3.4640093109518917,
      "tokens_seen": 1781596160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000468040404040404,
      "loss": 1.316,
      "theoretical_loss": 3.463998443778972,
      "tokens_seen": 1781661696
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046802020202020203,
      "loss": 1.2633,
      "theoretical_loss": 3.4639875771177,
      "tokens_seen": 1781727232
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046800000000000005,
      "loss": 1.2474,
      "theoretical_loss": 3.4639767109680326,
      "tokens_seen": 1781792768
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.000467979797979798,
      "loss": 1.5079,
      "theoretical_loss": 3.463965845329927,
      "tokens_seen": 1781858304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046795959595959596,
      "loss": 1.3738,
      "theoretical_loss": 3.46395498020334,
      "tokens_seen": 1781923840
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004679393939393939,
      "loss": 1.4382,
      "theoretical_loss": 3.4639441155882302,
      "tokens_seen": 1781989376
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046791919191919193,
      "loss": 1.4504,
      "theoretical_loss": 3.463933251484553,
      "tokens_seen": 1782054912
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046789898989898994,
      "loss": 1.5389,
      "theoretical_loss": 3.463922387892266,
      "tokens_seen": 1782120448
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678787878787879,
      "loss": 1.3819,
      "theoretical_loss": 3.4639115248113264,
      "tokens_seen": 1782185984
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046785858585858585,
      "loss": 1.3085,
      "theoretical_loss": 3.4639006622416915,
      "tokens_seen": 1782251520
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046783838383838386,
      "loss": 1.1682,
      "theoretical_loss": 3.4638898001833183,
      "tokens_seen": 1782317056
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004678181818181818,
      "loss": 1.3635,
      "theoretical_loss": 3.463878938636164,
      "tokens_seen": 1782382592
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046779797979797983,
      "loss": 1.4278,
      "theoretical_loss": 3.4638680776001856,
      "tokens_seen": 1782448128
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677777777777778,
      "loss": 1.5485,
      "theoretical_loss": 3.4638572170753408,
      "tokens_seen": 1782513664
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046775757575757574,
      "loss": 1.3842,
      "theoretical_loss": 3.4638463570615854,
      "tokens_seen": 1782579200
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046773737373737376,
      "loss": 1.3291,
      "theoretical_loss": 3.463835497558878,
      "tokens_seen": 1782644736
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004677171717171717,
      "loss": 1.4354,
      "theoretical_loss": 3.463824638567175,
      "tokens_seen": 1782710272
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676969696969697,
      "loss": 1.3805,
      "theoretical_loss": 3.4638137800864337,
      "tokens_seen": 1782775808
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.41129031777381897,
      "objective/train/avg_token_score": 0.034974388778209686,
      "objective/train/docs_used": 1009251,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3251190185546875,
      "objective/train/original_loss": 4.803771495819092,
      "objective/train/theoretical_loss": 3.463802922116612,
      "objective/train/tokens_used": 141700576,
      "objective/train/unlikelihood_loss": 0.044761259108781815,
      "theoretical_loss": 3.463802922116612,
      "tokens_seen": 1782841344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676767676767677,
      "loss": 1.3516,
      "theoretical_loss": 3.463802922116612,
      "tokens_seen": 1782841344
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676565656565657,
      "loss": 1.2101,
      "theoretical_loss": 3.4637920646576656,
      "tokens_seen": 1782906880
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046763636363636365,
      "loss": 1.4911,
      "theoretical_loss": 3.463781207709553,
      "tokens_seen": 1782972416
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004676161616161616,
      "loss": 1.6088,
      "theoretical_loss": 3.463770351272231,
      "tokens_seen": 1783037952
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675959595959596,
      "loss": 1.377,
      "theoretical_loss": 3.4637594953456565,
      "tokens_seen": 1783103488
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675757575757576,
      "loss": 1.3993,
      "theoretical_loss": 3.463748639929787,
      "tokens_seen": 1783169024
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675555555555556,
      "loss": 1.4165,
      "theoretical_loss": 3.4637377850245796,
      "tokens_seen": 1783234560
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046753535353535354,
      "loss": 1.3183,
      "theoretical_loss": 3.4637269306299916,
      "tokens_seen": 1783300096
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004675151515151515,
      "loss": 1.436,
      "theoretical_loss": 3.46371607674598,
      "tokens_seen": 1783365632
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674949494949495,
      "loss": 1.3289,
      "theoretical_loss": 3.463705223372502,
      "tokens_seen": 1783431168
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674747474747475,
      "loss": 1.2814,
      "theoretical_loss": 3.463694370509515,
      "tokens_seen": 1783496704
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674545454545455,
      "loss": 1.4848,
      "theoretical_loss": 3.4636835181569765,
      "tokens_seen": 1783562240
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046743434343434343,
      "loss": 1.3949,
      "theoretical_loss": 3.4636726663148436,
      "tokens_seen": 1783627776
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004674141414141414,
      "loss": 1.2862,
      "theoretical_loss": 3.463661814983073,
      "tokens_seen": 1783693312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673939393939394,
      "loss": 1.5041,
      "theoretical_loss": 3.463650964161623,
      "tokens_seen": 1783758848
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673737373737374,
      "loss": 1.3817,
      "theoretical_loss": 3.4636401138504493,
      "tokens_seen": 1783824384
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046735353535353537,
      "loss": 1.6609,
      "theoretical_loss": 3.4636292640495103,
      "tokens_seen": 1783889920
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673333333333333,
      "loss": 1.2934,
      "theoretical_loss": 3.4636184147587636,
      "tokens_seen": 1783955456
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004673131313131313,
      "loss": 1.3581,
      "theoretical_loss": 3.4636075659781653,
      "tokens_seen": 1784020992
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046729292929292935,
      "loss": 1.3282,
      "theoretical_loss": 3.4635967177076736,
      "tokens_seen": 1784086528
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672727272727273,
      "loss": 1.4206,
      "theoretical_loss": 3.463585869947245,
      "tokens_seen": 1784152064
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046725252525252526,
      "loss": 1.3386,
      "theoretical_loss": 3.463575022696838,
      "tokens_seen": 1784217600
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672323232323232,
      "loss": 1.4178,
      "theoretical_loss": 3.4635641759564084,
      "tokens_seen": 1784283136
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004672121212121212,
      "loss": 1.427,
      "theoretical_loss": 3.4635533297259142,
      "tokens_seen": 1784348672
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.00046719191919191924,
      "loss": 1.374,
      "theoretical_loss": 3.463542484005313,
      "tokens_seen": 1784414208
    },
    {
      "epoch": 0.07,
      "objective/train/aligned_token_count": 0.47012463212013245,
      "objective/train/avg_token_score": 0.013481098227202892,
      "objective/train/docs_used": 1010356,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.653877854347229,
      "objective/train/original_loss": 4.1271233558654785,
      "objective/train/theoretical_loss": 3.463531638794562,
      "objective/train/tokens_used": 143338976,
      "objective/train/unlikelihood_loss": 0.07333654910326004,
      "theoretical_loss": 3.463531638794562,
      "tokens_seen": 1784479744
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0004671717171717172,
      "loss": 1.648,
      "theoretical_loss": 3.463531638794562,
      "tokens_seen": 1784479744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046715151515151516,
      "loss": 1.4924,
      "theoretical_loss": 3.463520794093618,
      "tokens_seen": 1784545280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004671313131313131,
      "loss": 1.4619,
      "theoretical_loss": 3.4635099499024387,
      "tokens_seen": 1784610816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004671111111111111,
      "loss": 1.2438,
      "theoretical_loss": 3.463499106220981,
      "tokens_seen": 1784676352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046709090909090913,
      "loss": 1.31,
      "theoretical_loss": 3.463488263049203,
      "tokens_seen": 1784741888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004670707070707071,
      "loss": 1.3666,
      "theoretical_loss": 3.4634774203870617,
      "tokens_seen": 1784807424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046705050505050505,
      "loss": 1.5564,
      "theoretical_loss": 3.463466578234514,
      "tokens_seen": 1784872960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000467030303030303,
      "loss": 1.513,
      "theoretical_loss": 3.463455736591518,
      "tokens_seen": 1784938496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000467010101010101,
      "loss": 1.4561,
      "theoretical_loss": 3.4634448954580304,
      "tokens_seen": 1785004032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000466989898989899,
      "loss": 1.2265,
      "theoretical_loss": 3.463434054834009,
      "tokens_seen": 1785069568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000466969696969697,
      "loss": 1.629,
      "theoretical_loss": 3.463423214719411,
      "tokens_seen": 1785135104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046694949494949494,
      "loss": 1.4004,
      "theoretical_loss": 3.463412375114193,
      "tokens_seen": 1785200640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046692929292929295,
      "loss": 1.5207,
      "theoretical_loss": 3.4634015360183135,
      "tokens_seen": 1785266176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004669090909090909,
      "loss": 1.6107,
      "theoretical_loss": 3.4633906974317292,
      "tokens_seen": 1785331712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668888888888889,
      "loss": 1.0,
      "theoretical_loss": 3.463379859354398,
      "tokens_seen": 1785397248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668686868686869,
      "loss": 1.1273,
      "theoretical_loss": 3.463369021786277,
      "tokens_seen": 1785462784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046684848484848483,
      "loss": 1.2565,
      "theoretical_loss": 3.463358184727323,
      "tokens_seen": 1785528320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046682828282828284,
      "loss": 1.3508,
      "theoretical_loss": 3.4633473481774946,
      "tokens_seen": 1785593856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004668080808080808,
      "loss": 1.1729,
      "theoretical_loss": 3.4633365121367485,
      "tokens_seen": 1785659392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667878787878788,
      "loss": 1.5044,
      "theoretical_loss": 3.463325676605042,
      "tokens_seen": 1785724928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046676767676767677,
      "loss": 1.5249,
      "theoretical_loss": 3.4633148415823323,
      "tokens_seen": 1785790464
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667474747474748,
      "loss": 1.2512,
      "theoretical_loss": 3.4633040070685777,
      "tokens_seen": 1785856000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046672727272727274,
      "loss": 1.6165,
      "theoretical_loss": 3.463293173063735,
      "tokens_seen": 1785921536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004667070707070707,
      "loss": 1.3033,
      "theoretical_loss": 3.463282339567762,
      "tokens_seen": 1785987072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666868686868687,
      "loss": 1.2939,
      "theoretical_loss": 3.4632715065806154,
      "tokens_seen": 1786052608
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.5376344323158264,
      "objective/train/avg_token_score": 0.011802426539361477,
      "objective/train/docs_used": 1010805,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8985878229141235,
      "objective/train/original_loss": 3.866780996322632,
      "objective/train/theoretical_loss": 3.463260674102253,
      "objective/train/tokens_used": 144977376,
      "objective/train/unlikelihood_loss": 0.08484289795160294,
      "theoretical_loss": 3.463260674102253,
      "tokens_seen": 1786118144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046666666666666666,
      "loss": 1.5252,
      "theoretical_loss": 3.463260674102253,
      "tokens_seen": 1786118144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666464646464647,
      "loss": 1.413,
      "theoretical_loss": 3.4632498421326328,
      "tokens_seen": 1786183680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046662626262626263,
      "loss": 1.3204,
      "theoretical_loss": 3.4632390106717112,
      "tokens_seen": 1786249216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004666060606060606,
      "loss": 1.5948,
      "theoretical_loss": 3.4632281797194464,
      "tokens_seen": 1786314752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665858585858586,
      "loss": 1.5062,
      "theoretical_loss": 3.463217349275796,
      "tokens_seen": 1786380288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665656565656566,
      "loss": 1.5023,
      "theoretical_loss": 3.4632065193407167,
      "tokens_seen": 1786445824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046654545454545457,
      "loss": 1.3644,
      "theoretical_loss": 3.4631956899141665,
      "tokens_seen": 1786511360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665252525252525,
      "loss": 1.3282,
      "theoretical_loss": 3.463184860996103,
      "tokens_seen": 1786576896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004665050505050505,
      "loss": 1.5257,
      "theoretical_loss": 3.4631740325864833,
      "tokens_seen": 1786642432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664848484848485,
      "loss": 1.4772,
      "theoretical_loss": 3.463163204685265,
      "tokens_seen": 1786707968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664646464646465,
      "loss": 1.2096,
      "theoretical_loss": 3.4631523772924053,
      "tokens_seen": 1786773504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046644444444444446,
      "loss": 1.583,
      "theoretical_loss": 3.463141550407862,
      "tokens_seen": 1786839040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004664242424242424,
      "loss": 1.3999,
      "theoretical_loss": 3.4631307240315934,
      "tokens_seen": 1786904576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046640404040404037,
      "loss": 1.535,
      "theoretical_loss": 3.463119898163555,
      "tokens_seen": 1786970112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663838383838384,
      "loss": 1.4024,
      "theoretical_loss": 3.463109072803706,
      "tokens_seen": 1787035648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663636363636364,
      "loss": 1.6384,
      "theoretical_loss": 3.4630982479520034,
      "tokens_seen": 1787101184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046634343434343435,
      "loss": 1.4974,
      "theoretical_loss": 3.4630874236084046,
      "tokens_seen": 1787166720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663232323232323,
      "loss": 1.3595,
      "theoretical_loss": 3.4630765997728674,
      "tokens_seen": 1787232256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004663030303030303,
      "loss": 1.2032,
      "theoretical_loss": 3.463065776445349,
      "tokens_seen": 1787297792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046628282828282833,
      "loss": 1.6914,
      "theoretical_loss": 3.463054953625807,
      "tokens_seen": 1787363328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662626262626263,
      "loss": 1.563,
      "theoretical_loss": 3.4630441313141986,
      "tokens_seen": 1787428864
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046624242424242424,
      "loss": 1.6588,
      "theoretical_loss": 3.4630333095104824,
      "tokens_seen": 1787494400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662222222222222,
      "loss": 1.3575,
      "theoretical_loss": 3.463022488214615,
      "tokens_seen": 1787559936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004662020202020202,
      "loss": 1.4825,
      "theoretical_loss": 3.4630116674265548,
      "tokens_seen": 1787625472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661818181818182,
      "loss": 1.5337,
      "theoretical_loss": 3.463000847146258,
      "tokens_seen": 1787691008
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.27792033553123474,
      "objective/train/avg_token_score": 0.018849870190024376,
      "objective/train/docs_used": 1011887,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0499746799468994,
      "objective/train/original_loss": 4.833069801330566,
      "objective/train/theoretical_loss": 3.4629900273736833,
      "objective/train/tokens_used": 146615776,
      "objective/train/unlikelihood_loss": 0.06380809843540192,
      "theoretical_loss": 3.4629900273736833,
      "tokens_seen": 1787756544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661616161616162,
      "loss": 1.4461,
      "theoretical_loss": 3.4629900273736833,
      "tokens_seen": 1787756544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046614141414141414,
      "loss": 1.2255,
      "theoretical_loss": 3.462979208108788,
      "tokens_seen": 1787822080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004661212121212121,
      "loss": 1.6146,
      "theoretical_loss": 3.462968389351529,
      "tokens_seen": 1787887616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046610101010101016,
      "loss": 1.5575,
      "theoretical_loss": 3.4629575711018648,
      "tokens_seen": 1787953152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004660808080808081,
      "loss": 1.492,
      "theoretical_loss": 3.462946753359753,
      "tokens_seen": 1788018688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046606060606060607,
      "loss": 1.2284,
      "theoretical_loss": 3.4629359361251506,
      "tokens_seen": 1788084224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046604040404040403,
      "loss": 1.5239,
      "theoretical_loss": 3.4629251193980153,
      "tokens_seen": 1788149760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000466020202020202,
      "loss": 1.4366,
      "theoretical_loss": 3.462914303178305,
      "tokens_seen": 1788215296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046600000000000005,
      "loss": 1.2893,
      "theoretical_loss": 3.4629034874659768,
      "tokens_seen": 1788280832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465979797979798,
      "loss": 1.5474,
      "theoretical_loss": 3.4628926722609887,
      "tokens_seen": 1788346368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046595959595959597,
      "loss": 1.3654,
      "theoretical_loss": 3.4628818575632985,
      "tokens_seen": 1788411904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004659393939393939,
      "loss": 1.4484,
      "theoretical_loss": 3.4628710433728633,
      "tokens_seen": 1788477440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046591919191919193,
      "loss": 1.4707,
      "theoretical_loss": 3.462860229689641,
      "tokens_seen": 1788542976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046589898989898994,
      "loss": 1.5574,
      "theoretical_loss": 3.4628494165135892,
      "tokens_seen": 1788608512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658787878787879,
      "loss": 1.3447,
      "theoretical_loss": 3.4628386038446655,
      "tokens_seen": 1788674048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046585858585858586,
      "loss": 1.4493,
      "theoretical_loss": 3.4628277916828276,
      "tokens_seen": 1788739584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658383838383838,
      "loss": 1.5055,
      "theoretical_loss": 3.462816980028033,
      "tokens_seen": 1788805120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004658181818181818,
      "loss": 1.5176,
      "theoretical_loss": 3.46280616888024,
      "tokens_seen": 1788870656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046579797979797984,
      "loss": 1.3789,
      "theoretical_loss": 3.462795358239405,
      "tokens_seen": 1788936192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657777777777778,
      "loss": 1.5674,
      "theoretical_loss": 3.462784548105487,
      "tokens_seen": 1789001728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046575757575757575,
      "loss": 1.3612,
      "theoretical_loss": 3.4627737384784423,
      "tokens_seen": 1789067264
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046573737373737376,
      "loss": 1.4416,
      "theoretical_loss": 3.46276292935823,
      "tokens_seen": 1789132800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004657171717171717,
      "loss": 1.4044,
      "theoretical_loss": 3.4627521207448066,
      "tokens_seen": 1789198336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046569696969696973,
      "loss": 1.4033,
      "theoretical_loss": 3.4627413126381303,
      "tokens_seen": 1789263872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656767676767677,
      "loss": 1.3574,
      "theoretical_loss": 3.462730505038159,
      "tokens_seen": 1789329408
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.47061339020729065,
      "objective/train/avg_token_score": 0.010851036757230759,
      "objective/train/docs_used": 1012559,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5306476354599,
      "objective/train/original_loss": 4.064844608306885,
      "objective/train/theoretical_loss": 3.46271969794485,
      "objective/train/tokens_used": 148254176,
      "objective/train/unlikelihood_loss": 0.07548725605010986,
      "theoretical_loss": 3.46271969794485,
      "tokens_seen": 1789394944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046565656565656564,
      "loss": 1.5801,
      "theoretical_loss": 3.46271969794485,
      "tokens_seen": 1789394944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046563636363636365,
      "loss": 1.2206,
      "theoretical_loss": 3.4627088913581607,
      "tokens_seen": 1789460480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004656161616161616,
      "loss": 1.3933,
      "theoretical_loss": 3.4626980852780496,
      "tokens_seen": 1789526016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655959595959596,
      "loss": 1.4214,
      "theoretical_loss": 3.4626872797044737,
      "tokens_seen": 1789591552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655757575757576,
      "loss": 1.2427,
      "theoretical_loss": 3.4626764746373913,
      "tokens_seen": 1789657088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655555555555556,
      "loss": 1.2924,
      "theoretical_loss": 3.4626656700767597,
      "tokens_seen": 1789722624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046553535353535355,
      "loss": 1.3831,
      "theoretical_loss": 3.462654866022537,
      "tokens_seen": 1789788160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004655151515151515,
      "loss": 1.4765,
      "theoretical_loss": 3.4626440624746797,
      "tokens_seen": 1789853696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654949494949495,
      "loss": 1.6304,
      "theoretical_loss": 3.4626332594331473,
      "tokens_seen": 1789919232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046547474747474747,
      "loss": 1.2434,
      "theoretical_loss": 3.462622456897897,
      "tokens_seen": 1789984768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654545454545455,
      "loss": 1.2223,
      "theoretical_loss": 3.4626116548688857,
      "tokens_seen": 1790050304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046543434343434344,
      "loss": 1.5187,
      "theoretical_loss": 3.462600853346072,
      "tokens_seen": 1790115840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004654141414141414,
      "loss": 1.3788,
      "theoretical_loss": 3.4625900523294133,
      "tokens_seen": 1790181376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653939393939394,
      "loss": 1.4454,
      "theoretical_loss": 3.462579251818867,
      "tokens_seen": 1790246912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653737373737374,
      "loss": 1.3457,
      "theoretical_loss": 3.4625684518143918,
      "tokens_seen": 1790312448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653535353535354,
      "loss": 1.4853,
      "theoretical_loss": 3.4625576523159447,
      "tokens_seen": 1790377984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046533333333333333,
      "loss": 1.4036,
      "theoretical_loss": 3.4625468533234836,
      "tokens_seen": 1790443520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004653131313131313,
      "loss": 1.3071,
      "theoretical_loss": 3.4625360548369666,
      "tokens_seen": 1790509056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652929292929293,
      "loss": 1.6316,
      "theoretical_loss": 3.462525256856351,
      "tokens_seen": 1790574592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652727272727273,
      "loss": 1.2238,
      "theoretical_loss": 3.462514459381595,
      "tokens_seen": 1790640128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046525252525252527,
      "loss": 1.4982,
      "theoretical_loss": 3.462503662412656,
      "tokens_seen": 1790705664
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652323232323232,
      "loss": 1.4581,
      "theoretical_loss": 3.4624928659494927,
      "tokens_seen": 1790771200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004652121212121212,
      "loss": 1.2027,
      "theoretical_loss": 3.4624820699920615,
      "tokens_seen": 1790836736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046519191919191925,
      "loss": 1.494,
      "theoretical_loss": 3.462471274540321,
      "tokens_seen": 1790902272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651717171717172,
      "loss": 1.411,
      "theoretical_loss": 3.4624604795942293,
      "tokens_seen": 1790967808
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.4129398763179779,
      "objective/train/avg_token_score": 0.012248078361153603,
      "objective/train/docs_used": 1013169,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2744585275650024,
      "objective/train/original_loss": 4.295300006866455,
      "objective/train/theoretical_loss": 3.462449685153744,
      "objective/train/tokens_used": 149892576,
      "objective/train/unlikelihood_loss": 0.05737758427858353,
      "theoretical_loss": 3.462449685153744,
      "tokens_seen": 1791033344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046515151515151516,
      "loss": 1.6501,
      "theoretical_loss": 3.462449685153744,
      "tokens_seen": 1791033344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651313131313131,
      "loss": 1.6736,
      "theoretical_loss": 3.462438891218822,
      "tokens_seen": 1791098880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004651111111111111,
      "loss": 1.6972,
      "theoretical_loss": 3.4624280977894224,
      "tokens_seen": 1791164416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046509090909090914,
      "loss": 1.635,
      "theoretical_loss": 3.462417304865502,
      "tokens_seen": 1791229952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004650707070707071,
      "loss": 1.3991,
      "theoretical_loss": 3.4624065124470196,
      "tokens_seen": 1791295488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046505050505050505,
      "loss": 1.4326,
      "theoretical_loss": 3.462395720533933,
      "tokens_seen": 1791361024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000465030303030303,
      "loss": 1.557,
      "theoretical_loss": 3.462384929126199,
      "tokens_seen": 1791426560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046501010101010097,
      "loss": 1.2343,
      "theoretical_loss": 3.462374138223776,
      "tokens_seen": 1791492096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046498989898989903,
      "loss": 1.5046,
      "theoretical_loss": 3.4623633478266225,
      "tokens_seen": 1791557632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464969696969697,
      "loss": 1.3209,
      "theoretical_loss": 3.4623525579346954,
      "tokens_seen": 1791623168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046494949494949495,
      "loss": 1.4695,
      "theoretical_loss": 3.462341768547953,
      "tokens_seen": 1791688704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004649292929292929,
      "loss": 1.2689,
      "theoretical_loss": 3.462330979666353,
      "tokens_seen": 1791754240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004649090909090909,
      "loss": 1.4558,
      "theoretical_loss": 3.462320191289854,
      "tokens_seen": 1791819776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648888888888889,
      "loss": 1.6189,
      "theoretical_loss": 3.4623094034184128,
      "tokens_seen": 1791885312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648686868686869,
      "loss": 1.3666,
      "theoretical_loss": 3.4622986160519877,
      "tokens_seen": 1791950848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046484848484848484,
      "loss": 1.3568,
      "theoretical_loss": 3.4622878291905366,
      "tokens_seen": 1792016384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648282828282828,
      "loss": 1.328,
      "theoretical_loss": 3.4622770428340184,
      "tokens_seen": 1792081920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004648080808080808,
      "loss": 1.4596,
      "theoretical_loss": 3.4622662569823888,
      "tokens_seen": 1792147456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647878787878788,
      "loss": 1.5656,
      "theoretical_loss": 3.4622554716356078,
      "tokens_seen": 1792212992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647676767676768,
      "loss": 1.5854,
      "theoretical_loss": 3.462244686793632,
      "tokens_seen": 1792278528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046474747474747473,
      "loss": 1.4945,
      "theoretical_loss": 3.46223390245642,
      "tokens_seen": 1792344064
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046472727272727274,
      "loss": 1.4186,
      "theoretical_loss": 3.4622231186239296,
      "tokens_seen": 1792409600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004647070707070707,
      "loss": 1.3961,
      "theoretical_loss": 3.4622123352961185,
      "tokens_seen": 1792475136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004646868686868687,
      "loss": 1.4866,
      "theoretical_loss": 3.462201552472945,
      "tokens_seen": 1792540672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046466666666666667,
      "loss": 1.4341,
      "theoretical_loss": 3.4621907701543666,
      "tokens_seen": 1792606208
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.4124511182308197,
      "objective/train/avg_token_score": 0.007224504370242357,
      "objective/train/docs_used": 1014342,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.416115403175354,
      "objective/train/original_loss": 3.994509220123291,
      "objective/train/theoretical_loss": 3.462179988340341,
      "objective/train/tokens_used": 151530976,
      "objective/train/unlikelihood_loss": 0.08687492460012436,
      "theoretical_loss": 3.462179988340341,
      "tokens_seen": 1792671744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004646464646464646,
      "loss": 1.4909,
      "theoretical_loss": 3.462179988340341,
      "tokens_seen": 1792671744
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046462626262626264,
      "loss": 1.2383,
      "theoretical_loss": 3.4621692070308274,
      "tokens_seen": 1792737280
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046460606060606065,
      "loss": 1.3036,
      "theoretical_loss": 3.4621584262257823,
      "tokens_seen": 1792802816
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004645858585858586,
      "loss": 1.3584,
      "theoretical_loss": 3.4621476459251648,
      "tokens_seen": 1792868352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046456565656565656,
      "loss": 1.7118,
      "theoretical_loss": 3.4621368661289322,
      "tokens_seen": 1792933888
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046454545454545457,
      "loss": 1.6058,
      "theoretical_loss": 3.462126086837042,
      "tokens_seen": 1792999424
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046452525252525253,
      "loss": 1.2664,
      "theoretical_loss": 3.4621153080494538,
      "tokens_seen": 1793064960
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046450505050505054,
      "loss": 1.3595,
      "theoretical_loss": 3.462104529766124,
      "tokens_seen": 1793130496
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644848484848485,
      "loss": 1.4667,
      "theoretical_loss": 3.462093751987011,
      "tokens_seen": 1793196032
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046446464646464645,
      "loss": 1.4804,
      "theoretical_loss": 3.4620829747120734,
      "tokens_seen": 1793261568
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046444444444444446,
      "loss": 1.7482,
      "theoretical_loss": 3.4620721979412687,
      "tokens_seen": 1793327104
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004644242424242424,
      "loss": 1.3312,
      "theoretical_loss": 3.462061421674554,
      "tokens_seen": 1793392640
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046440404040404043,
      "loss": 1.4292,
      "theoretical_loss": 3.46205064591189,
      "tokens_seen": 1793458176
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004643838383838384,
      "loss": 1.4,
      "theoretical_loss": 3.462039870653231,
      "tokens_seen": 1793523712
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004643636363636364,
      "loss": 1.3416,
      "theoretical_loss": 3.4620290958985382,
      "tokens_seen": 1793589248
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046434343434343436,
      "loss": 1.5423,
      "theoretical_loss": 3.462018321647768,
      "tokens_seen": 1793654784
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004643232323232323,
      "loss": 1.3754,
      "theoretical_loss": 3.462007547900879,
      "tokens_seen": 1793720320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004643030303030303,
      "loss": 1.6072,
      "theoretical_loss": 3.4619967746578286,
      "tokens_seen": 1793785856
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642828282828283,
      "loss": 1.3222,
      "theoretical_loss": 3.4619860019185755,
      "tokens_seen": 1793851392
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642626262626263,
      "loss": 1.375,
      "theoretical_loss": 3.4619752296830772,
      "tokens_seen": 1793916928
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046424242424242425,
      "loss": 1.3657,
      "theoretical_loss": 3.461964457951292,
      "tokens_seen": 1793982464
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642222222222222,
      "loss": 1.4207,
      "theoretical_loss": 3.461953686723178,
      "tokens_seen": 1794048000
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004642020202020202,
      "loss": 1.2671,
      "theoretical_loss": 3.461942915998694,
      "tokens_seen": 1794113536
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046418181818181823,
      "loss": 1.4485,
      "theoretical_loss": 3.4619321457777965,
      "tokens_seen": 1794179072
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641616161616162,
      "loss": 1.4694,
      "theoretical_loss": 3.4619213760604444,
      "tokens_seen": 1794244608
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.4748900234699249,
      "objective/train/avg_token_score": 0.019929148256778717,
      "objective/train/docs_used": 1015045,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.42253839969635,
      "objective/train/original_loss": 4.371203422546387,
      "objective/train/theoretical_loss": 3.4619106068465957,
      "objective/train/tokens_used": 153169376,
      "objective/train/unlikelihood_loss": 0.07094332575798035,
      "theoretical_loss": 3.4619106068465957,
      "tokens_seen": 1794310144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046414141414141414,
      "loss": 1.4498,
      "theoretical_loss": 3.4619106068465957,
      "tokens_seen": 1794310144
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641212121212121,
      "loss": 1.1417,
      "theoretical_loss": 3.4618998381362087,
      "tokens_seen": 1794375680
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004641010101010101,
      "loss": 1.1921,
      "theoretical_loss": 3.4618890699292413,
      "tokens_seen": 1794441216
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640808080808081,
      "loss": 1.1941,
      "theoretical_loss": 3.461878302225651,
      "tokens_seen": 1794506752
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004640606060606061,
      "loss": 1.4643,
      "theoretical_loss": 3.4618675350253967,
      "tokens_seen": 1794572288
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046404040404040403,
      "loss": 1.4376,
      "theoretical_loss": 3.4618567683284365,
      "tokens_seen": 1794637824
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000464020202020202,
      "loss": 1.2493,
      "theoretical_loss": 3.4618460021347275,
      "tokens_seen": 1794703360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046400000000000006,
      "loss": 1.3667,
      "theoretical_loss": 3.4618352364442293,
      "tokens_seen": 1794768896
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463979797979798,
      "loss": 1.4107,
      "theoretical_loss": 3.4618244712568984,
      "tokens_seen": 1794834432
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046395959595959597,
      "loss": 1.3153,
      "theoretical_loss": 3.4618137065726944,
      "tokens_seen": 1794899968
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046393939393939393,
      "loss": 1.3565,
      "theoretical_loss": 3.4618029423915746,
      "tokens_seen": 1794965504
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004639191919191919,
      "loss": 1.4403,
      "theoretical_loss": 3.461792178713497,
      "tokens_seen": 1795031040
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046389898989898995,
      "loss": 1.356,
      "theoretical_loss": 3.46178141553842,
      "tokens_seen": 1795096576
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004638787878787879,
      "loss": 1.1425,
      "theoretical_loss": 3.461770652866302,
      "tokens_seen": 1795162112
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046385858585858586,
      "loss": 1.3359,
      "theoretical_loss": 3.4617598906971008,
      "tokens_seen": 1795227648
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004638383838383838,
      "loss": 1.337,
      "theoretical_loss": 3.4617491290307743,
      "tokens_seen": 1795293184
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046381818181818183,
      "loss": 1.6086,
      "theoretical_loss": 3.461738367867281,
      "tokens_seen": 1795358720
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046379797979797984,
      "loss": 1.4242,
      "theoretical_loss": 3.461727607206579,
      "tokens_seen": 1795424256
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637777777777778,
      "loss": 1.5684,
      "theoretical_loss": 3.461716847048627,
      "tokens_seen": 1795489792
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046375757575757576,
      "loss": 1.4422,
      "theoretical_loss": 3.461706087393382,
      "tokens_seen": 1795555328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637373737373737,
      "loss": 1.6148,
      "theoretical_loss": 3.461695328240803,
      "tokens_seen": 1795620864
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004637171717171717,
      "loss": 1.0624,
      "theoretical_loss": 3.461684569590848,
      "tokens_seen": 1795686400
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046369696969696974,
      "loss": 1.5246,
      "theoretical_loss": 3.461673811443475,
      "tokens_seen": 1795751936
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636767676767677,
      "loss": 1.4155,
      "theoretical_loss": 3.4616630537986426,
      "tokens_seen": 1795817472
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046365656565656565,
      "loss": 1.3022,
      "theoretical_loss": 3.4616522966563084,
      "tokens_seen": 1795883008
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.38929617404937744,
      "objective/train/avg_token_score": 0.010316792875528336,
      "objective/train/docs_used": 1016473,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3159542083740234,
      "objective/train/original_loss": 4.452950954437256,
      "objective/train/theoretical_loss": 3.461641540016431,
      "objective/train/tokens_used": 154807776,
      "objective/train/unlikelihood_loss": 0.060373466461896896,
      "theoretical_loss": 3.461641540016431,
      "tokens_seen": 1795948544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046363636363636366,
      "loss": 1.411,
      "theoretical_loss": 3.461641540016431,
      "tokens_seen": 1795948544
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004636161616161616,
      "loss": 1.3186,
      "theoretical_loss": 3.4616307838789684,
      "tokens_seen": 1796014080
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046359595959595963,
      "loss": 1.4337,
      "theoretical_loss": 3.4616200282438783,
      "tokens_seen": 1796079616
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635757575757576,
      "loss": 1.3708,
      "theoretical_loss": 3.4616092731111205,
      "tokens_seen": 1796145152
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046355555555555554,
      "loss": 1.41,
      "theoretical_loss": 3.4615985184806517,
      "tokens_seen": 1796210688
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046353535353535355,
      "loss": 1.3782,
      "theoretical_loss": 3.461587764352431,
      "tokens_seen": 1796276224
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004635151515151515,
      "loss": 1.3656,
      "theoretical_loss": 3.4615770107264154,
      "tokens_seen": 1796341760
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634949494949495,
      "loss": 1.314,
      "theoretical_loss": 3.461566257602565,
      "tokens_seen": 1796407296
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634747474747475,
      "loss": 1.2375,
      "theoretical_loss": 3.4615555049808364,
      "tokens_seen": 1796472832
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634545454545455,
      "loss": 1.5319,
      "theoretical_loss": 3.4615447528611885,
      "tokens_seen": 1796538368
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046343434343434345,
      "loss": 1.5685,
      "theoretical_loss": 3.46153400124358,
      "tokens_seen": 1796603904
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004634141414141414,
      "loss": 1.321,
      "theoretical_loss": 3.461523250127968,
      "tokens_seen": 1796669440
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633939393939394,
      "loss": 1.2743,
      "theoretical_loss": 3.4615124995143116,
      "tokens_seen": 1796734976
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046337373737373737,
      "loss": 1.3477,
      "theoretical_loss": 3.4615017494025686,
      "tokens_seen": 1796800512
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633535353535354,
      "loss": 1.7626,
      "theoretical_loss": 3.461490999792698,
      "tokens_seen": 1796866048
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046333333333333334,
      "loss": 1.7518,
      "theoretical_loss": 3.461480250684657,
      "tokens_seen": 1796931584
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004633131313131313,
      "loss": 1.3579,
      "theoretical_loss": 3.4614695020784048,
      "tokens_seen": 1796997120
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632929292929293,
      "loss": 1.2737,
      "theoretical_loss": 3.4614587539738992,
      "tokens_seen": 1797062656
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632727272727273,
      "loss": 1.5971,
      "theoretical_loss": 3.4614480063710986,
      "tokens_seen": 1797128192
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632525252525253,
      "loss": 1.3421,
      "theoretical_loss": 3.461437259269961,
      "tokens_seen": 1797193728
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046323232323232323,
      "loss": 1.3835,
      "theoretical_loss": 3.4614265126704455,
      "tokens_seen": 1797259264
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004632121212121212,
      "loss": 1.3964,
      "theoretical_loss": 3.46141576657251,
      "tokens_seen": 1797324800
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631919191919192,
      "loss": 1.3049,
      "theoretical_loss": 3.461405020976112,
      "tokens_seen": 1797390336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631717171717172,
      "loss": 1.5157,
      "theoretical_loss": 3.4613942758812106,
      "tokens_seen": 1797455872
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046315151515151517,
      "loss": 1.4886,
      "theoretical_loss": 3.4613835312877645,
      "tokens_seen": 1797521408
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.45002442598342896,
      "objective/train/avg_token_score": 0.027033532038331032,
      "objective/train/docs_used": 1017203,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.404165506362915,
      "objective/train/original_loss": 4.197830677032471,
      "objective/train/theoretical_loss": 3.461372787195731,
      "objective/train/tokens_used": 156446176,
      "objective/train/unlikelihood_loss": 0.05572577565908432,
      "theoretical_loss": 3.461372787195731,
      "tokens_seen": 1797586944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631313131313131,
      "loss": 1.5094,
      "theoretical_loss": 3.461372787195731,
      "tokens_seen": 1797586944
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004631111111111111,
      "loss": 1.0144,
      "theoretical_loss": 3.461362043605069,
      "tokens_seen": 1797652480
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630909090909091,
      "loss": 1.6355,
      "theoretical_loss": 3.4613513005157373,
      "tokens_seen": 1797718016
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004630707070707071,
      "loss": 1.6244,
      "theoretical_loss": 3.4613405579276932,
      "tokens_seen": 1797783552
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046305050505050506,
      "loss": 1.4438,
      "theoretical_loss": 3.4613298158408954,
      "tokens_seen": 1797849088
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000463030303030303,
      "loss": 1.1163,
      "theoretical_loss": 3.461319074255303,
      "tokens_seen": 1797914624
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046301010101010097,
      "loss": 1.3346,
      "theoretical_loss": 3.461308333170873,
      "tokens_seen": 1797980160
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046298989898989904,
      "loss": 1.5162,
      "theoretical_loss": 3.461297592587565,
      "tokens_seen": 1798045696
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.000462969696969697,
      "loss": 1.3135,
      "theoretical_loss": 3.461286852505336,
      "tokens_seen": 1798111232
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046294949494949495,
      "loss": 1.5227,
      "theoretical_loss": 3.4612761129241463,
      "tokens_seen": 1798176768
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629292929292929,
      "loss": 1.4748,
      "theoretical_loss": 3.4612653738439523,
      "tokens_seen": 1798242304
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004629090909090909,
      "loss": 1.2214,
      "theoretical_loss": 3.4612546352647136,
      "tokens_seen": 1798307840
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046288888888888893,
      "loss": 1.1131,
      "theoretical_loss": 3.461243897186388,
      "tokens_seen": 1798373376
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004628686868686869,
      "loss": 1.3627,
      "theoretical_loss": 3.461233159608934,
      "tokens_seen": 1798438912
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046284848484848484,
      "loss": 1.5596,
      "theoretical_loss": 3.4612224225323103,
      "tokens_seen": 1798504448
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004628282828282828,
      "loss": 1.3134,
      "theoretical_loss": 3.461211685956475,
      "tokens_seen": 1798569984
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046280808080808087,
      "loss": 1.3904,
      "theoretical_loss": 3.4612009498813867,
      "tokens_seen": 1798635520
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627878787878788,
      "loss": 1.3134,
      "theoretical_loss": 3.4611902143070035,
      "tokens_seen": 1798701056
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627676767676768,
      "loss": 1.2739,
      "theoretical_loss": 3.4611794792332837,
      "tokens_seen": 1798766592
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046274747474747474,
      "loss": 1.2228,
      "theoretical_loss": 3.4611687446601866,
      "tokens_seen": 1798832128
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004627272727272727,
      "loss": 1.2976,
      "theoretical_loss": 3.4611580105876696,
      "tokens_seen": 1798897664
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046270707070707076,
      "loss": 1.3079,
      "theoretical_loss": 3.4611472770156917,
      "tokens_seen": 1798963200
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626868686868687,
      "loss": 1.3461,
      "theoretical_loss": 3.4611365439442103,
      "tokens_seen": 1799028736
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004626666666666667,
      "loss": 1.4401,
      "theoretical_loss": 3.4611258113731855,
      "tokens_seen": 1799094272
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046264646464646463,
      "loss": 1.3903,
      "theoretical_loss": 3.4611150793025747,
      "tokens_seen": 1799159808
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.5161290168762207,
      "objective/train/avg_token_score": 0.012231235392391682,
      "objective/train/docs_used": 1018428,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5293020009994507,
      "objective/train/original_loss": 3.6187143325805664,
      "objective/train/theoretical_loss": 3.461104347732337,
      "objective/train/tokens_used": 158084576,
      "objective/train/unlikelihood_loss": 0.06525947153568268,
      "theoretical_loss": 3.461104347732337,
      "tokens_seen": 1799225344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046262626262626264,
      "loss": 1.4842,
      "theoretical_loss": 3.461104347732337,
      "tokens_seen": 1799225344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046260606060606065,
      "loss": 1.5779,
      "theoretical_loss": 3.4610936166624295,
      "tokens_seen": 1799290880
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625858585858586,
      "loss": 1.5093,
      "theoretical_loss": 3.4610828860928122,
      "tokens_seen": 1799356416
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046256565656565657,
      "loss": 1.3926,
      "theoretical_loss": 3.4610721560234428,
      "tokens_seen": 1799421952
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004625454545454545,
      "loss": 1.3863,
      "theoretical_loss": 3.46106142645428,
      "tokens_seen": 1799487488
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046252525252525253,
      "loss": 1.3946,
      "theoretical_loss": 3.4610506973852813,
      "tokens_seen": 1799553024
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046250505050505054,
      "loss": 1.5555,
      "theoretical_loss": 3.4610399688164066,
      "tokens_seen": 1799618560
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624848484848485,
      "loss": 1.5781,
      "theoretical_loss": 3.4610292407476138,
      "tokens_seen": 1799684096
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046246464646464646,
      "loss": 1.3303,
      "theoretical_loss": 3.4610185131788613,
      "tokens_seen": 1799749632
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046244444444444447,
      "loss": 1.4305,
      "theoretical_loss": 3.461007786110108,
      "tokens_seen": 1799815168
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004624242424242424,
      "loss": 1.3423,
      "theoretical_loss": 3.4609970595413113,
      "tokens_seen": 1799880704
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046240404040404044,
      "loss": 1.2689,
      "theoretical_loss": 3.4609863334724307,
      "tokens_seen": 1799946240
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623838383838384,
      "loss": 1.3048,
      "theoretical_loss": 3.460975607903425,
      "tokens_seen": 1800011776
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046236363636363635,
      "loss": 1.4274,
      "theoretical_loss": 3.460964882834251,
      "tokens_seen": 1800077312
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046234343434343436,
      "loss": 1.4462,
      "theoretical_loss": 3.460954158264869,
      "tokens_seen": 1800142848
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004623232323232323,
      "loss": 1.5049,
      "theoretical_loss": 3.460943434195237,
      "tokens_seen": 1800208384
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046230303030303033,
      "loss": 1.5564,
      "theoretical_loss": 3.460932710625313,
      "tokens_seen": 1800273920
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622828282828283,
      "loss": 1.6183,
      "theoretical_loss": 3.4609219875550563,
      "tokens_seen": 1800339456
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622626262626263,
      "loss": 1.3066,
      "theoretical_loss": 3.460911264984425,
      "tokens_seen": 1800404992
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046224242424242425,
      "loss": 1.4609,
      "theoretical_loss": 3.460900542913377,
      "tokens_seen": 1800470528
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622222222222222,
      "loss": 1.0031,
      "theoretical_loss": 3.460889821341872,
      "tokens_seen": 1800536064
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004622020202020202,
      "loss": 1.5134,
      "theoretical_loss": 3.4608791002698682,
      "tokens_seen": 1800601600
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621818181818182,
      "loss": 1.3774,
      "theoretical_loss": 3.460868379697324,
      "tokens_seen": 1800667136
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0004621616161616162,
      "loss": 1.4704,
      "theoretical_loss": 3.460857659624198,
      "tokens_seen": 1800732672
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00046214141414141415,
      "loss": 1.6048,
      "theoretical_loss": 3.4608469400504482,
      "tokens_seen": 1800798208
    },
    {
      "epoch": 0.08,
      "objective/train/aligned_token_count": 0.4208821952342987,
      "objective/train/avg_token_score": 0.024829581379890442,
      "objective/train/docs_used": 1018922,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4344236850738525,
      "objective/train/original_loss": 4.255611419677734,
      "objective/train/theoretical_loss": 3.4608362209760335,
      "objective/train/tokens_used": 159722976,
      "objective/train/unlikelihood_loss": 0.07284124940633774,
      "theoretical_loss": 3.4608362209760335,
      "tokens_seen": 1800863744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004621212121212121,
      "loss": 1.3354,
      "theoretical_loss": 3.4608362209760335,
      "tokens_seen": 1800863744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004621010101010101,
      "loss": 1.73,
      "theoretical_loss": 3.4608255024009136,
      "tokens_seen": 1800929280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620808080808081,
      "loss": 1.5529,
      "theoretical_loss": 3.4608147843250454,
      "tokens_seen": 1800994816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004620606060606061,
      "loss": 1.3241,
      "theoretical_loss": 3.4608040667483886,
      "tokens_seen": 1801060352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046204040404040404,
      "loss": 1.3847,
      "theoretical_loss": 3.460793349670901,
      "tokens_seen": 1801125888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000462020202020202,
      "loss": 1.603,
      "theoretical_loss": 3.460782633092542,
      "tokens_seen": 1801191424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000462,
      "loss": 1.4548,
      "theoretical_loss": 3.4607719170132696,
      "tokens_seen": 1801256960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461979797979798,
      "loss": 1.5245,
      "theoretical_loss": 3.4607612014330424,
      "tokens_seen": 1801322496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461959595959596,
      "loss": 1.2273,
      "theoretical_loss": 3.4607504863518193,
      "tokens_seen": 1801388032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046193939393939393,
      "loss": 1.6071,
      "theoretical_loss": 3.460739771769559,
      "tokens_seen": 1801453568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004619191919191919,
      "loss": 1.4026,
      "theoretical_loss": 3.46072905768622,
      "tokens_seen": 1801519104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046189898989898996,
      "loss": 1.4037,
      "theoretical_loss": 3.4607183441017604,
      "tokens_seen": 1801584640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618787878787879,
      "loss": 1.3166,
      "theoretical_loss": 3.460707631016139,
      "tokens_seen": 1801650176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046185858585858587,
      "loss": 1.3045,
      "theoretical_loss": 3.4606969184293153,
      "tokens_seen": 1801715712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618383838383838,
      "loss": 1.7083,
      "theoretical_loss": 3.460686206341247,
      "tokens_seen": 1801781248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004618181818181818,
      "loss": 1.531,
      "theoretical_loss": 3.4606754947518934,
      "tokens_seen": 1801846784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046179797979797985,
      "loss": 1.3199,
      "theoretical_loss": 3.4606647836612128,
      "tokens_seen": 1801912320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617777777777778,
      "loss": 1.3986,
      "theoretical_loss": 3.460654073069163,
      "tokens_seen": 1801977856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046175757575757576,
      "loss": 1.3911,
      "theoretical_loss": 3.4606433629757043,
      "tokens_seen": 1802043392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617373737373737,
      "loss": 1.279,
      "theoretical_loss": 3.4606326533807943,
      "tokens_seen": 1802108928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004617171717171717,
      "loss": 1.4534,
      "theoretical_loss": 3.4606219442843917,
      "tokens_seen": 1802174464
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046169696969696974,
      "loss": 1.5236,
      "theoretical_loss": 3.460611235686456,
      "tokens_seen": 1802240000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616767676767677,
      "loss": 1.5087,
      "theoretical_loss": 3.4606005275869447,
      "tokens_seen": 1802305536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046165656565656565,
      "loss": 1.4131,
      "theoretical_loss": 3.460589819985817,
      "tokens_seen": 1802371072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616363636363636,
      "loss": 1.4507,
      "theoretical_loss": 3.460579112883032,
      "tokens_seen": 1802436608
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.33669355511665344,
      "objective/train/avg_token_score": 0.009447429329156876,
      "objective/train/docs_used": 1019593,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2540216445922852,
      "objective/train/original_loss": 4.323973655700684,
      "objective/train/theoretical_loss": 3.460568406278548,
      "objective/train/tokens_used": 161361376,
      "objective/train/unlikelihood_loss": 0.07911861687898636,
      "theoretical_loss": 3.460568406278548,
      "tokens_seen": 1802502144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004616161616161616,
      "loss": 1.3935,
      "theoretical_loss": 3.460568406278548,
      "tokens_seen": 1802502144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046159595959595963,
      "loss": 1.336,
      "theoretical_loss": 3.4605577001723233,
      "tokens_seen": 1802567680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615757575757576,
      "loss": 1.4096,
      "theoretical_loss": 3.460546994564317,
      "tokens_seen": 1802633216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046155555555555555,
      "loss": 1.3497,
      "theoretical_loss": 3.460536289454488,
      "tokens_seen": 1802698752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615353535353535,
      "loss": 1.5146,
      "theoretical_loss": 3.460525584842795,
      "tokens_seen": 1802764288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004615151515151515,
      "loss": 1.3389,
      "theoretical_loss": 3.4605148807291966,
      "tokens_seen": 1802829824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614949494949495,
      "loss": 1.3034,
      "theoretical_loss": 3.460504177113651,
      "tokens_seen": 1802895360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614747474747475,
      "loss": 1.542,
      "theoretical_loss": 3.4604934739961175,
      "tokens_seen": 1802960896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046145454545454544,
      "loss": 1.5322,
      "theoretical_loss": 3.460482771376555,
      "tokens_seen": 1803026432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046143434343434345,
      "loss": 1.4973,
      "theoretical_loss": 3.460472069254922,
      "tokens_seen": 1803091968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004614141414141414,
      "loss": 1.2724,
      "theoretical_loss": 3.4604613676311766,
      "tokens_seen": 1803157504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613939393939394,
      "loss": 1.3662,
      "theoretical_loss": 3.4604506665052783,
      "tokens_seen": 1803223040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613737373737374,
      "loss": 1.4956,
      "theoretical_loss": 3.4604399658771863,
      "tokens_seen": 1803288576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046135353535353533,
      "loss": 1.1918,
      "theoretical_loss": 3.460429265746858,
      "tokens_seen": 1803354112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046133333333333334,
      "loss": 1.2887,
      "theoretical_loss": 3.460418566114253,
      "tokens_seen": 1803419648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004613131313131313,
      "loss": 1.2515,
      "theoretical_loss": 3.4604078669793297,
      "tokens_seen": 1803485184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612929292929293,
      "loss": 1.1434,
      "theoretical_loss": 3.4603971683420474,
      "tokens_seen": 1803550720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046127272727272727,
      "loss": 1.5023,
      "theoretical_loss": 3.4603864702023643,
      "tokens_seen": 1803616256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612525252525253,
      "loss": 1.495,
      "theoretical_loss": 3.4603757725602398,
      "tokens_seen": 1803681792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046123232323232324,
      "loss": 1.493,
      "theoretical_loss": 3.4603650754156323,
      "tokens_seen": 1803747328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004612121212121212,
      "loss": 1.5071,
      "theoretical_loss": 3.4603543787685,
      "tokens_seen": 1803812864
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004611919191919192,
      "loss": 1.3662,
      "theoretical_loss": 3.460343682618803,
      "tokens_seen": 1803878400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046117171717171716,
      "loss": 1.3852,
      "theoretical_loss": 3.460332986966499,
      "tokens_seen": 1803943936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046115151515151517,
      "loss": 1.3354,
      "theoretical_loss": 3.4603222918115475,
      "tokens_seen": 1804009472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046113131313131313,
      "loss": 1.2229,
      "theoretical_loss": 3.4603115971539067,
      "tokens_seen": 1804075008
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.36766862869262695,
      "objective/train/avg_token_score": 0.01413616631180048,
      "objective/train/docs_used": 1020148,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.183333396911621,
      "objective/train/original_loss": 4.116270065307617,
      "objective/train/theoretical_loss": 3.460300902993536,
      "objective/train/tokens_used": 162999776,
      "objective/train/unlikelihood_loss": 0.08400173485279083,
      "theoretical_loss": 3.460300902993536,
      "tokens_seen": 1804140544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046111111111111114,
      "loss": 1.4689,
      "theoretical_loss": 3.460300902993536,
      "tokens_seen": 1804140544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610909090909091,
      "loss": 1.4244,
      "theoretical_loss": 3.4602902093303936,
      "tokens_seen": 1804206080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004610707070707071,
      "loss": 1.6157,
      "theoretical_loss": 3.460279516164439,
      "tokens_seen": 1804271616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046105050505050506,
      "loss": 1.3807,
      "theoretical_loss": 3.460268823495631,
      "tokens_seen": 1804337152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000461030303030303,
      "loss": 1.3098,
      "theoretical_loss": 3.4602581313239273,
      "tokens_seen": 1804402688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046101010101010103,
      "loss": 1.4589,
      "theoretical_loss": 3.460247439649288,
      "tokens_seen": 1804468224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460989898989899,
      "loss": 1.2463,
      "theoretical_loss": 3.4602367484716714,
      "tokens_seen": 1804533760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460969696969697,
      "loss": 1.5439,
      "theoretical_loss": 3.4602260577910364,
      "tokens_seen": 1804599296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046094949494949496,
      "loss": 1.3021,
      "theoretical_loss": 3.4602153676073417,
      "tokens_seen": 1804664832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609292929292929,
      "loss": 1.4273,
      "theoretical_loss": 3.4602046779205464,
      "tokens_seen": 1804730368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004609090909090909,
      "loss": 1.587,
      "theoretical_loss": 3.4601939887306097,
      "tokens_seen": 1804795904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046088888888888894,
      "loss": 1.2694,
      "theoretical_loss": 3.4601833000374898,
      "tokens_seen": 1804861440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608686868686869,
      "loss": 1.3032,
      "theoretical_loss": 3.4601726118411458,
      "tokens_seen": 1804926976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046084848484848485,
      "loss": 1.4727,
      "theoretical_loss": 3.460161924141537,
      "tokens_seen": 1804992512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608282828282828,
      "loss": 1.4011,
      "theoretical_loss": 3.4601512369386214,
      "tokens_seen": 1805058048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004608080808080808,
      "loss": 1.2257,
      "theoretical_loss": 3.4601405502323583,
      "tokens_seen": 1805123584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046078787878787883,
      "loss": 1.373,
      "theoretical_loss": 3.460129864022707,
      "tokens_seen": 1805189120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607676767676768,
      "loss": 1.2235,
      "theoretical_loss": 3.460119178309626,
      "tokens_seen": 1805254656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046074747474747474,
      "loss": 1.5574,
      "theoretical_loss": 3.4601084930930743,
      "tokens_seen": 1805320192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004607272727272727,
      "loss": 1.3258,
      "theoretical_loss": 3.4600978083730105,
      "tokens_seen": 1805385728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046070707070707077,
      "loss": 1.4703,
      "theoretical_loss": 3.460087124149394,
      "tokens_seen": 1805451264
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606868686868687,
      "loss": 1.2853,
      "theoretical_loss": 3.4600764404221835,
      "tokens_seen": 1805516800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606666666666667,
      "loss": 1.5211,
      "theoretical_loss": 3.460065757191338,
      "tokens_seen": 1805582336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046064646464646464,
      "loss": 1.5547,
      "theoretical_loss": 3.460055074456816,
      "tokens_seen": 1805647872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004606262626262626,
      "loss": 1.3416,
      "theoretical_loss": 3.4600443922185766,
      "tokens_seen": 1805713408
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.40169844031333923,
      "objective/train/avg_token_score": 0.0685867890715599,
      "objective/train/docs_used": 1021470,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3303582668304443,
      "objective/train/original_loss": 4.088999271392822,
      "objective/train/theoretical_loss": 3.4600337104765795,
      "objective/train/tokens_used": 164638176,
      "objective/train/unlikelihood_loss": 0.06166958063840866,
      "theoretical_loss": 3.4600337104765795,
      "tokens_seen": 1805778944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046060606060606066,
      "loss": 1.422,
      "theoretical_loss": 3.4600337104765795,
      "tokens_seen": 1805778944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004605858585858586,
      "loss": 1.3202,
      "theoretical_loss": 3.4600230292307828,
      "tokens_seen": 1805844480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046056565656565657,
      "loss": 1.2661,
      "theoretical_loss": 3.4600123484811456,
      "tokens_seen": 1805910016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046054545454545453,
      "loss": 1.5477,
      "theoretical_loss": 3.460001668227627,
      "tokens_seen": 1805975552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046052525252525254,
      "loss": 1.2371,
      "theoretical_loss": 3.459990988470186,
      "tokens_seen": 1806041088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046050505050505055,
      "loss": 1.3883,
      "theoretical_loss": 3.459980309208781,
      "tokens_seen": 1806106624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604848484848485,
      "loss": 1.4474,
      "theoretical_loss": 3.459969630443372,
      "tokens_seen": 1806172160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046046464646464646,
      "loss": 1.4572,
      "theoretical_loss": 3.459958952173917,
      "tokens_seen": 1806237696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004604444444444444,
      "loss": 1.3687,
      "theoretical_loss": 3.459948274400375,
      "tokens_seen": 1806303232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046042424242424243,
      "loss": 1.5883,
      "theoretical_loss": 3.459937597122706,
      "tokens_seen": 1806368768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046040404040404044,
      "loss": 1.113,
      "theoretical_loss": 3.459926920340868,
      "tokens_seen": 1806434304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603838383838384,
      "loss": 1.3119,
      "theoretical_loss": 3.4599162440548197,
      "tokens_seen": 1806499840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046036363636363636,
      "loss": 1.1213,
      "theoretical_loss": 3.4599055682645217,
      "tokens_seen": 1806565376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046034343434343437,
      "loss": 1.5036,
      "theoretical_loss": 3.459894892969931,
      "tokens_seen": 1806630912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004603232323232323,
      "loss": 1.4614,
      "theoretical_loss": 3.459884218171008,
      "tokens_seen": 1806696448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046030303030303034,
      "loss": 1.2264,
      "theoretical_loss": 3.4598735438677117,
      "tokens_seen": 1806761984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602828282828283,
      "loss": 1.2862,
      "theoretical_loss": 3.4598628700600003,
      "tokens_seen": 1806827520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046026262626262625,
      "loss": 1.4969,
      "theoretical_loss": 3.4598521967478337,
      "tokens_seen": 1806893056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046024242424242426,
      "loss": 1.3143,
      "theoretical_loss": 3.4598415239311695,
      "tokens_seen": 1806958592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004602222222222222,
      "loss": 1.3717,
      "theoretical_loss": 3.4598308516099685,
      "tokens_seen": 1807024128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046020202020202023,
      "loss": 1.2754,
      "theoretical_loss": 3.4598201797841885,
      "tokens_seen": 1807089664
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601818181818182,
      "loss": 1.3765,
      "theoretical_loss": 3.459809508453789,
      "tokens_seen": 1807155200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601616161616162,
      "loss": 1.4214,
      "theoretical_loss": 3.459798837618729,
      "tokens_seen": 1807220736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046014141414141415,
      "loss": 1.2824,
      "theoretical_loss": 3.4597881672789677,
      "tokens_seen": 1807286272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601212121212121,
      "loss": 1.2776,
      "theoretical_loss": 3.4597774974344637,
      "tokens_seen": 1807351808
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.5522360801696777,
      "objective/train/avg_token_score": 0.005671007093042135,
      "objective/train/docs_used": 1022105,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.969338059425354,
      "objective/train/original_loss": 3.9141592979431152,
      "objective/train/theoretical_loss": 3.4597668280851765,
      "objective/train/tokens_used": 166276576,
      "objective/train/unlikelihood_loss": 0.07195361703634262,
      "theoretical_loss": 3.4597668280851765,
      "tokens_seen": 1807417344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004601010101010101,
      "loss": 1.4951,
      "theoretical_loss": 3.4597668280851765,
      "tokens_seen": 1807417344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600808080808081,
      "loss": 1.4208,
      "theoretical_loss": 3.4597561592310644,
      "tokens_seen": 1807482880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004600606060606061,
      "loss": 1.4108,
      "theoretical_loss": 3.4597454908720877,
      "tokens_seen": 1807548416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046004040404040405,
      "loss": 1.5542,
      "theoretical_loss": 3.459734823008205,
      "tokens_seen": 1807613952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000460020202020202,
      "loss": 1.2047,
      "theoretical_loss": 3.4597241556393747,
      "tokens_seen": 1807679488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00046,
      "loss": 1.4753,
      "theoretical_loss": 3.4597134887655567,
      "tokens_seen": 1807745024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459979797979798,
      "loss": 1.4563,
      "theoretical_loss": 3.459702822386709,
      "tokens_seen": 1807810560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000459959595959596,
      "loss": 1.4975,
      "theoretical_loss": 3.4596921565027925,
      "tokens_seen": 1807876096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045993939393939394,
      "loss": 1.5808,
      "theoretical_loss": 3.4596814911137646,
      "tokens_seen": 1807941632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004599191919191919,
      "loss": 1.3581,
      "theoretical_loss": 3.459670826219585,
      "tokens_seen": 1808007168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598989898989899,
      "loss": 1.4058,
      "theoretical_loss": 3.459660161820213,
      "tokens_seen": 1808072704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598787878787879,
      "loss": 1.4865,
      "theoretical_loss": 3.459649497915608,
      "tokens_seen": 1808138240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598585858585859,
      "loss": 1.3384,
      "theoretical_loss": 3.459638834505728,
      "tokens_seen": 1808203776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045983838383838383,
      "loss": 1.2745,
      "theoretical_loss": 3.4596281715905333,
      "tokens_seen": 1808269312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004598181818181818,
      "loss": 1.2995,
      "theoretical_loss": 3.4596175091699823,
      "tokens_seen": 1808334848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597979797979798,
      "loss": 1.2344,
      "theoretical_loss": 3.459606847244034,
      "tokens_seen": 1808400384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597777777777778,
      "loss": 1.2524,
      "theoretical_loss": 3.4595961858126483,
      "tokens_seen": 1808465920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045975757575757577,
      "loss": 1.3612,
      "theoretical_loss": 3.459585524875784,
      "tokens_seen": 1808531456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597373737373737,
      "loss": 1.3054,
      "theoretical_loss": 3.4595748644334,
      "tokens_seen": 1808596992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004597171717171717,
      "loss": 1.4338,
      "theoretical_loss": 3.4595642044854555,
      "tokens_seen": 1808662528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045969696969696975,
      "loss": 1.3573,
      "theoretical_loss": 3.45955354503191,
      "tokens_seen": 1808728064
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004596767676767677,
      "loss": 1.2754,
      "theoretical_loss": 3.4595428860727218,
      "tokens_seen": 1808793600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045965656565656566,
      "loss": 1.45,
      "theoretical_loss": 3.4595322276078515,
      "tokens_seen": 1808859136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004596363636363636,
      "loss": 1.3942,
      "theoretical_loss": 3.459521569637257,
      "tokens_seen": 1808924672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004596161616161616,
      "loss": 1.5969,
      "theoretical_loss": 3.459510912160898,
      "tokens_seen": 1808990208
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.4491080045700073,
      "objective/train/avg_token_score": 0.009337084367871284,
      "objective/train/docs_used": 1022607,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.514532446861267,
      "objective/train/original_loss": 3.918682336807251,
      "objective/train/theoretical_loss": 3.4595002551787335,
      "objective/train/tokens_used": 167914976,
      "objective/train/unlikelihood_loss": 0.08269507437944412,
      "theoretical_loss": 3.4595002551787335,
      "tokens_seen": 1809055744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045959595959595964,
      "loss": 1.38,
      "theoretical_loss": 3.4595002551787335,
      "tokens_seen": 1809055744
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004595757575757576,
      "loss": 1.4447,
      "theoretical_loss": 3.4594895986907233,
      "tokens_seen": 1809121280
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045955555555555555,
      "loss": 1.2231,
      "theoretical_loss": 3.4594789426968253,
      "tokens_seen": 1809186816
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004595353535353535,
      "loss": 1.5118,
      "theoretical_loss": 3.4594682871969997,
      "tokens_seen": 1809252352
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004595151515151515,
      "loss": 1.3982,
      "theoretical_loss": 3.4594576321912056,
      "tokens_seen": 1809317888
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045949494949494953,
      "loss": 1.3473,
      "theoretical_loss": 3.459446977679402,
      "tokens_seen": 1809383424
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594747474747475,
      "loss": 1.4469,
      "theoretical_loss": 3.4594363236615484,
      "tokens_seen": 1809448960
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045945454545454545,
      "loss": 1.1953,
      "theoretical_loss": 3.4594256701376036,
      "tokens_seen": 1809514496
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004594343434343434,
      "loss": 1.4947,
      "theoretical_loss": 3.459415017107527,
      "tokens_seen": 1809580032
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045941414141414147,
      "loss": 1.1828,
      "theoretical_loss": 3.4594043645712773,
      "tokens_seen": 1809645568
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004593939393939394,
      "loss": 1.5579,
      "theoretical_loss": 3.459393712528815,
      "tokens_seen": 1809711104
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004593737373737374,
      "loss": 1.2093,
      "theoretical_loss": 3.459383060980098,
      "tokens_seen": 1809776640
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045935353535353534,
      "loss": 1.3925,
      "theoretical_loss": 3.459372409925087,
      "tokens_seen": 1809842176
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045933333333333335,
      "loss": 1.5784,
      "theoretical_loss": 3.4593617593637394,
      "tokens_seen": 1809907712
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045931313131313136,
      "loss": 1.4518,
      "theoretical_loss": 3.4593511092960156,
      "tokens_seen": 1809973248
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592929292929293,
      "loss": 1.2988,
      "theoretical_loss": 3.459340459721875,
      "tokens_seen": 1810038784
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004592727272727273,
      "loss": 1.4289,
      "theoretical_loss": 3.459329810641276,
      "tokens_seen": 1810104320
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045925252525252523,
      "loss": 1.4281,
      "theoretical_loss": 3.4593191620541783,
      "tokens_seen": 1810169856
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045923232323232324,
      "loss": 1.2985,
      "theoretical_loss": 3.4593085139605417,
      "tokens_seen": 1810235392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045921212121212125,
      "loss": 1.3204,
      "theoretical_loss": 3.459297866360325,
      "tokens_seen": 1810300928
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004591919191919192,
      "loss": 1.4494,
      "theoretical_loss": 3.459287219253487,
      "tokens_seen": 1810366464
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045917171717171717,
      "loss": 1.5598,
      "theoretical_loss": 3.459276572639988,
      "tokens_seen": 1810432000
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004591515151515152,
      "loss": 1.3901,
      "theoretical_loss": 3.459265926519786,
      "tokens_seen": 1810497536
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045913131313131313,
      "loss": 1.3048,
      "theoretical_loss": 3.459255280892841,
      "tokens_seen": 1810563072
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045911111111111115,
      "loss": 1.4061,
      "theoretical_loss": 3.459244635759113,
      "tokens_seen": 1810628608
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.2624022364616394,
      "objective/train/avg_token_score": 0.027153268456459045,
      "objective/train/docs_used": 1024104,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0158097743988037,
      "objective/train/original_loss": 5.203606128692627,
      "objective/train/theoretical_loss": 3.459233991118561,
      "objective/train/tokens_used": 169553376,
      "objective/train/unlikelihood_loss": 0.043073952198028564,
      "theoretical_loss": 3.459233991118561,
      "tokens_seen": 1810694144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004590909090909091,
      "loss": 1.26,
      "theoretical_loss": 3.459233991118561,
      "tokens_seen": 1810694144
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045907070707070706,
      "loss": 1.2706,
      "theoretical_loss": 3.4592233469711426,
      "tokens_seen": 1810759680
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045905050505050507,
      "loss": 1.0934,
      "theoretical_loss": 3.459212703316819,
      "tokens_seen": 1810825216
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045903030303030303,
      "loss": 1.3351,
      "theoretical_loss": 3.4592020601555493,
      "tokens_seen": 1810890752
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045901010101010104,
      "loss": 1.2675,
      "theoretical_loss": 3.4591914174872915,
      "tokens_seen": 1810956288
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458989898989899,
      "loss": 1.3725,
      "theoretical_loss": 3.4591807753120065,
      "tokens_seen": 1811021824
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458969696969697,
      "loss": 1.4893,
      "theoretical_loss": 3.4591701336296534,
      "tokens_seen": 1811087360
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045894949494949496,
      "loss": 1.347,
      "theoretical_loss": 3.4591594924401905,
      "tokens_seen": 1811152896
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004589292929292929,
      "loss": 1.4972,
      "theoretical_loss": 3.459148851743578,
      "tokens_seen": 1811218432
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045890909090909093,
      "loss": 1.2867,
      "theoretical_loss": 3.459138211539775,
      "tokens_seen": 1811283968
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004588888888888889,
      "loss": 1.4704,
      "theoretical_loss": 3.4591275718287404,
      "tokens_seen": 1811349504
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004588686868686869,
      "loss": 1.27,
      "theoretical_loss": 3.459116932610435,
      "tokens_seen": 1811415040
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045884848484848486,
      "loss": 1.3998,
      "theoretical_loss": 3.4591062938848163,
      "tokens_seen": 1811480576
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004588282828282828,
      "loss": 1.4274,
      "theoretical_loss": 3.4590956556518444,
      "tokens_seen": 1811546112
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004588080808080808,
      "loss": 1.5243,
      "theoretical_loss": 3.4590850179114794,
      "tokens_seen": 1811611648
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045878787878787883,
      "loss": 1.3508,
      "theoretical_loss": 3.45907438066368,
      "tokens_seen": 1811677184
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587676767676768,
      "loss": 1.4852,
      "theoretical_loss": 3.4590637439084055,
      "tokens_seen": 1811742720
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045874747474747475,
      "loss": 1.3903,
      "theoretical_loss": 3.459053107645615,
      "tokens_seen": 1811808256
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587272727272727,
      "loss": 1.5327,
      "theoretical_loss": 3.459042471875269,
      "tokens_seen": 1811873792
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004587070707070707,
      "loss": 1.202,
      "theoretical_loss": 3.4590318365973256,
      "tokens_seen": 1811939328
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045868686868686873,
      "loss": 1.2374,
      "theoretical_loss": 3.459021201811745,
      "tokens_seen": 1812004864
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004586666666666667,
      "loss": 1.3289,
      "theoretical_loss": 3.4590105675184866,
      "tokens_seen": 1812070400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045864646464646464,
      "loss": 1.4467,
      "theoretical_loss": 3.458999933717509,
      "tokens_seen": 1812135936
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004586262626262626,
      "loss": 1.1531,
      "theoretical_loss": 3.4589893004087733,
      "tokens_seen": 1812201472
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045860606060606066,
      "loss": 1.4899,
      "theoretical_loss": 3.4589786675922367,
      "tokens_seen": 1812267008
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.4100073277950287,
      "objective/train/avg_token_score": 0.012066290713846684,
      "objective/train/docs_used": 1024786,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3427298069000244,
      "objective/train/original_loss": 4.287503242492676,
      "objective/train/theoretical_loss": 3.4589680352678602,
      "objective/train/tokens_used": 171191776,
      "objective/train/unlikelihood_loss": 0.05709630250930786,
      "theoretical_loss": 3.4589680352678602,
      "tokens_seen": 1812332544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585858585858586,
      "loss": 1.2515,
      "theoretical_loss": 3.4589680352678602,
      "tokens_seen": 1812332544
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585656565656566,
      "loss": 1.2303,
      "theoretical_loss": 3.4589574034356025,
      "tokens_seen": 1812398080
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045854545454545453,
      "loss": 1.6304,
      "theoretical_loss": 3.4589467720954232,
      "tokens_seen": 1812463616
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004585252525252525,
      "loss": 1.3892,
      "theoretical_loss": 3.4589361412472828,
      "tokens_seen": 1812529152
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045850505050505056,
      "loss": 1.4854,
      "theoretical_loss": 3.4589255108911385,
      "tokens_seen": 1812594688
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004584848484848485,
      "loss": 1.5415,
      "theoretical_loss": 3.458914881026952,
      "tokens_seen": 1812660224
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045846464646464647,
      "loss": 1.4956,
      "theoretical_loss": 3.458904251654681,
      "tokens_seen": 1812725760
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004584444444444444,
      "loss": 1.3248,
      "theoretical_loss": 3.458893622774286,
      "tokens_seen": 1812791296
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004584242424242424,
      "loss": 1.3537,
      "theoretical_loss": 3.4588829943857258,
      "tokens_seen": 1812856832
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045840404040404045,
      "loss": 1.4755,
      "theoretical_loss": 3.4588723664889605,
      "tokens_seen": 1812922368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583838383838384,
      "loss": 1.2218,
      "theoretical_loss": 3.4588617390839493,
      "tokens_seen": 1812987904
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045836363636363636,
      "loss": 1.3251,
      "theoretical_loss": 3.4588511121706516,
      "tokens_seen": 1813053440
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004583434343434343,
      "loss": 1.4598,
      "theoretical_loss": 3.458840485749027,
      "tokens_seen": 1813118976
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045832323232323233,
      "loss": 1.6122,
      "theoretical_loss": 3.4588298598190352,
      "tokens_seen": 1813184512
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045830303030303034,
      "loss": 1.1923,
      "theoretical_loss": 3.458819234380635,
      "tokens_seen": 1813250048
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582828282828283,
      "loss": 1.3857,
      "theoretical_loss": 3.4588086094337864,
      "tokens_seen": 1813315584
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045826262626262626,
      "loss": 1.4029,
      "theoretical_loss": 3.4587979849784487,
      "tokens_seen": 1813381120
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582424242424242,
      "loss": 1.3769,
      "theoretical_loss": 3.458787361014582,
      "tokens_seen": 1813446656
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004582222222222222,
      "loss": 1.3538,
      "theoretical_loss": 3.4587767375421445,
      "tokens_seen": 1813512192
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045820202020202023,
      "loss": 1.2557,
      "theoretical_loss": 3.458766114561097,
      "tokens_seen": 1813577728
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004581818181818182,
      "loss": 1.4583,
      "theoretical_loss": 3.458755492071398,
      "tokens_seen": 1813643264
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045816161616161615,
      "loss": 1.3099,
      "theoretical_loss": 3.458744870073008,
      "tokens_seen": 1813708800
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045814141414141416,
      "loss": 1.2544,
      "theoretical_loss": 3.458734248565886,
      "tokens_seen": 1813774336
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004581212121212121,
      "loss": 1.461,
      "theoretical_loss": 3.4587236275499915,
      "tokens_seen": 1813839872
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004581010101010101,
      "loss": 1.1234,
      "theoretical_loss": 3.458713007025284,
      "tokens_seen": 1813905408
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.4995112419128418,
      "objective/train/avg_token_score": 0.0071546295657753944,
      "objective/train/docs_used": 1026014,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7556525468826294,
      "objective/train/original_loss": 3.839704990386963,
      "objective/train/theoretical_loss": 3.4587023869917237,
      "objective/train/tokens_used": 172830176,
      "objective/train/unlikelihood_loss": 0.09379331767559052,
      "theoretical_loss": 3.4587023869917237,
      "tokens_seen": 1813970944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004580808080808081,
      "loss": 1.4608,
      "theoretical_loss": 3.4587023869917237,
      "tokens_seen": 1813970944
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045806060606060604,
      "loss": 1.3605,
      "theoretical_loss": 3.458691767449269,
      "tokens_seen": 1814036480
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045804040404040405,
      "loss": 1.1711,
      "theoretical_loss": 3.45868114839788,
      "tokens_seen": 1814102016
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458020202020202,
      "loss": 1.3351,
      "theoretical_loss": 3.4586705298375167,
      "tokens_seen": 1814167552
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000458,
      "loss": 1.5054,
      "theoretical_loss": 3.4586599117681383,
      "tokens_seen": 1814233088
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457979797979798,
      "loss": 1.6297,
      "theoretical_loss": 3.458649294189704,
      "tokens_seen": 1814298624
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.000457959595959596,
      "loss": 1.4436,
      "theoretical_loss": 3.458638677102174,
      "tokens_seen": 1814364160
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045793939393939394,
      "loss": 1.3552,
      "theoretical_loss": 3.4586280605055073,
      "tokens_seen": 1814429696
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004579191919191919,
      "loss": 1.5167,
      "theoretical_loss": 3.458617444399664,
      "tokens_seen": 1814495232
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578989898989899,
      "loss": 1.4367,
      "theoretical_loss": 3.4586068287846032,
      "tokens_seen": 1814560768
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045787878787878787,
      "loss": 1.4934,
      "theoretical_loss": 3.458596213660285,
      "tokens_seen": 1814626304
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578585858585859,
      "loss": 1.3668,
      "theoretical_loss": 3.4585855990266685,
      "tokens_seen": 1814691840
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045783838383838384,
      "loss": 1.2387,
      "theoretical_loss": 3.4585749848837133,
      "tokens_seen": 1814757376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004578181818181818,
      "loss": 1.2531,
      "theoretical_loss": 3.45856437123138,
      "tokens_seen": 1814822912
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004577979797979798,
      "loss": 1.4417,
      "theoretical_loss": 3.4585537580696264,
      "tokens_seen": 1814888448
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004577777777777778,
      "loss": 1.6412,
      "theoretical_loss": 3.458543145398414,
      "tokens_seen": 1814953984
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045775757575757577,
      "loss": 1.6142,
      "theoretical_loss": 3.458532533217701,
      "tokens_seen": 1815019520
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045773737373737373,
      "loss": 1.5157,
      "theoretical_loss": 3.4585219215274474,
      "tokens_seen": 1815085056
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045771717171717174,
      "loss": 1.43,
      "theoretical_loss": 3.458511310327613,
      "tokens_seen": 1815150592
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576969696969697,
      "loss": 1.2891,
      "theoretical_loss": 3.458500699618158,
      "tokens_seen": 1815216128
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576767676767677,
      "loss": 1.4889,
      "theoretical_loss": 3.458490089399041,
      "tokens_seen": 1815281664
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045765656565656567,
      "loss": 1.5082,
      "theoretical_loss": 3.4584794796702223,
      "tokens_seen": 1815347200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004576363636363636,
      "loss": 1.3993,
      "theoretical_loss": 3.458468870431661,
      "tokens_seen": 1815412736
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045761616161616163,
      "loss": 1.4451,
      "theoretical_loss": 3.4584582616833175,
      "tokens_seen": 1815478272
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045759595959595964,
      "loss": 1.3518,
      "theoretical_loss": 3.458447653425151,
      "tokens_seen": 1815543808
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.34109237790107727,
      "objective/train/avg_token_score": 0.014077051542699337,
      "objective/train/docs_used": 1026828,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3032065629959106,
      "objective/train/original_loss": 4.5359601974487305,
      "objective/train/theoretical_loss": 3.4584370456571207,
      "objective/train/tokens_used": 174468576,
      "objective/train/unlikelihood_loss": 0.057071033865213394,
      "theoretical_loss": 3.4584370456571207,
      "tokens_seen": 1815609344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004575757575757576,
      "loss": 1.4086,
      "theoretical_loss": 3.4584370456571207,
      "tokens_seen": 1815609344
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045755555555555556,
      "loss": 1.3255,
      "theoretical_loss": 3.4584264383791874,
      "tokens_seen": 1815674880
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004575353535353535,
      "loss": 1.5165,
      "theoretical_loss": 3.4584158315913096,
      "tokens_seen": 1815740416
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004575151515151515,
      "loss": 1.4522,
      "theoretical_loss": 3.458405225293448,
      "tokens_seen": 1815805952
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045749494949494954,
      "loss": 1.4464,
      "theoretical_loss": 3.458394619485561,
      "tokens_seen": 1815871488
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004574747474747475,
      "loss": 1.495,
      "theoretical_loss": 3.4583840141676094,
      "tokens_seen": 1815937024
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045745454545454545,
      "loss": 1.6631,
      "theoretical_loss": 3.458373409339553,
      "tokens_seen": 1816002560
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004574343434343434,
      "loss": 1.4861,
      "theoretical_loss": 3.4583628050013506,
      "tokens_seen": 1816068096
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004574141414141415,
      "loss": 1.312,
      "theoretical_loss": 3.4583522011529624,
      "tokens_seen": 1816133632
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045739393939393943,
      "loss": 1.3567,
      "theoretical_loss": 3.458341597794348,
      "tokens_seen": 1816199168
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004573737373737374,
      "loss": 1.4321,
      "theoretical_loss": 3.4583309949254675,
      "tokens_seen": 1816264704
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045735353535353534,
      "loss": 1.3804,
      "theoretical_loss": 3.4583203925462795,
      "tokens_seen": 1816330240
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004573333333333333,
      "loss": 1.292,
      "theoretical_loss": 3.4583097906567453,
      "tokens_seen": 1816395776
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045731313131313137,
      "loss": 1.3965,
      "theoretical_loss": 3.458299189256823,
      "tokens_seen": 1816461312
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004572929292929293,
      "loss": 1.532,
      "theoretical_loss": 3.458288588346474,
      "tokens_seen": 1816526848
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004572727272727273,
      "loss": 1.5313,
      "theoretical_loss": 3.4582779879256567,
      "tokens_seen": 1816592384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045725252525252524,
      "loss": 1.314,
      "theoretical_loss": 3.458267387994331,
      "tokens_seen": 1816657920
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045723232323232325,
      "loss": 1.4601,
      "theoretical_loss": 3.458256788552457,
      "tokens_seen": 1816723456
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045721212121212126,
      "loss": 1.169,
      "theoretical_loss": 3.4582461895999947,
      "tokens_seen": 1816788992
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004571919191919192,
      "loss": 1.7361,
      "theoretical_loss": 3.458235591136903,
      "tokens_seen": 1816854528
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045717171717171717,
      "loss": 1.4227,
      "theoretical_loss": 3.458224993163143,
      "tokens_seen": 1816920064
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045715151515151513,
      "loss": 1.4055,
      "theoretical_loss": 3.458214395678673,
      "tokens_seen": 1816985600
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045713131313131314,
      "loss": 1.37,
      "theoretical_loss": 3.4582037986834537,
      "tokens_seen": 1817051136
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00045711111111111115,
      "loss": 1.6846,
      "theoretical_loss": 3.458193202177444,
      "tokens_seen": 1817116672
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0004570909090909091,
      "loss": 1.2868,
      "theoretical_loss": 3.4581826061606047,
      "tokens_seen": 1817182208
    },
    {
      "epoch": 0.09,
      "objective/train/aligned_token_count": 0.3228861093521118,
      "objective/train/avg_token_score": 0.038064464926719666,
      "objective/train/docs_used": 1028029,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0878517627716064,
      "objective/train/original_loss": 4.760405540466309,
      "objective/train/theoretical_loss": 3.458172010632895,
      "objective/train/tokens_used": 176106976,
      "objective/train/unlikelihood_loss": 0.04580136016011238,
      "theoretical_loss": 3.458172010632895,
      "tokens_seen": 1817247744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045707070707070706,
      "loss": 1.159,
      "theoretical_loss": 3.458172010632895,
      "tokens_seen": 1817247744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004570505050505051,
      "loss": 1.4579,
      "theoretical_loss": 3.458161415594275,
      "tokens_seen": 1817313280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045703030303030303,
      "loss": 1.3157,
      "theoretical_loss": 3.458150821044704,
      "tokens_seen": 1817378816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045701010101010104,
      "loss": 1.324,
      "theoretical_loss": 3.458140226984142,
      "tokens_seen": 1817444352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000456989898989899,
      "loss": 1.2537,
      "theoretical_loss": 3.458129633412549,
      "tokens_seen": 1817509888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045696969696969696,
      "loss": 1.6337,
      "theoretical_loss": 3.4581190403298847,
      "tokens_seen": 1817575424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045694949494949497,
      "loss": 1.3909,
      "theoretical_loss": 3.458108447736109,
      "tokens_seen": 1817640960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004569292929292929,
      "loss": 1.3148,
      "theoretical_loss": 3.4580978556311814,
      "tokens_seen": 1817706496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045690909090909094,
      "loss": 1.2169,
      "theoretical_loss": 3.458087264015062,
      "tokens_seen": 1817772032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004568888888888889,
      "loss": 1.5658,
      "theoretical_loss": 3.45807667288771,
      "tokens_seen": 1817837568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004568686868686869,
      "loss": 1.5285,
      "theoretical_loss": 3.4580660822490863,
      "tokens_seen": 1817903104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045684848484848486,
      "loss": 1.2645,
      "theoretical_loss": 3.45805549209915,
      "tokens_seen": 1817968640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004568282828282828,
      "loss": 1.5466,
      "theoretical_loss": 3.4580449024378614,
      "tokens_seen": 1818034176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045680808080808083,
      "loss": 1.4763,
      "theoretical_loss": 3.4580343132651796,
      "tokens_seen": 1818099712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567878787878788,
      "loss": 1.3013,
      "theoretical_loss": 3.458023724581065,
      "tokens_seen": 1818165248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567676767676768,
      "loss": 1.4691,
      "theoretical_loss": 3.4580131363854782,
      "tokens_seen": 1818230784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045674747474747475,
      "loss": 1.3829,
      "theoretical_loss": 3.458002548678377,
      "tokens_seen": 1818296320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567272727272727,
      "loss": 1.4288,
      "theoretical_loss": 3.457991961459723,
      "tokens_seen": 1818361856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004567070707070707,
      "loss": 1.7387,
      "theoretical_loss": 3.4579813747294756,
      "tokens_seen": 1818427392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045668686868686873,
      "loss": 1.2347,
      "theoretical_loss": 3.4579707884875948,
      "tokens_seen": 1818492928
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566666666666667,
      "loss": 1.3851,
      "theoretical_loss": 3.4579602027340393,
      "tokens_seen": 1818558464
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045664646464646465,
      "loss": 1.4735,
      "theoretical_loss": 3.457949617468771,
      "tokens_seen": 1818624000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566262626262626,
      "loss": 1.4051,
      "theoretical_loss": 3.457939032691748,
      "tokens_seen": 1818689536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004566060606060606,
      "loss": 1.4462,
      "theoretical_loss": 3.4579284484029316,
      "tokens_seen": 1818755072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565858585858586,
      "loss": 1.2984,
      "theoretical_loss": 3.45791786460228,
      "tokens_seen": 1818820608
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.464809387922287,
      "objective/train/avg_token_score": 0.016029013320803642,
      "objective/train/docs_used": 1028792,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6073720455169678,
      "objective/train/original_loss": 3.9546051025390625,
      "objective/train/theoretical_loss": 3.4579072812897547,
      "objective/train/tokens_used": 177745376,
      "objective/train/unlikelihood_loss": 0.07868336141109467,
      "theoretical_loss": 3.4579072812897547,
      "tokens_seen": 1818886144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565656565656566,
      "loss": 1.5138,
      "theoretical_loss": 3.4579072812897547,
      "tokens_seen": 1818886144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045654545454545454,
      "loss": 1.4139,
      "theoretical_loss": 3.457896698465315,
      "tokens_seen": 1818951680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565252525252525,
      "loss": 1.5099,
      "theoretical_loss": 3.457886116128921,
      "tokens_seen": 1819017216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004565050505050505,
      "loss": 1.3184,
      "theoretical_loss": 3.4578755342805323,
      "tokens_seen": 1819082752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004564848484848485,
      "loss": 1.4564,
      "theoretical_loss": 3.457864952920109,
      "tokens_seen": 1819148288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004564646464646465,
      "loss": 1.4435,
      "theoretical_loss": 3.4578543720476103,
      "tokens_seen": 1819213824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045644444444444443,
      "loss": 1.6601,
      "theoretical_loss": 3.4578437916629974,
      "tokens_seen": 1819279360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004564242424242424,
      "loss": 1.3769,
      "theoretical_loss": 3.4578332117662294,
      "tokens_seen": 1819344896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045640404040404045,
      "loss": 1.4961,
      "theoretical_loss": 3.4578226323572663,
      "tokens_seen": 1819410432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004563838383838384,
      "loss": 1.2399,
      "theoretical_loss": 3.4578120534360686,
      "tokens_seen": 1819475968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045636363636363637,
      "loss": 1.5018,
      "theoretical_loss": 3.4578014750025954,
      "tokens_seen": 1819541504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004563434343434343,
      "loss": 1.4262,
      "theoretical_loss": 3.457790897056807,
      "tokens_seen": 1819607040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004563232323232323,
      "loss": 1.4682,
      "theoretical_loss": 3.4577803195986636,
      "tokens_seen": 1819672576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045630303030303035,
      "loss": 1.5494,
      "theoretical_loss": 3.457769742628125,
      "tokens_seen": 1819738112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562828282828283,
      "loss": 1.4778,
      "theoretical_loss": 3.4577591661451517,
      "tokens_seen": 1819803648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045626262626262626,
      "loss": 1.459,
      "theoretical_loss": 3.457748590149703,
      "tokens_seen": 1819869184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004562424242424242,
      "loss": 1.5946,
      "theoretical_loss": 3.457738014641738,
      "tokens_seen": 1819934720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045622222222222223,
      "loss": 1.3618,
      "theoretical_loss": 3.4577274396212183,
      "tokens_seen": 1820000256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045620202020202024,
      "loss": 1.2889,
      "theoretical_loss": 3.457716865088103,
      "tokens_seen": 1820065792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561818181818182,
      "loss": 1.3204,
      "theoretical_loss": 3.457706291042353,
      "tokens_seen": 1820131328
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045616161616161615,
      "loss": 1.4486,
      "theoretical_loss": 3.4576957174839267,
      "tokens_seen": 1820196864
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561414141414141,
      "loss": 1.2579,
      "theoretical_loss": 3.4576851444127854,
      "tokens_seen": 1820262400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004561212121212121,
      "loss": 1.2631,
      "theoretical_loss": 3.457674571828889,
      "tokens_seen": 1820327936
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045610101010101013,
      "loss": 1.2376,
      "theoretical_loss": 3.457663999732197,
      "tokens_seen": 1820393472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004560808080808081,
      "loss": 1.3863,
      "theoretical_loss": 3.4576534281226694,
      "tokens_seen": 1820459008
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.5623167157173157,
      "objective/train/avg_token_score": 0.014118832536041737,
      "objective/train/docs_used": 1029550,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7937812805175781,
      "objective/train/original_loss": 3.694382429122925,
      "objective/train/theoretical_loss": 3.457642857000267,
      "objective/train/tokens_used": 179383776,
      "objective/train/unlikelihood_loss": 0.07253962755203247,
      "theoretical_loss": 3.457642857000267,
      "tokens_seen": 1820524544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045606060606060605,
      "loss": 1.5616,
      "theoretical_loss": 3.457642857000267,
      "tokens_seen": 1820524544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045604040404040406,
      "loss": 1.4623,
      "theoretical_loss": 3.457632286364949,
      "tokens_seen": 1820590080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045602020202020207,
      "loss": 1.5794,
      "theoretical_loss": 3.4576217162166754,
      "tokens_seen": 1820655616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000456,
      "loss": 1.4721,
      "theoretical_loss": 3.457611146555407,
      "tokens_seen": 1820721152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000455979797979798,
      "loss": 1.5718,
      "theoretical_loss": 3.457600577381103,
      "tokens_seen": 1820786688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045595959595959594,
      "loss": 1.3938,
      "theoretical_loss": 3.4575900086937237,
      "tokens_seen": 1820852224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045593939393939395,
      "loss": 1.5038,
      "theoretical_loss": 3.4575794404932294,
      "tokens_seen": 1820917760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045591919191919196,
      "loss": 1.4119,
      "theoretical_loss": 3.4575688727795804,
      "tokens_seen": 1820983296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004558989898989899,
      "loss": 1.3345,
      "theoretical_loss": 3.4575583055527357,
      "tokens_seen": 1821048832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004558787878787879,
      "loss": 1.2874,
      "theoretical_loss": 3.4575477388126563,
      "tokens_seen": 1821114368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004558585858585859,
      "loss": 1.1633,
      "theoretical_loss": 3.457537172559302,
      "tokens_seen": 1821179904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045583838383838384,
      "loss": 1.4255,
      "theoretical_loss": 3.4575266067926327,
      "tokens_seen": 1821245440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045581818181818185,
      "loss": 1.4251,
      "theoretical_loss": 3.4575160415126085,
      "tokens_seen": 1821310976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004557979797979798,
      "loss": 1.4595,
      "theoretical_loss": 3.4575054767191897,
      "tokens_seen": 1821376512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045577777777777777,
      "loss": 1.4112,
      "theoretical_loss": 3.4574949124123364,
      "tokens_seen": 1821442048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004557575757575758,
      "loss": 1.3731,
      "theoretical_loss": 3.457484348592008,
      "tokens_seen": 1821507584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045573737373737374,
      "loss": 1.6005,
      "theoretical_loss": 3.4574737852581654,
      "tokens_seen": 1821573120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045571717171717175,
      "loss": 1.4437,
      "theoretical_loss": 3.4574632224107686,
      "tokens_seen": 1821638656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556969696969697,
      "loss": 1.6006,
      "theoretical_loss": 3.4574526600497775,
      "tokens_seen": 1821704192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004556767676767677,
      "loss": 1.3622,
      "theoretical_loss": 3.457442098175152,
      "tokens_seen": 1821769728
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045565656565656567,
      "loss": 1.4977,
      "theoretical_loss": 3.457431536786853,
      "tokens_seen": 1821835264
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045563636363636363,
      "loss": 1.5852,
      "theoretical_loss": 3.4574209758848395,
      "tokens_seen": 1821900800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045561616161616164,
      "loss": 1.3857,
      "theoretical_loss": 3.457410415469072,
      "tokens_seen": 1821966336
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555959595959596,
      "loss": 1.5078,
      "theoretical_loss": 3.4573998555395113,
      "tokens_seen": 1822031872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555757575757576,
      "loss": 1.4179,
      "theoretical_loss": 3.4573892960961166,
      "tokens_seen": 1822097408
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.4111681282520294,
      "objective/train/avg_token_score": 0.017115775495767593,
      "objective/train/docs_used": 1030259,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2357131242752075,
      "objective/train/original_loss": 3.788912296295166,
      "objective/train/theoretical_loss": 3.457378737138849,
      "objective/train/tokens_used": 181022176,
      "objective/train/unlikelihood_loss": 0.0951184332370758,
      "theoretical_loss": 3.457378737138849,
      "tokens_seen": 1822162944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045555555555555556,
      "loss": 1.2408,
      "theoretical_loss": 3.457378737138849,
      "tokens_seen": 1822162944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004555353535353535,
      "loss": 1.3708,
      "theoretical_loss": 3.4573681786676675,
      "tokens_seen": 1822228480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045551515151515153,
      "loss": 1.5096,
      "theoretical_loss": 3.457357620682533,
      "tokens_seen": 1822294016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045549494949494954,
      "loss": 1.3569,
      "theoretical_loss": 3.4573470631834056,
      "tokens_seen": 1822359552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004554747474747475,
      "loss": 1.3497,
      "theoretical_loss": 3.457336506170245,
      "tokens_seen": 1822425088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045545454545454546,
      "loss": 1.5165,
      "theoretical_loss": 3.457325949643012,
      "tokens_seen": 1822490624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004554343434343434,
      "loss": 1.5752,
      "theoretical_loss": 3.4573153936016663,
      "tokens_seen": 1822556160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004554141414141414,
      "loss": 1.2614,
      "theoretical_loss": 3.4573048380461686,
      "tokens_seen": 1822621696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045539393939393944,
      "loss": 1.4106,
      "theoretical_loss": 3.4572942829764783,
      "tokens_seen": 1822687232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553737373737374,
      "loss": 1.4145,
      "theoretical_loss": 3.4572837283925564,
      "tokens_seen": 1822752768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045535353535353535,
      "loss": 1.4457,
      "theoretical_loss": 3.457273174294362,
      "tokens_seen": 1822818304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004553333333333333,
      "loss": 1.418,
      "theoretical_loss": 3.457262620681856,
      "tokens_seen": 1822883840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045531313131313137,
      "loss": 1.3548,
      "theoretical_loss": 3.4572520675549985,
      "tokens_seen": 1822949376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045529292929292933,
      "loss": 1.4232,
      "theoretical_loss": 3.45724151491375,
      "tokens_seen": 1823014912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552727272727273,
      "loss": 1.5722,
      "theoretical_loss": 3.4572309627580706,
      "tokens_seen": 1823080448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045525252525252524,
      "loss": 1.6458,
      "theoretical_loss": 3.4572204110879197,
      "tokens_seen": 1823145984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004552323232323232,
      "loss": 1.352,
      "theoretical_loss": 3.4572098599032586,
      "tokens_seen": 1823211520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045521212121212126,
      "loss": 1.6786,
      "theoretical_loss": 3.4571993092040465,
      "tokens_seen": 1823277056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551919191919192,
      "loss": 1.6506,
      "theoretical_loss": 3.4571887589902444,
      "tokens_seen": 1823342592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551717171717172,
      "loss": 1.2894,
      "theoretical_loss": 3.457178209261812,
      "tokens_seen": 1823408128
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045515151515151513,
      "loss": 1.3039,
      "theoretical_loss": 3.457167660018711,
      "tokens_seen": 1823473664
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004551313131313131,
      "loss": 1.1383,
      "theoretical_loss": 3.4571571112608988,
      "tokens_seen": 1823539200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045511111111111116,
      "loss": 1.3979,
      "theoretical_loss": 3.4571465629883384,
      "tokens_seen": 1823604736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004550909090909091,
      "loss": 1.3313,
      "theoretical_loss": 3.4571360152009882,
      "tokens_seen": 1823670272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045507070707070707,
      "loss": 1.5223,
      "theoretical_loss": 3.4571254678988095,
      "tokens_seen": 1823735808
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.38483625650405884,
      "objective/train/avg_token_score": 0.019340477883815765,
      "objective/train/docs_used": 1031478,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2517280578613281,
      "objective/train/original_loss": 4.015500068664551,
      "objective/train/theoretical_loss": 3.4571149210817618,
      "objective/train/tokens_used": 182660576,
      "objective/train/unlikelihood_loss": 0.09212837368249893,
      "theoretical_loss": 3.4571149210817618,
      "tokens_seen": 1823801344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045505050505050503,
      "loss": 1.3643,
      "theoretical_loss": 3.4571149210817618,
      "tokens_seen": 1823801344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045503030303030304,
      "loss": 1.4704,
      "theoretical_loss": 3.4571043747498065,
      "tokens_seen": 1823866880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045501010101010105,
      "loss": 1.5384,
      "theoretical_loss": 3.4570938289029023,
      "tokens_seen": 1823932416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454989898989899,
      "loss": 1.518,
      "theoretical_loss": 3.457083283541011,
      "tokens_seen": 1823997952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045496969696969696,
      "loss": 1.1422,
      "theoretical_loss": 3.4570727386640914,
      "tokens_seen": 1824063488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004549494949494949,
      "loss": 1.3168,
      "theoretical_loss": 3.4570621942721047,
      "tokens_seen": 1824129024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045492929292929293,
      "loss": 1.3456,
      "theoretical_loss": 3.4570516503650115,
      "tokens_seen": 1824194560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045490909090909094,
      "loss": 1.4976,
      "theoretical_loss": 3.4570411069427713,
      "tokens_seen": 1824260096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004548888888888889,
      "loss": 1.4185,
      "theoretical_loss": 3.4570305640053443,
      "tokens_seen": 1824325632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045486868686868686,
      "loss": 1.3554,
      "theoretical_loss": 3.457020021552691,
      "tokens_seen": 1824391168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045484848484848487,
      "loss": 1.4911,
      "theoretical_loss": 3.4570094795847726,
      "tokens_seen": 1824456704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004548282828282828,
      "loss": 1.3135,
      "theoretical_loss": 3.456998938101548,
      "tokens_seen": 1824522240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045480808080808083,
      "loss": 1.0882,
      "theoretical_loss": 3.456988397102979,
      "tokens_seen": 1824587776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004547878787878788,
      "loss": 1.3685,
      "theoretical_loss": 3.4569778565890243,
      "tokens_seen": 1824653312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045476767676767675,
      "loss": 1.4268,
      "theoretical_loss": 3.456967316559645,
      "tokens_seen": 1824718848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045474747474747476,
      "loss": 1.5668,
      "theoretical_loss": 3.4569567770148018,
      "tokens_seen": 1824784384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004547272727272727,
      "loss": 1.6636,
      "theoretical_loss": 3.4569462379544547,
      "tokens_seen": 1824849920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045470707070707073,
      "loss": 1.4786,
      "theoretical_loss": 3.4569356993785636,
      "tokens_seen": 1824915456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546868686868687,
      "loss": 1.3914,
      "theoretical_loss": 3.456925161287089,
      "tokens_seen": 1824980992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546666666666667,
      "loss": 1.2798,
      "theoretical_loss": 3.456914623679992,
      "tokens_seen": 1825046528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045464646464646465,
      "loss": 1.4608,
      "theoretical_loss": 3.4569040865572322,
      "tokens_seen": 1825112064
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546262626262626,
      "loss": 1.6782,
      "theoretical_loss": 3.4568935499187696,
      "tokens_seen": 1825177600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004546060606060606,
      "loss": 1.4233,
      "theoretical_loss": 3.456883013764566,
      "tokens_seen": 1825243136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545858585858586,
      "loss": 1.5035,
      "theoretical_loss": 3.4568724780945796,
      "tokens_seen": 1825308672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545656565656566,
      "loss": 1.4682,
      "theoretical_loss": 3.456861942908773,
      "tokens_seen": 1825374208
    },
    {
      "debugging/Self-BLEU-5": 0.383412367211192,
      "debugging/distinct-1-grams": 0.7526443544193546,
      "debugging/distinct-2-grams": 0.9328595808568966,
      "debugging/entropy-1-grams": 5.657827037788948,
      "debugging/entropy-2-grams": 6.397750202501998,
      "debugging/length": 484.77777777777777,
      "debugging/num_segments": 9,
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.5272482633590698,
      "objective/train/avg_token_score": 0.014473862946033478,
      "objective/train/docs_used": 1032081,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6364436149597168,
      "objective/train/original_loss": 3.784334897994995,
      "objective/train/theoretical_loss": 3.456851408207105,
      "objective/train/tokens_used": 184298976,
      "objective/train/unlikelihood_loss": 0.08270136266946793,
      "theoretical_loss": 3.456851408207105,
      "tokens_seen": 1825439744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045454545454545455,
      "loss": 1.618,
      "theoretical_loss": 3.456851408207105,
      "tokens_seen": 1825439744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545252525252525,
      "loss": 1.2274,
      "theoretical_loss": 3.456840873989537,
      "tokens_seen": 1825505280
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004545050505050505,
      "loss": 1.3231,
      "theoretical_loss": 3.4568303402560288,
      "tokens_seen": 1825570816
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544848484848485,
      "loss": 1.3608,
      "theoretical_loss": 3.4568198070065406,
      "tokens_seen": 1825636352
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544646464646465,
      "loss": 1.4591,
      "theoretical_loss": 3.4568092742410332,
      "tokens_seen": 1825701888
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045444444444444444,
      "loss": 1.2171,
      "theoretical_loss": 3.456798741959467,
      "tokens_seen": 1825767424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544242424242424,
      "loss": 1.4365,
      "theoretical_loss": 3.456788210161802,
      "tokens_seen": 1825832960
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004544040404040404,
      "loss": 1.4114,
      "theoretical_loss": 3.456777678847999,
      "tokens_seen": 1825898496
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543838383838384,
      "loss": 1.3937,
      "theoretical_loss": 3.4567671480180184,
      "tokens_seen": 1825964032
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543636363636364,
      "loss": 1.4272,
      "theoretical_loss": 3.4567566176718203,
      "tokens_seen": 1826029568
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045434343434343433,
      "loss": 1.5479,
      "theoretical_loss": 3.456746087809365,
      "tokens_seen": 1826095104
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004543232323232323,
      "loss": 1.1661,
      "theoretical_loss": 3.456735558430614,
      "tokens_seen": 1826160640
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045430303030303035,
      "loss": 1.1125,
      "theoretical_loss": 3.4567250295355265,
      "tokens_seen": 1826226176
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004542828282828283,
      "loss": 1.4769,
      "theoretical_loss": 3.456714501124063,
      "tokens_seen": 1826291712
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045426262626262627,
      "loss": 1.3936,
      "theoretical_loss": 3.4567039731961846,
      "tokens_seen": 1826357248
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004542424242424242,
      "loss": 1.2966,
      "theoretical_loss": 3.4566934457518514,
      "tokens_seen": 1826422784
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045422222222222223,
      "loss": 1.341,
      "theoretical_loss": 3.456682918791024,
      "tokens_seen": 1826488320
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045420202020202025,
      "loss": 1.4049,
      "theoretical_loss": 3.4566723923136626,
      "tokens_seen": 1826553856
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541818181818182,
      "loss": 1.4241,
      "theoretical_loss": 3.4566618663197275,
      "tokens_seen": 1826619392
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045416161616161616,
      "loss": 1.3875,
      "theoretical_loss": 3.45665134080918,
      "tokens_seen": 1826684928
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541414141414141,
      "loss": 1.5876,
      "theoretical_loss": 3.4566408157819795,
      "tokens_seen": 1826750464
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004541212121212122,
      "loss": 1.3572,
      "theoretical_loss": 3.456630291238087,
      "tokens_seen": 1826816000
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045410101010101014,
      "loss": 1.5581,
      "theoretical_loss": 3.4566197671774628,
      "tokens_seen": 1826881536
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004540808080808081,
      "loss": 1.4737,
      "theoretical_loss": 3.456609243600068,
      "tokens_seen": 1826947072
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045406060606060605,
      "loss": 1.428,
      "theoretical_loss": 3.456598720505862,
      "tokens_seen": 1827012608
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.3259408473968506,
      "objective/train/avg_token_score": 0.02392873726785183,
      "objective/train/docs_used": 1033277,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0870252847671509,
      "objective/train/original_loss": 3.9227612018585205,
      "objective/train/theoretical_loss": 3.4565881978948063,
      "objective/train/tokens_used": 185937376,
      "objective/train/unlikelihood_loss": 0.08083102852106094,
      "theoretical_loss": 3.4565881978948063,
      "tokens_seen": 1827078144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000454040404040404,
      "loss": 1.1697,
      "theoretical_loss": 3.4565881978948063,
      "tokens_seen": 1827078144
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004540202020202021,
      "loss": 1.3098,
      "theoretical_loss": 3.4565776757668605,
      "tokens_seen": 1827143680
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045400000000000003,
      "loss": 1.3512,
      "theoretical_loss": 3.456567154121986,
      "tokens_seen": 1827209216
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000453979797979798,
      "loss": 1.438,
      "theoretical_loss": 3.4565566329601425,
      "tokens_seen": 1827274752
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045395959595959594,
      "loss": 1.6695,
      "theoretical_loss": 3.4565461122812904,
      "tokens_seen": 1827340288
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045393939393939396,
      "loss": 1.1057,
      "theoretical_loss": 3.4565355920853915,
      "tokens_seen": 1827405824
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045391919191919197,
      "loss": 1.5515,
      "theoretical_loss": 3.456525072372405,
      "tokens_seen": 1827471360
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004538989898989899,
      "loss": 1.2735,
      "theoretical_loss": 3.456514553142292,
      "tokens_seen": 1827536896
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004538787878787879,
      "loss": 1.3733,
      "theoretical_loss": 3.456504034395013,
      "tokens_seen": 1827602432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045385858585858584,
      "loss": 1.3649,
      "theoretical_loss": 3.4564935161305277,
      "tokens_seen": 1827667968
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045383838383838385,
      "loss": 1.7118,
      "theoretical_loss": 3.4564829983487977,
      "tokens_seen": 1827733504
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045381818181818186,
      "loss": 1.3295,
      "theoretical_loss": 3.4564724810497838,
      "tokens_seen": 1827799040
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537979797979798,
      "loss": 1.2664,
      "theoretical_loss": 3.456461964233445,
      "tokens_seen": 1827864576
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537777777777778,
      "loss": 1.5361,
      "theoretical_loss": 3.4564514478997435,
      "tokens_seen": 1827930112
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004537575757575758,
      "loss": 1.5431,
      "theoretical_loss": 3.456440932048639,
      "tokens_seen": 1827995648
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045373737373737374,
      "loss": 1.377,
      "theoretical_loss": 3.4564304166800914,
      "tokens_seen": 1828061184
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045371717171717175,
      "loss": 1.3454,
      "theoretical_loss": 3.456419901794063,
      "tokens_seen": 1828126720
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536969696969697,
      "loss": 1.4446,
      "theoretical_loss": 3.4564093873905124,
      "tokens_seen": 1828192256
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045367676767676767,
      "loss": 1.4514,
      "theoretical_loss": 3.4563988734694018,
      "tokens_seen": 1828257792
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004536565656565657,
      "loss": 1.3498,
      "theoretical_loss": 3.456388360030691,
      "tokens_seen": 1828323328
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045363636363636363,
      "loss": 1.4994,
      "theoretical_loss": 3.4563778470743403,
      "tokens_seen": 1828388864
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045361616161616164,
      "loss": 1.1171,
      "theoretical_loss": 3.456367334600311,
      "tokens_seen": 1828454400
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535959595959596,
      "loss": 1.37,
      "theoretical_loss": 3.4563568226085635,
      "tokens_seen": 1828519936
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535757575757576,
      "loss": 1.4101,
      "theoretical_loss": 3.456346311099058,
      "tokens_seen": 1828585472
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045355555555555557,
      "loss": 1.2472,
      "theoretical_loss": 3.4563358000717552,
      "tokens_seen": 1828651008
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.3474462330341339,
      "objective/train/avg_token_score": 0.01256563700735569,
      "objective/train/docs_used": 1034021,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1423338651657104,
      "objective/train/original_loss": 4.592188835144043,
      "objective/train/theoretical_loss": 3.456325289526616,
      "objective/train/tokens_used": 187575776,
      "objective/train/unlikelihood_loss": 0.06064225360751152,
      "theoretical_loss": 3.456325289526616,
      "tokens_seen": 1828716544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004535353535353535,
      "loss": 1.4799,
      "theoretical_loss": 3.456325289526616,
      "tokens_seen": 1828716544
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045351515151515154,
      "loss": 1.2678,
      "theoretical_loss": 3.456314779463601,
      "tokens_seen": 1828782080
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534949494949495,
      "loss": 1.3876,
      "theoretical_loss": 3.4563042698826707,
      "tokens_seen": 1828847616
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534747474747475,
      "loss": 1.4213,
      "theoretical_loss": 3.456293760783786,
      "tokens_seen": 1828913152
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045345454545454546,
      "loss": 1.596,
      "theoretical_loss": 3.456283252166906,
      "tokens_seen": 1828978688
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004534343434343434,
      "loss": 1.253,
      "theoretical_loss": 3.4562727440319936,
      "tokens_seen": 1829044224
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045341414141414143,
      "loss": 1.4315,
      "theoretical_loss": 3.4562622363790076,
      "tokens_seen": 1829109760
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045339393939393944,
      "loss": 1.1965,
      "theoretical_loss": 3.45625172920791,
      "tokens_seen": 1829175296
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533737373737374,
      "loss": 1.3171,
      "theoretical_loss": 3.45624122251866,
      "tokens_seen": 1829240832
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045335353535353535,
      "loss": 1.2845,
      "theoretical_loss": 3.45623071631122,
      "tokens_seen": 1829306368
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533333333333333,
      "loss": 1.6198,
      "theoretical_loss": 3.456220210585549,
      "tokens_seen": 1829371904
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004533131313131313,
      "loss": 1.268,
      "theoretical_loss": 3.4562097053416085,
      "tokens_seen": 1829437440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045329292929292933,
      "loss": 1.3782,
      "theoretical_loss": 3.456199200579359,
      "tokens_seen": 1829502976
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532727272727273,
      "loss": 1.4489,
      "theoretical_loss": 3.456188696298761,
      "tokens_seen": 1829568512
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045325252525252525,
      "loss": 1.3057,
      "theoretical_loss": 3.4561781924997756,
      "tokens_seen": 1829634048
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532323232323232,
      "loss": 1.5815,
      "theoretical_loss": 3.456167689182363,
      "tokens_seen": 1829699584
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004532121212121212,
      "loss": 1.5887,
      "theoretical_loss": 3.456157186346484,
      "tokens_seen": 1829765120
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004531919191919192,
      "loss": 1.5131,
      "theoretical_loss": 3.4561466839921,
      "tokens_seen": 1829830656
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004531717171717172,
      "loss": 1.6467,
      "theoretical_loss": 3.4561361821191703,
      "tokens_seen": 1829896192
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045315151515151514,
      "loss": 1.54,
      "theoretical_loss": 3.456125680727656,
      "tokens_seen": 1829961728
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004531313131313131,
      "loss": 1.6346,
      "theoretical_loss": 3.4561151798175187,
      "tokens_seen": 1830027264
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045311111111111116,
      "loss": 1.2857,
      "theoretical_loss": 3.4561046793887185,
      "tokens_seen": 1830092800
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004530909090909091,
      "loss": 1.1951,
      "theoretical_loss": 3.456094179441216,
      "tokens_seen": 1830158336
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004530707070707071,
      "loss": 1.2465,
      "theoretical_loss": 3.456083679974972,
      "tokens_seen": 1830223872
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045305050505050503,
      "loss": 1.3245,
      "theoretical_loss": 3.456073180989947,
      "tokens_seen": 1830289408
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.4240591526031494,
      "objective/train/avg_token_score": 0.011298497207462788,
      "objective/train/docs_used": 1035168,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4181580543518066,
      "objective/train/original_loss": 4.181857585906982,
      "objective/train/theoretical_loss": 3.456062682486102,
      "objective/train/tokens_used": 189214176,
      "objective/train/unlikelihood_loss": 0.05961797386407852,
      "theoretical_loss": 3.456062682486102,
      "tokens_seen": 1830354944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000453030303030303,
      "loss": 1.5355,
      "theoretical_loss": 3.456062682486102,
      "tokens_seen": 1830354944
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045301010101010106,
      "loss": 1.466,
      "theoretical_loss": 3.4560521844633976,
      "tokens_seen": 1830420480
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000452989898989899,
      "loss": 1.5066,
      "theoretical_loss": 3.456041686921795,
      "tokens_seen": 1830486016
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045296969696969697,
      "loss": 1.3351,
      "theoretical_loss": 3.456031189861254,
      "tokens_seen": 1830551552
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004529494949494949,
      "loss": 1.2997,
      "theoretical_loss": 3.4560206932817357,
      "tokens_seen": 1830617088
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045292929292929294,
      "loss": 1.5116,
      "theoretical_loss": 3.4560101971832013,
      "tokens_seen": 1830682624
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045290909090909095,
      "loss": 1.0942,
      "theoretical_loss": 3.455999701565611,
      "tokens_seen": 1830748160
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528888888888889,
      "loss": 1.3721,
      "theoretical_loss": 3.455989206428926,
      "tokens_seen": 1830813696
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045286868686868686,
      "loss": 1.2505,
      "theoretical_loss": 3.455978711773107,
      "tokens_seen": 1830879232
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004528484848484848,
      "loss": 1.6109,
      "theoretical_loss": 3.455968217598114,
      "tokens_seen": 1830944768
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045282828282828283,
      "loss": 1.3838,
      "theoretical_loss": 3.455957723903909,
      "tokens_seen": 1831010304
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045280808080808084,
      "loss": 1.2037,
      "theoretical_loss": 3.4559472306904517,
      "tokens_seen": 1831075840
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004527878787878788,
      "loss": 1.3973,
      "theoretical_loss": 3.4559367379577033,
      "tokens_seen": 1831141376
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045276767676767675,
      "loss": 1.4359,
      "theoretical_loss": 3.4559262457056246,
      "tokens_seen": 1831206912
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045274747474747477,
      "loss": 1.6251,
      "theoretical_loss": 3.4559157539341765,
      "tokens_seen": 1831272448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004527272727272727,
      "loss": 1.4879,
      "theoretical_loss": 3.455905262643319,
      "tokens_seen": 1831337984
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045270707070707073,
      "loss": 1.7081,
      "theoretical_loss": 3.4558947718330137,
      "tokens_seen": 1831403520
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526868686868687,
      "loss": 1.4078,
      "theoretical_loss": 3.4558842815032214,
      "tokens_seen": 1831469056
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045266666666666665,
      "loss": 1.5619,
      "theoretical_loss": 3.455873791653903,
      "tokens_seen": 1831534592
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045264646464646466,
      "loss": 1.4385,
      "theoretical_loss": 3.4558633022850183,
      "tokens_seen": 1831600128
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526262626262626,
      "loss": 1.1827,
      "theoretical_loss": 3.455852813396529,
      "tokens_seen": 1831665664
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004526060606060606,
      "loss": 1.5327,
      "theoretical_loss": 3.455842324988396,
      "tokens_seen": 1831731200
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004525858585858586,
      "loss": 1.6505,
      "theoretical_loss": 3.455831837060579,
      "tokens_seen": 1831796736
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004525656565656566,
      "loss": 1.2537,
      "theoretical_loss": 3.45582134961304,
      "tokens_seen": 1831862272
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045254545454545455,
      "loss": 1.2179,
      "theoretical_loss": 3.4558108626457402,
      "tokens_seen": 1831927808
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.2812194526195526,
      "objective/train/avg_token_score": 0.04614793136715889,
      "objective/train/docs_used": 1035908,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9016737937927246,
      "objective/train/original_loss": 4.561851978302002,
      "objective/train/theoretical_loss": 3.455800376158639,
      "objective/train/tokens_used": 190852576,
      "objective/train/unlikelihood_loss": 0.06523799896240234,
      "theoretical_loss": 3.455800376158639,
      "tokens_seen": 1831993344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045252525252525256,
      "loss": 1.376,
      "theoretical_loss": 3.455800376158639,
      "tokens_seen": 1831993344
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004525050505050505,
      "loss": 1.4152,
      "theoretical_loss": 3.4557898901516984,
      "tokens_seen": 1832058880
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524848484848485,
      "loss": 1.565,
      "theoretical_loss": 3.455779404624878,
      "tokens_seen": 1832124416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524646464646465,
      "loss": 1.4454,
      "theoretical_loss": 3.45576891957814,
      "tokens_seen": 1832189952
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045244444444444444,
      "loss": 1.6646,
      "theoretical_loss": 3.4557584350114445,
      "tokens_seen": 1832255488
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045242424242424245,
      "loss": 1.2104,
      "theoretical_loss": 3.4557479509247524,
      "tokens_seen": 1832321024
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004524040404040404,
      "loss": 1.6734,
      "theoretical_loss": 3.455737467318025,
      "tokens_seen": 1832386560
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523838383838384,
      "loss": 1.4283,
      "theoretical_loss": 3.4557269841912226,
      "tokens_seen": 1832452096
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523636363636364,
      "loss": 1.5698,
      "theoretical_loss": 3.4557165015443063,
      "tokens_seen": 1832517632
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045234343434343434,
      "loss": 1.1731,
      "theoretical_loss": 3.455706019377237,
      "tokens_seen": 1832583168
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045232323232323235,
      "loss": 1.2459,
      "theoretical_loss": 3.4556955376899756,
      "tokens_seen": 1832648704
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004523030303030303,
      "loss": 1.3951,
      "theoretical_loss": 3.4556850564824826,
      "tokens_seen": 1832714240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004522828282828283,
      "loss": 1.247,
      "theoretical_loss": 3.4556745757547196,
      "tokens_seen": 1832779776
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045226262626262627,
      "loss": 1.4864,
      "theoretical_loss": 3.455664095506647,
      "tokens_seen": 1832845312
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045224242424242423,
      "loss": 1.4557,
      "theoretical_loss": 3.4556536157382256,
      "tokens_seen": 1832910848
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045222222222222224,
      "loss": 1.4369,
      "theoretical_loss": 3.455643136449417,
      "tokens_seen": 1832976384
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045220202020202025,
      "loss": 1.4263,
      "theoretical_loss": 3.455632657640181,
      "tokens_seen": 1833041920
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004521818181818182,
      "loss": 1.458,
      "theoretical_loss": 3.4556221793104793,
      "tokens_seen": 1833107456
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045216161616161616,
      "loss": 1.6213,
      "theoretical_loss": 3.4556117014602727,
      "tokens_seen": 1833172992
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004521414141414141,
      "loss": 1.3254,
      "theoretical_loss": 3.455601224089522,
      "tokens_seen": 1833238528
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045212121212121213,
      "loss": 1.5437,
      "theoretical_loss": 3.4555907471981886,
      "tokens_seen": 1833304064
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045210101010101014,
      "loss": 1.3216,
      "theoretical_loss": 3.455580270786232,
      "tokens_seen": 1833369600
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004520808080808081,
      "loss": 1.459,
      "theoretical_loss": 3.455569794853615,
      "tokens_seen": 1833435136
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.00045206060606060606,
      "loss": 1.4695,
      "theoretical_loss": 3.4555593194002974,
      "tokens_seen": 1833500672
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.000452040404040404,
      "loss": 1.392,
      "theoretical_loss": 3.4555488444262403,
      "tokens_seen": 1833566208
    },
    {
      "epoch": 0.1,
      "objective/train/aligned_token_count": 0.48057183623313904,
      "objective/train/avg_token_score": 0.015353728085756302,
      "objective/train/docs_used": 1036404,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4945865869522095,
      "objective/train/original_loss": 4.170429706573486,
      "objective/train/theoretical_loss": 3.4555383699314044,
      "objective/train/tokens_used": 192490976,
      "objective/train/unlikelihood_loss": 0.0738421231508255,
      "theoretical_loss": 3.4555383699314044,
      "tokens_seen": 1833631744
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0004520202020202021,
      "loss": 1.5297,
      "theoretical_loss": 3.4555383699314044,
      "tokens_seen": 1833631744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045200000000000004,
      "loss": 1.5953,
      "theoretical_loss": 3.4555278959157514,
      "tokens_seen": 1833697280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000451979797979798,
      "loss": 1.4807,
      "theoretical_loss": 3.455517422379242,
      "tokens_seen": 1833762816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045195959595959595,
      "loss": 1.4942,
      "theoretical_loss": 3.455506949321837,
      "tokens_seen": 1833828352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004519393939393939,
      "loss": 1.3145,
      "theoretical_loss": 3.455496476743497,
      "tokens_seen": 1833893888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045191919191919197,
      "loss": 1.5973,
      "theoretical_loss": 3.4554860046441833,
      "tokens_seen": 1833959424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045189898989898993,
      "loss": 1.3642,
      "theoretical_loss": 3.455475533023857,
      "tokens_seen": 1834024960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518787878787879,
      "loss": 1.2289,
      "theoretical_loss": 3.455465061882479,
      "tokens_seen": 1834090496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045185858585858584,
      "loss": 1.5492,
      "theoretical_loss": 3.4554545912200103,
      "tokens_seen": 1834156032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004518383838383838,
      "loss": 1.2268,
      "theoretical_loss": 3.455444121036412,
      "tokens_seen": 1834221568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045181818181818186,
      "loss": 1.3147,
      "theoretical_loss": 3.455433651331645,
      "tokens_seen": 1834287104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004517979797979798,
      "loss": 1.4772,
      "theoretical_loss": 3.45542318210567,
      "tokens_seen": 1834352640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004517777777777778,
      "loss": 1.4504,
      "theoretical_loss": 3.455412713358448,
      "tokens_seen": 1834418176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045175757575757574,
      "loss": 1.5777,
      "theoretical_loss": 3.45540224508994,
      "tokens_seen": 1834483712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045173737373737375,
      "loss": 1.4718,
      "theoretical_loss": 3.455391777300108,
      "tokens_seen": 1834549248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045171717171717176,
      "loss": 1.5121,
      "theoretical_loss": 3.455381309988912,
      "tokens_seen": 1834614784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004516969696969697,
      "loss": 1.544,
      "theoretical_loss": 3.455370843156313,
      "tokens_seen": 1834680320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045167676767676767,
      "loss": 1.6024,
      "theoretical_loss": 3.4553603768022727,
      "tokens_seen": 1834745856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045165656565656563,
      "loss": 1.4055,
      "theoretical_loss": 3.4553499109267514,
      "tokens_seen": 1834811392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045163636363636364,
      "loss": 1.3813,
      "theoretical_loss": 3.4553394455297104,
      "tokens_seen": 1834876928
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045161616161616165,
      "loss": 1.4046,
      "theoretical_loss": 3.455328980611111,
      "tokens_seen": 1834942464
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004515959595959596,
      "loss": 1.5633,
      "theoretical_loss": 3.455318516170914,
      "tokens_seen": 1835008000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045157575757575756,
      "loss": 1.2908,
      "theoretical_loss": 3.4553080522090807,
      "tokens_seen": 1835073536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004515555555555556,
      "loss": 1.2658,
      "theoretical_loss": 3.4552975887255717,
      "tokens_seen": 1835139072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045153535353535353,
      "loss": 1.0296,
      "theoretical_loss": 3.455287125720348,
      "tokens_seen": 1835204608
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.46584799885749817,
      "objective/train/avg_token_score": 0.026970431208610535,
      "objective/train/docs_used": 1037817,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6634101867675781,
      "objective/train/original_loss": 4.356515884399414,
      "objective/train/theoretical_loss": 3.455276663193372,
      "objective/train/tokens_used": 194129376,
      "objective/train/unlikelihood_loss": 0.06059441342949867,
      "theoretical_loss": 3.455276663193372,
      "tokens_seen": 1835270144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045151515151515154,
      "loss": 1.435,
      "theoretical_loss": 3.455276663193372,
      "tokens_seen": 1835270144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004514949494949495,
      "loss": 1.6169,
      "theoretical_loss": 3.4552662011446023,
      "tokens_seen": 1835335680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045147474747474746,
      "loss": 1.4235,
      "theoretical_loss": 3.4552557395740022,
      "tokens_seen": 1835401216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045145454545454547,
      "loss": 1.4122,
      "theoretical_loss": 3.4552452784815317,
      "tokens_seen": 1835466752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004514343434343434,
      "loss": 1.5246,
      "theoretical_loss": 3.455234817867152,
      "tokens_seen": 1835532288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045141414141414144,
      "loss": 1.4689,
      "theoretical_loss": 3.455224357730825,
      "tokens_seen": 1835597824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513939393939394,
      "loss": 1.2776,
      "theoretical_loss": 3.4552138980725102,
      "tokens_seen": 1835663360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513737373737374,
      "loss": 1.3385,
      "theoretical_loss": 3.4552034388921697,
      "tokens_seen": 1835728896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045135353535353536,
      "loss": 1.3866,
      "theoretical_loss": 3.4551929801897647,
      "tokens_seen": 1835794432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004513333333333333,
      "loss": 1.5819,
      "theoretical_loss": 3.455182521965256,
      "tokens_seen": 1835859968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045131313131313133,
      "loss": 1.3853,
      "theoretical_loss": 3.455172064218605,
      "tokens_seen": 1835925504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512929292929293,
      "loss": 1.3696,
      "theoretical_loss": 3.4551616069497726,
      "tokens_seen": 1835991040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512727272727273,
      "loss": 1.2499,
      "theoretical_loss": 3.4551511501587195,
      "tokens_seen": 1836056576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045125252525252525,
      "loss": 1.5628,
      "theoretical_loss": 3.4551406938454075,
      "tokens_seen": 1836122112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512323232323232,
      "loss": 1.6587,
      "theoretical_loss": 3.4551302380097972,
      "tokens_seen": 1836187648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004512121212121212,
      "loss": 1.4389,
      "theoretical_loss": 3.4551197826518507,
      "tokens_seen": 1836253184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045119191919191923,
      "loss": 1.3724,
      "theoretical_loss": 3.4551093277715275,
      "tokens_seen": 1836318720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004511717171717172,
      "loss": 1.3688,
      "theoretical_loss": 3.45509887336879,
      "tokens_seen": 1836384256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045115151515151515,
      "loss": 1.4629,
      "theoretical_loss": 3.455088419443599,
      "tokens_seen": 1836449792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004511313131313131,
      "loss": 1.4199,
      "theoretical_loss": 3.4550779659959154,
      "tokens_seen": 1836515328
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004511111111111111,
      "loss": 1.4308,
      "theoretical_loss": 3.455067513025701,
      "tokens_seen": 1836580864
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510909090909091,
      "loss": 1.4939,
      "theoretical_loss": 3.4550570605329165,
      "tokens_seen": 1836646400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004510707070707071,
      "loss": 1.5078,
      "theoretical_loss": 3.455046608517523,
      "tokens_seen": 1836711936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045105050505050504,
      "loss": 1.5747,
      "theoretical_loss": 3.4550361569794816,
      "tokens_seen": 1836777472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000451030303030303,
      "loss": 1.4927,
      "theoretical_loss": 3.4550257059187537,
      "tokens_seen": 1836843008
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.5,
      "objective/train/avg_token_score": 0.013788322918117046,
      "objective/train/docs_used": 1038484,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6846576929092407,
      "objective/train/original_loss": 3.9736297130584717,
      "objective/train/theoretical_loss": 3.4550152553353,
      "objective/train/tokens_used": 195767776,
      "objective/train/unlikelihood_loss": 0.05097836256027222,
      "theoretical_loss": 3.4550152553353,
      "tokens_seen": 1836908544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045101010101010106,
      "loss": 1.3611,
      "theoretical_loss": 3.4550152553353,
      "tokens_seen": 1836908544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000450989898989899,
      "loss": 1.3674,
      "theoretical_loss": 3.4550048052290823,
      "tokens_seen": 1836974080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000450969696969697,
      "loss": 1.3659,
      "theoretical_loss": 3.454994355600062,
      "tokens_seen": 1837039616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045094949494949493,
      "loss": 1.2901,
      "theoretical_loss": 3.4549839064481995,
      "tokens_seen": 1837105152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004509292929292929,
      "loss": 1.4305,
      "theoretical_loss": 3.4549734577734563,
      "tokens_seen": 1837170688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045090909090909095,
      "loss": 1.45,
      "theoretical_loss": 3.4549630095757937,
      "tokens_seen": 1837236224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508888888888889,
      "loss": 1.379,
      "theoretical_loss": 3.4549525618551726,
      "tokens_seen": 1837301760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045086868686868687,
      "loss": 1.6066,
      "theoretical_loss": 3.454942114611555,
      "tokens_seen": 1837367296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508484848484848,
      "loss": 1.4182,
      "theoretical_loss": 3.454931667844901,
      "tokens_seen": 1837432832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004508282828282829,
      "loss": 1.3809,
      "theoretical_loss": 3.4549212215551726,
      "tokens_seen": 1837498368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045080808080808085,
      "loss": 1.4271,
      "theoretical_loss": 3.4549107757423307,
      "tokens_seen": 1837563904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507878787878788,
      "loss": 1.4629,
      "theoretical_loss": 3.4549003304063364,
      "tokens_seen": 1837629440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045076767676767676,
      "loss": 1.6347,
      "theoretical_loss": 3.4548898855471517,
      "tokens_seen": 1837694976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507474747474747,
      "loss": 1.4147,
      "theoretical_loss": 3.4548794411647368,
      "tokens_seen": 1837760512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004507272727272728,
      "loss": 1.561,
      "theoretical_loss": 3.4548689972590534,
      "tokens_seen": 1837826048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045070707070707074,
      "loss": 1.7009,
      "theoretical_loss": 3.4548585538300625,
      "tokens_seen": 1837891584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004506868686868687,
      "loss": 1.5077,
      "theoretical_loss": 3.454848110877726,
      "tokens_seen": 1837957120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045066666666666665,
      "loss": 1.3909,
      "theoretical_loss": 3.4548376684020043,
      "tokens_seen": 1838022656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045064646464646466,
      "loss": 1.4584,
      "theoretical_loss": 3.45482722640286,
      "tokens_seen": 1838088192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004506262626262627,
      "loss": 1.5085,
      "theoretical_loss": 3.4548167848802525,
      "tokens_seen": 1838153728
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045060606060606063,
      "loss": 1.4115,
      "theoretical_loss": 3.454806343834144,
      "tokens_seen": 1838219264
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505858585858586,
      "loss": 1.4466,
      "theoretical_loss": 3.454795903264496,
      "tokens_seen": 1838284800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045056565656565655,
      "loss": 1.7046,
      "theoretical_loss": 3.4547854631712696,
      "tokens_seen": 1838350336
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045054545454545456,
      "loss": 1.6425,
      "theoretical_loss": 3.454775023554426,
      "tokens_seen": 1838415872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045052525252525257,
      "loss": 1.3357,
      "theoretical_loss": 3.454764584413927,
      "tokens_seen": 1838481408
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.28757330775260925,
      "objective/train/avg_token_score": 0.022113921120762825,
      "objective/train/docs_used": 1039769,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9949387907981873,
      "objective/train/original_loss": 4.7078962326049805,
      "objective/train/theoretical_loss": 3.4547541457497326,
      "objective/train/tokens_used": 197406176,
      "objective/train/unlikelihood_loss": 0.05575425177812576,
      "theoretical_loss": 3.4547541457497326,
      "tokens_seen": 1838546944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004505050505050505,
      "loss": 1.4208,
      "theoretical_loss": 3.4547541457497326,
      "tokens_seen": 1838546944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004504848484848485,
      "loss": 1.3145,
      "theoretical_loss": 3.4547437075618053,
      "tokens_seen": 1838612480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004504646464646465,
      "loss": 1.4106,
      "theoretical_loss": 3.454733269850106,
      "tokens_seen": 1838678016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045044444444444445,
      "loss": 1.3905,
      "theoretical_loss": 3.4547228326145953,
      "tokens_seen": 1838743552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045042424242424246,
      "loss": 1.6694,
      "theoretical_loss": 3.454712395855236,
      "tokens_seen": 1838809088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004504040404040404,
      "loss": 1.5321,
      "theoretical_loss": 3.454701959571988,
      "tokens_seen": 1838874624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503838383838384,
      "loss": 1.6927,
      "theoretical_loss": 3.4546915237648137,
      "tokens_seen": 1838940160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503636363636364,
      "loss": 1.3334,
      "theoretical_loss": 3.4546810884336736,
      "tokens_seen": 1839005696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045034343434343434,
      "loss": 1.4439,
      "theoretical_loss": 3.454670653578529,
      "tokens_seen": 1839071232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045032323232323235,
      "loss": 1.5006,
      "theoretical_loss": 3.4546602191993427,
      "tokens_seen": 1839136768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004503030303030303,
      "loss": 1.4973,
      "theoretical_loss": 3.4546497852960742,
      "tokens_seen": 1839202304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502828282828283,
      "loss": 1.3218,
      "theoretical_loss": 3.4546393518686855,
      "tokens_seen": 1839267840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502626262626263,
      "loss": 1.3238,
      "theoretical_loss": 3.4546289189171375,
      "tokens_seen": 1839333376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045024242424242423,
      "loss": 1.2593,
      "theoretical_loss": 3.454618486441393,
      "tokens_seen": 1839398912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045022222222222225,
      "loss": 1.3184,
      "theoretical_loss": 3.4546080544414117,
      "tokens_seen": 1839464448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004502020202020202,
      "loss": 1.264,
      "theoretical_loss": 3.4545976229171558,
      "tokens_seen": 1839529984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004501818181818182,
      "loss": 1.0797,
      "theoretical_loss": 3.454587191868587,
      "tokens_seen": 1839595520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045016161616161617,
      "loss": 1.6226,
      "theoretical_loss": 3.454576761295665,
      "tokens_seen": 1839661056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004501414141414141,
      "loss": 1.6046,
      "theoretical_loss": 3.454566331198353,
      "tokens_seen": 1839726592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045012121212121214,
      "loss": 1.2764,
      "theoretical_loss": 3.4545559015766116,
      "tokens_seen": 1839792128
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045010101010101015,
      "loss": 1.4158,
      "theoretical_loss": 3.4545454724304028,
      "tokens_seen": 1839857664
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004500808080808081,
      "loss": 1.5471,
      "theoretical_loss": 3.4545350437596865,
      "tokens_seen": 1839923200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045006060606060606,
      "loss": 1.4801,
      "theoretical_loss": 3.4545246155644254,
      "tokens_seen": 1839988736
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000450040404040404,
      "loss": 1.4026,
      "theoretical_loss": 3.4545141878445804,
      "tokens_seen": 1840054272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045002020202020203,
      "loss": 1.6834,
      "theoretical_loss": 3.4545037606001134,
      "tokens_seen": 1840119808
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.6422287225723267,
      "objective/train/avg_token_score": 0.03135533630847931,
      "objective/train/docs_used": 1040587,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.251701593399048,
      "objective/train/original_loss": 3.8327317237854004,
      "objective/train/theoretical_loss": 3.4544933338309853,
      "objective/train/tokens_used": 199044576,
      "objective/train/unlikelihood_loss": 0.05906657129526138,
      "theoretical_loss": 3.4544933338309853,
      "tokens_seen": 1840185344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00045000000000000004,
      "loss": 1.7251,
      "theoretical_loss": 3.4544933338309853,
      "tokens_seen": 1840185344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000449979797979798,
      "loss": 1.4279,
      "theoretical_loss": 3.4544829075371575,
      "tokens_seen": 1840250880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044995959595959596,
      "loss": 1.5526,
      "theoretical_loss": 3.4544724817185912,
      "tokens_seen": 1840316416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004499393939393939,
      "loss": 1.3022,
      "theoretical_loss": 3.454462056375249,
      "tokens_seen": 1840381952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004499191919191919,
      "loss": 1.6946,
      "theoretical_loss": 3.454451631507091,
      "tokens_seen": 1840447488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044989898989898993,
      "loss": 1.3439,
      "theoretical_loss": 3.454441207114079,
      "tokens_seen": 1840513024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004498787878787879,
      "loss": 1.5178,
      "theoretical_loss": 3.454430783196174,
      "tokens_seen": 1840578560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044985858585858585,
      "loss": 1.328,
      "theoretical_loss": 3.4544203597533385,
      "tokens_seen": 1840644096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004498383838383838,
      "loss": 1.3865,
      "theoretical_loss": 3.454409936785533,
      "tokens_seen": 1840709632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044981818181818187,
      "loss": 1.4214,
      "theoretical_loss": 3.45439951429272,
      "tokens_seen": 1840775168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044979797979797983,
      "loss": 1.4532,
      "theoretical_loss": 3.4543890922748597,
      "tokens_seen": 1840840704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497777777777778,
      "loss": 1.7572,
      "theoretical_loss": 3.454378670731914,
      "tokens_seen": 1840906240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044975757575757574,
      "loss": 1.2789,
      "theoretical_loss": 3.454368249663845,
      "tokens_seen": 1840971776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004497373737373737,
      "loss": 1.4031,
      "theoretical_loss": 3.454357829070613,
      "tokens_seen": 1841037312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044971717171717176,
      "loss": 1.2913,
      "theoretical_loss": 3.4543474089521804,
      "tokens_seen": 1841102848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004496969696969697,
      "loss": 1.3861,
      "theoretical_loss": 3.4543369893085085,
      "tokens_seen": 1841168384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004496767676767677,
      "loss": 1.2665,
      "theoretical_loss": 3.4543265701395582,
      "tokens_seen": 1841233920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044965656565656563,
      "loss": 1.4635,
      "theoretical_loss": 3.4543161514452914,
      "tokens_seen": 1841299456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044963636363636364,
      "loss": 1.3889,
      "theoretical_loss": 3.45430573322567,
      "tokens_seen": 1841364992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044961616161616166,
      "loss": 1.5038,
      "theoretical_loss": 3.4542953154806546,
      "tokens_seen": 1841430528
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004495959595959596,
      "loss": 1.4529,
      "theoretical_loss": 3.4542848982102075,
      "tokens_seen": 1841496064
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044957575757575757,
      "loss": 1.3667,
      "theoretical_loss": 3.4542744814142896,
      "tokens_seen": 1841561600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004495555555555555,
      "loss": 1.4548,
      "theoretical_loss": 3.454264065092863,
      "tokens_seen": 1841627136
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044953535353535354,
      "loss": 1.4561,
      "theoretical_loss": 3.454253649245888,
      "tokens_seen": 1841692672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044951515151515155,
      "loss": 1.3452,
      "theoretical_loss": 3.4542432338733278,
      "tokens_seen": 1841758208
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.3169599175453186,
      "objective/train/avg_token_score": 0.005386400036513805,
      "objective/train/docs_used": 1041999,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0743252038955688,
      "objective/train/original_loss": 4.4719977378845215,
      "objective/train/theoretical_loss": 3.4542328189751426,
      "objective/train/tokens_used": 200682976,
      "objective/train/unlikelihood_loss": 0.08110522478818893,
      "theoretical_loss": 3.4542328189751426,
      "tokens_seen": 1841823744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004494949494949495,
      "loss": 1.1803,
      "theoretical_loss": 3.4542328189751426,
      "tokens_seen": 1841823744
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044947474747474746,
      "loss": 1.474,
      "theoretical_loss": 3.4542224045512944,
      "tokens_seen": 1841889280
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004494545454545455,
      "loss": 1.5223,
      "theoretical_loss": 3.4542119906017446,
      "tokens_seen": 1841954816
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044943434343434343,
      "loss": 1.4056,
      "theoretical_loss": 3.4542015771264554,
      "tokens_seen": 1842020352
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044941414141414144,
      "loss": 1.4223,
      "theoretical_loss": 3.454191164125387,
      "tokens_seen": 1842085888
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004493939393939394,
      "loss": 1.4174,
      "theoretical_loss": 3.4541807515985017,
      "tokens_seen": 1842151424
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044937373737373735,
      "loss": 1.5226,
      "theoretical_loss": 3.4541703395457617,
      "tokens_seen": 1842216960
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044935353535353537,
      "loss": 1.4965,
      "theoretical_loss": 3.4541599279671273,
      "tokens_seen": 1842282496
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004493333333333333,
      "loss": 1.5692,
      "theoretical_loss": 3.454149516862561,
      "tokens_seen": 1842348032
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044931313131313133,
      "loss": 1.3349,
      "theoretical_loss": 3.4541391062320237,
      "tokens_seen": 1842413568
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492929292929293,
      "loss": 1.4055,
      "theoretical_loss": 3.454128696075477,
      "tokens_seen": 1842479104
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492727272727273,
      "loss": 1.3759,
      "theoretical_loss": 3.454118286392883,
      "tokens_seen": 1842544640
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044925252525252526,
      "loss": 1.5596,
      "theoretical_loss": 3.454107877184203,
      "tokens_seen": 1842610176
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492323232323232,
      "loss": 1.3979,
      "theoretical_loss": 3.4540974684493984,
      "tokens_seen": 1842675712
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004492121212121212,
      "loss": 1.6526,
      "theoretical_loss": 3.4540870601884306,
      "tokens_seen": 1842741248
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004491919191919192,
      "loss": 1.5212,
      "theoretical_loss": 3.4540766524012616,
      "tokens_seen": 1842806784
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004491717171717172,
      "loss": 1.5532,
      "theoretical_loss": 3.454066245087853,
      "tokens_seen": 1842872320
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044915151515151515,
      "loss": 1.5621,
      "theoretical_loss": 3.4540558382481663,
      "tokens_seen": 1842937856
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044913131313131316,
      "loss": 1.3275,
      "theoretical_loss": 3.4540454318821627,
      "tokens_seen": 1843003392
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004491111111111111,
      "loss": 1.4118,
      "theoretical_loss": 3.4540350259898043,
      "tokens_seen": 1843068928
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044909090909090913,
      "loss": 1.5738,
      "theoretical_loss": 3.4540246205710528,
      "tokens_seen": 1843134464
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004490707070707071,
      "loss": 1.5702,
      "theoretical_loss": 3.454014215625869,
      "tokens_seen": 1843200000
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044905050505050504,
      "loss": 1.5978,
      "theoretical_loss": 3.4540038111542155,
      "tokens_seen": 1843265536
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044903030303030306,
      "loss": 1.4861,
      "theoretical_loss": 3.4539934071560534,
      "tokens_seen": 1843331072
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000449010101010101,
      "loss": 1.5306,
      "theoretical_loss": 3.453983003631344,
      "tokens_seen": 1843396608
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.30321359634399414,
      "objective/train/avg_token_score": 0.03322860598564148,
      "objective/train/docs_used": 1042423,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0053961277008057,
      "objective/train/original_loss": 4.474796772003174,
      "objective/train/theoretical_loss": 3.45397260058005,
      "objective/train/tokens_used": 202321376,
      "objective/train/unlikelihood_loss": 0.05541876330971718,
      "theoretical_loss": 3.45397260058005,
      "tokens_seen": 1843462144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000448989898989899,
      "loss": 1.5138,
      "theoretical_loss": 3.45397260058005,
      "tokens_seen": 1843462144
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000448969696969697,
      "loss": 1.5161,
      "theoretical_loss": 3.453962198002132,
      "tokens_seen": 1843527680
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044894949494949494,
      "loss": 1.2524,
      "theoretical_loss": 3.4539517958975514,
      "tokens_seen": 1843593216
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044892929292929295,
      "loss": 1.5945,
      "theoretical_loss": 3.4539413942662707,
      "tokens_seen": 1843658752
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044890909090909096,
      "loss": 1.5995,
      "theoretical_loss": 3.4539309931082514,
      "tokens_seen": 1843724288
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004488888888888889,
      "loss": 1.4001,
      "theoretical_loss": 3.4539205924234553,
      "tokens_seen": 1843789824
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044886868686868687,
      "loss": 1.5303,
      "theoretical_loss": 3.4539101922118434,
      "tokens_seen": 1843855360
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044884848484848483,
      "loss": 1.3048,
      "theoretical_loss": 3.4538997924733774,
      "tokens_seen": 1843920896
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044882828282828284,
      "loss": 1.4622,
      "theoretical_loss": 3.4538893932080197,
      "tokens_seen": 1843986432
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044880808080808085,
      "loss": 1.3001,
      "theoretical_loss": 3.4538789944157315,
      "tokens_seen": 1844051968
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487878787878788,
      "loss": 1.232,
      "theoretical_loss": 3.4538685960964743,
      "tokens_seen": 1844117504
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044876767676767677,
      "loss": 1.5215,
      "theoretical_loss": 3.45385819825021,
      "tokens_seen": 1844183040
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487474747474747,
      "loss": 1.4533,
      "theoretical_loss": 3.4538478008769,
      "tokens_seen": 1844248576
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004487272727272728,
      "loss": 1.2763,
      "theoretical_loss": 3.4538374039765065,
      "tokens_seen": 1844314112
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044870707070707074,
      "loss": 1.6231,
      "theoretical_loss": 3.4538270075489907,
      "tokens_seen": 1844379648
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486868686868687,
      "loss": 1.4151,
      "theoretical_loss": 3.4538166115943145,
      "tokens_seen": 1844445184
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044866666666666666,
      "loss": 1.5396,
      "theoretical_loss": 3.45380621611244,
      "tokens_seen": 1844510720
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486464646464646,
      "loss": 1.4981,
      "theoretical_loss": 3.453795821103328,
      "tokens_seen": 1844576256
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004486262626262627,
      "loss": 1.2735,
      "theoretical_loss": 3.4537854265669408,
      "tokens_seen": 1844641792
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044860606060606064,
      "loss": 1.4933,
      "theoretical_loss": 3.4537750325032404,
      "tokens_seen": 1844707328
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004485858585858586,
      "loss": 1.3423,
      "theoretical_loss": 3.4537646389121877,
      "tokens_seen": 1844772864
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044856565656565655,
      "loss": 1.4131,
      "theoretical_loss": 3.4537542457937453,
      "tokens_seen": 1844838400
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004485454545454545,
      "loss": 1.3256,
      "theoretical_loss": 3.4537438531478735,
      "tokens_seen": 1844903936
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004485252525252526,
      "loss": 1.3916,
      "theoretical_loss": 3.4537334609745356,
      "tokens_seen": 1844969472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044850505050505053,
      "loss": 1.5918,
      "theoretical_loss": 3.4537230692736927,
      "tokens_seen": 1845035008
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.3615591526031494,
      "objective/train/avg_token_score": 0.012851543724536896,
      "objective/train/docs_used": 1043833,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1701569557189941,
      "objective/train/original_loss": 4.571152687072754,
      "objective/train/theoretical_loss": 3.4537126780453065,
      "objective/train/tokens_used": 203959776,
      "objective/train/unlikelihood_loss": 0.05402657762169838,
      "theoretical_loss": 3.4537126780453065,
      "tokens_seen": 1845100544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004484848484848485,
      "loss": 1.4921,
      "theoretical_loss": 3.4537126780453065,
      "tokens_seen": 1845100544
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044846464646464644,
      "loss": 1.4343,
      "theoretical_loss": 3.453702287289339,
      "tokens_seen": 1845166080
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044844444444444445,
      "loss": 1.4802,
      "theoretical_loss": 3.4536918970057515,
      "tokens_seen": 1845231616
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044842424242424247,
      "loss": 1.4453,
      "theoretical_loss": 3.453681507194506,
      "tokens_seen": 1845297152
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004484040404040404,
      "loss": 1.6961,
      "theoretical_loss": 3.4536711178555644,
      "tokens_seen": 1845362688
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483838383838384,
      "loss": 1.4932,
      "theoretical_loss": 3.453660728988888,
      "tokens_seen": 1845428224
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044836363636363634,
      "loss": 1.4792,
      "theoretical_loss": 3.453650340594439,
      "tokens_seen": 1845493760
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044834343434343435,
      "loss": 1.454,
      "theoretical_loss": 3.453639952672179,
      "tokens_seen": 1845559296
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044832323232323236,
      "loss": 1.5949,
      "theoretical_loss": 3.4536295652220694,
      "tokens_seen": 1845624832
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004483030303030303,
      "loss": 1.4518,
      "theoretical_loss": 3.453619178244073,
      "tokens_seen": 1845690368
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044828282828282827,
      "loss": 1.5919,
      "theoretical_loss": 3.4536087917381506,
      "tokens_seen": 1845755904
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482626262626263,
      "loss": 1.6802,
      "theoretical_loss": 3.4535984057042644,
      "tokens_seen": 1845821440
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044824242424242424,
      "loss": 1.3382,
      "theoretical_loss": 3.4535880201423765,
      "tokens_seen": 1845886976
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044822222222222225,
      "loss": 1.5491,
      "theoretical_loss": 3.4535776350524476,
      "tokens_seen": 1845952512
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004482020202020202,
      "loss": 1.3793,
      "theoretical_loss": 3.453567250434441,
      "tokens_seen": 1846018048
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044818181818181816,
      "loss": 1.5925,
      "theoretical_loss": 3.453556866288317,
      "tokens_seen": 1846083584
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004481616161616162,
      "loss": 1.5228,
      "theoretical_loss": 3.453546482614038,
      "tokens_seen": 1846149120
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044814141414141413,
      "loss": 1.5559,
      "theoretical_loss": 3.4535360994115667,
      "tokens_seen": 1846214656
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044812121212121214,
      "loss": 1.5943,
      "theoretical_loss": 3.453525716680863,
      "tokens_seen": 1846280192
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004481010101010101,
      "loss": 1.4601,
      "theoretical_loss": 3.453515334421891,
      "tokens_seen": 1846345728
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004480808080808081,
      "loss": 1.3743,
      "theoretical_loss": 3.453504952634611,
      "tokens_seen": 1846411264
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044806060606060607,
      "loss": 1.4953,
      "theoretical_loss": 3.453494571318985,
      "tokens_seen": 1846476800
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000448040404040404,
      "loss": 1.5075,
      "theoretical_loss": 3.453484190474976,
      "tokens_seen": 1846542336
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044802020202020204,
      "loss": 1.4881,
      "theoretical_loss": 3.4534738101025435,
      "tokens_seen": 1846607872
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000448,
      "loss": 1.2744,
      "theoretical_loss": 3.4534634302016514,
      "tokens_seen": 1846673408
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.4943181872367859,
      "objective/train/avg_token_score": 0.03060787171125412,
      "objective/train/docs_used": 1044447,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6434940099716187,
      "objective/train/original_loss": 4.084818363189697,
      "objective/train/theoretical_loss": 3.4534530507722607,
      "objective/train/tokens_used": 205598176,
      "objective/train/unlikelihood_loss": 0.05918062478303909,
      "theoretical_loss": 3.4534530507722607,
      "tokens_seen": 1846738944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447979797979798,
      "loss": 1.5188,
      "theoretical_loss": 3.4534530507722607,
      "tokens_seen": 1846738944
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044795959595959596,
      "loss": 1.5925,
      "theoretical_loss": 3.453442671814334,
      "tokens_seen": 1846804480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004479393939393939,
      "loss": 1.3562,
      "theoretical_loss": 3.4534322933278316,
      "tokens_seen": 1846870016
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044791919191919193,
      "loss": 1.4972,
      "theoretical_loss": 3.453421915312717,
      "tokens_seen": 1846935552
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044789898989898994,
      "loss": 1.306,
      "theoretical_loss": 3.453411537768951,
      "tokens_seen": 1847001088
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478787878787879,
      "loss": 1.3592,
      "theoretical_loss": 3.4534011606964965,
      "tokens_seen": 1847066624
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044785858585858585,
      "loss": 1.6231,
      "theoretical_loss": 3.4533907840953146,
      "tokens_seen": 1847132160
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478383838383838,
      "loss": 1.5106,
      "theoretical_loss": 3.453380407965367,
      "tokens_seen": 1847197696
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004478181818181818,
      "loss": 1.438,
      "theoretical_loss": 3.4533700323066157,
      "tokens_seen": 1847263232
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044779797979797983,
      "loss": 1.3438,
      "theoretical_loss": 3.4533596571190235,
      "tokens_seen": 1847328768
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477777777777778,
      "loss": 1.3831,
      "theoretical_loss": 3.4533492824025513,
      "tokens_seen": 1847394304
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044775757575757575,
      "loss": 1.423,
      "theoretical_loss": 3.453338908157161,
      "tokens_seen": 1847459840
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004477373737373737,
      "loss": 1.5044,
      "theoretical_loss": 3.4533285343828153,
      "tokens_seen": 1847525376
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044771717171717177,
      "loss": 1.5387,
      "theoretical_loss": 3.453318161079475,
      "tokens_seen": 1847590912
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476969696969697,
      "loss": 1.2799,
      "theoretical_loss": 3.453307788247103,
      "tokens_seen": 1847656448
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476767676767677,
      "loss": 1.5208,
      "theoretical_loss": 3.45329741588566,
      "tokens_seen": 1847721984
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044765656565656564,
      "loss": 1.6915,
      "theoretical_loss": 3.4532870439951098,
      "tokens_seen": 1847787520
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004476363636363636,
      "loss": 1.7828,
      "theoretical_loss": 3.453276672575413,
      "tokens_seen": 1847853056
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044761616161616166,
      "loss": 1.515,
      "theoretical_loss": 3.4532663016265315,
      "tokens_seen": 1847918592
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004475959595959596,
      "loss": 1.3178,
      "theoretical_loss": 3.4532559311484277,
      "tokens_seen": 1847984128
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004475757575757576,
      "loss": 1.4015,
      "theoretical_loss": 3.4532455611410633,
      "tokens_seen": 1848049664
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044755555555555553,
      "loss": 1.3752,
      "theoretical_loss": 3.4532351916044,
      "tokens_seen": 1848115200
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044753535353535354,
      "loss": 1.4995,
      "theoretical_loss": 3.4532248225384006,
      "tokens_seen": 1848180736
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044751515151515155,
      "loss": 1.2319,
      "theoretical_loss": 3.453214453943026,
      "tokens_seen": 1848246272
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474949494949495,
      "loss": 1.3379,
      "theoretical_loss": 3.4532040858182387,
      "tokens_seen": 1848311808
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.3084677457809448,
      "objective/train/avg_token_score": 0.008858302608132362,
      "objective/train/docs_used": 1045763,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8898696899414062,
      "objective/train/original_loss": 4.509169101715088,
      "objective/train/theoretical_loss": 3.453193718164001,
      "objective/train/tokens_used": 207236576,
      "objective/train/unlikelihood_loss": 0.06134689226746559,
      "theoretical_loss": 3.453193718164001,
      "tokens_seen": 1848377344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044747474747474747,
      "loss": 1.286,
      "theoretical_loss": 3.453193718164001,
      "tokens_seen": 1848377344
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004474545454545454,
      "loss": 1.4064,
      "theoretical_loss": 3.4531833509802734,
      "tokens_seen": 1848442880
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044743434343434344,
      "loss": 1.3498,
      "theoretical_loss": 3.45317298426702,
      "tokens_seen": 1848508416
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044741414141414145,
      "loss": 1.1875,
      "theoretical_loss": 3.453162618024201,
      "tokens_seen": 1848573952
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004473939393939394,
      "loss": 1.5256,
      "theoretical_loss": 3.4531522522517797,
      "tokens_seen": 1848639488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044737373737373736,
      "loss": 1.5795,
      "theoretical_loss": 3.453141886949717,
      "tokens_seen": 1848705024
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044735353535353537,
      "loss": 1.4829,
      "theoretical_loss": 3.453131522117975,
      "tokens_seen": 1848770560
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004473333333333334,
      "loss": 1.6749,
      "theoretical_loss": 3.4531211577565166,
      "tokens_seen": 1848836096
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044731313131313134,
      "loss": 1.5497,
      "theoretical_loss": 3.453110793865303,
      "tokens_seen": 1848901632
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472929292929293,
      "loss": 1.431,
      "theoretical_loss": 3.4531004304442963,
      "tokens_seen": 1848967168
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044727272727272725,
      "loss": 1.686,
      "theoretical_loss": 3.4530900674934584,
      "tokens_seen": 1849032704
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044725252525252526,
      "loss": 1.3995,
      "theoretical_loss": 3.453079705012752,
      "tokens_seen": 1849098240
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004472323232323233,
      "loss": 1.5956,
      "theoretical_loss": 3.4530693430021384,
      "tokens_seen": 1849163776
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044721212121212123,
      "loss": 1.6193,
      "theoretical_loss": 3.45305898146158,
      "tokens_seen": 1849229312
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471919191919192,
      "loss": 1.3201,
      "theoretical_loss": 3.453048620391038,
      "tokens_seen": 1849294848
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471717171717172,
      "loss": 1.2248,
      "theoretical_loss": 3.4530382597904756,
      "tokens_seen": 1849360384
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044715151515151516,
      "loss": 1.2997,
      "theoretical_loss": 3.4530278996598547,
      "tokens_seen": 1849425920
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044713131313131317,
      "loss": 1.3588,
      "theoretical_loss": 3.453017539999136,
      "tokens_seen": 1849491456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004471111111111111,
      "loss": 1.5867,
      "theoretical_loss": 3.4530071808082834,
      "tokens_seen": 1849556992
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470909090909091,
      "loss": 1.5271,
      "theoretical_loss": 3.4529968220872576,
      "tokens_seen": 1849622528
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0004470707070707071,
      "loss": 1.3612,
      "theoretical_loss": 3.4529864638360213,
      "tokens_seen": 1849688064
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044705050505050505,
      "loss": 1.7033,
      "theoretical_loss": 3.452976106054536,
      "tokens_seen": 1849753600
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044703030303030306,
      "loss": 1.5997,
      "theoretical_loss": 3.4529657487427636,
      "tokens_seen": 1849819136
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.000447010101010101,
      "loss": 1.2692,
      "theoretical_loss": 3.4529553919006672,
      "tokens_seen": 1849884672
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.00044698989898989903,
      "loss": 1.4866,
      "theoretical_loss": 3.4529450355282085,
      "tokens_seen": 1849950208
    },
    {
      "epoch": 0.11,
      "objective/train/aligned_token_count": 0.5390396118164062,
      "objective/train/avg_token_score": 0.004928861279040575,
      "objective/train/docs_used": 1046397,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7157353162765503,
      "objective/train/original_loss": 3.4299583435058594,
      "objective/train/theoretical_loss": 3.452934679625349,
      "objective/train/tokens_used": 208874976,
      "objective/train/unlikelihood_loss": 0.08482911437749863,
      "theoretical_loss": 3.452934679625349,
      "tokens_seen": 1850015744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000446969696969697,
      "loss": 1.5823,
      "theoretical_loss": 3.452934679625349,
      "tokens_seen": 1850015744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044694949494949494,
      "loss": 1.4851,
      "theoretical_loss": 3.4529243241920513,
      "tokens_seen": 1850081280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044692929292929295,
      "loss": 1.5261,
      "theoretical_loss": 3.452913969228277,
      "tokens_seen": 1850146816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004469090909090909,
      "loss": 1.4711,
      "theoretical_loss": 3.4529036147339887,
      "tokens_seen": 1850212352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468888888888889,
      "loss": 1.5649,
      "theoretical_loss": 3.452893260709148,
      "tokens_seen": 1850277888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468686868686869,
      "loss": 1.5514,
      "theoretical_loss": 3.4528829071537177,
      "tokens_seen": 1850343424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044684848484848484,
      "loss": 1.3146,
      "theoretical_loss": 3.4528725540676595,
      "tokens_seen": 1850408960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044682828282828285,
      "loss": 1.5784,
      "theoretical_loss": 3.452862201450935,
      "tokens_seen": 1850474496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004468080808080808,
      "loss": 1.7336,
      "theoretical_loss": 3.452851849303507,
      "tokens_seen": 1850540032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004467878787878788,
      "loss": 1.4407,
      "theoretical_loss": 3.452841497625337,
      "tokens_seen": 1850605568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044676767676767677,
      "loss": 1.1921,
      "theoretical_loss": 3.452831146416388,
      "tokens_seen": 1850671104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044674747474747473,
      "loss": 1.4424,
      "theoretical_loss": 3.4528207956766215,
      "tokens_seen": 1850736640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044672727272727274,
      "loss": 1.6409,
      "theoretical_loss": 3.4528104454059996,
      "tokens_seen": 1850802176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044670707070707075,
      "loss": 1.4714,
      "theoretical_loss": 3.4528000956044846,
      "tokens_seen": 1850867712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004466868686868687,
      "loss": 1.3572,
      "theoretical_loss": 3.4527897462720385,
      "tokens_seen": 1850933248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044666666666666666,
      "loss": 1.2197,
      "theoretical_loss": 3.452779397408624,
      "tokens_seen": 1850998784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004466464646464646,
      "loss": 1.4201,
      "theoretical_loss": 3.452769049014202,
      "tokens_seen": 1851064320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044662626262626263,
      "loss": 1.4719,
      "theoretical_loss": 3.4527587010887357,
      "tokens_seen": 1851129856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044660606060606064,
      "loss": 1.4856,
      "theoretical_loss": 3.4527483536321864,
      "tokens_seen": 1851195392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465858585858586,
      "loss": 1.5586,
      "theoretical_loss": 3.4527380066445175,
      "tokens_seen": 1851260928
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044656565656565656,
      "loss": 1.4921,
      "theoretical_loss": 3.4527276601256904,
      "tokens_seen": 1851326464
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465454545454545,
      "loss": 1.3876,
      "theoretical_loss": 3.4527173140756666,
      "tokens_seen": 1851392000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004465252525252526,
      "loss": 1.6662,
      "theoretical_loss": 3.4527069684944096,
      "tokens_seen": 1851457536
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044650505050505054,
      "loss": 1.3775,
      "theoretical_loss": 3.4526966233818808,
      "tokens_seen": 1851523072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004464848484848485,
      "loss": 1.3225,
      "theoretical_loss": 3.4526862787380423,
      "tokens_seen": 1851588608
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.2721163332462311,
      "objective/train/avg_token_score": 0.026380645111203194,
      "objective/train/docs_used": 1047383,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.913923442363739,
      "objective/train/original_loss": 4.239612102508545,
      "objective/train/theoretical_loss": 3.4526759345628566,
      "objective/train/tokens_used": 210513376,
      "objective/train/unlikelihood_loss": 0.079360231757164,
      "theoretical_loss": 3.4526759345628566,
      "tokens_seen": 1851654144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044646464646464645,
      "loss": 1.4303,
      "theoretical_loss": 3.4526759345628566,
      "tokens_seen": 1851654144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004464444444444444,
      "loss": 1.2667,
      "theoretical_loss": 3.452665590856286,
      "tokens_seen": 1851719680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044642424242424247,
      "loss": 1.4178,
      "theoretical_loss": 3.452655247618292,
      "tokens_seen": 1851785216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044640404040404043,
      "loss": 1.3794,
      "theoretical_loss": 3.452644904848838,
      "tokens_seen": 1851850752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004463838383838384,
      "loss": 1.5685,
      "theoretical_loss": 3.4526345625478845,
      "tokens_seen": 1851916288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044636363636363634,
      "loss": 1.553,
      "theoretical_loss": 3.4526242207153954,
      "tokens_seen": 1851981824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044634343434343435,
      "loss": 1.7237,
      "theoretical_loss": 3.4526138793513317,
      "tokens_seen": 1852047360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044632323232323236,
      "loss": 1.753,
      "theoretical_loss": 3.452603538455656,
      "tokens_seen": 1852112896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004463030303030303,
      "loss": 1.4571,
      "theoretical_loss": 3.452593198028331,
      "tokens_seen": 1852178432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462828282828283,
      "loss": 1.31,
      "theoretical_loss": 3.4525828580693183,
      "tokens_seen": 1852243968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044626262626262623,
      "loss": 1.5186,
      "theoretical_loss": 3.45257251857858,
      "tokens_seen": 1852309504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044624242424242425,
      "loss": 1.6351,
      "theoretical_loss": 3.452562179556079,
      "tokens_seen": 1852375040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044622222222222226,
      "loss": 1.1539,
      "theoretical_loss": 3.452551841001777,
      "tokens_seen": 1852440576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004462020202020202,
      "loss": 1.3132,
      "theoretical_loss": 3.4525415029156363,
      "tokens_seen": 1852506112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044618181818181817,
      "loss": 1.2812,
      "theoretical_loss": 3.452531165297619,
      "tokens_seen": 1852571648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461616161616162,
      "loss": 1.606,
      "theoretical_loss": 3.452520828147688,
      "tokens_seen": 1852637184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044614141414141414,
      "loss": 1.4788,
      "theoretical_loss": 3.452510491465805,
      "tokens_seen": 1852702720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044612121212121215,
      "loss": 1.5314,
      "theoretical_loss": 3.4525001552519323,
      "tokens_seen": 1852768256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004461010101010101,
      "loss": 1.309,
      "theoretical_loss": 3.4524898195060323,
      "tokens_seen": 1852833792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044608080808080806,
      "loss": 1.384,
      "theoretical_loss": 3.4524794842280673,
      "tokens_seen": 1852899328
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004460606060606061,
      "loss": 1.462,
      "theoretical_loss": 3.4524691494179995,
      "tokens_seen": 1852964864
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044604040404040403,
      "loss": 1.6896,
      "theoretical_loss": 3.452458815075791,
      "tokens_seen": 1853030400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044602020202020204,
      "loss": 1.4372,
      "theoretical_loss": 3.452448481201404,
      "tokens_seen": 1853095936
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000446,
      "loss": 1.4209,
      "theoretical_loss": 3.452438147794801,
      "tokens_seen": 1853161472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000445979797979798,
      "loss": 1.422,
      "theoretical_loss": 3.4524278148559446,
      "tokens_seen": 1853227008
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.31262218952178955,
      "objective/train/avg_token_score": 0.019027452915906906,
      "objective/train/docs_used": 1048172,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1232824325561523,
      "objective/train/original_loss": 4.744579792022705,
      "objective/train/theoretical_loss": 3.4524174823847966,
      "objective/train/tokens_used": 212151776,
      "objective/train/unlikelihood_loss": 0.04645653814077377,
      "theoretical_loss": 3.4524174823847966,
      "tokens_seen": 1853292544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044595959595959597,
      "loss": 1.1746,
      "theoretical_loss": 3.4524174823847966,
      "tokens_seen": 1853292544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004459393939393939,
      "loss": 1.4012,
      "theoretical_loss": 3.452407150381319,
      "tokens_seen": 1853358080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044591919191919193,
      "loss": 1.4507,
      "theoretical_loss": 3.4523968188454752,
      "tokens_seen": 1853423616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458989898989899,
      "loss": 1.463,
      "theoretical_loss": 3.452386487777227,
      "tokens_seen": 1853489152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458787878787879,
      "loss": 1.4316,
      "theoretical_loss": 3.452376157176536,
      "tokens_seen": 1853554688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044585858585858586,
      "loss": 1.7792,
      "theoretical_loss": 3.452365827043365,
      "tokens_seen": 1853620224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004458383838383838,
      "loss": 1.6054,
      "theoretical_loss": 3.4523554973776767,
      "tokens_seen": 1853685760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044581818181818183,
      "loss": 1.3471,
      "theoretical_loss": 3.452345168179433,
      "tokens_seen": 1853751296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044579797979797984,
      "loss": 1.4412,
      "theoretical_loss": 3.4523348394485964,
      "tokens_seen": 1853816832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457777777777778,
      "loss": 1.3501,
      "theoretical_loss": 3.452324511185129,
      "tokens_seen": 1853882368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044575757575757575,
      "loss": 1.5304,
      "theoretical_loss": 3.452314183388993,
      "tokens_seen": 1853947904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457373737373737,
      "loss": 1.4022,
      "theoretical_loss": 3.4523038560601513,
      "tokens_seen": 1854013440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004457171717171717,
      "loss": 1.3444,
      "theoretical_loss": 3.452293529198566,
      "tokens_seen": 1854078976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044569696969696973,
      "loss": 1.2717,
      "theoretical_loss": 3.4522832028041988,
      "tokens_seen": 1854144512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004456767676767677,
      "loss": 1.3764,
      "theoretical_loss": 3.4522728768770135,
      "tokens_seen": 1854210048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044565656565656564,
      "loss": 1.3803,
      "theoretical_loss": 3.452262551416971,
      "tokens_seen": 1854275584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044563636363636366,
      "loss": 1.4479,
      "theoretical_loss": 3.4522522264240347,
      "tokens_seen": 1854341120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044561616161616167,
      "loss": 1.6955,
      "theoretical_loss": 3.452241901898166,
      "tokens_seen": 1854406656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455959595959596,
      "loss": 1.6218,
      "theoretical_loss": 3.452231577839328,
      "tokens_seen": 1854472192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004455757575757576,
      "loss": 1.3389,
      "theoretical_loss": 3.452221254247483,
      "tokens_seen": 1854537728
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044555555555555554,
      "loss": 1.521,
      "theoretical_loss": 3.4522109311225933,
      "tokens_seen": 1854603264
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044553535353535355,
      "loss": 1.2394,
      "theoretical_loss": 3.4522006084646204,
      "tokens_seen": 1854668800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044551515151515156,
      "loss": 1.4995,
      "theoretical_loss": 3.4521902862735283,
      "tokens_seen": 1854734336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454949494949495,
      "loss": 1.3931,
      "theoretical_loss": 3.4521799645492783,
      "tokens_seen": 1854799872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454747474747475,
      "loss": 1.5762,
      "theoretical_loss": 3.452169643291833,
      "tokens_seen": 1854865408
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.40560850501060486,
      "objective/train/avg_token_score": 0.008102362044155598,
      "objective/train/docs_used": 1049323,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3408201932907104,
      "objective/train/original_loss": 4.186563968658447,
      "objective/train/theoretical_loss": 3.452159322501155,
      "objective/train/tokens_used": 213790176,
      "objective/train/unlikelihood_loss": 0.057345665991306305,
      "theoretical_loss": 3.452159322501155,
      "tokens_seen": 1854930944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044545454545454543,
      "loss": 1.5162,
      "theoretical_loss": 3.452159322501155,
      "tokens_seen": 1854930944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004454343434343435,
      "loss": 1.5745,
      "theoretical_loss": 3.4521490021772063,
      "tokens_seen": 1854996480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044541414141414145,
      "loss": 1.3886,
      "theoretical_loss": 3.45213868231995,
      "tokens_seen": 1855062016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453939393939394,
      "loss": 1.381,
      "theoretical_loss": 3.452128362929348,
      "tokens_seen": 1855127552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044537373737373737,
      "loss": 1.703,
      "theoretical_loss": 3.4521180440053625,
      "tokens_seen": 1855193088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453535353535353,
      "loss": 1.569,
      "theoretical_loss": 3.452107725547956,
      "tokens_seen": 1855258624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004453333333333334,
      "loss": 1.2288,
      "theoretical_loss": 3.4520974075570914,
      "tokens_seen": 1855324160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044531313131313135,
      "loss": 1.5825,
      "theoretical_loss": 3.452087090032731,
      "tokens_seen": 1855389696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004452929292929293,
      "loss": 1.5339,
      "theoretical_loss": 3.452076772974837,
      "tokens_seen": 1855455232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044527272727272726,
      "loss": 1.173,
      "theoretical_loss": 3.4520664563833723,
      "tokens_seen": 1855520768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004452525252525252,
      "loss": 1.5338,
      "theoretical_loss": 3.452056140258298,
      "tokens_seen": 1855586304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004452323232323233,
      "loss": 1.2947,
      "theoretical_loss": 3.4520458245995784,
      "tokens_seen": 1855651840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044521212121212124,
      "loss": 1.4635,
      "theoretical_loss": 3.4520355094071746,
      "tokens_seen": 1855717376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004451919191919192,
      "loss": 1.2195,
      "theoretical_loss": 3.45202519468105,
      "tokens_seen": 1855782912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044517171717171715,
      "loss": 1.5204,
      "theoretical_loss": 3.452014880421166,
      "tokens_seen": 1855848448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044515151515151516,
      "loss": 1.2559,
      "theoretical_loss": 3.452004566627486,
      "tokens_seen": 1855913984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004451313131313132,
      "loss": 1.3033,
      "theoretical_loss": 3.4519942532999712,
      "tokens_seen": 1855979520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044511111111111113,
      "loss": 1.5715,
      "theoretical_loss": 3.451983940438586,
      "tokens_seen": 1856045056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004450909090909091,
      "loss": 1.506,
      "theoretical_loss": 3.4519736280432918,
      "tokens_seen": 1856110592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044507070707070704,
      "loss": 1.4881,
      "theoretical_loss": 3.4519633161140506,
      "tokens_seen": 1856176128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044505050505050506,
      "loss": 1.3911,
      "theoretical_loss": 3.4519530046508256,
      "tokens_seen": 1856241664
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044503030303030307,
      "loss": 1.3362,
      "theoretical_loss": 3.451942693653579,
      "tokens_seen": 1856307200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000445010101010101,
      "loss": 1.4433,
      "theoretical_loss": 3.4519323831222737,
      "tokens_seen": 1856372736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444989898989899,
      "loss": 1.2808,
      "theoretical_loss": 3.4519220730568714,
      "tokens_seen": 1856438272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444969696969697,
      "loss": 1.43,
      "theoretical_loss": 3.4519117634573355,
      "tokens_seen": 1856503808
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.33345553278923035,
      "objective/train/avg_token_score": 0.06058519333600998,
      "objective/train/docs_used": 1049929,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.12827730178833,
      "objective/train/original_loss": 4.973564624786377,
      "objective/train/theoretical_loss": 3.451901454323628,
      "objective/train/tokens_used": 215428576,
      "objective/train/unlikelihood_loss": 0.046354275196790695,
      "theoretical_loss": 3.451901454323628,
      "tokens_seen": 1856569344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044494949494949495,
      "loss": 1.4263,
      "theoretical_loss": 3.451901454323628,
      "tokens_seen": 1856569344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044492929292929296,
      "loss": 1.2963,
      "theoretical_loss": 3.4518911456557113,
      "tokens_seen": 1856634880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004449090909090909,
      "loss": 1.6288,
      "theoretical_loss": 3.4518808374535483,
      "tokens_seen": 1856700416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044488888888888887,
      "loss": 1.3875,
      "theoretical_loss": 3.4518705297171013,
      "tokens_seen": 1856765952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004448686868686869,
      "loss": 1.3577,
      "theoretical_loss": 3.4518602224463324,
      "tokens_seen": 1856831488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044484848484848484,
      "loss": 1.5654,
      "theoretical_loss": 3.4518499156412052,
      "tokens_seen": 1856897024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044482828282828285,
      "loss": 1.2628,
      "theoretical_loss": 3.4518396093016817,
      "tokens_seen": 1856962560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004448080808080808,
      "loss": 1.3431,
      "theoretical_loss": 3.451829303427724,
      "tokens_seen": 1857028096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004447878787878788,
      "loss": 1.4391,
      "theoretical_loss": 3.451818998019295,
      "tokens_seen": 1857093632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004447676767676768,
      "loss": 1.4135,
      "theoretical_loss": 3.451808693076357,
      "tokens_seen": 1857159168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044474747474747473,
      "loss": 1.5886,
      "theoretical_loss": 3.4517983885988732,
      "tokens_seen": 1857224704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044472727272727274,
      "loss": 1.4285,
      "theoretical_loss": 3.4517880845868056,
      "tokens_seen": 1857290240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004447070707070707,
      "loss": 1.4832,
      "theoretical_loss": 3.4517777810401165,
      "tokens_seen": 1857355776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004446868686868687,
      "loss": 1.3972,
      "theoretical_loss": 3.4517674779587693,
      "tokens_seen": 1857421312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044466666666666667,
      "loss": 1.4377,
      "theoretical_loss": 3.451757175342726,
      "tokens_seen": 1857486848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004446464646464646,
      "loss": 1.5561,
      "theoretical_loss": 3.451746873191949,
      "tokens_seen": 1857552384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044462626262626264,
      "loss": 1.3046,
      "theoretical_loss": 3.4517365715064017,
      "tokens_seen": 1857617920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044460606060606065,
      "loss": 1.5743,
      "theoretical_loss": 3.451726270286046,
      "tokens_seen": 1857683456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445858585858586,
      "loss": 1.238,
      "theoretical_loss": 3.451715969530844,
      "tokens_seen": 1857748992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044456565656565656,
      "loss": 1.6605,
      "theoretical_loss": 3.4517056692407597,
      "tokens_seen": 1857814528
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004445454545454545,
      "loss": 1.5474,
      "theoretical_loss": 3.451695369415755,
      "tokens_seen": 1857880064
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044452525252525253,
      "loss": 1.4325,
      "theoretical_loss": 3.4516850700557917,
      "tokens_seen": 1857945600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044450505050505054,
      "loss": 1.4822,
      "theoretical_loss": 3.451674771160833,
      "tokens_seen": 1858011136
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444848484848485,
      "loss": 1.3059,
      "theoretical_loss": 3.451664472730842,
      "tokens_seen": 1858076672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044446464646464645,
      "loss": 1.3968,
      "theoretical_loss": 3.451654174765781,
      "tokens_seen": 1858142208
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.40243157744407654,
      "objective/train/avg_token_score": 0.01305119413882494,
      "objective/train/docs_used": 1050597,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3890475034713745,
      "objective/train/original_loss": 4.159512042999268,
      "objective/train/theoretical_loss": 3.451643877265613,
      "objective/train/tokens_used": 217066976,
      "objective/train/unlikelihood_loss": 0.06591649353504181,
      "theoretical_loss": 3.451643877265613,
      "tokens_seen": 1858207744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444444444444444,
      "loss": 1.5441,
      "theoretical_loss": 3.451643877265613,
      "tokens_seen": 1858207744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004444242424242425,
      "loss": 1.4623,
      "theoretical_loss": 3.451633580230299,
      "tokens_seen": 1858273280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044440404040404043,
      "loss": 1.5315,
      "theoretical_loss": 3.4516232836598033,
      "tokens_seen": 1858338816
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004443838383838384,
      "loss": 1.2582,
      "theoretical_loss": 3.451612987554088,
      "tokens_seen": 1858404352
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044436363636363635,
      "loss": 1.5215,
      "theoretical_loss": 3.451602691913116,
      "tokens_seen": 1858469888
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004443434343434343,
      "loss": 1.3035,
      "theoretical_loss": 3.4515923967368494,
      "tokens_seen": 1858535424
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044432323232323237,
      "loss": 1.342,
      "theoretical_loss": 3.4515821020252506,
      "tokens_seen": 1858600960
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004443030303030303,
      "loss": 1.3374,
      "theoretical_loss": 3.4515718077782838,
      "tokens_seen": 1858666496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004442828282828283,
      "loss": 1.3353,
      "theoretical_loss": 3.4515615139959097,
      "tokens_seen": 1858732032
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044426262626262624,
      "loss": 1.4248,
      "theoretical_loss": 3.451551220678092,
      "tokens_seen": 1858797568
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044424242424242425,
      "loss": 1.47,
      "theoretical_loss": 3.4515409278247935,
      "tokens_seen": 1858863104
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044422222222222226,
      "loss": 1.1401,
      "theoretical_loss": 3.4515306354359767,
      "tokens_seen": 1858928640
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004442020202020202,
      "loss": 1.5552,
      "theoretical_loss": 3.4515203435116035,
      "tokens_seen": 1858994176
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441818181818182,
      "loss": 1.4739,
      "theoretical_loss": 3.4515100520516375,
      "tokens_seen": 1859059712
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044416161616161613,
      "loss": 1.4808,
      "theoretical_loss": 3.4514997610560414,
      "tokens_seen": 1859125248
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044414141414141414,
      "loss": 1.3743,
      "theoretical_loss": 3.451489470524777,
      "tokens_seen": 1859190784
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044412121212121215,
      "loss": 1.4206,
      "theoretical_loss": 3.4514791804578078,
      "tokens_seen": 1859256320
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004441010101010101,
      "loss": 1.5322,
      "theoretical_loss": 3.4514688908550966,
      "tokens_seen": 1859321856
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044408080808080807,
      "loss": 1.4228,
      "theoretical_loss": 3.451458601716605,
      "tokens_seen": 1859387392
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004440606060606061,
      "loss": 1.7084,
      "theoretical_loss": 3.451448313042297,
      "tokens_seen": 1859452928
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044404040404040404,
      "loss": 1.1568,
      "theoretical_loss": 3.451438024832134,
      "tokens_seen": 1859518464
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044402020202020205,
      "loss": 1.3594,
      "theoretical_loss": 3.45142773708608,
      "tokens_seen": 1859584000
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000444,
      "loss": 1.5296,
      "theoretical_loss": 3.451417449804097,
      "tokens_seen": 1859649536
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044397979797979796,
      "loss": 1.243,
      "theoretical_loss": 3.451407162986148,
      "tokens_seen": 1859715072
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044395959595959597,
      "loss": 1.3292,
      "theoretical_loss": 3.451396876632195,
      "tokens_seen": 1859780608
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.3087732195854187,
      "objective/train/avg_token_score": 0.015457415953278542,
      "objective/train/docs_used": 1051522,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0185426473617554,
      "objective/train/original_loss": 4.571277618408203,
      "objective/train/theoretical_loss": 3.4513865907422017,
      "objective/train/tokens_used": 218705376,
      "objective/train/unlikelihood_loss": 0.05881958827376366,
      "theoretical_loss": 3.4513865907422017,
      "tokens_seen": 1859846144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000443939393939394,
      "loss": 1.3964,
      "theoretical_loss": 3.4513865907422017,
      "tokens_seen": 1859846144
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044391919191919194,
      "loss": 1.4011,
      "theoretical_loss": 3.4513763053161304,
      "tokens_seen": 1859911680
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438989898989899,
      "loss": 1.6454,
      "theoretical_loss": 3.4513660203539436,
      "tokens_seen": 1859977216
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438787878787879,
      "loss": 1.4499,
      "theoretical_loss": 3.4513557358556044,
      "tokens_seen": 1860042752
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044385858585858587,
      "loss": 1.4571,
      "theoretical_loss": 3.4513454518210755,
      "tokens_seen": 1860108288
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004438383838383839,
      "loss": 1.1652,
      "theoretical_loss": 3.4513351682503193,
      "tokens_seen": 1860173824
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044381818181818183,
      "loss": 1.4164,
      "theoretical_loss": 3.4513248851432987,
      "tokens_seen": 1860239360
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437979797979798,
      "loss": 1.2335,
      "theoretical_loss": 3.451314602499977,
      "tokens_seen": 1860304896
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437777777777778,
      "loss": 1.4228,
      "theoretical_loss": 3.451304320320316,
      "tokens_seen": 1860370432
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044375757575757576,
      "loss": 1.3746,
      "theoretical_loss": 3.451294038604279,
      "tokens_seen": 1860435968
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044373737373737377,
      "loss": 1.6736,
      "theoretical_loss": 3.4512837573518294,
      "tokens_seen": 1860501504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004437171717171717,
      "loss": 1.1772,
      "theoretical_loss": 3.4512734765629283,
      "tokens_seen": 1860567040
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044369696969696974,
      "loss": 1.406,
      "theoretical_loss": 3.45126319623754,
      "tokens_seen": 1860632576
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004436767676767677,
      "loss": 1.4388,
      "theoretical_loss": 3.451252916375627,
      "tokens_seen": 1860698112
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044365656565656565,
      "loss": 1.6811,
      "theoretical_loss": 3.4512426369771516,
      "tokens_seen": 1860763648
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044363636363636366,
      "loss": 1.3693,
      "theoretical_loss": 3.4512323580420765,
      "tokens_seen": 1860829184
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004436161616161616,
      "loss": 1.4087,
      "theoretical_loss": 3.451222079570365,
      "tokens_seen": 1860894720
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044359595959595963,
      "loss": 1.3867,
      "theoretical_loss": 3.4512118015619797,
      "tokens_seen": 1860960256
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004435757575757576,
      "loss": 1.566,
      "theoretical_loss": 3.4512015240168834,
      "tokens_seen": 1861025792
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044355555555555554,
      "loss": 1.513,
      "theoretical_loss": 3.451191246935039,
      "tokens_seen": 1861091328
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044353535353535355,
      "loss": 1.492,
      "theoretical_loss": 3.4511809703164094,
      "tokens_seen": 1861156864
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004435151515151515,
      "loss": 1.5802,
      "theoretical_loss": 3.4511706941609566,
      "tokens_seen": 1861222400
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434949494949495,
      "loss": 1.3848,
      "theoretical_loss": 3.4511604184686444,
      "tokens_seen": 1861287936
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004434747474747475,
      "loss": 1.4916,
      "theoretical_loss": 3.4511501432394356,
      "tokens_seen": 1861353472
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044345454545454544,
      "loss": 1.1949,
      "theoretical_loss": 3.451139868473292,
      "tokens_seen": 1861419008
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.37634408473968506,
      "objective/train/avg_token_score": 0.01542601641267538,
      "objective/train/docs_used": 1052031,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.181724190711975,
      "objective/train/original_loss": 4.738986968994141,
      "objective/train/theoretical_loss": 3.4511295941701774,
      "objective/train/tokens_used": 220343776,
      "objective/train/unlikelihood_loss": 0.061755482107400894,
      "theoretical_loss": 3.4511295941701774,
      "tokens_seen": 1861484544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044343434343434345,
      "loss": 1.2778,
      "theoretical_loss": 3.4511295941701774,
      "tokens_seen": 1861484544
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044341414141414146,
      "loss": 1.4889,
      "theoretical_loss": 3.4511193203300543,
      "tokens_seen": 1861550080
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004433939393939394,
      "loss": 1.6773,
      "theoretical_loss": 3.451109046952886,
      "tokens_seen": 1861615616
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044337373737373737,
      "loss": 1.4959,
      "theoretical_loss": 3.451098774038634,
      "tokens_seen": 1861681152
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044335353535353533,
      "loss": 1.5357,
      "theoretical_loss": 3.4510885015872628,
      "tokens_seen": 1861746688
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044333333333333334,
      "loss": 1.1331,
      "theoretical_loss": 3.4510782295987346,
      "tokens_seen": 1861812224
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044331313131313135,
      "loss": 1.1775,
      "theoretical_loss": 3.4510679580730117,
      "tokens_seen": 1861877760
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004432929292929293,
      "loss": 1.4438,
      "theoretical_loss": 3.4510576870100573,
      "tokens_seen": 1861943296
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044327272727272726,
      "loss": 1.2656,
      "theoretical_loss": 3.4510474164098346,
      "tokens_seen": 1862008832
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004432525252525252,
      "loss": 1.3183,
      "theoretical_loss": 3.4510371462723066,
      "tokens_seen": 1862074368
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004432323232323233,
      "loss": 1.4267,
      "theoretical_loss": 3.4510268765974357,
      "tokens_seen": 1862139904
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044321212121212124,
      "loss": 1.3859,
      "theoretical_loss": 3.451016607385185,
      "tokens_seen": 1862205440
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431919191919192,
      "loss": 1.2745,
      "theoretical_loss": 3.451006338635517,
      "tokens_seen": 1862270976
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044317171717171716,
      "loss": 1.264,
      "theoretical_loss": 3.4509960703483946,
      "tokens_seen": 1862336512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431515151515151,
      "loss": 1.4957,
      "theoretical_loss": 3.450985802523782,
      "tokens_seen": 1862402048
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004431313131313132,
      "loss": 1.1714,
      "theoretical_loss": 3.45097553516164,
      "tokens_seen": 1862467584
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044311111111111114,
      "loss": 1.6347,
      "theoretical_loss": 3.4509652682619327,
      "tokens_seen": 1862533120
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004430909090909091,
      "loss": 1.456,
      "theoretical_loss": 3.4509550018246236,
      "tokens_seen": 1862598656
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044307070707070705,
      "loss": 1.48,
      "theoretical_loss": 3.4509447358496743,
      "tokens_seen": 1862664192
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044305050505050506,
      "loss": 1.5629,
      "theoretical_loss": 3.450934470337048,
      "tokens_seen": 1862729728
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044303030303030307,
      "loss": 1.2863,
      "theoretical_loss": 3.450924205286708,
      "tokens_seen": 1862795264
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044301010101010103,
      "loss": 1.5154,
      "theoretical_loss": 3.4509139406986176,
      "tokens_seen": 1862860800
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000442989898989899,
      "loss": 1.4515,
      "theoretical_loss": 3.450903676572739,
      "tokens_seen": 1862926336
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044296969696969694,
      "loss": 1.3576,
      "theoretical_loss": 3.4508934129090356,
      "tokens_seen": 1862991872
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044294949494949495,
      "loss": 1.4711,
      "theoretical_loss": 3.4508831497074697,
      "tokens_seen": 1863057408
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.35129520297050476,
      "objective/train/avg_token_score": 0.03377072885632515,
      "objective/train/docs_used": 1053498,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3309406042099,
      "objective/train/original_loss": 4.976852893829346,
      "objective/train/theoretical_loss": 3.4508728869680043,
      "objective/train/tokens_used": 221982176,
      "objective/train/unlikelihood_loss": 0.04581369832158089,
      "theoretical_loss": 3.4508728869680043,
      "tokens_seen": 1863122944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044292929292929296,
      "loss": 1.6527,
      "theoretical_loss": 3.4508728869680043,
      "tokens_seen": 1863122944
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004429090909090909,
      "loss": 1.2831,
      "theoretical_loss": 3.450862624690603,
      "tokens_seen": 1863188480
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004428888888888889,
      "loss": 1.2713,
      "theoretical_loss": 3.450852362875229,
      "tokens_seen": 1863254016
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004428686868686869,
      "loss": 1.5471,
      "theoretical_loss": 3.450842101521844,
      "tokens_seen": 1863319552
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044284848484848485,
      "loss": 1.3051,
      "theoretical_loss": 3.4508318406304115,
      "tokens_seen": 1863385088
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044282828282828286,
      "loss": 1.401,
      "theoretical_loss": 3.4508215802008952,
      "tokens_seen": 1863450624
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004428080808080808,
      "loss": 1.2803,
      "theoretical_loss": 3.450811320233257,
      "tokens_seen": 1863516160
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044278787878787877,
      "loss": 1.3002,
      "theoretical_loss": 3.4508010607274606,
      "tokens_seen": 1863581696
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427676767676768,
      "loss": 1.478,
      "theoretical_loss": 3.4507908016834685,
      "tokens_seen": 1863647232
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044274747474747474,
      "loss": 1.4375,
      "theoretical_loss": 3.4507805431012435,
      "tokens_seen": 1863712768
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044272727272727275,
      "loss": 1.3306,
      "theoretical_loss": 3.4507702849807496,
      "tokens_seen": 1863778304
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004427070707070707,
      "loss": 1.4733,
      "theoretical_loss": 3.450760027321948,
      "tokens_seen": 1863843840
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004426868686868687,
      "loss": 1.5835,
      "theoretical_loss": 3.450749770124804,
      "tokens_seen": 1863909376
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004426666666666667,
      "loss": 1.4993,
      "theoretical_loss": 3.4507395133892786,
      "tokens_seen": 1863974912
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044264646464646463,
      "loss": 1.423,
      "theoretical_loss": 3.450729257115336,
      "tokens_seen": 1864040448
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044262626262626264,
      "loss": 1.493,
      "theoretical_loss": 3.4507190013029385,
      "tokens_seen": 1864105984
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004426060606060606,
      "loss": 1.5426,
      "theoretical_loss": 3.45070874595205,
      "tokens_seen": 1864171520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004425858585858586,
      "loss": 1.5896,
      "theoretical_loss": 3.4506984910626324,
      "tokens_seen": 1864237056
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044256565656565657,
      "loss": 1.5533,
      "theoretical_loss": 3.450688236634649,
      "tokens_seen": 1864302592
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004425454545454545,
      "loss": 1.2419,
      "theoretical_loss": 3.450677982668063,
      "tokens_seen": 1864368128
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044252525252525254,
      "loss": 1.2114,
      "theoretical_loss": 3.450667729162838,
      "tokens_seen": 1864433664
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044250505050505055,
      "loss": 1.3821,
      "theoretical_loss": 3.450657476118936,
      "tokens_seen": 1864499200
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004424848484848485,
      "loss": 1.4739,
      "theoretical_loss": 3.450647223536321,
      "tokens_seen": 1864564736
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044246464646464646,
      "loss": 1.5295,
      "theoretical_loss": 3.450636971414955,
      "tokens_seen": 1864630272
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004424444444444444,
      "loss": 1.2789,
      "theoretical_loss": 3.4506267197548017,
      "tokens_seen": 1864695808
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.39766618609428406,
      "objective/train/avg_token_score": 0.03243307024240494,
      "objective/train/docs_used": 1054143,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4144164323806763,
      "objective/train/original_loss": 5.110235691070557,
      "objective/train/theoretical_loss": 3.4506164685558245,
      "objective/train/tokens_used": 223620576,
      "objective/train/unlikelihood_loss": 0.04839473217725754,
      "theoretical_loss": 3.4506164685558245,
      "tokens_seen": 1864761344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044242424242424243,
      "loss": 1.4935,
      "theoretical_loss": 3.4506164685558245,
      "tokens_seen": 1864761344
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044240404040404044,
      "loss": 1.3372,
      "theoretical_loss": 3.4506062178179855,
      "tokens_seen": 1864826880
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423838383838384,
      "loss": 1.4895,
      "theoretical_loss": 3.450595967541248,
      "tokens_seen": 1864892416
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044236363636363635,
      "loss": 1.3013,
      "theoretical_loss": 3.450585717725576,
      "tokens_seen": 1864957952
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423434343434343,
      "loss": 1.5954,
      "theoretical_loss": 3.450575468370931,
      "tokens_seen": 1865023488
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004423232323232324,
      "loss": 1.5314,
      "theoretical_loss": 3.4505652194772773,
      "tokens_seen": 1865089024
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044230303030303033,
      "loss": 1.386,
      "theoretical_loss": 3.4505549710445775,
      "tokens_seen": 1865154560
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422828282828283,
      "loss": 1.3209,
      "theoretical_loss": 3.450544723072795,
      "tokens_seen": 1865220096
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044226262626262625,
      "loss": 1.5006,
      "theoretical_loss": 3.4505344755618923,
      "tokens_seen": 1865285632
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422424242424242,
      "loss": 1.6062,
      "theoretical_loss": 3.450524228511833,
      "tokens_seen": 1865351168
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044222222222222227,
      "loss": 1.4058,
      "theoretical_loss": 3.4505139819225796,
      "tokens_seen": 1865416704
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004422020202020202,
      "loss": 1.5056,
      "theoretical_loss": 3.450503735794096,
      "tokens_seen": 1865482240
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421818181818182,
      "loss": 1.4215,
      "theoretical_loss": 3.4504934901263447,
      "tokens_seen": 1865547776
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044216161616161614,
      "loss": 1.2773,
      "theoretical_loss": 3.450483244919289,
      "tokens_seen": 1865613312
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421414141414142,
      "loss": 1.7203,
      "theoretical_loss": 3.450473000172892,
      "tokens_seen": 1865678848
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044212121212121216,
      "loss": 1.412,
      "theoretical_loss": 3.450462755887117,
      "tokens_seen": 1865744384
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004421010101010101,
      "loss": 1.3453,
      "theoretical_loss": 3.4504525120619265,
      "tokens_seen": 1865809920
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004420808080808081,
      "loss": 1.4669,
      "theoretical_loss": 3.4504422686972838,
      "tokens_seen": 1865875456
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044206060606060603,
      "loss": 1.4215,
      "theoretical_loss": 3.4504320257931527,
      "tokens_seen": 1865940992
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004420404040404041,
      "loss": 1.5517,
      "theoretical_loss": 3.450421783349496,
      "tokens_seen": 1866006528
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044202020202020205,
      "loss": 1.2322,
      "theoretical_loss": 3.450411541366276,
      "tokens_seen": 1866072064
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000442,
      "loss": 1.578,
      "theoretical_loss": 3.450401299843457,
      "tokens_seen": 1866137600
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044197979797979797,
      "loss": 1.392,
      "theoretical_loss": 3.450391058781001,
      "tokens_seen": 1866203136
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0004419595959595959,
      "loss": 1.3039,
      "theoretical_loss": 3.450380818178872,
      "tokens_seen": 1866268672
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.000441939393939394,
      "loss": 1.6445,
      "theoretical_loss": 3.4503705780370337,
      "tokens_seen": 1866334208
    },
    {
      "epoch": 0.12,
      "objective/train/aligned_token_count": 0.35972630977630615,
      "objective/train/avg_token_score": 0.007249823771417141,
      "objective/train/docs_used": 1055457,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1949776411056519,
      "objective/train/original_loss": 4.569217681884766,
      "objective/train/theoretical_loss": 3.4503603383554475,
      "objective/train/tokens_used": 225258976,
      "objective/train/unlikelihood_loss": 0.057365208864212036,
      "theoretical_loss": 3.4503603383554475,
      "tokens_seen": 1866399744
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.00044191919191919195,
      "loss": 1.4593,
      "theoretical_loss": 3.4503603383554475,
      "tokens_seen": 1866399744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418989898989899,
      "loss": 1.3262,
      "theoretical_loss": 3.450350099134078,
      "tokens_seen": 1866465280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044187878787878786,
      "loss": 1.3928,
      "theoretical_loss": 3.450339860372888,
      "tokens_seen": 1866530816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044185858585858587,
      "loss": 1.4387,
      "theoretical_loss": 3.4503296220718402,
      "tokens_seen": 1866596352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004418383838383839,
      "loss": 1.4918,
      "theoretical_loss": 3.4503193842308986,
      "tokens_seen": 1866661888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044181818181818184,
      "loss": 1.2656,
      "theoretical_loss": 3.4503091468500253,
      "tokens_seen": 1866727424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004417979797979798,
      "loss": 1.4871,
      "theoretical_loss": 3.4502989099291845,
      "tokens_seen": 1866792960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044177777777777775,
      "loss": 1.5025,
      "theoretical_loss": 3.450288673468339,
      "tokens_seen": 1866858496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044175757575757576,
      "loss": 1.3684,
      "theoretical_loss": 3.4502784374674516,
      "tokens_seen": 1866924032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004417373737373738,
      "loss": 1.2633,
      "theoretical_loss": 3.4502682019264856,
      "tokens_seen": 1866989568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044171717171717173,
      "loss": 1.432,
      "theoretical_loss": 3.450257966845405,
      "tokens_seen": 1867055104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004416969696969697,
      "loss": 1.385,
      "theoretical_loss": 3.450247732224172,
      "tokens_seen": 1867120640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004416767676767677,
      "loss": 1.3333,
      "theoretical_loss": 3.4502374980627506,
      "tokens_seen": 1867186176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044165656565656566,
      "loss": 1.5311,
      "theoretical_loss": 3.4502272643611027,
      "tokens_seen": 1867251712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044163636363636367,
      "loss": 1.2273,
      "theoretical_loss": 3.4502170311191933,
      "tokens_seen": 1867317248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004416161616161616,
      "loss": 1.3906,
      "theoretical_loss": 3.4502067983369846,
      "tokens_seen": 1867382784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004415959595959596,
      "loss": 1.4789,
      "theoretical_loss": 3.45019656601444,
      "tokens_seen": 1867448320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004415757575757576,
      "loss": 1.4171,
      "theoretical_loss": 3.450186334151522,
      "tokens_seen": 1867513856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044155555555555555,
      "loss": 1.6178,
      "theoretical_loss": 3.4501761027481948,
      "tokens_seen": 1867579392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044153535353535356,
      "loss": 1.3581,
      "theoretical_loss": 3.4501658718044217,
      "tokens_seen": 1867644928
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004415151515151515,
      "loss": 1.2223,
      "theoretical_loss": 3.4501556413201646,
      "tokens_seen": 1867710464
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044149494949494953,
      "loss": 1.1382,
      "theoretical_loss": 3.450145411295389,
      "tokens_seen": 1867776000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004414747474747475,
      "loss": 1.4542,
      "theoretical_loss": 3.450135181730056,
      "tokens_seen": 1867841536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044145454545454544,
      "loss": 1.5126,
      "theoretical_loss": 3.4501249526241295,
      "tokens_seen": 1867907072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044143434343434345,
      "loss": 1.5464,
      "theoretical_loss": 3.450114723977573,
      "tokens_seen": 1867972608
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.5038489699363708,
      "objective/train/avg_token_score": 0.0115775465965271,
      "objective/train/docs_used": 1056292,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7018463611602783,
      "objective/train/original_loss": 4.302921295166016,
      "objective/train/theoretical_loss": 3.4501044957903506,
      "objective/train/tokens_used": 226897376,
      "objective/train/unlikelihood_loss": 0.07719095051288605,
      "theoretical_loss": 3.4501044957903506,
      "tokens_seen": 1868038144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004414141414141414,
      "loss": 1.5629,
      "theoretical_loss": 3.4501044957903506,
      "tokens_seen": 1868038144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413939393939394,
      "loss": 1.3905,
      "theoretical_loss": 3.450094268062424,
      "tokens_seen": 1868103680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004413737373737374,
      "loss": 1.514,
      "theoretical_loss": 3.4500840407937567,
      "tokens_seen": 1868169216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044135353535353533,
      "loss": 1.4503,
      "theoretical_loss": 3.450073813984313,
      "tokens_seen": 1868234752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044133333333333335,
      "loss": 1.3241,
      "theoretical_loss": 3.4500635876340553,
      "tokens_seen": 1868300288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044131313131313136,
      "loss": 1.3679,
      "theoretical_loss": 3.450053361742947,
      "tokens_seen": 1868365824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004412929292929293,
      "loss": 1.3124,
      "theoretical_loss": 3.450043136310952,
      "tokens_seen": 1868431360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044127272727272727,
      "loss": 1.4361,
      "theoretical_loss": 3.450032911338032,
      "tokens_seen": 1868496896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004412525252525252,
      "loss": 1.5088,
      "theoretical_loss": 3.4500226868241524,
      "tokens_seen": 1868562432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044123232323232324,
      "loss": 1.5715,
      "theoretical_loss": 3.450012462769275,
      "tokens_seen": 1868627968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044121212121212125,
      "loss": 1.3653,
      "theoretical_loss": 3.450002239173364,
      "tokens_seen": 1868693504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411919191919192,
      "loss": 1.6154,
      "theoretical_loss": 3.449992016036382,
      "tokens_seen": 1868759040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044117171717171716,
      "loss": 1.5006,
      "theoretical_loss": 3.449981793358292,
      "tokens_seen": 1868824576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411515151515151,
      "loss": 1.5315,
      "theoretical_loss": 3.4499715711390584,
      "tokens_seen": 1868890112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004411313131313132,
      "loss": 1.416,
      "theoretical_loss": 3.4499613493786443,
      "tokens_seen": 1868955648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044111111111111114,
      "loss": 1.7077,
      "theoretical_loss": 3.4499511280770125,
      "tokens_seen": 1869021184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004410909090909091,
      "loss": 1.4733,
      "theoretical_loss": 3.4499409072341263,
      "tokens_seen": 1869086720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044107070707070706,
      "loss": 1.2313,
      "theoretical_loss": 3.4499306868499495,
      "tokens_seen": 1869152256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000441050505050505,
      "loss": 1.3513,
      "theoretical_loss": 3.449920466924445,
      "tokens_seen": 1869217792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004410303030303031,
      "loss": 1.4508,
      "theoretical_loss": 3.4499102474575762,
      "tokens_seen": 1869283328
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044101010101010103,
      "loss": 1.4852,
      "theoretical_loss": 3.4499000284493073,
      "tokens_seen": 1869348864
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000440989898989899,
      "loss": 1.4368,
      "theoretical_loss": 3.4498898098996,
      "tokens_seen": 1869414400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044096969696969695,
      "loss": 1.3618,
      "theoretical_loss": 3.4498795918084193,
      "tokens_seen": 1869479936
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044094949494949496,
      "loss": 1.3093,
      "theoretical_loss": 3.4498693741757274,
      "tokens_seen": 1869545472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044092929292929297,
      "loss": 1.3623,
      "theoretical_loss": 3.4498591570014883,
      "tokens_seen": 1869611008
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.49095797538757324,
      "objective/train/avg_token_score": 0.0304701067507267,
      "objective/train/docs_used": 1057648,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6677335500717163,
      "objective/train/original_loss": 4.48096227645874,
      "objective/train/theoretical_loss": 3.4498489402856647,
      "objective/train/tokens_used": 228535776,
      "objective/train/unlikelihood_loss": 0.04762193188071251,
      "theoretical_loss": 3.4498489402856647,
      "tokens_seen": 1869676544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044090909090909093,
      "loss": 1.5815,
      "theoretical_loss": 3.4498489402856647,
      "tokens_seen": 1869676544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408888888888889,
      "loss": 1.3001,
      "theoretical_loss": 3.4498387240282207,
      "tokens_seen": 1869742080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044086868686868684,
      "loss": 1.6185,
      "theoretical_loss": 3.4498285082291194,
      "tokens_seen": 1869807616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044084848484848485,
      "loss": 1.5549,
      "theoretical_loss": 3.449818292888324,
      "tokens_seen": 1869873152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044082828282828286,
      "loss": 1.3115,
      "theoretical_loss": 3.4498080780057983,
      "tokens_seen": 1869938688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004408080808080808,
      "loss": 1.3658,
      "theoretical_loss": 3.449797863581505,
      "tokens_seen": 1870004224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407878787878788,
      "loss": 1.5948,
      "theoretical_loss": 3.4497876496154083,
      "tokens_seen": 1870069760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407676767676768,
      "loss": 1.399,
      "theoretical_loss": 3.4497774361074707,
      "tokens_seen": 1870135296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044074747474747474,
      "loss": 1.3492,
      "theoretical_loss": 3.4497672230576564,
      "tokens_seen": 1870200832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044072727272727276,
      "loss": 1.3206,
      "theoretical_loss": 3.4497570104659285,
      "tokens_seen": 1870266368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004407070707070707,
      "loss": 1.6108,
      "theoretical_loss": 3.4497467983322503,
      "tokens_seen": 1870331904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044068686868686867,
      "loss": 1.559,
      "theoretical_loss": 3.4497365866565852,
      "tokens_seen": 1870397440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004406666666666667,
      "loss": 1.3866,
      "theoretical_loss": 3.4497263754388965,
      "tokens_seen": 1870462976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044064646464646464,
      "loss": 1.2945,
      "theoretical_loss": 3.449716164679148,
      "tokens_seen": 1870528512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044062626262626265,
      "loss": 1.5109,
      "theoretical_loss": 3.449705954377303,
      "tokens_seen": 1870594048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004406060606060606,
      "loss": 1.4244,
      "theoretical_loss": 3.449695744533325,
      "tokens_seen": 1870659584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004405858585858586,
      "loss": 1.4769,
      "theoretical_loss": 3.449685535147177,
      "tokens_seen": 1870725120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004405656565656566,
      "loss": 1.5157,
      "theoretical_loss": 3.4496753262188227,
      "tokens_seen": 1870790656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044054545454545453,
      "loss": 1.2994,
      "theoretical_loss": 3.4496651177482254,
      "tokens_seen": 1870856192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044052525252525254,
      "loss": 1.4714,
      "theoretical_loss": 3.449654909735349,
      "tokens_seen": 1870921728
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004405050505050505,
      "loss": 1.3804,
      "theoretical_loss": 3.449644702180156,
      "tokens_seen": 1870987264
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404848484848485,
      "loss": 1.4414,
      "theoretical_loss": 3.4496344950826114,
      "tokens_seen": 1871052800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044046464646464647,
      "loss": 1.3298,
      "theoretical_loss": 3.4496242884426773,
      "tokens_seen": 1871118336
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004404444444444445,
      "loss": 1.3629,
      "theoretical_loss": 3.449614082260317,
      "tokens_seen": 1871183872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044042424242424243,
      "loss": 1.442,
      "theoretical_loss": 3.4496038765354955,
      "tokens_seen": 1871249408
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.4029814302921295,
      "objective/train/avg_token_score": 0.022246306762099266,
      "objective/train/docs_used": 1058498,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2657883167266846,
      "objective/train/original_loss": 4.2054595947265625,
      "objective/train/theoretical_loss": 3.449593671268175,
      "objective/train/tokens_used": 230174176,
      "objective/train/unlikelihood_loss": 0.0596332885324955,
      "theoretical_loss": 3.449593671268175,
      "tokens_seen": 1871314944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044040404040404044,
      "loss": 1.2121,
      "theoretical_loss": 3.449593671268175,
      "tokens_seen": 1871314944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403838383838384,
      "loss": 1.4529,
      "theoretical_loss": 3.449583466458319,
      "tokens_seen": 1871380480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044036363636363636,
      "loss": 1.3306,
      "theoretical_loss": 3.4495732621058917,
      "tokens_seen": 1871446016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044034343434343437,
      "loss": 1.7467,
      "theoretical_loss": 3.449563058210856,
      "tokens_seen": 1871511552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004403232323232323,
      "loss": 1.2752,
      "theoretical_loss": 3.4495528547731755,
      "tokens_seen": 1871577088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044030303030303034,
      "loss": 1.337,
      "theoretical_loss": 3.4495426517928136,
      "tokens_seen": 1871642624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402828282828283,
      "loss": 1.2836,
      "theoretical_loss": 3.449532449269734,
      "tokens_seen": 1871708160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044026262626262625,
      "loss": 1.3727,
      "theoretical_loss": 3.4495222472039,
      "tokens_seen": 1871773696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044024242424242426,
      "loss": 1.3912,
      "theoretical_loss": 3.449512045595275,
      "tokens_seen": 1871839232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004402222222222222,
      "loss": 1.4787,
      "theoretical_loss": 3.449501844443823,
      "tokens_seen": 1871904768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044020202020202023,
      "loss": 1.5051,
      "theoretical_loss": 3.449491643749507,
      "tokens_seen": 1871970304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004401818181818182,
      "loss": 1.6093,
      "theoretical_loss": 3.449481443512291,
      "tokens_seen": 1872035840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044016161616161614,
      "loss": 1.4234,
      "theoretical_loss": 3.449471243732138,
      "tokens_seen": 1872101376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044014141414141416,
      "loss": 1.6637,
      "theoretical_loss": 3.449461044409012,
      "tokens_seen": 1872166912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044012121212121217,
      "loss": 1.4772,
      "theoretical_loss": 3.449450845542876,
      "tokens_seen": 1872232448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004401010101010101,
      "loss": 1.491,
      "theoretical_loss": 3.4494406471336934,
      "tokens_seen": 1872297984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004400808080808081,
      "loss": 1.6623,
      "theoretical_loss": 3.4494304491814285,
      "tokens_seen": 1872363520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044006060606060604,
      "loss": 1.5462,
      "theoretical_loss": 3.449420251686045,
      "tokens_seen": 1872429056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044004040404040405,
      "loss": 1.4287,
      "theoretical_loss": 3.4494100546475055,
      "tokens_seen": 1872494592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044002020202020206,
      "loss": 1.532,
      "theoretical_loss": 3.4493998580657736,
      "tokens_seen": 1872560128
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00044,
      "loss": 1.2202,
      "theoretical_loss": 3.4493896619408138,
      "tokens_seen": 1872625664
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043997979797979797,
      "loss": 1.4907,
      "theoretical_loss": 3.449379466272589,
      "tokens_seen": 1872691200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043995959595959593,
      "loss": 1.61,
      "theoretical_loss": 3.4493692710610624,
      "tokens_seen": 1872756736
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439939393939394,
      "loss": 1.4321,
      "theoretical_loss": 3.449359076306198,
      "tokens_seen": 1872822272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043991919191919195,
      "loss": 1.608,
      "theoretical_loss": 3.4493488820079596,
      "tokens_seen": 1872887808
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.39846041798591614,
      "objective/train/avg_token_score": 0.02917778491973877,
      "objective/train/docs_used": 1059136,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3530430793762207,
      "objective/train/original_loss": 4.573091506958008,
      "objective/train/theoretical_loss": 3.4493386881663106,
      "objective/train/tokens_used": 231812576,
      "objective/train/unlikelihood_loss": 0.08011636883020401,
      "theoretical_loss": 3.4493386881663106,
      "tokens_seen": 1872953344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398989898989899,
      "loss": 1.4504,
      "theoretical_loss": 3.4493386881663106,
      "tokens_seen": 1872953344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043987878787878787,
      "loss": 1.2369,
      "theoretical_loss": 3.4493284947812146,
      "tokens_seen": 1873018880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398585858585858,
      "loss": 1.5826,
      "theoretical_loss": 3.4493183018526348,
      "tokens_seen": 1873084416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004398383838383839,
      "loss": 1.2305,
      "theoretical_loss": 3.449308109380535,
      "tokens_seen": 1873149952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043981818181818184,
      "loss": 1.5691,
      "theoretical_loss": 3.449297917364879,
      "tokens_seen": 1873215488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397979797979798,
      "loss": 1.5117,
      "theoretical_loss": 3.44928772580563,
      "tokens_seen": 1873281024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043977777777777776,
      "loss": 1.4064,
      "theoretical_loss": 3.449277534702752,
      "tokens_seen": 1873346560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043975757575757577,
      "loss": 1.6067,
      "theoretical_loss": 3.4492673440562083,
      "tokens_seen": 1873412096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004397373737373738,
      "loss": 1.4002,
      "theoretical_loss": 3.449257153865963,
      "tokens_seen": 1873477632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043971717171717174,
      "loss": 1.6485,
      "theoretical_loss": 3.449246964131979,
      "tokens_seen": 1873543168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004396969696969697,
      "loss": 1.4152,
      "theoretical_loss": 3.4492367748542203,
      "tokens_seen": 1873608704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043967676767676765,
      "loss": 1.572,
      "theoretical_loss": 3.4492265860326508,
      "tokens_seen": 1873674240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043965656565656566,
      "loss": 1.545,
      "theoretical_loss": 3.449216397667233,
      "tokens_seen": 1873739776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043963636363636367,
      "loss": 1.3972,
      "theoretical_loss": 3.449206209757932,
      "tokens_seen": 1873805312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043961616161616163,
      "loss": 1.3669,
      "theoretical_loss": 3.4491960223047107,
      "tokens_seen": 1873870848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395959595959596,
      "loss": 1.3002,
      "theoretical_loss": 3.4491858353075324,
      "tokens_seen": 1873936384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395757575757576,
      "loss": 1.4357,
      "theoretical_loss": 3.449175648766362,
      "tokens_seen": 1874001920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043955555555555555,
      "loss": 1.432,
      "theoretical_loss": 3.449165462681161,
      "tokens_seen": 1874067456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043953535353535357,
      "loss": 1.3794,
      "theoretical_loss": 3.4491552770518954,
      "tokens_seen": 1874132992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004395151515151515,
      "loss": 1.5028,
      "theoretical_loss": 3.449145091878527,
      "tokens_seen": 1874198528
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394949494949495,
      "loss": 1.5145,
      "theoretical_loss": 3.4491349071610204,
      "tokens_seen": 1874264064
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394747474747475,
      "loss": 1.3535,
      "theoretical_loss": 3.449124722899339,
      "tokens_seen": 1874329600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043945454545454545,
      "loss": 1.4699,
      "theoretical_loss": 3.449114539093447,
      "tokens_seen": 1874395136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043943434343434346,
      "loss": 1.3388,
      "theoretical_loss": 3.449104355743307,
      "tokens_seen": 1874460672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004394141414141414,
      "loss": 1.2571,
      "theoretical_loss": 3.4490941728488838,
      "tokens_seen": 1874526208
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.3365102708339691,
      "objective/train/avg_token_score": 0.021979324519634247,
      "objective/train/docs_used": 1059915,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1313660144805908,
      "objective/train/original_loss": 4.815589904785156,
      "objective/train/theoretical_loss": 3.44908399041014,
      "objective/train/tokens_used": 233450976,
      "objective/train/unlikelihood_loss": 0.04281238839030266,
      "theoretical_loss": 3.44908399041014,
      "tokens_seen": 1874591744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004393939393939394,
      "loss": 1.4928,
      "theoretical_loss": 3.44908399041014,
      "tokens_seen": 1874591744
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004393737373737374,
      "loss": 1.3874,
      "theoretical_loss": 3.4490738084270403,
      "tokens_seen": 1874657280
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043935353535353534,
      "loss": 1.4704,
      "theoretical_loss": 3.4490636268995476,
      "tokens_seen": 1874722816
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043933333333333335,
      "loss": 1.5677,
      "theoretical_loss": 3.4490534458276256,
      "tokens_seen": 1874788352
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004393131313131313,
      "loss": 1.499,
      "theoretical_loss": 3.449043265211239,
      "tokens_seen": 1874853888
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004392929292929293,
      "loss": 1.5894,
      "theoretical_loss": 3.4490330850503503,
      "tokens_seen": 1874919424
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004392727272727273,
      "loss": 1.4306,
      "theoretical_loss": 3.4490229053449237,
      "tokens_seen": 1874984960
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043925252525252523,
      "loss": 1.3304,
      "theoretical_loss": 3.449012726094923,
      "tokens_seen": 1875050496
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043923232323232324,
      "loss": 1.5022,
      "theoretical_loss": 3.449002547300312,
      "tokens_seen": 1875116032
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043921212121212125,
      "loss": 1.3642,
      "theoretical_loss": 3.4489923689610538,
      "tokens_seen": 1875181568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004391919191919192,
      "loss": 1.3564,
      "theoretical_loss": 3.4489821910771123,
      "tokens_seen": 1875247104
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043917171717171717,
      "loss": 1.4577,
      "theoretical_loss": 3.448972013648452,
      "tokens_seen": 1875312640
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004391515151515151,
      "loss": 1.3256,
      "theoretical_loss": 3.4489618366750356,
      "tokens_seen": 1875378176
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043913131313131314,
      "loss": 1.3953,
      "theoretical_loss": 3.448951660156828,
      "tokens_seen": 1875443712
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043911111111111115,
      "loss": 1.44,
      "theoretical_loss": 3.4489414840937913,
      "tokens_seen": 1875509248
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390909090909091,
      "loss": 1.3612,
      "theoretical_loss": 3.448931308485891,
      "tokens_seen": 1875574784
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043907070707070706,
      "loss": 1.6487,
      "theoretical_loss": 3.4489211333330894,
      "tokens_seen": 1875640320
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000439050505050505,
      "loss": 1.4251,
      "theoretical_loss": 3.4489109586353512,
      "tokens_seen": 1875705856
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004390303030303031,
      "loss": 1.2871,
      "theoretical_loss": 3.44890078439264,
      "tokens_seen": 1875771392
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043901010101010104,
      "loss": 1.5798,
      "theoretical_loss": 3.448890610604918,
      "tokens_seen": 1875836928
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438989898989899,
      "loss": 1.3093,
      "theoretical_loss": 3.4488804372721518,
      "tokens_seen": 1875902464
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043896969696969695,
      "loss": 1.6002,
      "theoretical_loss": 3.448870264394303,
      "tokens_seen": 1875968000
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004389494949494949,
      "loss": 1.5905,
      "theoretical_loss": 3.448860091971336,
      "tokens_seen": 1876033536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438929292929293,
      "loss": 1.3809,
      "theoretical_loss": 3.448849920003215,
      "tokens_seen": 1876099072
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043890909090909093,
      "loss": 1.5747,
      "theoretical_loss": 3.4488397484899034,
      "tokens_seen": 1876164608
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.3816593289375305,
      "objective/train/avg_token_score": 0.014148054644465446,
      "objective/train/docs_used": 1060624,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0186139345169067,
      "objective/train/original_loss": 3.7407007217407227,
      "objective/train/theoretical_loss": 3.448829577431365,
      "objective/train/tokens_used": 235089376,
      "objective/train/unlikelihood_loss": 0.09927164018154144,
      "theoretical_loss": 3.448829577431365,
      "tokens_seen": 1876230144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004388888888888889,
      "loss": 1.2479,
      "theoretical_loss": 3.448829577431365,
      "tokens_seen": 1876230144
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043886868686868685,
      "loss": 1.4699,
      "theoretical_loss": 3.448819406827563,
      "tokens_seen": 1876295680
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004388484848484848,
      "loss": 1.3708,
      "theoretical_loss": 3.448809236678462,
      "tokens_seen": 1876361216
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043882828282828287,
      "loss": 1.4538,
      "theoretical_loss": 3.4487990669840256,
      "tokens_seen": 1876426752
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004388080808080808,
      "loss": 1.7556,
      "theoretical_loss": 3.448788897744217,
      "tokens_seen": 1876492288
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387878787878788,
      "loss": 1.3391,
      "theoretical_loss": 3.4487787289590015,
      "tokens_seen": 1876557824
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043876767676767674,
      "loss": 1.5276,
      "theoretical_loss": 3.4487685606283414,
      "tokens_seen": 1876623360
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387474747474748,
      "loss": 1.6076,
      "theoretical_loss": 3.4487583927522008,
      "tokens_seen": 1876688896
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043872727272727276,
      "loss": 1.3362,
      "theoretical_loss": 3.4487482253305437,
      "tokens_seen": 1876754432
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004387070707070707,
      "loss": 1.4142,
      "theoretical_loss": 3.4487380583633342,
      "tokens_seen": 1876819968
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386868686868687,
      "loss": 1.6026,
      "theoretical_loss": 3.4487278918505364,
      "tokens_seen": 1876885504
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043866666666666663,
      "loss": 1.4723,
      "theoretical_loss": 3.448717725792113,
      "tokens_seen": 1876951040
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386464646464647,
      "loss": 1.6469,
      "theoretical_loss": 3.448707560188028,
      "tokens_seen": 1877016576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043862626262626265,
      "loss": 1.5528,
      "theoretical_loss": 3.4486973950382467,
      "tokens_seen": 1877082112
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004386060606060606,
      "loss": 1.4923,
      "theoretical_loss": 3.448687230342731,
      "tokens_seen": 1877147648
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043858585858585857,
      "loss": 1.1962,
      "theoretical_loss": 3.448677066101446,
      "tokens_seen": 1877213184
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385656565656566,
      "loss": 1.3383,
      "theoretical_loss": 3.448666902314355,
      "tokens_seen": 1877278720
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385454545454546,
      "loss": 1.354,
      "theoretical_loss": 3.4486567389814224,
      "tokens_seen": 1877344256
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043852525252525255,
      "loss": 1.4967,
      "theoretical_loss": 3.4486465761026115,
      "tokens_seen": 1877409792
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004385050505050505,
      "loss": 1.362,
      "theoretical_loss": 3.448636413677886,
      "tokens_seen": 1877475328
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043848484848484846,
      "loss": 1.5587,
      "theoretical_loss": 3.4486262517072106,
      "tokens_seen": 1877540864
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043846464646464647,
      "loss": 1.4487,
      "theoretical_loss": 3.4486160901905487,
      "tokens_seen": 1877606400
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004384444444444445,
      "loss": 1.242,
      "theoretical_loss": 3.448605929127864,
      "tokens_seen": 1877671936
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043842424242424244,
      "loss": 1.5602,
      "theoretical_loss": 3.4485957685191204,
      "tokens_seen": 1877737472
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004384040404040404,
      "loss": 1.3082,
      "theoretical_loss": 3.448585608364282,
      "tokens_seen": 1877803008
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.49346286058425903,
      "objective/train/avg_token_score": 0.00607198616489768,
      "objective/train/docs_used": 1062058,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6114000082015991,
      "objective/train/original_loss": 3.4861137866973877,
      "objective/train/theoretical_loss": 3.4485754486633127,
      "objective/train/tokens_used": 236727776,
      "objective/train/unlikelihood_loss": 0.09068997949361801,
      "theoretical_loss": 3.4485754486633127,
      "tokens_seen": 1877868544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004383838383838384,
      "loss": 1.4889,
      "theoretical_loss": 3.4485754486633127,
      "tokens_seen": 1877868544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043836363636363636,
      "loss": 1.4036,
      "theoretical_loss": 3.448565289416176,
      "tokens_seen": 1877934080
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004383434343434344,
      "loss": 1.5098,
      "theoretical_loss": 3.448555130622836,
      "tokens_seen": 1877999616
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043832323232323233,
      "loss": 1.429,
      "theoretical_loss": 3.4485449722832566,
      "tokens_seen": 1878065152
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004383030303030303,
      "loss": 1.2384,
      "theoretical_loss": 3.4485348143974024,
      "tokens_seen": 1878130688
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382828282828283,
      "loss": 1.3225,
      "theoretical_loss": 3.448524656965236,
      "tokens_seen": 1878196224
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043826262626262626,
      "loss": 1.5303,
      "theoretical_loss": 3.448514499986722,
      "tokens_seen": 1878261760
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043824242424242427,
      "loss": 1.4846,
      "theoretical_loss": 3.448504343461824,
      "tokens_seen": 1878327296
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004382222222222222,
      "loss": 1.4366,
      "theoretical_loss": 3.448494187390507,
      "tokens_seen": 1878392832
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043820202020202024,
      "loss": 1.2521,
      "theoretical_loss": 3.448484031772734,
      "tokens_seen": 1878458368
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004381818181818182,
      "loss": 1.2788,
      "theoretical_loss": 3.4484738766084684,
      "tokens_seen": 1878523904
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043816161616161615,
      "loss": 1.4862,
      "theoretical_loss": 3.4484637218976752,
      "tokens_seen": 1878589440
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043814141414141416,
      "loss": 1.5265,
      "theoretical_loss": 3.4484535676403176,
      "tokens_seen": 1878654976
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004381212121212121,
      "loss": 1.734,
      "theoretical_loss": 3.44844341383636,
      "tokens_seen": 1878720512
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043810101010101013,
      "loss": 1.465,
      "theoretical_loss": 3.448433260485766,
      "tokens_seen": 1878786048
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004380808080808081,
      "loss": 1.44,
      "theoretical_loss": 3.4484231075885,
      "tokens_seen": 1878851584
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043806060606060604,
      "loss": 1.6416,
      "theoretical_loss": 3.4484129551445255,
      "tokens_seen": 1878917120
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043804040404040405,
      "loss": 1.3597,
      "theoretical_loss": 3.4484028031538063,
      "tokens_seen": 1878982656
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043802020202020206,
      "loss": 1.7944,
      "theoretical_loss": 3.4483926516163073,
      "tokens_seen": 1879048192
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000438,
      "loss": 1.4988,
      "theoretical_loss": 3.4483825005319915,
      "tokens_seen": 1879113728
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000437979797979798,
      "loss": 1.2804,
      "theoretical_loss": 3.448372349900823,
      "tokens_seen": 1879179264
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043795959595959593,
      "loss": 1.3335,
      "theoretical_loss": 3.448362199722766,
      "tokens_seen": 1879244800
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043793939393939395,
      "loss": 1.5275,
      "theoretical_loss": 3.4483520499977844,
      "tokens_seen": 1879310336
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043791919191919196,
      "loss": 1.4088,
      "theoretical_loss": 3.4483419007258425,
      "tokens_seen": 1879375872
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004378989898989899,
      "loss": 1.5582,
      "theoretical_loss": 3.4483317519069043,
      "tokens_seen": 1879441408
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.5173509120941162,
      "objective/train/avg_token_score": 0.04670140892267227,
      "objective/train/docs_used": 1062685,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7349787950515747,
      "objective/train/original_loss": 3.844883918762207,
      "objective/train/theoretical_loss": 3.4483216035409328,
      "objective/train/tokens_used": 238366176,
      "objective/train/unlikelihood_loss": 0.05952979996800423,
      "theoretical_loss": 3.4483216035409328,
      "tokens_seen": 1879506944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043787878787878787,
      "loss": 1.3996,
      "theoretical_loss": 3.4483216035409328,
      "tokens_seen": 1879506944
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043785858585858583,
      "loss": 1.4732,
      "theoretical_loss": 3.4483114556278927,
      "tokens_seen": 1879572480
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004378383838383839,
      "loss": 1.5456,
      "theoretical_loss": 3.448301308167748,
      "tokens_seen": 1879638016
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043781818181818185,
      "loss": 1.732,
      "theoretical_loss": 3.448291161160463,
      "tokens_seen": 1879703552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004377979797979798,
      "loss": 1.5276,
      "theoretical_loss": 3.4482810146060014,
      "tokens_seen": 1879769088
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043777777777777776,
      "loss": 1.6654,
      "theoretical_loss": 3.4482708685043266,
      "tokens_seen": 1879834624
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004377575757575757,
      "loss": 1.4529,
      "theoretical_loss": 3.4482607228554034,
      "tokens_seen": 1879900160
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004377373737373738,
      "loss": 1.367,
      "theoretical_loss": 3.448250577659196,
      "tokens_seen": 1879965696
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043771717171717174,
      "loss": 1.3748,
      "theoretical_loss": 3.448240432915667,
      "tokens_seen": 1880031232
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376969696969697,
      "loss": 1.3824,
      "theoretical_loss": 3.4482302886247824,
      "tokens_seen": 1880096768
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043767676767676766,
      "loss": 1.4715,
      "theoretical_loss": 3.448220144786505,
      "tokens_seen": 1880162304
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043765656565656567,
      "loss": 1.3517,
      "theoretical_loss": 3.448210001400799,
      "tokens_seen": 1880227840
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004376363636363637,
      "loss": 1.3135,
      "theoretical_loss": 3.4481998584676288,
      "tokens_seen": 1880293376
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043761616161616164,
      "loss": 1.5466,
      "theoretical_loss": 3.448189715986958,
      "tokens_seen": 1880358912
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004375959595959596,
      "loss": 1.5333,
      "theoretical_loss": 3.4481795739587504,
      "tokens_seen": 1880424448
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043757575757575755,
      "loss": 1.3885,
      "theoretical_loss": 3.448169432382971,
      "tokens_seen": 1880489984
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043755555555555556,
      "loss": 1.6391,
      "theoretical_loss": 3.448159291259583,
      "tokens_seen": 1880555520
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043753535353535357,
      "loss": 1.7587,
      "theoretical_loss": 3.448149150588551,
      "tokens_seen": 1880621056
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043751515151515153,
      "loss": 1.4384,
      "theoretical_loss": 3.4481390103698386,
      "tokens_seen": 1880686592
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374949494949495,
      "loss": 1.3425,
      "theoretical_loss": 3.44812887060341,
      "tokens_seen": 1880752128
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374747474747475,
      "loss": 1.597,
      "theoretical_loss": 3.448118731289229,
      "tokens_seen": 1880817664
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043745454545454545,
      "loss": 1.2963,
      "theoretical_loss": 3.448108592427261,
      "tokens_seen": 1880883200
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043743434343434346,
      "loss": 1.4757,
      "theoretical_loss": 3.448098454017469,
      "tokens_seen": 1880948736
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004374141414141414,
      "loss": 1.3524,
      "theoretical_loss": 3.4480883160598164,
      "tokens_seen": 1881014272
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004373939393939394,
      "loss": 1.672,
      "theoretical_loss": 3.448078178554268,
      "tokens_seen": 1881079808
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.41361191868782043,
      "objective/train/avg_token_score": 0.008350033313035965,
      "objective/train/docs_used": 1064316,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4659239053726196,
      "objective/train/original_loss": 4.062493801116943,
      "objective/train/theoretical_loss": 3.4480680415007887,
      "objective/train/tokens_used": 240004576,
      "objective/train/unlikelihood_loss": 0.06435056775808334,
      "theoretical_loss": 3.4480680415007887,
      "tokens_seen": 1881145344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004373737373737374,
      "loss": 1.3753,
      "theoretical_loss": 3.4480680415007887,
      "tokens_seen": 1881145344
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043735353535353535,
      "loss": 1.301,
      "theoretical_loss": 3.4480579048993416,
      "tokens_seen": 1881210880
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043733333333333336,
      "loss": 1.3698,
      "theoretical_loss": 3.448047768749891,
      "tokens_seen": 1881276416
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004373131313131313,
      "loss": 1.6219,
      "theoretical_loss": 3.448037633052401,
      "tokens_seen": 1881341952
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004372929292929293,
      "loss": 1.4685,
      "theoretical_loss": 3.4480274978068355,
      "tokens_seen": 1881407488
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004372727272727273,
      "loss": 1.3703,
      "theoretical_loss": 3.448017363013159,
      "tokens_seen": 1881473024
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043725252525252524,
      "loss": 1.3487,
      "theoretical_loss": 3.4480072286713357,
      "tokens_seen": 1881538560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043723232323232325,
      "loss": 1.1748,
      "theoretical_loss": 3.4479970947813294,
      "tokens_seen": 1881604096
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004372121212121212,
      "loss": 1.4412,
      "theoretical_loss": 3.447986961343104,
      "tokens_seen": 1881669632
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004371919191919192,
      "loss": 1.5357,
      "theoretical_loss": 3.4479768283566243,
      "tokens_seen": 1881735168
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004371717171717172,
      "loss": 1.6326,
      "theoretical_loss": 3.447966695821854,
      "tokens_seen": 1881800704
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043715151515151513,
      "loss": 1.5148,
      "theoretical_loss": 3.447956563738757,
      "tokens_seen": 1881866240
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043713131313131314,
      "loss": 1.6197,
      "theoretical_loss": 3.4479464321072983,
      "tokens_seen": 1881931776
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043711111111111115,
      "loss": 1.3488,
      "theoretical_loss": 3.4479363009274406,
      "tokens_seen": 1881997312
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370909090909091,
      "loss": 1.5204,
      "theoretical_loss": 3.4479261701991497,
      "tokens_seen": 1882062848
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043707070707070707,
      "loss": 1.4517,
      "theoretical_loss": 3.4479160399223887,
      "tokens_seen": 1882128384
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004370505050505051,
      "loss": 1.5259,
      "theoretical_loss": 3.4479059100971217,
      "tokens_seen": 1882193920
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043703030303030303,
      "loss": 1.4554,
      "theoretical_loss": 3.4478957807233135,
      "tokens_seen": 1882259456
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043701010101010105,
      "loss": 1.4525,
      "theoretical_loss": 3.447885651800928,
      "tokens_seen": 1882324992
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.000436989898989899,
      "loss": 1.2467,
      "theoretical_loss": 3.4478755233299294,
      "tokens_seen": 1882390528
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043696969696969696,
      "loss": 1.3973,
      "theoretical_loss": 3.4478653953102816,
      "tokens_seen": 1882456064
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043694949494949497,
      "loss": 1.3898,
      "theoretical_loss": 3.4478552677419487,
      "tokens_seen": 1882521600
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043692929292929293,
      "loss": 1.2577,
      "theoretical_loss": 3.4478451406248953,
      "tokens_seen": 1882587136
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00043690909090909094,
      "loss": 1.4622,
      "theoretical_loss": 3.4478350139590854,
      "tokens_seen": 1882652672
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0004368888888888889,
      "loss": 1.5288,
      "theoretical_loss": 3.447824887744483,
      "tokens_seen": 1882718208
    },
    {
      "epoch": 0.13,
      "objective/train/aligned_token_count": 0.41856059432029724,
      "objective/train/avg_token_score": 0.06959132105112076,
      "objective/train/docs_used": 1064917,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4428768157958984,
      "objective/train/original_loss": 4.215625286102295,
      "objective/train/theoretical_loss": 3.447814761981053,
      "objective/train/tokens_used": 241642976,
      "objective/train/unlikelihood_loss": 0.06671393662691116,
      "theoretical_loss": 3.447814761981053,
      "tokens_seen": 1882783744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043686868686868685,
      "loss": 1.6306,
      "theoretical_loss": 3.447814761981053,
      "tokens_seen": 1882783744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043684848484848486,
      "loss": 1.304,
      "theoretical_loss": 3.447804636668759,
      "tokens_seen": 1882849280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004368282828282829,
      "loss": 1.347,
      "theoretical_loss": 3.4477945118075652,
      "tokens_seen": 1882914816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043680808080808083,
      "loss": 1.4824,
      "theoretical_loss": 3.4477843873974354,
      "tokens_seen": 1882980352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004367878787878788,
      "loss": 1.3546,
      "theoretical_loss": 3.447774263438335,
      "tokens_seen": 1883045888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043676767676767674,
      "loss": 1.3092,
      "theoretical_loss": 3.447764139930227,
      "tokens_seen": 1883111424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043674747474747476,
      "loss": 1.3361,
      "theoretical_loss": 3.4477540168730765,
      "tokens_seen": 1883176960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043672727272727277,
      "loss": 1.2515,
      "theoretical_loss": 3.4477438942668472,
      "tokens_seen": 1883242496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004367070707070707,
      "loss": 1.3784,
      "theoretical_loss": 3.447733772111503,
      "tokens_seen": 1883308032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366868686868687,
      "loss": 1.4131,
      "theoretical_loss": 3.447723650407009,
      "tokens_seen": 1883373568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043666666666666664,
      "loss": 1.3234,
      "theoretical_loss": 3.447713529153329,
      "tokens_seen": 1883439104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366464646464647,
      "loss": 1.3278,
      "theoretical_loss": 3.447703408350427,
      "tokens_seen": 1883504640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043662626262626266,
      "loss": 1.3996,
      "theoretical_loss": 3.447693287998268,
      "tokens_seen": 1883570176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004366060606060606,
      "loss": 1.3433,
      "theoretical_loss": 3.447683168096815,
      "tokens_seen": 1883635712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365858585858586,
      "loss": 1.5963,
      "theoretical_loss": 3.447673048646034,
      "tokens_seen": 1883701248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043656565656565653,
      "loss": 1.5976,
      "theoretical_loss": 3.4476629296458876,
      "tokens_seen": 1883766784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365454545454546,
      "loss": 1.4234,
      "theoretical_loss": 3.447652811096341,
      "tokens_seen": 1883832320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043652525252525255,
      "loss": 1.4693,
      "theoretical_loss": 3.447642692997358,
      "tokens_seen": 1883897856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004365050505050505,
      "loss": 1.3536,
      "theoretical_loss": 3.4476325753489028,
      "tokens_seen": 1883963392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043648484848484847,
      "loss": 1.5021,
      "theoretical_loss": 3.4476224581509403,
      "tokens_seen": 1884028928
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004364646464646465,
      "loss": 1.4861,
      "theoretical_loss": 3.447612341403434,
      "tokens_seen": 1884094464
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004364444444444445,
      "loss": 1.4886,
      "theoretical_loss": 3.4476022251063485,
      "tokens_seen": 1884160000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043642424242424244,
      "loss": 1.007,
      "theoretical_loss": 3.447592109259648,
      "tokens_seen": 1884225536
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004364040404040404,
      "loss": 1.3627,
      "theoretical_loss": 3.447581993863297,
      "tokens_seen": 1884291072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043638383838383836,
      "loss": 1.3705,
      "theoretical_loss": 3.44757187891726,
      "tokens_seen": 1884356608
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.40738025307655334,
      "objective/train/avg_token_score": 0.015582275576889515,
      "objective/train/docs_used": 1065907,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.418578028678894,
      "objective/train/original_loss": 4.257914066314697,
      "objective/train/theoretical_loss": 3.4475617644215006,
      "objective/train/tokens_used": 243281376,
      "objective/train/unlikelihood_loss": 0.0637265294790268,
      "theoretical_loss": 3.4475617644215006,
      "tokens_seen": 1884422144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043636363636363637,
      "loss": 1.5404,
      "theoretical_loss": 3.4475617644215006,
      "tokens_seen": 1884422144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004363434343434344,
      "loss": 1.2178,
      "theoretical_loss": 3.447551650375984,
      "tokens_seen": 1884487680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043632323232323234,
      "loss": 1.4998,
      "theoretical_loss": 3.4475415367806734,
      "tokens_seen": 1884553216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004363030303030303,
      "loss": 1.3251,
      "theoretical_loss": 3.4475314236355343,
      "tokens_seen": 1884618752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004362828282828283,
      "loss": 1.4469,
      "theoretical_loss": 3.44752131094053,
      "tokens_seen": 1884684288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043626262626262626,
      "loss": 1.5319,
      "theoretical_loss": 3.447511198695625,
      "tokens_seen": 1884749824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004362424242424243,
      "loss": 1.3746,
      "theoretical_loss": 3.4475010869007843,
      "tokens_seen": 1884815360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043622222222222223,
      "loss": 1.2662,
      "theoretical_loss": 3.4474909755559713,
      "tokens_seen": 1884880896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004362020202020202,
      "loss": 1.4889,
      "theoretical_loss": 3.4474808646611512,
      "tokens_seen": 1884946432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361818181818182,
      "loss": 1.2985,
      "theoretical_loss": 3.4474707542162877,
      "tokens_seen": 1885011968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043616161616161616,
      "loss": 1.4024,
      "theoretical_loss": 3.447460644221345,
      "tokens_seen": 1885077504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043614141414141417,
      "loss": 1.6368,
      "theoretical_loss": 3.447450534676288,
      "tokens_seen": 1885143040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004361212121212121,
      "loss": 1.4649,
      "theoretical_loss": 3.447440425581081,
      "tokens_seen": 1885208576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043610101010101013,
      "loss": 1.3478,
      "theoretical_loss": 3.447430316935688,
      "tokens_seen": 1885274112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004360808080808081,
      "loss": 1.1866,
      "theoretical_loss": 3.4474202087400734,
      "tokens_seen": 1885339648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043606060606060605,
      "loss": 1.4361,
      "theoretical_loss": 3.447410100994202,
      "tokens_seen": 1885405184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043604040404040406,
      "loss": 1.5223,
      "theoretical_loss": 3.4473999936980375,
      "tokens_seen": 1885470720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000436020202020202,
      "loss": 1.549,
      "theoretical_loss": 3.4473898868515445,
      "tokens_seen": 1885536256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000436,
      "loss": 1.2515,
      "theoretical_loss": 3.447379780454688,
      "tokens_seen": 1885601792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000435979797979798,
      "loss": 1.5817,
      "theoretical_loss": 3.4473696745074314,
      "tokens_seen": 1885667328
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043595959595959594,
      "loss": 1.5611,
      "theoretical_loss": 3.4473595690097394,
      "tokens_seen": 1885732864
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043593939393939395,
      "loss": 1.6062,
      "theoretical_loss": 3.4473494639615767,
      "tokens_seen": 1885798400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043591919191919196,
      "loss": 1.3081,
      "theoretical_loss": 3.447339359362907,
      "tokens_seen": 1885863936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004358989898989899,
      "loss": 1.5584,
      "theoretical_loss": 3.447329255213696,
      "tokens_seen": 1885929472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004358787878787879,
      "loss": 1.362,
      "theoretical_loss": 3.4473191515139066,
      "tokens_seen": 1885995008
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.4603494703769684,
      "objective/train/avg_token_score": 0.004963160492479801,
      "objective/train/docs_used": 1066425,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4913145303726196,
      "objective/train/original_loss": 4.012363910675049,
      "objective/train/theoretical_loss": 3.447309048263504,
      "objective/train/tokens_used": 244919776,
      "objective/train/unlikelihood_loss": 0.06790236383676529,
      "theoretical_loss": 3.447309048263504,
      "tokens_seen": 1886060544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043585858585858583,
      "loss": 1.3731,
      "theoretical_loss": 3.447309048263504,
      "tokens_seen": 1886060544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043583838383838384,
      "loss": 1.3272,
      "theoretical_loss": 3.4472989454624523,
      "tokens_seen": 1886126080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043581818181818186,
      "loss": 1.4449,
      "theoretical_loss": 3.447288843110716,
      "tokens_seen": 1886191616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357979797979798,
      "loss": 1.4432,
      "theoretical_loss": 3.4472787412082595,
      "tokens_seen": 1886257152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043577777777777777,
      "loss": 1.5336,
      "theoretical_loss": 3.4472686397550474,
      "tokens_seen": 1886322688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357575757575757,
      "loss": 1.4763,
      "theoretical_loss": 3.4472585387510435,
      "tokens_seen": 1886388224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004357373737373738,
      "loss": 1.3966,
      "theoretical_loss": 3.447248438196213,
      "tokens_seen": 1886453760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043571717171717175,
      "loss": 1.4625,
      "theoretical_loss": 3.4472383380905205,
      "tokens_seen": 1886519296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004356969696969697,
      "loss": 1.4471,
      "theoretical_loss": 3.4472282384339294,
      "tokens_seen": 1886584832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043567676767676766,
      "loss": 1.3535,
      "theoretical_loss": 3.4472181392264045,
      "tokens_seen": 1886650368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004356565656565656,
      "loss": 1.3392,
      "theoretical_loss": 3.4472080404679106,
      "tokens_seen": 1886715904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004356363636363637,
      "loss": 1.4189,
      "theoretical_loss": 3.447197942158412,
      "tokens_seen": 1886781440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043561616161616164,
      "loss": 1.5119,
      "theoretical_loss": 3.447187844297873,
      "tokens_seen": 1886846976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004355959595959596,
      "loss": 1.6033,
      "theoretical_loss": 3.447177746886258,
      "tokens_seen": 1886912512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043557575757575755,
      "loss": 1.5703,
      "theoretical_loss": 3.4471676499235318,
      "tokens_seen": 1886978048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004355555555555555,
      "loss": 1.4541,
      "theoretical_loss": 3.4471575534096583,
      "tokens_seen": 1887043584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004355353535353536,
      "loss": 1.4322,
      "theoretical_loss": 3.4471474573446024,
      "tokens_seen": 1887109120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043551515151515153,
      "loss": 1.485,
      "theoretical_loss": 3.4471373617283287,
      "tokens_seen": 1887174656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354949494949495,
      "loss": 1.4842,
      "theoretical_loss": 3.447127266560801,
      "tokens_seen": 1887240192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043547474747474745,
      "loss": 1.4015,
      "theoretical_loss": 3.4471171718419846,
      "tokens_seen": 1887305728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043545454545454546,
      "loss": 1.5789,
      "theoretical_loss": 3.447107077571843,
      "tokens_seen": 1887371264
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043543434343434347,
      "loss": 1.5174,
      "theoretical_loss": 3.447096983750342,
      "tokens_seen": 1887436800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004354141414141414,
      "loss": 1.2054,
      "theoretical_loss": 3.447086890377445,
      "tokens_seen": 1887502336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004353939393939394,
      "loss": 1.3724,
      "theoretical_loss": 3.447076797453117,
      "tokens_seen": 1887567872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043537373737373734,
      "loss": 1.6524,
      "theoretical_loss": 3.4470667049773214,
      "tokens_seen": 1887633408
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.4772116243839264,
      "objective/train/avg_token_score": 0.018777167424559593,
      "objective/train/docs_used": 1067085,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5127924680709839,
      "objective/train/original_loss": 4.050947666168213,
      "objective/train/theoretical_loss": 3.4470566129500244,
      "objective/train/tokens_used": 246558176,
      "objective/train/unlikelihood_loss": 0.05088551715016365,
      "theoretical_loss": 3.4470566129500244,
      "tokens_seen": 1887698944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043535353535353535,
      "loss": 1.3626,
      "theoretical_loss": 3.4470566129500244,
      "tokens_seen": 1887698944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043533333333333336,
      "loss": 1.3871,
      "theoretical_loss": 3.4470465213711896,
      "tokens_seen": 1887764480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004353131313131313,
      "loss": 1.5453,
      "theoretical_loss": 3.4470364302407814,
      "tokens_seen": 1887830016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352929292929293,
      "loss": 1.4407,
      "theoretical_loss": 3.447026339558765,
      "tokens_seen": 1887895552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352727272727273,
      "loss": 1.4609,
      "theoretical_loss": 3.4470162493251038,
      "tokens_seen": 1887961088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352525252525253,
      "loss": 1.4642,
      "theoretical_loss": 3.4470061595397627,
      "tokens_seen": 1888026624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043523232323232325,
      "loss": 1.2666,
      "theoretical_loss": 3.446996070202707,
      "tokens_seen": 1888092160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004352121212121212,
      "loss": 1.2846,
      "theoretical_loss": 3.446985981313901,
      "tokens_seen": 1888157696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043519191919191917,
      "loss": 1.4019,
      "theoretical_loss": 3.4469758928733083,
      "tokens_seen": 1888223232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004351717171717172,
      "loss": 1.4626,
      "theoretical_loss": 3.446965804880894,
      "tokens_seen": 1888288768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004351515151515152,
      "loss": 1.1625,
      "theoretical_loss": 3.446955717336623,
      "tokens_seen": 1888354304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043513131313131315,
      "loss": 1.5753,
      "theoretical_loss": 3.4469456302404597,
      "tokens_seen": 1888419840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004351111111111111,
      "loss": 1.4806,
      "theoretical_loss": 3.4469355435923683,
      "tokens_seen": 1888485376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004350909090909091,
      "loss": 1.5697,
      "theoretical_loss": 3.4469254573923136,
      "tokens_seen": 1888550912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043507070707070707,
      "loss": 1.4095,
      "theoretical_loss": 3.4469153716402596,
      "tokens_seen": 1888616448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004350505050505051,
      "loss": 1.5738,
      "theoretical_loss": 3.4469052863361713,
      "tokens_seen": 1888681984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043503030303030304,
      "loss": 1.4386,
      "theoretical_loss": 3.4468952014800136,
      "tokens_seen": 1888747520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000435010101010101,
      "loss": 1.4282,
      "theoretical_loss": 3.4468851170717505,
      "tokens_seen": 1888813056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434989898989899,
      "loss": 1.4533,
      "theoretical_loss": 3.446875033111347,
      "tokens_seen": 1888878592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043496969696969696,
      "loss": 1.4092,
      "theoretical_loss": 3.4468649495987678,
      "tokens_seen": 1888944128
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434949494949495,
      "loss": 1.4225,
      "theoretical_loss": 3.4468548665339767,
      "tokens_seen": 1889009664
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043492929292929293,
      "loss": 1.1841,
      "theoretical_loss": 3.446844783916939,
      "tokens_seen": 1889075200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043490909090909094,
      "loss": 1.4895,
      "theoretical_loss": 3.4468347017476186,
      "tokens_seen": 1889140736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004348888888888889,
      "loss": 1.4007,
      "theoretical_loss": 3.446824620025981,
      "tokens_seen": 1889206272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043486868686868686,
      "loss": 1.3433,
      "theoretical_loss": 3.44681453875199,
      "tokens_seen": 1889271808
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.4841764271259308,
      "objective/train/avg_token_score": 0.00916412752121687,
      "objective/train/docs_used": 1068281,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6944667100906372,
      "objective/train/original_loss": 4.594786643981934,
      "objective/train/theoretical_loss": 3.4468044579256105,
      "objective/train/tokens_used": 248196576,
      "objective/train/unlikelihood_loss": 0.028151439502835274,
      "theoretical_loss": 3.4468044579256105,
      "tokens_seen": 1889337344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043484848484848487,
      "loss": 1.437,
      "theoretical_loss": 3.4468044579256105,
      "tokens_seen": 1889337344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004348282828282828,
      "loss": 1.4151,
      "theoretical_loss": 3.446794377546807,
      "tokens_seen": 1889402880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043480808080808084,
      "loss": 1.3273,
      "theoretical_loss": 3.4467842976155447,
      "tokens_seen": 1889468416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004347878787878788,
      "loss": 1.5779,
      "theoretical_loss": 3.4467742181317873,
      "tokens_seen": 1889533952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043476767676767675,
      "loss": 1.5525,
      "theoretical_loss": 3.4467641390955,
      "tokens_seen": 1889599488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043474747474747476,
      "loss": 1.4035,
      "theoretical_loss": 3.446754060506647,
      "tokens_seen": 1889665024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043472727272727277,
      "loss": 1.4794,
      "theoretical_loss": 3.4467439823651933,
      "tokens_seen": 1889730560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043470707070707073,
      "loss": 1.2579,
      "theoretical_loss": 3.4467339046711034,
      "tokens_seen": 1889796096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004346868686868687,
      "loss": 1.3739,
      "theoretical_loss": 3.446723827424342,
      "tokens_seen": 1889861632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043466666666666664,
      "loss": 1.4137,
      "theoretical_loss": 3.4467137506248733,
      "tokens_seen": 1889927168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043464646464646465,
      "loss": 1.4186,
      "theoretical_loss": 3.446703674272663,
      "tokens_seen": 1889992704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043462626262626267,
      "loss": 1.4476,
      "theoretical_loss": 3.4466935983676743,
      "tokens_seen": 1890058240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004346060606060606,
      "loss": 1.5298,
      "theoretical_loss": 3.446683522909873,
      "tokens_seen": 1890123776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345858585858586,
      "loss": 1.4053,
      "theoretical_loss": 3.446673447899223,
      "tokens_seen": 1890189312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043456565656565654,
      "loss": 1.4032,
      "theoretical_loss": 3.4466633733356895,
      "tokens_seen": 1890254848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345454545454546,
      "loss": 1.488,
      "theoretical_loss": 3.4466532992192365,
      "tokens_seen": 1890320384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043452525252525256,
      "loss": 1.3714,
      "theoretical_loss": 3.4466432255498294,
      "tokens_seen": 1890385920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004345050505050505,
      "loss": 1.2779,
      "theoretical_loss": 3.4466331523274327,
      "tokens_seen": 1890451456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043448484848484847,
      "loss": 1.6398,
      "theoretical_loss": 3.446623079552011,
      "tokens_seen": 1890516992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043446464646464643,
      "loss": 1.319,
      "theoretical_loss": 3.4466130072235286,
      "tokens_seen": 1890582528
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344444444444445,
      "loss": 1.5699,
      "theoretical_loss": 3.4466029353419505,
      "tokens_seen": 1890648064
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043442424242424245,
      "loss": 1.4971,
      "theoretical_loss": 3.4465928639072416,
      "tokens_seen": 1890713600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004344040404040404,
      "loss": 1.5592,
      "theoretical_loss": 3.446582792919366,
      "tokens_seen": 1890779136
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043438383838383836,
      "loss": 1.4512,
      "theoretical_loss": 3.446572722378289,
      "tokens_seen": 1890844672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343636363636364,
      "loss": 1.365,
      "theoretical_loss": 3.4465626522839745,
      "tokens_seen": 1890910208
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.34188660979270935,
      "objective/train/avg_token_score": 0.036958470940589905,
      "objective/train/docs_used": 1069021,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.070196270942688,
      "objective/train/original_loss": 4.393176555633545,
      "objective/train/theoretical_loss": 3.4465525826363885,
      "objective/train/tokens_used": 249834976,
      "objective/train/unlikelihood_loss": 0.0641588643193245,
      "theoretical_loss": 3.4465525826363885,
      "tokens_seen": 1890975744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343434343434344,
      "loss": 1.2818,
      "theoretical_loss": 3.4465525826363885,
      "tokens_seen": 1890975744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043432323232323234,
      "loss": 1.4453,
      "theoretical_loss": 3.4465425134354946,
      "tokens_seen": 1891041280
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004343030303030303,
      "loss": 1.2402,
      "theoretical_loss": 3.4465324446812575,
      "tokens_seen": 1891106816
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043428282828282826,
      "loss": 1.1511,
      "theoretical_loss": 3.4465223763736423,
      "tokens_seen": 1891172352
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043426262626262627,
      "loss": 1.3436,
      "theoretical_loss": 3.446512308512614,
      "tokens_seen": 1891237888
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004342424242424243,
      "loss": 1.5911,
      "theoretical_loss": 3.446502241098137,
      "tokens_seen": 1891303424
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043422222222222224,
      "loss": 1.6752,
      "theoretical_loss": 3.4464921741301757,
      "tokens_seen": 1891368960
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004342020202020202,
      "loss": 1.6683,
      "theoretical_loss": 3.4464821076086953,
      "tokens_seen": 1891434496
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004341818181818182,
      "loss": 1.3651,
      "theoretical_loss": 3.44647204153366,
      "tokens_seen": 1891500032
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043416161616161616,
      "loss": 1.2549,
      "theoretical_loss": 3.4464619759050352,
      "tokens_seen": 1891565568
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043414141414141417,
      "loss": 1.3456,
      "theoretical_loss": 3.4464519107227853,
      "tokens_seen": 1891631104
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043412121212121213,
      "loss": 1.3359,
      "theoretical_loss": 3.4464418459868753,
      "tokens_seen": 1891696640
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004341010101010101,
      "loss": 1.3899,
      "theoretical_loss": 3.446431781697269,
      "tokens_seen": 1891762176
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004340808080808081,
      "loss": 1.3479,
      "theoretical_loss": 3.4464217178539327,
      "tokens_seen": 1891827712
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043406060606060605,
      "loss": 1.3248,
      "theoretical_loss": 3.4464116544568295,
      "tokens_seen": 1891893248
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043404040404040406,
      "loss": 1.4866,
      "theoretical_loss": 3.446401591505925,
      "tokens_seen": 1891958784
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000434020202020202,
      "loss": 1.5089,
      "theoretical_loss": 3.4463915290011844,
      "tokens_seen": 1892024320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043400000000000003,
      "loss": 1.3789,
      "theoretical_loss": 3.446381466942572,
      "tokens_seen": 1892089856
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433979797979798,
      "loss": 1.4955,
      "theoretical_loss": 3.4463714053300523,
      "tokens_seen": 1892155392
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043395959595959595,
      "loss": 1.3215,
      "theoretical_loss": 3.44636134416359,
      "tokens_seen": 1892220928
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043393939393939396,
      "loss": 1.3518,
      "theoretical_loss": 3.4463512834431507,
      "tokens_seen": 1892286464
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004339191919191919,
      "loss": 1.3355,
      "theoretical_loss": 3.446341223168698,
      "tokens_seen": 1892352000
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338989898989899,
      "loss": 1.2708,
      "theoretical_loss": 3.446331163340198,
      "tokens_seen": 1892417536
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004338787878787879,
      "loss": 1.2829,
      "theoretical_loss": 3.446321103957615,
      "tokens_seen": 1892483072
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043385858585858584,
      "loss": 1.3394,
      "theoretical_loss": 3.4463110450209133,
      "tokens_seen": 1892548608
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.34042033553123474,
      "objective/train/avg_token_score": 0.015821443870663643,
      "objective/train/docs_used": 1070317,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1134051084518433,
      "objective/train/original_loss": 4.933949947357178,
      "objective/train/theoretical_loss": 3.4463009865300585,
      "objective/train/tokens_used": 251473376,
      "objective/train/unlikelihood_loss": 0.06462860852479935,
      "theoretical_loss": 3.4463009865300585,
      "tokens_seen": 1892614144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043383838383838385,
      "loss": 1.2695,
      "theoretical_loss": 3.4463009865300585,
      "tokens_seen": 1892614144
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043381818181818186,
      "loss": 1.4255,
      "theoretical_loss": 3.4462909284850145,
      "tokens_seen": 1892679680
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004337979797979798,
      "loss": 1.2214,
      "theoretical_loss": 3.4462808708857464,
      "tokens_seen": 1892745216
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004337777777777778,
      "loss": 1.2181,
      "theoretical_loss": 3.4462708137322196,
      "tokens_seen": 1892810752
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043375757575757573,
      "loss": 1.5435,
      "theoretical_loss": 3.446260757024398,
      "tokens_seen": 1892876288
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043373737373737374,
      "loss": 1.3984,
      "theoretical_loss": 3.4462507007622474,
      "tokens_seen": 1892941824
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043371717171717175,
      "loss": 1.3259,
      "theoretical_loss": 3.4462406449457315,
      "tokens_seen": 1893007360
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004336969696969697,
      "loss": 1.3695,
      "theoretical_loss": 3.4462305895748164,
      "tokens_seen": 1893072896
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043367676767676767,
      "loss": 1.2128,
      "theoretical_loss": 3.4462205346494654,
      "tokens_seen": 1893138432
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004336565656565656,
      "loss": 1.6229,
      "theoretical_loss": 3.446210480169645,
      "tokens_seen": 1893203968
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043363636363636364,
      "loss": 1.2829,
      "theoretical_loss": 3.446200426135319,
      "tokens_seen": 1893269504
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043361616161616165,
      "loss": 1.6587,
      "theoretical_loss": 3.4461903725464524,
      "tokens_seen": 1893335040
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004335959595959596,
      "loss": 1.1699,
      "theoretical_loss": 3.4461803194030103,
      "tokens_seen": 1893400576
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043357575757575756,
      "loss": 1.2126,
      "theoretical_loss": 3.446170266704957,
      "tokens_seen": 1893466112
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043355555555555557,
      "loss": 1.2627,
      "theoretical_loss": 3.446160214452258,
      "tokens_seen": 1893531648
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004335353535353536,
      "loss": 1.3907,
      "theoretical_loss": 3.4461501626448783,
      "tokens_seen": 1893597184
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043351515151515154,
      "loss": 1.4824,
      "theoretical_loss": 3.446140111282782,
      "tokens_seen": 1893662720
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334949494949495,
      "loss": 1.3936,
      "theoretical_loss": 3.4461300603659337,
      "tokens_seen": 1893728256
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043347474747474745,
      "loss": 1.3941,
      "theoretical_loss": 3.4461200098942997,
      "tokens_seen": 1893793792
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043345454545454546,
      "loss": 1.3449,
      "theoretical_loss": 3.4461099598678437,
      "tokens_seen": 1893859328
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004334343434343435,
      "loss": 1.7087,
      "theoretical_loss": 3.446099910286531,
      "tokens_seen": 1893924864
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043341414141414143,
      "loss": 1.8037,
      "theoretical_loss": 3.446089861150326,
      "tokens_seen": 1893990400
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333939393939394,
      "loss": 1.5121,
      "theoretical_loss": 3.4460798124591947,
      "tokens_seen": 1894055936
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043337373737373735,
      "loss": 1.377,
      "theoretical_loss": 3.446069764213101,
      "tokens_seen": 1894121472
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333535353535354,
      "loss": 1.3279,
      "theoretical_loss": 3.4460597164120097,
      "tokens_seen": 1894187008
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.36241447925567627,
      "objective/train/avg_token_score": 0.02387678064405918,
      "objective/train/docs_used": 1070823,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2387202978134155,
      "objective/train/original_loss": 4.809488296508789,
      "objective/train/theoretical_loss": 3.4460496690558866,
      "objective/train/tokens_used": 253111776,
      "objective/train/unlikelihood_loss": 0.04541734606027603,
      "theoretical_loss": 3.4460496690558866,
      "tokens_seen": 1894252544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043333333333333337,
      "loss": 1.2874,
      "theoretical_loss": 3.4460496690558866,
      "tokens_seen": 1894252544
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004333131313131313,
      "loss": 1.5159,
      "theoretical_loss": 3.4460396221446956,
      "tokens_seen": 1894318080
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004332929292929293,
      "loss": 1.2886,
      "theoretical_loss": 3.4460295756784025,
      "tokens_seen": 1894383616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043327272727272724,
      "loss": 1.3834,
      "theoretical_loss": 3.446019529656972,
      "tokens_seen": 1894449152
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004332525252525253,
      "loss": 1.3738,
      "theoretical_loss": 3.446009484080368,
      "tokens_seen": 1894514688
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043323232323232326,
      "loss": 1.4346,
      "theoretical_loss": 3.445999438948557,
      "tokens_seen": 1894580224
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004332121212121212,
      "loss": 1.415,
      "theoretical_loss": 3.4459893942615025,
      "tokens_seen": 1894645760
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004331919191919192,
      "loss": 1.3809,
      "theoretical_loss": 3.4459793500191704,
      "tokens_seen": 1894711296
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004331717171717172,
      "loss": 1.5446,
      "theoretical_loss": 3.4459693062215253,
      "tokens_seen": 1894776832
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004331515151515152,
      "loss": 1.5414,
      "theoretical_loss": 3.445959262868532,
      "tokens_seen": 1894842368
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043313131313131315,
      "loss": 1.1756,
      "theoretical_loss": 3.4459492199601556,
      "tokens_seen": 1894907904
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004331111111111111,
      "loss": 1.5137,
      "theoretical_loss": 3.445939177496361,
      "tokens_seen": 1894973440
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043309090909090907,
      "loss": 1.2745,
      "theoretical_loss": 3.445929135477113,
      "tokens_seen": 1895038976
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004330707070707071,
      "loss": 1.6113,
      "theoretical_loss": 3.445919093902377,
      "tokens_seen": 1895104512
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004330505050505051,
      "loss": 1.4817,
      "theoretical_loss": 3.4459090527721177,
      "tokens_seen": 1895170048
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043303030303030305,
      "loss": 1.6125,
      "theoretical_loss": 3.4458990120863,
      "tokens_seen": 1895235584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000433010101010101,
      "loss": 1.345,
      "theoretical_loss": 3.445888971844888,
      "tokens_seen": 1895301120
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000432989898989899,
      "loss": 1.4279,
      "theoretical_loss": 3.4458789320478487,
      "tokens_seen": 1895366656
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043296969696969697,
      "loss": 1.4733,
      "theoretical_loss": 3.4458688926951453,
      "tokens_seen": 1895432192
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000432949494949495,
      "loss": 1.2651,
      "theoretical_loss": 3.4458588537867434,
      "tokens_seen": 1895497728
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043292929292929294,
      "loss": 1.2785,
      "theoretical_loss": 3.4458488153226083,
      "tokens_seen": 1895563264
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004329090909090909,
      "loss": 1.3361,
      "theoretical_loss": 3.445838777302704,
      "tokens_seen": 1895628800
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328888888888889,
      "loss": 1.4636,
      "theoretical_loss": 3.4458287397269967,
      "tokens_seen": 1895694336
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043286868686868686,
      "loss": 1.257,
      "theoretical_loss": 3.4458187025954503,
      "tokens_seen": 1895759872
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004328484848484849,
      "loss": 1.486,
      "theoretical_loss": 3.4458086659080305,
      "tokens_seen": 1895825408
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.3908846378326416,
      "objective/train/avg_token_score": 0.010636522434651852,
      "objective/train/docs_used": 1071554,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2562838792800903,
      "objective/train/original_loss": 4.368358612060547,
      "objective/train/theoretical_loss": 3.4457986296647025,
      "objective/train/tokens_used": 254750176,
      "objective/train/unlikelihood_loss": 0.06785128265619278,
      "theoretical_loss": 3.4457986296647025,
      "tokens_seen": 1895890944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043282828282828283,
      "loss": 1.4474,
      "theoretical_loss": 3.4457986296647025,
      "tokens_seen": 1895890944
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043280808080808084,
      "loss": 1.3487,
      "theoretical_loss": 3.4457885938654305,
      "tokens_seen": 1895956480
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004327878787878788,
      "loss": 1.4548,
      "theoretical_loss": 3.44577855851018,
      "tokens_seen": 1896022016
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043276767676767676,
      "loss": 1.3746,
      "theoretical_loss": 3.4457685235989155,
      "tokens_seen": 1896087552
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043274747474747477,
      "loss": 1.4116,
      "theoretical_loss": 3.445758489131603,
      "tokens_seen": 1896153088
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004327272727272727,
      "loss": 1.5319,
      "theoretical_loss": 3.4457484551082063,
      "tokens_seen": 1896218624
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043270707070707073,
      "loss": 1.5303,
      "theoretical_loss": 3.4457384215286915,
      "tokens_seen": 1896284160
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004326868686868687,
      "loss": 1.2872,
      "theoretical_loss": 3.4457283883930234,
      "tokens_seen": 1896349696
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043266666666666665,
      "loss": 1.2291,
      "theoretical_loss": 3.4457183557011666,
      "tokens_seen": 1896415232
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043264646464646466,
      "loss": 1.1889,
      "theoretical_loss": 3.4457083234530863,
      "tokens_seen": 1896480768
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043262626262626267,
      "loss": 1.3665,
      "theoretical_loss": 3.445698291648747,
      "tokens_seen": 1896546304
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043260606060606063,
      "loss": 1.3058,
      "theoretical_loss": 3.445688260288115,
      "tokens_seen": 1896611840
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004325858585858586,
      "loss": 1.4865,
      "theoretical_loss": 3.445678229371154,
      "tokens_seen": 1896677376
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043256565656565654,
      "loss": 1.4227,
      "theoretical_loss": 3.4456681988978306,
      "tokens_seen": 1896742912
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043254545454545455,
      "loss": 1.0513,
      "theoretical_loss": 3.4456581688681087,
      "tokens_seen": 1896808448
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043252525252525256,
      "loss": 1.4629,
      "theoretical_loss": 3.4456481392819533,
      "tokens_seen": 1896873984
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004325050505050505,
      "loss": 1.2892,
      "theoretical_loss": 3.44563811013933,
      "tokens_seen": 1896939520
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324848484848485,
      "loss": 1.5847,
      "theoretical_loss": 3.445628081440203,
      "tokens_seen": 1897005056
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043246464646464643,
      "loss": 1.2676,
      "theoretical_loss": 3.445618053184539,
      "tokens_seen": 1897070592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324444444444445,
      "loss": 1.4628,
      "theoretical_loss": 3.4456080253723016,
      "tokens_seen": 1897136128
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043242424242424246,
      "loss": 1.5347,
      "theoretical_loss": 3.445597998003456,
      "tokens_seen": 1897201664
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004324040404040404,
      "loss": 1.38,
      "theoretical_loss": 3.445587971077968,
      "tokens_seen": 1897267200
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043238383838383837,
      "loss": 1.2716,
      "theoretical_loss": 3.4455779445958026,
      "tokens_seen": 1897332736
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323636363636363,
      "loss": 1.4612,
      "theoretical_loss": 3.4455679185569243,
      "tokens_seen": 1897398272
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323434343434344,
      "loss": 1.404,
      "theoretical_loss": 3.4455578929612987,
      "tokens_seen": 1897463808
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.2730327546596527,
      "objective/train/avg_token_score": 0.044983237981796265,
      "objective/train/docs_used": 1072893,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9510472416877747,
      "objective/train/original_loss": 5.534091949462891,
      "objective/train/theoretical_loss": 3.44554786780889,
      "objective/train/tokens_used": 256388576,
      "objective/train/unlikelihood_loss": 0.05126022920012474,
      "theoretical_loss": 3.44554786780889,
      "tokens_seen": 1897529344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043232323232323235,
      "loss": 1.0224,
      "theoretical_loss": 3.44554786780889,
      "tokens_seen": 1897529344
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004323030303030303,
      "loss": 1.3145,
      "theoretical_loss": 3.4455378430996646,
      "tokens_seen": 1897594880
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043228282828282826,
      "loss": 1.354,
      "theoretical_loss": 3.445527818833587,
      "tokens_seen": 1897660416
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004322626262626262,
      "loss": 1.3433,
      "theoretical_loss": 3.4455177950106215,
      "tokens_seen": 1897725952
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004322424242424243,
      "loss": 1.1633,
      "theoretical_loss": 3.445507771630735,
      "tokens_seen": 1897791488
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043222222222222224,
      "loss": 1.7054,
      "theoretical_loss": 3.4454977486938914,
      "tokens_seen": 1897857024
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004322020202020202,
      "loss": 1.3594,
      "theoretical_loss": 3.4454877262000556,
      "tokens_seen": 1897922560
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043218181818181816,
      "loss": 1.3057,
      "theoretical_loss": 3.4454777041491935,
      "tokens_seen": 1897988096
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043216161616161617,
      "loss": 1.1239,
      "theoretical_loss": 3.44546768254127,
      "tokens_seen": 1898053632
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321414141414142,
      "loss": 1.3014,
      "theoretical_loss": 3.44545766137625,
      "tokens_seen": 1898119168
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043212121212121213,
      "loss": 1.5642,
      "theoretical_loss": 3.4454476406540984,
      "tokens_seen": 1898184704
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004321010101010101,
      "loss": 1.1932,
      "theoretical_loss": 3.445437620374781,
      "tokens_seen": 1898250240
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043208080808080805,
      "loss": 1.3947,
      "theoretical_loss": 3.445427600538263,
      "tokens_seen": 1898315776
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043206060606060606,
      "loss": 1.2837,
      "theoretical_loss": 3.4454175811445085,
      "tokens_seen": 1898381312
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043204040404040407,
      "loss": 1.4581,
      "theoretical_loss": 3.445407562193484,
      "tokens_seen": 1898446848
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000432020202020202,
      "loss": 1.3282,
      "theoretical_loss": 3.4453975436851536,
      "tokens_seen": 1898512384
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000432,
      "loss": 1.2793,
      "theoretical_loss": 3.445387525619483,
      "tokens_seen": 1898577920
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.000431979797979798,
      "loss": 1.1152,
      "theoretical_loss": 3.445377507996437,
      "tokens_seen": 1898643456
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043195959595959595,
      "loss": 1.4725,
      "theoretical_loss": 3.445367490815981,
      "tokens_seen": 1898708992
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043193939393939396,
      "loss": 1.2346,
      "theoretical_loss": 3.4453574740780804,
      "tokens_seen": 1898774528
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004319191919191919,
      "loss": 1.4218,
      "theoretical_loss": 3.4453474577826997,
      "tokens_seen": 1898840064
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004318989898989899,
      "loss": 1.2056,
      "theoretical_loss": 3.445337441929805,
      "tokens_seen": 1898905600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004318787878787879,
      "loss": 1.2807,
      "theoretical_loss": 3.445327426519361,
      "tokens_seen": 1898971136
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004318585858585859,
      "loss": 1.4656,
      "theoretical_loss": 3.4453174115513323,
      "tokens_seen": 1899036672
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00043183838383838386,
      "loss": 1.4633,
      "theoretical_loss": 3.445307397025685,
      "tokens_seen": 1899102208
    },
    {
      "epoch": 0.14,
      "objective/train/aligned_token_count": 0.42344820499420166,
      "objective/train/avg_token_score": 0.01560823805630207,
      "objective/train/docs_used": 1073632,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4783105850219727,
      "objective/train/original_loss": 4.294405937194824,
      "objective/train/theoretical_loss": 3.445297382942384,
      "objective/train/tokens_used": 258026976,
      "objective/train/unlikelihood_loss": 0.07553756982088089,
      "theoretical_loss": 3.445297382942384,
      "tokens_seen": 1899167744
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0004318181818181818,
      "loss": 1.2705,
      "theoretical_loss": 3.445297382942384,
      "tokens_seen": 1899167744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317979797979798,
      "loss": 1.622,
      "theoretical_loss": 3.4452873693013943,
      "tokens_seen": 1899233280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317777777777778,
      "loss": 1.4875,
      "theoretical_loss": 3.4452773561026815,
      "tokens_seen": 1899298816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317575757575758,
      "loss": 1.5592,
      "theoretical_loss": 3.4452673433462104,
      "tokens_seen": 1899364352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043173737373737375,
      "loss": 1.3319,
      "theoretical_loss": 3.4452573310319465,
      "tokens_seen": 1899429888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004317171717171717,
      "loss": 1.4507,
      "theoretical_loss": 3.4452473191598547,
      "tokens_seen": 1899495424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316969696969697,
      "loss": 1.4265,
      "theoretical_loss": 3.4452373077299003,
      "tokens_seen": 1899560960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316767676767677,
      "loss": 1.4393,
      "theoretical_loss": 3.4452272967420488,
      "tokens_seen": 1899626496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004316565656565657,
      "loss": 1.5585,
      "theoretical_loss": 3.4452172861962653,
      "tokens_seen": 1899692032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043163636363636364,
      "loss": 1.3088,
      "theoretical_loss": 3.445207276092515,
      "tokens_seen": 1899757568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043161616161616165,
      "loss": 1.4769,
      "theoretical_loss": 3.445197266430763,
      "tokens_seen": 1899823104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315959595959596,
      "loss": 1.2856,
      "theoretical_loss": 3.4451872572109745,
      "tokens_seen": 1899888640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043157575757575757,
      "loss": 1.4165,
      "theoretical_loss": 3.445177248433115,
      "tokens_seen": 1899954176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004315555555555556,
      "loss": 1.5391,
      "theoretical_loss": 3.44516724009715,
      "tokens_seen": 1900019712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043153535353535353,
      "loss": 1.4838,
      "theoretical_loss": 3.445157232203044,
      "tokens_seen": 1900085248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043151515151515154,
      "loss": 1.2202,
      "theoretical_loss": 3.4451472247507624,
      "tokens_seen": 1900150784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314949494949495,
      "loss": 1.3836,
      "theoretical_loss": 3.445137217740271,
      "tokens_seen": 1900216320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043147474747474746,
      "loss": 1.0373,
      "theoretical_loss": 3.4451272111715348,
      "tokens_seen": 1900281856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043145454545454547,
      "loss": 1.2555,
      "theoretical_loss": 3.445117205044519,
      "tokens_seen": 1900347392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004314343434343435,
      "loss": 1.3939,
      "theoretical_loss": 3.445107199359189,
      "tokens_seen": 1900412928
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043141414141414144,
      "loss": 1.4978,
      "theoretical_loss": 3.44509719411551,
      "tokens_seen": 1900478464
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313939393939394,
      "loss": 1.5079,
      "theoretical_loss": 3.4450871893134467,
      "tokens_seen": 1900544000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043137373737373735,
      "loss": 1.2719,
      "theoretical_loss": 3.4450771849529653,
      "tokens_seen": 1900609536
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043135353535353536,
      "loss": 1.3437,
      "theoretical_loss": 3.4450671810340308,
      "tokens_seen": 1900675072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004313333333333334,
      "loss": 1.3644,
      "theoretical_loss": 3.4450571775566083,
      "tokens_seen": 1900740608
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.41379520297050476,
      "objective/train/avg_token_score": 0.014362385496497154,
      "objective/train/docs_used": 1075027,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3654704093933105,
      "objective/train/original_loss": 4.367673397064209,
      "objective/train/theoretical_loss": 3.445047174520663,
      "objective/train/tokens_used": 259665376,
      "objective/train/unlikelihood_loss": 0.07200692594051361,
      "theoretical_loss": 3.445047174520663,
      "tokens_seen": 1900806144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043131313131313133,
      "loss": 1.5024,
      "theoretical_loss": 3.445047174520663,
      "tokens_seen": 1900806144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312929292929293,
      "loss": 1.1957,
      "theoretical_loss": 3.4450371719261605,
      "tokens_seen": 1900871680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043127272727272724,
      "loss": 1.3585,
      "theoretical_loss": 3.4450271697730663,
      "tokens_seen": 1900937216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312525252525253,
      "loss": 1.4242,
      "theoretical_loss": 3.445017168061345,
      "tokens_seen": 1901002752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043123232323232327,
      "loss": 1.5416,
      "theoretical_loss": 3.4450071667909623,
      "tokens_seen": 1901068288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004312121212121212,
      "loss": 1.5098,
      "theoretical_loss": 3.4449971659618837,
      "tokens_seen": 1901133824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311919191919192,
      "loss": 1.3824,
      "theoretical_loss": 3.4449871655740743,
      "tokens_seen": 1901199360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043117171717171714,
      "loss": 1.3807,
      "theoretical_loss": 3.444977165627499,
      "tokens_seen": 1901264896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311515151515152,
      "loss": 1.4715,
      "theoretical_loss": 3.444967166122124,
      "tokens_seen": 1901330432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043113131313131316,
      "loss": 1.228,
      "theoretical_loss": 3.444957167057914,
      "tokens_seen": 1901395968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004311111111111111,
      "loss": 1.4939,
      "theoretical_loss": 3.444947168434835,
      "tokens_seen": 1901461504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043109090909090907,
      "loss": 1.4963,
      "theoretical_loss": 3.4449371702528513,
      "tokens_seen": 1901527040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310707070707071,
      "loss": 1.336,
      "theoretical_loss": 3.444927172511929,
      "tokens_seen": 1901592576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004310505050505051,
      "loss": 1.2116,
      "theoretical_loss": 3.4449171752120336,
      "tokens_seen": 1901658112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043103030303030305,
      "loss": 1.2582,
      "theoretical_loss": 3.4449071783531293,
      "tokens_seen": 1901723648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000431010101010101,
      "loss": 1.5321,
      "theoretical_loss": 3.4448971819351826,
      "tokens_seen": 1901789184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043098989898989897,
      "loss": 1.5561,
      "theoretical_loss": 3.4448871859581587,
      "tokens_seen": 1901854720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000430969696969697,
      "loss": 1.478,
      "theoretical_loss": 3.444877190422023,
      "tokens_seen": 1901920256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000430949494949495,
      "loss": 1.5553,
      "theoretical_loss": 3.44486719532674,
      "tokens_seen": 1901985792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043092929292929294,
      "loss": 1.5061,
      "theoretical_loss": 3.444857200672276,
      "tokens_seen": 1902051328
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004309090909090909,
      "loss": 1.4788,
      "theoretical_loss": 3.4448472064585958,
      "tokens_seen": 1902116864
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308888888888889,
      "loss": 1.5469,
      "theoretical_loss": 3.4448372126856652,
      "tokens_seen": 1902182400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043086868686868687,
      "loss": 1.4974,
      "theoretical_loss": 3.4448272193534493,
      "tokens_seen": 1902247936
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308484848484849,
      "loss": 1.3048,
      "theoretical_loss": 3.4448172264619137,
      "tokens_seen": 1902313472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043082828282828284,
      "loss": 1.5208,
      "theoretical_loss": 3.444807234011024,
      "tokens_seen": 1902379008
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.4122678339481354,
      "objective/train/avg_token_score": 0.011399907991290092,
      "objective/train/docs_used": 1075669,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.282902479171753,
      "objective/train/original_loss": 4.224909782409668,
      "objective/train/theoretical_loss": 3.4447972420007447,
      "objective/train/tokens_used": 261303776,
      "objective/train/unlikelihood_loss": 0.0683131068944931,
      "theoretical_loss": 3.4447972420007447,
      "tokens_seen": 1902444544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004308080808080808,
      "loss": 1.3168,
      "theoretical_loss": 3.4447972420007447,
      "tokens_seen": 1902444544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004307878787878788,
      "loss": 1.4082,
      "theoretical_loss": 3.444787250431042,
      "tokens_seen": 1902510080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043076767676767676,
      "loss": 1.5043,
      "theoretical_loss": 3.444777259301881,
      "tokens_seen": 1902575616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043074747474747477,
      "loss": 1.6162,
      "theoretical_loss": 3.4447672686132274,
      "tokens_seen": 1902641152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043072727272727273,
      "loss": 1.3416,
      "theoretical_loss": 3.4447572783650457,
      "tokens_seen": 1902706688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043070707070707074,
      "loss": 1.2339,
      "theoretical_loss": 3.444747288557303,
      "tokens_seen": 1902772224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306868686868687,
      "loss": 1.5255,
      "theoretical_loss": 3.444737299189963,
      "tokens_seen": 1902837760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043066666666666665,
      "loss": 1.3823,
      "theoretical_loss": 3.444727310262992,
      "tokens_seen": 1902903296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043064646464646467,
      "loss": 1.3543,
      "theoretical_loss": 3.444717321776355,
      "tokens_seen": 1902968832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004306262626262626,
      "loss": 1.4874,
      "theoretical_loss": 3.444707333730018,
      "tokens_seen": 1903034368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043060606060606063,
      "loss": 1.5816,
      "theoretical_loss": 3.444697346123946,
      "tokens_seen": 1903099904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305858585858586,
      "loss": 1.3016,
      "theoretical_loss": 3.444687358958104,
      "tokens_seen": 1903165440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043056565656565655,
      "loss": 1.3505,
      "theoretical_loss": 3.4446773722324586,
      "tokens_seen": 1903230976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043054545454545456,
      "loss": 1.4311,
      "theoretical_loss": 3.4446673859469747,
      "tokens_seen": 1903296512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043052525252525257,
      "loss": 1.5355,
      "theoretical_loss": 3.444657400101617,
      "tokens_seen": 1903362048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004305050505050505,
      "loss": 1.3983,
      "theoretical_loss": 3.444647414696352,
      "tokens_seen": 1903427584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004304848484848485,
      "loss": 1.2973,
      "theoretical_loss": 3.444637429731145,
      "tokens_seen": 1903493120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043046464646464644,
      "loss": 1.5914,
      "theoretical_loss": 3.4446274452059606,
      "tokens_seen": 1903558656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043044444444444445,
      "loss": 1.3239,
      "theoretical_loss": 3.444617461120765,
      "tokens_seen": 1903624192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043042424242424246,
      "loss": 1.2012,
      "theoretical_loss": 3.444607477475524,
      "tokens_seen": 1903689728
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004304040404040404,
      "loss": 1.4485,
      "theoretical_loss": 3.4445974942702025,
      "tokens_seen": 1903755264
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004303838383838384,
      "loss": 1.4015,
      "theoretical_loss": 3.4445875115047655,
      "tokens_seen": 1903820800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043036363636363633,
      "loss": 1.5879,
      "theoretical_loss": 3.4445775291791794,
      "tokens_seen": 1903886336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043034343434343434,
      "loss": 1.4073,
      "theoretical_loss": 3.444567547293409,
      "tokens_seen": 1903951872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043032323232323235,
      "loss": 1.3852,
      "theoretical_loss": 3.444557565847421,
      "tokens_seen": 1904017408
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.3672409653663635,
      "objective/train/avg_token_score": 0.005817563273012638,
      "objective/train/docs_used": 1076839,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2039426565170288,
      "objective/train/original_loss": 3.8227851390838623,
      "objective/train/theoretical_loss": 3.444547584841179,
      "objective/train/tokens_used": 262942176,
      "objective/train/unlikelihood_loss": 0.13508202135562897,
      "theoretical_loss": 3.444547584841179,
      "tokens_seen": 1904082944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004303030303030303,
      "loss": 1.2922,
      "theoretical_loss": 3.444547584841179,
      "tokens_seen": 1904082944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043028282828282827,
      "loss": 1.2261,
      "theoretical_loss": 3.4445376042746503,
      "tokens_seen": 1904148480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004302626262626262,
      "loss": 1.3418,
      "theoretical_loss": 3.444527624147799,
      "tokens_seen": 1904214016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004302424242424243,
      "loss": 1.402,
      "theoretical_loss": 3.444517644460591,
      "tokens_seen": 1904279552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043022222222222225,
      "loss": 1.406,
      "theoretical_loss": 3.4445076652129925,
      "tokens_seen": 1904345088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004302020202020202,
      "loss": 1.3695,
      "theoretical_loss": 3.4444976864049686,
      "tokens_seen": 1904410624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043018181818181816,
      "loss": 1.5536,
      "theoretical_loss": 3.444487708036484,
      "tokens_seen": 1904476160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043016161616161617,
      "loss": 1.4343,
      "theoretical_loss": 3.444477730107505,
      "tokens_seen": 1904541696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004301414141414142,
      "loss": 1.1865,
      "theoretical_loss": 3.4444677526179976,
      "tokens_seen": 1904607232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043012121212121214,
      "loss": 1.6126,
      "theoretical_loss": 3.444457775567927,
      "tokens_seen": 1904672768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004301010101010101,
      "loss": 1.6826,
      "theoretical_loss": 3.444447798957258,
      "tokens_seen": 1904738304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043008080808080805,
      "loss": 1.335,
      "theoretical_loss": 3.4444378227859564,
      "tokens_seen": 1904803840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004300606060606061,
      "loss": 1.4046,
      "theoretical_loss": 3.4444278470539884,
      "tokens_seen": 1904869376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004300404040404041,
      "loss": 1.5168,
      "theoretical_loss": 3.4444178717613188,
      "tokens_seen": 1904934912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043002020202020203,
      "loss": 1.4559,
      "theoretical_loss": 3.4444078969079133,
      "tokens_seen": 1905000448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00043,
      "loss": 1.3272,
      "theoretical_loss": 3.4443979224937378,
      "tokens_seen": 1905065984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042997979797979795,
      "loss": 1.3365,
      "theoretical_loss": 3.4443879485187576,
      "tokens_seen": 1905131520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000429959595959596,
      "loss": 1.4219,
      "theoretical_loss": 3.4443779749829386,
      "tokens_seen": 1905197056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042993939393939397,
      "loss": 1.5434,
      "theoretical_loss": 3.4443680018862457,
      "tokens_seen": 1905262592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004299191919191919,
      "loss": 1.5114,
      "theoretical_loss": 3.4443580292286446,
      "tokens_seen": 1905328128
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298989898989899,
      "loss": 1.5505,
      "theoretical_loss": 3.4443480570101017,
      "tokens_seen": 1905393664
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298787878787879,
      "loss": 1.3701,
      "theoretical_loss": 3.444338085230582,
      "tokens_seen": 1905459200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298585858585859,
      "loss": 1.3049,
      "theoretical_loss": 3.44432811389005,
      "tokens_seen": 1905524736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042983838383838386,
      "loss": 1.4485,
      "theoretical_loss": 3.444318142988473,
      "tokens_seen": 1905590272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004298181818181818,
      "loss": 1.4768,
      "theoretical_loss": 3.444308172525816,
      "tokens_seen": 1905655808
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.35905423760414124,
      "objective/train/avg_token_score": 0.026558930054306984,
      "objective/train/docs_used": 1077487,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1562389135360718,
      "objective/train/original_loss": 4.9117536544799805,
      "objective/train/theoretical_loss": 3.444298202502044,
      "objective/train/tokens_used": 264580576,
      "objective/train/unlikelihood_loss": 0.038373254239559174,
      "theoretical_loss": 3.444298202502044,
      "tokens_seen": 1905721344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297979797979798,
      "loss": 1.3541,
      "theoretical_loss": 3.444298202502044,
      "tokens_seen": 1905721344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297777777777778,
      "loss": 1.3404,
      "theoretical_loss": 3.444288232917123,
      "tokens_seen": 1905786880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297575757575758,
      "loss": 1.3965,
      "theoretical_loss": 3.444278263771019,
      "tokens_seen": 1905852416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042973737373737375,
      "loss": 1.3603,
      "theoretical_loss": 3.444268295063697,
      "tokens_seen": 1905917952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004297171717171717,
      "loss": 1.5391,
      "theoretical_loss": 3.444258326795123,
      "tokens_seen": 1905983488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296969696969697,
      "loss": 1.5438,
      "theoretical_loss": 3.444248358965263,
      "tokens_seen": 1906049024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296767676767677,
      "loss": 1.3518,
      "theoretical_loss": 3.444238391574081,
      "tokens_seen": 1906114560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296565656565657,
      "loss": 1.1733,
      "theoretical_loss": 3.4442284246215444,
      "tokens_seen": 1906180096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042963636363636365,
      "loss": 1.4349,
      "theoretical_loss": 3.444218458107618,
      "tokens_seen": 1906245632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004296161616161616,
      "loss": 1.436,
      "theoretical_loss": 3.444208492032267,
      "tokens_seen": 1906311168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295959595959596,
      "loss": 1.3842,
      "theoretical_loss": 3.444198526395458,
      "tokens_seen": 1906376704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042957575757575757,
      "loss": 1.3577,
      "theoretical_loss": 3.444188561197156,
      "tokens_seen": 1906442240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004295555555555556,
      "loss": 1.4073,
      "theoretical_loss": 3.4441785964373266,
      "tokens_seen": 1906507776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042953535353535354,
      "loss": 1.3905,
      "theoretical_loss": 3.444168632115936,
      "tokens_seen": 1906573312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042951515151515155,
      "loss": 1.3146,
      "theoretical_loss": 3.444158668232949,
      "tokens_seen": 1906638848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004294949494949495,
      "loss": 1.4867,
      "theoretical_loss": 3.444148704788332,
      "tokens_seen": 1906704384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042947474747474746,
      "loss": 1.5472,
      "theoretical_loss": 3.44413874178205,
      "tokens_seen": 1906769920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004294545454545455,
      "loss": 1.4212,
      "theoretical_loss": 3.4441287792140693,
      "tokens_seen": 1906835456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042943434343434343,
      "loss": 1.2909,
      "theoretical_loss": 3.4441188170843553,
      "tokens_seen": 1906900992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042941414141414144,
      "loss": 1.4079,
      "theoretical_loss": 3.444108855392874,
      "tokens_seen": 1906966528
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004293939393939394,
      "loss": 1.4177,
      "theoretical_loss": 3.44409889413959,
      "tokens_seen": 1907032064
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042937373737373736,
      "loss": 1.4536,
      "theoretical_loss": 3.4440889333244695,
      "tokens_seen": 1907097600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042935353535353537,
      "loss": 1.4919,
      "theoretical_loss": 3.444078972947479,
      "tokens_seen": 1907163136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004293333333333334,
      "loss": 1.3301,
      "theoretical_loss": 3.4440690130085825,
      "tokens_seen": 1907228672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042931313131313134,
      "loss": 1.3173,
      "theoretical_loss": 3.4440590535077478,
      "tokens_seen": 1907294208
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.3209921717643738,
      "objective/train/avg_token_score": 0.06907281279563904,
      "objective/train/docs_used": 1078017,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0279090404510498,
      "objective/train/original_loss": 4.60770320892334,
      "objective/train/theoretical_loss": 3.4440490944449387,
      "objective/train/tokens_used": 266218976,
      "objective/train/unlikelihood_loss": 0.046941839158535004,
      "theoretical_loss": 3.4440490944449387,
      "tokens_seen": 1907359744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004292929292929293,
      "loss": 1.2658,
      "theoretical_loss": 3.4440490944449387,
      "tokens_seen": 1907359744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042927272727272725,
      "loss": 1.43,
      "theoretical_loss": 3.444039135820122,
      "tokens_seen": 1907425280
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042925252525252526,
      "loss": 1.4694,
      "theoretical_loss": 3.444029177633263,
      "tokens_seen": 1907490816
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042923232323232327,
      "loss": 1.2123,
      "theoretical_loss": 3.4440192198843276,
      "tokens_seen": 1907556352
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042921212121212123,
      "loss": 1.3888,
      "theoretical_loss": 3.4440092625732808,
      "tokens_seen": 1907621888
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291919191919192,
      "loss": 1.3047,
      "theoretical_loss": 3.443999305700089,
      "tokens_seen": 1907687424
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042917171717171714,
      "loss": 1.2473,
      "theoretical_loss": 3.443989349264718,
      "tokens_seen": 1907752960
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291515151515152,
      "loss": 1.4243,
      "theoretical_loss": 3.4439793932671323,
      "tokens_seen": 1907818496
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042913131313131316,
      "loss": 1.4856,
      "theoretical_loss": 3.4439694377073,
      "tokens_seen": 1907884032
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004291111111111111,
      "loss": 1.5347,
      "theoretical_loss": 3.4439594825851843,
      "tokens_seen": 1907949568
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004290909090909091,
      "loss": 1.5656,
      "theoretical_loss": 3.4439495279007524,
      "tokens_seen": 1908015104
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042907070707070703,
      "loss": 1.3704,
      "theoretical_loss": 3.443939573653969,
      "tokens_seen": 1908080640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004290505050505051,
      "loss": 1.2791,
      "theoretical_loss": 3.443929619844801,
      "tokens_seen": 1908146176
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042903030303030306,
      "loss": 1.1324,
      "theoretical_loss": 3.4439196664732137,
      "tokens_seen": 1908211712
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000429010101010101,
      "loss": 1.5249,
      "theoretical_loss": 3.4439097135391723,
      "tokens_seen": 1908277248
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042898989898989897,
      "loss": 1.5048,
      "theoretical_loss": 3.443899761042643,
      "tokens_seen": 1908342784
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042896969696969693,
      "loss": 1.5334,
      "theoretical_loss": 3.4438898089835916,
      "tokens_seen": 1908408320
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000428949494949495,
      "loss": 1.6011,
      "theoretical_loss": 3.443879857361984,
      "tokens_seen": 1908473856
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042892929292929295,
      "loss": 1.4087,
      "theoretical_loss": 3.4438699061777855,
      "tokens_seen": 1908539392
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004289090909090909,
      "loss": 1.3535,
      "theoretical_loss": 3.4438599554309617,
      "tokens_seen": 1908604928
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042888888888888886,
      "loss": 1.4036,
      "theoretical_loss": 3.4438500051214787,
      "tokens_seen": 1908670464
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288686868686869,
      "loss": 1.2635,
      "theoretical_loss": 3.4438400552493027,
      "tokens_seen": 1908736000
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288484848484849,
      "loss": 1.3974,
      "theoretical_loss": 3.4438301058143987,
      "tokens_seen": 1908801536
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042882828282828284,
      "loss": 1.5568,
      "theoretical_loss": 3.443820156816733,
      "tokens_seen": 1908867072
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004288080808080808,
      "loss": 1.4838,
      "theoretical_loss": 3.443810208256271,
      "tokens_seen": 1908932608
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.37292277812957764,
      "objective/train/avg_token_score": 0.02143978700041771,
      "objective/train/docs_used": 1079511,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.331836462020874,
      "objective/train/original_loss": 4.501890659332275,
      "objective/train/theoretical_loss": 3.4438002601329787,
      "objective/train/tokens_used": 267857376,
      "objective/train/unlikelihood_loss": 0.05701581388711929,
      "theoretical_loss": 3.4438002601329787,
      "tokens_seen": 1908998144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042878787878787876,
      "loss": 1.5142,
      "theoretical_loss": 3.4438002601329787,
      "tokens_seen": 1908998144
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042876767676767677,
      "loss": 1.4552,
      "theoretical_loss": 3.443790312446822,
      "tokens_seen": 1909063680
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004287474747474748,
      "loss": 1.1977,
      "theoretical_loss": 3.4437803651977665,
      "tokens_seen": 1909129216
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042872727272727274,
      "loss": 1.5264,
      "theoretical_loss": 3.4437704183857782,
      "tokens_seen": 1909194752
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004287070707070707,
      "loss": 1.3258,
      "theoretical_loss": 3.443760472010822,
      "tokens_seen": 1909260288
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004286868686868687,
      "loss": 1.3133,
      "theoretical_loss": 3.443750526072865,
      "tokens_seen": 1909325824
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042866666666666666,
      "loss": 1.2519,
      "theoretical_loss": 3.4437405805718724,
      "tokens_seen": 1909391360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042864646464646467,
      "loss": 1.282,
      "theoretical_loss": 3.4437306355078094,
      "tokens_seen": 1909456896
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042862626262626263,
      "loss": 1.3893,
      "theoretical_loss": 3.4437206908806433,
      "tokens_seen": 1909522432
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004286060606060606,
      "loss": 1.2844,
      "theoretical_loss": 3.4437107466903383,
      "tokens_seen": 1909587968
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004285858585858586,
      "loss": 1.5531,
      "theoretical_loss": 3.443700802936861,
      "tokens_seen": 1909653504
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042856565656565655,
      "loss": 1.5443,
      "theoretical_loss": 3.443690859620178,
      "tokens_seen": 1909719040
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042854545454545456,
      "loss": 1.4003,
      "theoretical_loss": 3.443680916740253,
      "tokens_seen": 1909784576
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004285252525252525,
      "loss": 1.4069,
      "theoretical_loss": 3.443670974297054,
      "tokens_seen": 1909850112
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042850505050505053,
      "loss": 1.574,
      "theoretical_loss": 3.443661032290546,
      "tokens_seen": 1909915648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004284848484848485,
      "loss": 1.5845,
      "theoretical_loss": 3.4436510907206945,
      "tokens_seen": 1909981184
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042846464646464645,
      "loss": 1.2667,
      "theoretical_loss": 3.4436411495874655,
      "tokens_seen": 1910046720
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042844444444444446,
      "loss": 1.1777,
      "theoretical_loss": 3.4436312088908254,
      "tokens_seen": 1910112256
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004284242424242424,
      "loss": 1.3442,
      "theoretical_loss": 3.4436212686307393,
      "tokens_seen": 1910177792
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004284040404040404,
      "loss": 1.3153,
      "theoretical_loss": 3.4436113288071732,
      "tokens_seen": 1910243328
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004283838383838384,
      "loss": 1.5457,
      "theoretical_loss": 3.4436013894200936,
      "tokens_seen": 1910308864
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004283636363636364,
      "loss": 1.2672,
      "theoretical_loss": 3.443591450469466,
      "tokens_seen": 1910374400
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042834343434343435,
      "loss": 1.084,
      "theoretical_loss": 3.4435815119552555,
      "tokens_seen": 1910439936
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042832323232323236,
      "loss": 1.3617,
      "theoretical_loss": 3.443571573877429,
      "tokens_seen": 1910505472
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004283030303030303,
      "loss": 1.3787,
      "theoretical_loss": 3.443561636235952,
      "tokens_seen": 1910571008
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.5610948204994202,
      "objective/train/avg_token_score": 0.003200286766514182,
      "objective/train/docs_used": 1080108,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.690970540046692,
      "objective/train/original_loss": 3.2980706691741943,
      "objective/train/theoretical_loss": 3.4435516990307904,
      "objective/train/tokens_used": 269495776,
      "objective/train/unlikelihood_loss": 0.08375748991966248,
      "theoretical_loss": 3.4435516990307904,
      "tokens_seen": 1910636544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282828282828283,
      "loss": 1.4718,
      "theoretical_loss": 3.4435516990307904,
      "tokens_seen": 1910636544
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282626262626263,
      "loss": 1.5774,
      "theoretical_loss": 3.44354176226191,
      "tokens_seen": 1910702080
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042824242424242424,
      "loss": 1.6015,
      "theoretical_loss": 3.4435318259292766,
      "tokens_seen": 1910767616
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042822222222222225,
      "loss": 1.3461,
      "theoretical_loss": 3.443521890032856,
      "tokens_seen": 1910833152
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004282020202020202,
      "loss": 1.394,
      "theoretical_loss": 3.4435119545726147,
      "tokens_seen": 1910898688
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042818181818181817,
      "loss": 1.3769,
      "theoretical_loss": 3.443502019548518,
      "tokens_seen": 1910964224
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281616161616162,
      "loss": 1.1004,
      "theoretical_loss": 3.4434920849605324,
      "tokens_seen": 1911029760
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281414141414142,
      "loss": 1.5169,
      "theoretical_loss": 3.443482150808623,
      "tokens_seen": 1911095296
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042812121212121215,
      "loss": 1.4129,
      "theoretical_loss": 3.443472217092756,
      "tokens_seen": 1911160832
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004281010101010101,
      "loss": 1.3754,
      "theoretical_loss": 3.443462283812898,
      "tokens_seen": 1911226368
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042808080808080806,
      "loss": 1.3636,
      "theoretical_loss": 3.4434523509690136,
      "tokens_seen": 1911291904
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042806060606060607,
      "loss": 1.2775,
      "theoretical_loss": 3.44344241856107,
      "tokens_seen": 1911357440
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004280404040404041,
      "loss": 1.6295,
      "theoretical_loss": 3.4434324865890322,
      "tokens_seen": 1911422976
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042802020202020204,
      "loss": 1.5199,
      "theoretical_loss": 3.4434225550528668,
      "tokens_seen": 1911488512
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000428,
      "loss": 1.3486,
      "theoretical_loss": 3.443412623952539,
      "tokens_seen": 1911554048
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042797979797979795,
      "loss": 1.4287,
      "theoretical_loss": 3.4434026932880153,
      "tokens_seen": 1911619584
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000427959595959596,
      "loss": 1.4952,
      "theoretical_loss": 3.4433927630592613,
      "tokens_seen": 1911685120
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000427939393939394,
      "loss": 1.6055,
      "theoretical_loss": 3.4433828332662437,
      "tokens_seen": 1911750656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042791919191919193,
      "loss": 1.3703,
      "theoretical_loss": 3.4433729039089274,
      "tokens_seen": 1911816192
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004278989898989899,
      "loss": 1.3665,
      "theoretical_loss": 3.443362974987279,
      "tokens_seen": 1911881728
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042787878787878784,
      "loss": 1.3683,
      "theoretical_loss": 3.4433530465012643,
      "tokens_seen": 1911947264
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004278585858585859,
      "loss": 1.3692,
      "theoretical_loss": 3.443343118450849,
      "tokens_seen": 1912012800
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042783838383838387,
      "loss": 1.4151,
      "theoretical_loss": 3.4433331908359994,
      "tokens_seen": 1912078336
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004278181818181818,
      "loss": 1.4771,
      "theoretical_loss": 3.4433232636566817,
      "tokens_seen": 1912143872
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277979797979798,
      "loss": 1.4327,
      "theoretical_loss": 3.443313336912861,
      "tokens_seen": 1912209408
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.33864858746528625,
      "objective/train/avg_token_score": 0.007881486788392067,
      "objective/train/docs_used": 1081319,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1766334772109985,
      "objective/train/original_loss": 4.45033597946167,
      "objective/train/theoretical_loss": 3.443303410604504,
      "objective/train/tokens_used": 271134176,
      "objective/train/unlikelihood_loss": 0.05650121346116066,
      "theoretical_loss": 3.443303410604504,
      "tokens_seen": 1912274944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277777777777778,
      "loss": 1.4087,
      "theoretical_loss": 3.443303410604504,
      "tokens_seen": 1912274944
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277575757575758,
      "loss": 1.6281,
      "theoretical_loss": 3.4432934847315764,
      "tokens_seen": 1912340480
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042773737373737376,
      "loss": 1.4231,
      "theoretical_loss": 3.443283559294044,
      "tokens_seen": 1912406016
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004277171717171717,
      "loss": 1.3221,
      "theoretical_loss": 3.4432736342918737,
      "tokens_seen": 1912471552
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004276969696969697,
      "loss": 1.3538,
      "theoretical_loss": 3.44326370972503,
      "tokens_seen": 1912537088
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004276767676767677,
      "loss": 1.3519,
      "theoretical_loss": 3.4432537855934804,
      "tokens_seen": 1912602624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004276565656565657,
      "loss": 1.2999,
      "theoretical_loss": 3.44324386189719,
      "tokens_seen": 1912668160
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042763636363636365,
      "loss": 1.4608,
      "theoretical_loss": 3.4432339386361246,
      "tokens_seen": 1912733696
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004276161616161616,
      "loss": 1.5103,
      "theoretical_loss": 3.4432240158102507,
      "tokens_seen": 1912799232
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275959595959596,
      "loss": 1.3832,
      "theoretical_loss": 3.443214093419534,
      "tokens_seen": 1912864768
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275757575757576,
      "loss": 1.319,
      "theoretical_loss": 3.4432041714639414,
      "tokens_seen": 1912930304
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275555555555556,
      "loss": 1.2985,
      "theoretical_loss": 3.4431942499434376,
      "tokens_seen": 1912995840
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042753535353535354,
      "loss": 1.2791,
      "theoretical_loss": 3.4431843288579893,
      "tokens_seen": 1913061376
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004275151515151515,
      "loss": 1.3884,
      "theoretical_loss": 3.443174408207563,
      "tokens_seen": 1913126912
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274949494949495,
      "loss": 1.3111,
      "theoretical_loss": 3.4431644879921235,
      "tokens_seen": 1913192448
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042747474747474747,
      "loss": 1.4369,
      "theoretical_loss": 3.4431545682116376,
      "tokens_seen": 1913257984
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004274545454545455,
      "loss": 1.2989,
      "theoretical_loss": 3.4431446488660713,
      "tokens_seen": 1913323520
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042743434343434344,
      "loss": 1.5759,
      "theoretical_loss": 3.44313472995539,
      "tokens_seen": 1913389056
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042741414141414145,
      "loss": 1.5069,
      "theoretical_loss": 3.443124811479561,
      "tokens_seen": 1913454592
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004273939393939394,
      "loss": 1.3957,
      "theoretical_loss": 3.44311489343855,
      "tokens_seen": 1913520128
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042737373737373736,
      "loss": 1.4563,
      "theoretical_loss": 3.443104975832322,
      "tokens_seen": 1913585664
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004273535353535354,
      "loss": 1.1546,
      "theoretical_loss": 3.4430950586608438,
      "tokens_seen": 1913651200
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042733333333333333,
      "loss": 1.3923,
      "theoretical_loss": 3.4430851419240818,
      "tokens_seen": 1913716736
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042731313131313134,
      "loss": 1.3045,
      "theoretical_loss": 3.4430752256220014,
      "tokens_seen": 1913782272
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272929292929293,
      "loss": 1.2177,
      "theoretical_loss": 3.443065309754569,
      "tokens_seen": 1913847808
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.29844820499420166,
      "objective/train/avg_token_score": 0.011274569667875767,
      "objective/train/docs_used": 1081978,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.975776731967926,
      "objective/train/original_loss": 4.575207710266113,
      "objective/train/theoretical_loss": 3.4430553943217506,
      "objective/train/tokens_used": 272772576,
      "objective/train/unlikelihood_loss": 0.051495835185050964,
      "theoretical_loss": 3.4430553943217506,
      "tokens_seen": 1913913344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042727272727272726,
      "loss": 1.4363,
      "theoretical_loss": 3.4430553943217506,
      "tokens_seen": 1913913344
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042725252525252527,
      "loss": 1.283,
      "theoretical_loss": 3.443045479323512,
      "tokens_seen": 1913978880
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004272323232323233,
      "loss": 1.5076,
      "theoretical_loss": 3.4430355647598194,
      "tokens_seen": 1914044416
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042721212121212123,
      "loss": 1.2358,
      "theoretical_loss": 3.4430256506306396,
      "tokens_seen": 1914109952
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004271919191919192,
      "loss": 1.4152,
      "theoretical_loss": 3.443015736935938,
      "tokens_seen": 1914175488
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042717171717171715,
      "loss": 1.2006,
      "theoretical_loss": 3.4430058236756804,
      "tokens_seen": 1914241024
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042715151515151516,
      "loss": 1.5529,
      "theoretical_loss": 3.4429959108498336,
      "tokens_seen": 1914306560
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042713131313131317,
      "loss": 1.3738,
      "theoretical_loss": 3.4429859984583633,
      "tokens_seen": 1914372096
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004271111111111111,
      "loss": 1.4871,
      "theoretical_loss": 3.4429760865012353,
      "tokens_seen": 1914437632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004270909090909091,
      "loss": 1.3976,
      "theoretical_loss": 3.442966174978416,
      "tokens_seen": 1914503168
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042707070707070704,
      "loss": 1.1469,
      "theoretical_loss": 3.4429562638898723,
      "tokens_seen": 1914568704
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042705050505050505,
      "loss": 1.2692,
      "theoretical_loss": 3.442946353235569,
      "tokens_seen": 1914634240
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042703030303030306,
      "loss": 1.5126,
      "theoretical_loss": 3.4429364430154727,
      "tokens_seen": 1914699776
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000427010101010101,
      "loss": 1.4688,
      "theoretical_loss": 3.44292653322955,
      "tokens_seen": 1914765312
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000426989898989899,
      "loss": 1.4204,
      "theoretical_loss": 3.4429166238777658,
      "tokens_seen": 1914830848
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042696969696969693,
      "loss": 1.5229,
      "theoretical_loss": 3.442906714960088,
      "tokens_seen": 1914896384
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.000426949494949495,
      "loss": 1.504,
      "theoretical_loss": 3.4428968064764813,
      "tokens_seen": 1914961920
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042692929292929296,
      "loss": 1.3443,
      "theoretical_loss": 3.442886898426912,
      "tokens_seen": 1915027456
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004269090909090909,
      "loss": 1.2391,
      "theoretical_loss": 3.442876990811347,
      "tokens_seen": 1915092992
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042688888888888887,
      "loss": 1.1644,
      "theoretical_loss": 3.4428670836297517,
      "tokens_seen": 1915158528
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268686868686868,
      "loss": 1.4366,
      "theoretical_loss": 3.442857176882093,
      "tokens_seen": 1915224064
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268484848484849,
      "loss": 1.6832,
      "theoretical_loss": 3.442847270568336,
      "tokens_seen": 1915289600
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042682828282828285,
      "loss": 1.41,
      "theoretical_loss": 3.4428373646884474,
      "tokens_seen": 1915355136
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0004268080808080808,
      "loss": 1.37,
      "theoretical_loss": 3.4428274592423938,
      "tokens_seen": 1915420672
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042678787878787876,
      "loss": 1.3526,
      "theoretical_loss": 3.4428175542301407,
      "tokens_seen": 1915486208
    },
    {
      "epoch": 0.15,
      "objective/train/aligned_token_count": 0.41758307814598083,
      "objective/train/avg_token_score": 0.019459545612335205,
      "objective/train/docs_used": 1083212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3157575130462646,
      "objective/train/original_loss": 4.584113121032715,
      "objective/train/theoretical_loss": 3.442807649651654,
      "objective/train/tokens_used": 274410976,
      "objective/train/unlikelihood_loss": 0.04653491452336311,
      "theoretical_loss": 3.442807649651654,
      "tokens_seen": 1915551744
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.00042676767676767677,
      "loss": 1.4367,
      "theoretical_loss": 3.442807649651654,
      "tokens_seen": 1915551744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004267474747474748,
      "loss": 1.189,
      "theoretical_loss": 3.4427977455069008,
      "tokens_seen": 1915617280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042672727272727274,
      "loss": 1.3401,
      "theoretical_loss": 3.4427878417958464,
      "tokens_seen": 1915682816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004267070707070707,
      "loss": 1.2258,
      "theoretical_loss": 3.442777938518458,
      "tokens_seen": 1915748352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042668686868686865,
      "loss": 1.505,
      "theoretical_loss": 3.4427680356747006,
      "tokens_seen": 1915813888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004266666666666667,
      "loss": 1.6968,
      "theoretical_loss": 3.442758133264541,
      "tokens_seen": 1915879424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004266464646464647,
      "loss": 1.431,
      "theoretical_loss": 3.442748231287946,
      "tokens_seen": 1915944960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042662626262626263,
      "loss": 1.4045,
      "theoretical_loss": 3.4427383297448806,
      "tokens_seen": 1916010496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004266060606060606,
      "loss": 1.2996,
      "theoretical_loss": 3.4427284286353115,
      "tokens_seen": 1916076032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265858585858586,
      "loss": 1.403,
      "theoretical_loss": 3.442718527959205,
      "tokens_seen": 1916141568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265656565656566,
      "loss": 1.3892,
      "theoretical_loss": 3.4427086277165273,
      "tokens_seen": 1916207104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042654545454545457,
      "loss": 1.5321,
      "theoretical_loss": 3.442698727907244,
      "tokens_seen": 1916272640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265252525252525,
      "loss": 1.5272,
      "theoretical_loss": 3.442688828531322,
      "tokens_seen": 1916338176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004265050505050505,
      "loss": 1.5984,
      "theoretical_loss": 3.4426789295887277,
      "tokens_seen": 1916403712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264848484848485,
      "loss": 1.3496,
      "theoretical_loss": 3.4426690310794266,
      "tokens_seen": 1916469248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264646464646465,
      "loss": 1.3557,
      "theoretical_loss": 3.442659133003385,
      "tokens_seen": 1916534784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042644444444444446,
      "loss": 1.2853,
      "theoretical_loss": 3.4426492353605695,
      "tokens_seen": 1916600320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004264242424242424,
      "loss": 1.5258,
      "theoretical_loss": 3.442639338150946,
      "tokens_seen": 1916665856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042640404040404043,
      "loss": 1.2848,
      "theoretical_loss": 3.4426294413744816,
      "tokens_seen": 1916731392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263838383838384,
      "loss": 1.4453,
      "theoretical_loss": 3.4426195450311416,
      "tokens_seen": 1916796928
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263636363636364,
      "loss": 1.2794,
      "theoretical_loss": 3.442609649120892,
      "tokens_seen": 1916862464
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042634343434343435,
      "loss": 1.2195,
      "theoretical_loss": 3.4425997536437,
      "tokens_seen": 1916928000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263232323232323,
      "loss": 1.6612,
      "theoretical_loss": 3.4425898585995314,
      "tokens_seen": 1916993536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004263030303030303,
      "loss": 1.3784,
      "theoretical_loss": 3.4425799639883525,
      "tokens_seen": 1917059072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004262828282828283,
      "loss": 1.2362,
      "theoretical_loss": 3.442570069810129,
      "tokens_seen": 1917124608
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.3271627426147461,
      "objective/train/avg_token_score": 0.00935707613825798,
      "objective/train/docs_used": 1083753,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1364368200302124,
      "objective/train/original_loss": 4.213170051574707,
      "objective/train/theoretical_loss": 3.442560176064828,
      "objective/train/tokens_used": 276049376,
      "objective/train/unlikelihood_loss": 0.05666195973753929,
      "theoretical_loss": 3.442560176064828,
      "tokens_seen": 1917190144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004262626262626263,
      "loss": 1.2487,
      "theoretical_loss": 3.442560176064828,
      "tokens_seen": 1917190144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042624242424242425,
      "loss": 1.3266,
      "theoretical_loss": 3.442550282752415,
      "tokens_seen": 1917255680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042622222222222226,
      "loss": 1.4237,
      "theoretical_loss": 3.4425403898728573,
      "tokens_seen": 1917321216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004262020202020202,
      "loss": 1.3311,
      "theoretical_loss": 3.44253049742612,
      "tokens_seen": 1917386752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042618181818181817,
      "loss": 1.4461,
      "theoretical_loss": 3.44252060541217,
      "tokens_seen": 1917452288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261616161616162,
      "loss": 1.5083,
      "theoretical_loss": 3.4425107138309734,
      "tokens_seen": 1917517824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042614141414141414,
      "loss": 1.5493,
      "theoretical_loss": 3.4425008226824967,
      "tokens_seen": 1917583360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042612121212121215,
      "loss": 1.3401,
      "theoretical_loss": 3.442490931966706,
      "tokens_seen": 1917648896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004261010101010101,
      "loss": 1.521,
      "theoretical_loss": 3.4424810416835676,
      "tokens_seen": 1917714432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042608080808080806,
      "loss": 1.3547,
      "theoretical_loss": 3.4424711518330477,
      "tokens_seen": 1917779968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004260606060606061,
      "loss": 1.4561,
      "theoretical_loss": 3.442461262415113,
      "tokens_seen": 1917845504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004260404040404041,
      "loss": 1.4914,
      "theoretical_loss": 3.442451373429729,
      "tokens_seen": 1917911040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042602020202020204,
      "loss": 1.2794,
      "theoretical_loss": 3.4424414848768627,
      "tokens_seen": 1917976576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000426,
      "loss": 1.2327,
      "theoretical_loss": 3.4424315967564807,
      "tokens_seen": 1918042112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042597979797979796,
      "loss": 1.5526,
      "theoretical_loss": 3.4424217090685483,
      "tokens_seen": 1918107648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042595959595959597,
      "loss": 1.5492,
      "theoretical_loss": 3.4424118218130326,
      "tokens_seen": 1918173184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425939393939394,
      "loss": 1.3568,
      "theoretical_loss": 3.442401934989899,
      "tokens_seen": 1918238720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042591919191919194,
      "loss": 1.3991,
      "theoretical_loss": 3.442392048599115,
      "tokens_seen": 1918304256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258989898989899,
      "loss": 1.3583,
      "theoretical_loss": 3.442382162640646,
      "tokens_seen": 1918369792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042587878787878785,
      "loss": 1.5928,
      "theoretical_loss": 3.442372277114459,
      "tokens_seen": 1918435328
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004258585858585859,
      "loss": 1.1758,
      "theoretical_loss": 3.44236239202052,
      "tokens_seen": 1918500864
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042583838383838387,
      "loss": 1.4569,
      "theoretical_loss": 3.4423525073587955,
      "tokens_seen": 1918566400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042581818181818183,
      "loss": 1.4608,
      "theoretical_loss": 3.4423426231292518,
      "tokens_seen": 1918631936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257979797979798,
      "loss": 1.5675,
      "theoretical_loss": 3.442332739331855,
      "tokens_seen": 1918697472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042577777777777774,
      "loss": 1.2314,
      "theoretical_loss": 3.4423228559665713,
      "tokens_seen": 1918763008
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.4659701883792877,
      "objective/train/avg_token_score": 0.018756110221147537,
      "objective/train/docs_used": 1084888,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6133241653442383,
      "objective/train/original_loss": 4.159404277801514,
      "objective/train/theoretical_loss": 3.442312973033368,
      "objective/train/tokens_used": 277687776,
      "objective/train/unlikelihood_loss": 0.05935472249984741,
      "theoretical_loss": 3.442312973033368,
      "tokens_seen": 1918828544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257575757575758,
      "loss": 1.3151,
      "theoretical_loss": 3.442312973033368,
      "tokens_seen": 1918828544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042573737373737377,
      "loss": 1.4888,
      "theoretical_loss": 3.4423030905322105,
      "tokens_seen": 1918894080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004257171717171717,
      "loss": 1.4874,
      "theoretical_loss": 3.4422932084630653,
      "tokens_seen": 1918959616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004256969696969697,
      "loss": 1.4814,
      "theoretical_loss": 3.442283326825899,
      "tokens_seen": 1919025152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042567676767676764,
      "loss": 1.3381,
      "theoretical_loss": 3.4422734456206783,
      "tokens_seen": 1919090688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004256565656565657,
      "loss": 1.4287,
      "theoretical_loss": 3.4422635648473685,
      "tokens_seen": 1919156224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042563636363636366,
      "loss": 1.2571,
      "theoretical_loss": 3.442253684505937,
      "tokens_seen": 1919221760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004256161616161616,
      "loss": 1.3749,
      "theoretical_loss": 3.4422438045963495,
      "tokens_seen": 1919287296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042559595959595957,
      "loss": 1.6466,
      "theoretical_loss": 3.4422339251185727,
      "tokens_seen": 1919352832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255757575757576,
      "loss": 1.3934,
      "theoretical_loss": 3.4422240460725737,
      "tokens_seen": 1919418368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255555555555556,
      "loss": 1.2283,
      "theoretical_loss": 3.4422141674583173,
      "tokens_seen": 1919483904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042553535353535355,
      "loss": 1.3776,
      "theoretical_loss": 3.4422042892757716,
      "tokens_seen": 1919549440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004255151515151515,
      "loss": 1.4192,
      "theoretical_loss": 3.4421944115249015,
      "tokens_seen": 1919614976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042549494949494946,
      "loss": 1.4077,
      "theoretical_loss": 3.442184534205674,
      "tokens_seen": 1919680512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254747474747475,
      "loss": 1.3611,
      "theoretical_loss": 3.4421746573180556,
      "tokens_seen": 1919746048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254545454545455,
      "loss": 1.3919,
      "theoretical_loss": 3.4421647808620133,
      "tokens_seen": 1919811584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042543434343434344,
      "loss": 1.6059,
      "theoretical_loss": 3.442154904837512,
      "tokens_seen": 1919877120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004254141414141414,
      "loss": 1.4137,
      "theoretical_loss": 3.4421450292445197,
      "tokens_seen": 1919942656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253939393939394,
      "loss": 1.4409,
      "theoretical_loss": 3.442135154083002,
      "tokens_seen": 1920008192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042537373737373737,
      "loss": 1.2995,
      "theoretical_loss": 3.4421252793529247,
      "tokens_seen": 1920073728
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253535353535354,
      "loss": 1.457,
      "theoretical_loss": 3.4421154050542553,
      "tokens_seen": 1920139264
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042533333333333334,
      "loss": 1.3346,
      "theoretical_loss": 3.4421055311869604,
      "tokens_seen": 1920204800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004253131313131313,
      "loss": 1.4856,
      "theoretical_loss": 3.442095657751006,
      "tokens_seen": 1920270336
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004252929292929293,
      "loss": 1.4931,
      "theoretical_loss": 3.4420857847463573,
      "tokens_seen": 1920335872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042527272727272726,
      "loss": 1.4235,
      "theoretical_loss": 3.442075912172983,
      "tokens_seen": 1920401408
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.2597140669822693,
      "objective/train/avg_token_score": 0.00819999910891056,
      "objective/train/docs_used": 1085471,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8639290928840637,
      "objective/train/original_loss": 3.9054853916168213,
      "objective/train/theoretical_loss": 3.4420660400308476,
      "objective/train/tokens_used": 279326176,
      "objective/train/unlikelihood_loss": 0.08530934154987335,
      "theoretical_loss": 3.4420660400308476,
      "tokens_seen": 1920466944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042525252525252527,
      "loss": 1.29,
      "theoretical_loss": 3.4420660400308476,
      "tokens_seen": 1920466944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042523232323232323,
      "loss": 1.4392,
      "theoretical_loss": 3.442056168319919,
      "tokens_seen": 1920532480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042521212121212124,
      "loss": 1.5414,
      "theoretical_loss": 3.442046297040163,
      "tokens_seen": 1920598016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004251919191919192,
      "loss": 1.5928,
      "theoretical_loss": 3.442036426191546,
      "tokens_seen": 1920663552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042517171717171715,
      "loss": 1.5551,
      "theoretical_loss": 3.442026555774034,
      "tokens_seen": 1920729088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042515151515151516,
      "loss": 1.4526,
      "theoretical_loss": 3.442016685787595,
      "tokens_seen": 1920794624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004251313131313131,
      "loss": 1.3303,
      "theoretical_loss": 3.442006816232194,
      "tokens_seen": 1920860160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042511111111111113,
      "loss": 1.4716,
      "theoretical_loss": 3.441996947107798,
      "tokens_seen": 1920925696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004250909090909091,
      "loss": 1.7319,
      "theoretical_loss": 3.441987078414373,
      "tokens_seen": 1920991232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042507070707070705,
      "loss": 1.298,
      "theoretical_loss": 3.4419772101518866,
      "tokens_seen": 1921056768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042505050505050506,
      "loss": 1.4499,
      "theoretical_loss": 3.4419673423203037,
      "tokens_seen": 1921122304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042503030303030307,
      "loss": 1.4579,
      "theoretical_loss": 3.4419574749195925,
      "tokens_seen": 1921187840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000425010101010101,
      "loss": 1.3127,
      "theoretical_loss": 3.441947607949718,
      "tokens_seen": 1921253376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424989898989899,
      "loss": 1.3889,
      "theoretical_loss": 3.441937741410648,
      "tokens_seen": 1921318912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424969696969697,
      "loss": 1.3067,
      "theoretical_loss": 3.441927875302348,
      "tokens_seen": 1921384448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042494949494949495,
      "loss": 1.2578,
      "theoretical_loss": 3.441918009624785,
      "tokens_seen": 1921449984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042492929292929296,
      "loss": 1.2618,
      "theoretical_loss": 3.4419081443779254,
      "tokens_seen": 1921515520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004249090909090909,
      "loss": 1.5736,
      "theoretical_loss": 3.4418982795617357,
      "tokens_seen": 1921581056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248888888888889,
      "loss": 1.4578,
      "theoretical_loss": 3.441888415176182,
      "tokens_seen": 1921646592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248686868686869,
      "loss": 1.3573,
      "theoretical_loss": 3.4418785512212313,
      "tokens_seen": 1921712128
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248484848484849,
      "loss": 1.3088,
      "theoretical_loss": 3.44186868769685,
      "tokens_seen": 1921777664
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042482828282828285,
      "loss": 1.2699,
      "theoretical_loss": 3.441858824603005,
      "tokens_seen": 1921843200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004248080808080808,
      "loss": 1.4885,
      "theoretical_loss": 3.4418489619396624,
      "tokens_seen": 1921908736
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042478787878787877,
      "loss": 1.4237,
      "theoretical_loss": 3.4418390997067885,
      "tokens_seen": 1921974272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247676767676768,
      "loss": 1.4151,
      "theoretical_loss": 3.44182923790435,
      "tokens_seen": 1922039808
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.37664955854415894,
      "objective/train/avg_token_score": 0.027382804080843925,
      "objective/train/docs_used": 1086847,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2897065877914429,
      "objective/train/original_loss": 4.506113529205322,
      "objective/train/theoretical_loss": 3.4418193765323135,
      "objective/train/tokens_used": 280964576,
      "objective/train/unlikelihood_loss": 0.06149740517139435,
      "theoretical_loss": 3.4418193765323135,
      "tokens_seen": 1922105344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247474747474748,
      "loss": 1.3919,
      "theoretical_loss": 3.4418193765323135,
      "tokens_seen": 1922105344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042472727272727275,
      "loss": 1.1067,
      "theoretical_loss": 3.4418095155906463,
      "tokens_seen": 1922170880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004247070707070707,
      "loss": 1.3229,
      "theoretical_loss": 3.441799655079314,
      "tokens_seen": 1922236416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042468686868686866,
      "loss": 1.283,
      "theoretical_loss": 3.441789794998283,
      "tokens_seen": 1922301952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246666666666667,
      "loss": 1.3661,
      "theoretical_loss": 3.4417799353475202,
      "tokens_seen": 1922367488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246464646464647,
      "loss": 1.3491,
      "theoretical_loss": 3.4417700761269927,
      "tokens_seen": 1922433024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042462626262626264,
      "loss": 1.4425,
      "theoretical_loss": 3.4417602173366664,
      "tokens_seen": 1922498560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004246060606060606,
      "loss": 1.5416,
      "theoretical_loss": 3.441750358976508,
      "tokens_seen": 1922564096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042458585858585855,
      "loss": 1.2915,
      "theoretical_loss": 3.4417405010464837,
      "tokens_seen": 1922629632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245656565656566,
      "loss": 1.7483,
      "theoretical_loss": 3.441730643546561,
      "tokens_seen": 1922695168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245454545454546,
      "loss": 1.4832,
      "theoretical_loss": 3.441720786476706,
      "tokens_seen": 1922760704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042452525252525253,
      "loss": 1.5824,
      "theoretical_loss": 3.441710929836885,
      "tokens_seen": 1922826240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004245050505050505,
      "loss": 1.5308,
      "theoretical_loss": 3.4417010736270646,
      "tokens_seen": 1922891776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244848484848485,
      "loss": 1.2997,
      "theoretical_loss": 3.441691217847212,
      "tokens_seen": 1922957312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244646464646465,
      "loss": 1.4648,
      "theoretical_loss": 3.4416813624972935,
      "tokens_seen": 1923022848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042444444444444447,
      "loss": 1.1912,
      "theoretical_loss": 3.441671507577275,
      "tokens_seen": 1923088384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244242424242424,
      "loss": 1.3946,
      "theoretical_loss": 3.441661653087124,
      "tokens_seen": 1923153920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004244040404040404,
      "loss": 1.4577,
      "theoretical_loss": 3.4416517990268067,
      "tokens_seen": 1923219456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243838383838384,
      "loss": 1.2794,
      "theoretical_loss": 3.44164194539629,
      "tokens_seen": 1923284992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243636363636364,
      "loss": 1.4351,
      "theoretical_loss": 3.4416320921955403,
      "tokens_seen": 1923350528
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042434343434343436,
      "loss": 1.4165,
      "theoretical_loss": 3.4416222394245235,
      "tokens_seen": 1923416064
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004243232323232323,
      "loss": 1.4822,
      "theoretical_loss": 3.441612387083208,
      "tokens_seen": 1923481600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042430303030303033,
      "loss": 1.2961,
      "theoretical_loss": 3.4416025351715587,
      "tokens_seen": 1923547136
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242828282828283,
      "loss": 1.5846,
      "theoretical_loss": 3.441592683689543,
      "tokens_seen": 1923612672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242626262626263,
      "loss": 1.5942,
      "theoretical_loss": 3.4415828326371276,
      "tokens_seen": 1923678208
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.353128045797348,
      "objective/train/avg_token_score": 0.019363651052117348,
      "objective/train/docs_used": 1087440,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.210463047027588,
      "objective/train/original_loss": 4.447400093078613,
      "objective/train/theoretical_loss": 3.4415729820142786,
      "objective/train/tokens_used": 282602976,
      "objective/train/unlikelihood_loss": 0.08732765913009644,
      "theoretical_loss": 3.4415729820142786,
      "tokens_seen": 1923743744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042424242424242425,
      "loss": 1.3893,
      "theoretical_loss": 3.4415729820142786,
      "tokens_seen": 1923743744
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242222222222222,
      "loss": 1.3323,
      "theoretical_loss": 3.441563131820963,
      "tokens_seen": 1923809280
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004242020202020202,
      "loss": 1.3105,
      "theoretical_loss": 3.4415532820571473,
      "tokens_seen": 1923874816
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241818181818182,
      "loss": 1.3959,
      "theoretical_loss": 3.4415434327227983,
      "tokens_seen": 1923940352
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241616161616162,
      "loss": 1.5941,
      "theoretical_loss": 3.441533583817883,
      "tokens_seen": 1924005888
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042414141414141415,
      "loss": 1.3544,
      "theoretical_loss": 3.4415237353423676,
      "tokens_seen": 1924071424
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042412121212121216,
      "loss": 1.3194,
      "theoretical_loss": 3.4415138872962183,
      "tokens_seen": 1924136960
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004241010101010101,
      "loss": 1.4216,
      "theoretical_loss": 3.4415040396794025,
      "tokens_seen": 1924202496
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042408080808080807,
      "loss": 1.4374,
      "theoretical_loss": 3.441494192491887,
      "tokens_seen": 1924268032
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004240606060606061,
      "loss": 1.6304,
      "theoretical_loss": 3.4414843457336373,
      "tokens_seen": 1924333568
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042404040404040404,
      "loss": 1.2735,
      "theoretical_loss": 3.441474499404621,
      "tokens_seen": 1924399104
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042402020202020205,
      "loss": 1.325,
      "theoretical_loss": 3.441464653504805,
      "tokens_seen": 1924464640
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000424,
      "loss": 1.6185,
      "theoretical_loss": 3.4414548080341554,
      "tokens_seen": 1924530176
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042397979797979796,
      "loss": 1.3112,
      "theoretical_loss": 3.441444962992639,
      "tokens_seen": 1924595712
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423959595959596,
      "loss": 1.6121,
      "theoretical_loss": 3.441435118380223,
      "tokens_seen": 1924661248
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042393939393939393,
      "loss": 1.4117,
      "theoretical_loss": 3.4414252741968734,
      "tokens_seen": 1924726784
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042391919191919194,
      "loss": 1.3539,
      "theoretical_loss": 3.4414154304425573,
      "tokens_seen": 1924792320
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004238989898989899,
      "loss": 1.2487,
      "theoretical_loss": 3.4414055871172406,
      "tokens_seen": 1924857856
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042387878787878786,
      "loss": 1.4242,
      "theoretical_loss": 3.441395744220891,
      "tokens_seen": 1924923392
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042385858585858587,
      "loss": 1.4151,
      "theoretical_loss": 3.441385901753475,
      "tokens_seen": 1924988928
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004238383838383839,
      "loss": 1.4439,
      "theoretical_loss": 3.441376059714959,
      "tokens_seen": 1925054464
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042381818181818183,
      "loss": 1.5184,
      "theoretical_loss": 3.44136621810531,
      "tokens_seen": 1925120000
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004237979797979798,
      "loss": 1.5182,
      "theoretical_loss": 3.4413563769244946,
      "tokens_seen": 1925185536
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042377777777777775,
      "loss": 1.4028,
      "theoretical_loss": 3.441346536172479,
      "tokens_seen": 1925251072
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042375757575757576,
      "loss": 1.3353,
      "theoretical_loss": 3.441336695849231,
      "tokens_seen": 1925316608
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.48949170112609863,
      "objective/train/avg_token_score": 0.014318176545202732,
      "objective/train/docs_used": 1088918,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6369242668151855,
      "objective/train/original_loss": 4.034110069274902,
      "objective/train/theoretical_loss": 3.441326855954716,
      "objective/train/tokens_used": 284241376,
      "objective/train/unlikelihood_loss": 0.05818219855427742,
      "theoretical_loss": 3.441326855954716,
      "tokens_seen": 1925382144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042373737373737377,
      "loss": 1.5529,
      "theoretical_loss": 3.441326855954716,
      "tokens_seen": 1925382144
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042371717171717173,
      "loss": 1.3521,
      "theoretical_loss": 3.4413170164889024,
      "tokens_seen": 1925447680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236969696969697,
      "loss": 1.267,
      "theoretical_loss": 3.441307177451755,
      "tokens_seen": 1925513216
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042367676767676764,
      "loss": 1.6063,
      "theoretical_loss": 3.4412973388432424,
      "tokens_seen": 1925578752
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236565656565657,
      "loss": 1.4146,
      "theoretical_loss": 3.44128750066333,
      "tokens_seen": 1925644288
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042363636363636366,
      "loss": 1.6882,
      "theoretical_loss": 3.4412776629119852,
      "tokens_seen": 1925709824
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004236161616161616,
      "loss": 1.3858,
      "theoretical_loss": 3.441267825589174,
      "tokens_seen": 1925775360
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004235959595959596,
      "loss": 1.4042,
      "theoretical_loss": 3.4412579886948644,
      "tokens_seen": 1925840896
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042357575757575753,
      "loss": 1.3035,
      "theoretical_loss": 3.441248152229022,
      "tokens_seen": 1925906432
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004235555555555556,
      "loss": 1.4533,
      "theoretical_loss": 3.4412383161916145,
      "tokens_seen": 1925971968
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042353535353535356,
      "loss": 1.743,
      "theoretical_loss": 3.4412284805826077,
      "tokens_seen": 1926037504
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004235151515151515,
      "loss": 1.2599,
      "theoretical_loss": 3.441218645401969,
      "tokens_seen": 1926103040
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042349494949494947,
      "loss": 1.4499,
      "theoretical_loss": 3.441208810649665,
      "tokens_seen": 1926168576
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004234747474747475,
      "loss": 1.5833,
      "theoretical_loss": 3.4411989763256625,
      "tokens_seen": 1926234112
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004234545454545455,
      "loss": 1.3539,
      "theoretical_loss": 3.441189142429928,
      "tokens_seen": 1926299648
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042343434343434345,
      "loss": 1.4448,
      "theoretical_loss": 3.441179308962429,
      "tokens_seen": 1926365184
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004234141414141414,
      "loss": 1.485,
      "theoretical_loss": 3.4411694759231315,
      "tokens_seen": 1926430720
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042339393939393936,
      "loss": 1.6433,
      "theoretical_loss": 3.4411596433120026,
      "tokens_seen": 1926496256
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004233737373737374,
      "loss": 1.311,
      "theoretical_loss": 3.4411498111290095,
      "tokens_seen": 1926561792
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004233535353535354,
      "loss": 1.408,
      "theoretical_loss": 3.4411399793741184,
      "tokens_seen": 1926627328
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042333333333333334,
      "loss": 1.3597,
      "theoretical_loss": 3.441130148047296,
      "tokens_seen": 1926692864
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004233131313131313,
      "loss": 1.4085,
      "theoretical_loss": 3.44112031714851,
      "tokens_seen": 1926758400
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232929292929293,
      "loss": 1.2997,
      "theoretical_loss": 3.441110486677726,
      "tokens_seen": 1926823936
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232727272727273,
      "loss": 1.3375,
      "theoretical_loss": 3.4411006566349114,
      "tokens_seen": 1926889472
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232525252525253,
      "loss": 1.3776,
      "theoretical_loss": 3.441090827020033,
      "tokens_seen": 1926955008
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.4183773100376129,
      "objective/train/avg_token_score": 0.027611570432782173,
      "objective/train/docs_used": 1089550,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.40208101272583,
      "objective/train/original_loss": 3.778869867324829,
      "objective/train/theoretical_loss": 3.4410809978330583,
      "objective/train/tokens_used": 285879776,
      "objective/train/unlikelihood_loss": 0.07978764176368713,
      "theoretical_loss": 3.4410809978330583,
      "tokens_seen": 1927020544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042323232323232323,
      "loss": 1.1572,
      "theoretical_loss": 3.4410809978330583,
      "tokens_seen": 1927020544
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004232121212121212,
      "loss": 1.3984,
      "theoretical_loss": 3.441071169073953,
      "tokens_seen": 1927086080
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231919191919192,
      "loss": 1.5348,
      "theoretical_loss": 3.441061340742684,
      "tokens_seen": 1927151616
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231717171717172,
      "loss": 1.6712,
      "theoretical_loss": 3.4410515128392194,
      "tokens_seen": 1927217152
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042315151515151517,
      "loss": 1.43,
      "theoretical_loss": 3.4410416853635244,
      "tokens_seen": 1927282688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004231313131313131,
      "loss": 1.5115,
      "theoretical_loss": 3.441031858315567,
      "tokens_seen": 1927348224
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042311111111111114,
      "loss": 1.5486,
      "theoretical_loss": 3.4410220316953133,
      "tokens_seen": 1927413760
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004230909090909091,
      "loss": 1.408,
      "theoretical_loss": 3.441012205502731,
      "tokens_seen": 1927479296
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004230707070707071,
      "loss": 1.483,
      "theoretical_loss": 3.441002379737786,
      "tokens_seen": 1927544832
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042305050505050506,
      "loss": 1.5122,
      "theoretical_loss": 3.4409925544004456,
      "tokens_seen": 1927610368
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000423030303030303,
      "loss": 1.4116,
      "theoretical_loss": 3.4409827294906767,
      "tokens_seen": 1927675904
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042301010101010103,
      "loss": 1.561,
      "theoretical_loss": 3.4409729050084463,
      "tokens_seen": 1927741440
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422989898989899,
      "loss": 1.5038,
      "theoretical_loss": 3.440963080953721,
      "tokens_seen": 1927806976
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422969696969697,
      "loss": 1.3527,
      "theoretical_loss": 3.440953257326467,
      "tokens_seen": 1927872512
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042294949494949496,
      "loss": 1.3933,
      "theoretical_loss": 3.4409434341266527,
      "tokens_seen": 1927938048
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042292929292929297,
      "loss": 1.6347,
      "theoretical_loss": 3.440933611354244,
      "tokens_seen": 1928003584
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004229090909090909,
      "loss": 1.3893,
      "theoretical_loss": 3.4409237890092075,
      "tokens_seen": 1928069120
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004228888888888889,
      "loss": 1.0593,
      "theoretical_loss": 3.440913967091511,
      "tokens_seen": 1928134656
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004228686868686869,
      "loss": 1.2617,
      "theoretical_loss": 3.4409041456011202,
      "tokens_seen": 1928200192
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042284848484848485,
      "loss": 1.2376,
      "theoretical_loss": 3.4408943245380033,
      "tokens_seen": 1928265728
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042282828282828286,
      "loss": 1.4189,
      "theoretical_loss": 3.4408845039021263,
      "tokens_seen": 1928331264
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004228080808080808,
      "loss": 1.3413,
      "theoretical_loss": 3.4408746836934565,
      "tokens_seen": 1928396800
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042278787878787877,
      "loss": 1.3963,
      "theoretical_loss": 3.4408648639119606,
      "tokens_seen": 1928462336
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227676767676768,
      "loss": 1.2697,
      "theoretical_loss": 3.4408550445576056,
      "tokens_seen": 1928527872
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227474747474748,
      "loss": 1.7067,
      "theoretical_loss": 3.4408452256303583,
      "tokens_seen": 1928593408
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.30810117721557617,
      "objective/train/avg_token_score": 0.016539035364985466,
      "objective/train/docs_used": 1091004,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9405573606491089,
      "objective/train/original_loss": 4.469072341918945,
      "objective/train/theoretical_loss": 3.4408354071301854,
      "objective/train/tokens_used": 287518176,
      "objective/train/unlikelihood_loss": 0.0698821097612381,
      "theoretical_loss": 3.4408354071301854,
      "tokens_seen": 1928658944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042272727272727275,
      "loss": 1.2846,
      "theoretical_loss": 3.4408354071301854,
      "tokens_seen": 1928658944
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004227070707070707,
      "loss": 1.4077,
      "theoretical_loss": 3.4408255890570545,
      "tokens_seen": 1928724480
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042268686868686867,
      "loss": 1.5451,
      "theoretical_loss": 3.440815771410932,
      "tokens_seen": 1928790016
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226666666666667,
      "loss": 1.4673,
      "theoretical_loss": 3.440805954191785,
      "tokens_seen": 1928855552
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226464646464647,
      "loss": 1.5765,
      "theoretical_loss": 3.4407961373995803,
      "tokens_seen": 1928921088
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042262626262626264,
      "loss": 1.4946,
      "theoretical_loss": 3.4407863210342846,
      "tokens_seen": 1928986624
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004226060606060606,
      "loss": 1.6541,
      "theoretical_loss": 3.440776505095865,
      "tokens_seen": 1929052160
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042258585858585856,
      "loss": 1.4803,
      "theoretical_loss": 3.440766689584289,
      "tokens_seen": 1929117696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004225656565656566,
      "loss": 1.5839,
      "theoretical_loss": 3.4407568744995234,
      "tokens_seen": 1929183232
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004225454545454546,
      "loss": 1.4868,
      "theoretical_loss": 3.4407470598415344,
      "tokens_seen": 1929248768
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042252525252525254,
      "loss": 1.3395,
      "theoretical_loss": 3.440737245610289,
      "tokens_seen": 1929314304
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004225050505050505,
      "loss": 1.437,
      "theoretical_loss": 3.4407274318057555,
      "tokens_seen": 1929379840
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042248484848484845,
      "loss": 1.5503,
      "theoretical_loss": 3.440717618427899,
      "tokens_seen": 1929445376
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004224646464646465,
      "loss": 1.3331,
      "theoretical_loss": 3.4407078054766878,
      "tokens_seen": 1929510912
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004224444444444445,
      "loss": 1.4377,
      "theoretical_loss": 3.440697992952088,
      "tokens_seen": 1929576448
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042242424242424243,
      "loss": 1.5041,
      "theoretical_loss": 3.4406881808540675,
      "tokens_seen": 1929641984
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004224040404040404,
      "loss": 1.4972,
      "theoretical_loss": 3.440678369182592,
      "tokens_seen": 1929707520
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042238383838383834,
      "loss": 1.4901,
      "theoretical_loss": 3.44066855793763,
      "tokens_seen": 1929773056
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223636363636364,
      "loss": 1.3971,
      "theoretical_loss": 3.440658747119147,
      "tokens_seen": 1929838592
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042234343434343437,
      "loss": 1.6506,
      "theoretical_loss": 3.440648936727111,
      "tokens_seen": 1929904128
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223232323232323,
      "loss": 1.6115,
      "theoretical_loss": 3.4406391267614884,
      "tokens_seen": 1929969664
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004223030303030303,
      "loss": 1.4711,
      "theoretical_loss": 3.4406293172222466,
      "tokens_seen": 1930035200
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004222828282828283,
      "loss": 1.5098,
      "theoretical_loss": 3.4406195081093527,
      "tokens_seen": 1930100736
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004222626262626263,
      "loss": 1.3378,
      "theoretical_loss": 3.4406096994227733,
      "tokens_seen": 1930166272
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042224242424242426,
      "loss": 1.5783,
      "theoretical_loss": 3.440599891162475,
      "tokens_seen": 1930231808
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.6237170100212097,
      "objective/train/avg_token_score": 0.006665772292762995,
      "objective/train/docs_used": 1091522,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7739847898483276,
      "objective/train/original_loss": 3.4937384128570557,
      "objective/train/theoretical_loss": 3.4405900833284258,
      "objective/train/tokens_used": 289156576,
      "objective/train/unlikelihood_loss": 0.047466766089200974,
      "theoretical_loss": 3.4405900833284258,
      "tokens_seen": 1930297344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004222222222222222,
      "loss": 1.4705,
      "theoretical_loss": 3.4405900833284258,
      "tokens_seen": 1930297344
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042220202020202017,
      "loss": 1.3696,
      "theoretical_loss": 3.440580275920592,
      "tokens_seen": 1930362880
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221818181818182,
      "loss": 1.3453,
      "theoretical_loss": 3.440570468938941,
      "tokens_seen": 1930428416
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221616161616162,
      "loss": 1.6253,
      "theoretical_loss": 3.4405606623834393,
      "tokens_seen": 1930493952
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042214141414141415,
      "loss": 1.5188,
      "theoretical_loss": 3.4405508562540543,
      "tokens_seen": 1930559488
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221212121212121,
      "loss": 1.4635,
      "theoretical_loss": 3.4405410505507534,
      "tokens_seen": 1930625024
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004221010101010101,
      "loss": 1.3429,
      "theoretical_loss": 3.440531245273503,
      "tokens_seen": 1930690560
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004220808080808081,
      "loss": 1.319,
      "theoretical_loss": 3.4405214404222706,
      "tokens_seen": 1930756096
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004220606060606061,
      "loss": 1.2327,
      "theoretical_loss": 3.4405116359970225,
      "tokens_seen": 1930821632
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042204040404040404,
      "loss": 1.5396,
      "theoretical_loss": 3.4405018319977265,
      "tokens_seen": 1930887168
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422020202020202,
      "loss": 1.5807,
      "theoretical_loss": 3.440492028424349,
      "tokens_seen": 1930952704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000422,
      "loss": 1.3657,
      "theoretical_loss": 3.440482225276858,
      "tokens_seen": 1931018240
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042197979797979797,
      "loss": 1.2815,
      "theoretical_loss": 3.440472422555219,
      "tokens_seen": 1931083776
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.000421959595959596,
      "loss": 1.2827,
      "theoretical_loss": 3.4404626202594004,
      "tokens_seen": 1931149312
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042193939393939394,
      "loss": 1.4397,
      "theoretical_loss": 3.4404528183893692,
      "tokens_seen": 1931214848
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042191919191919195,
      "loss": 1.7767,
      "theoretical_loss": 3.4404430169450917,
      "tokens_seen": 1931280384
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004218989898989899,
      "loss": 1.6945,
      "theoretical_loss": 3.4404332159265354,
      "tokens_seen": 1931345920
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042187878787878786,
      "loss": 1.3724,
      "theoretical_loss": 3.440423415333667,
      "tokens_seen": 1931411456
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042185858585858587,
      "loss": 1.2764,
      "theoretical_loss": 3.440413615166454,
      "tokens_seen": 1931476992
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042183838383838383,
      "loss": 1.4126,
      "theoretical_loss": 3.440403815424864,
      "tokens_seen": 1931542528
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042181818181818184,
      "loss": 1.3462,
      "theoretical_loss": 3.440394016108863,
      "tokens_seen": 1931608064
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004217979797979798,
      "loss": 1.3443,
      "theoretical_loss": 3.4403842172184183,
      "tokens_seen": 1931673600
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042177777777777775,
      "loss": 1.5187,
      "theoretical_loss": 3.440374418753497,
      "tokens_seen": 1931739136
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00042175757575757577,
      "loss": 1.4158,
      "theoretical_loss": 3.4403646207140666,
      "tokens_seen": 1931804672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0004217373737373738,
      "loss": 1.5499,
      "theoretical_loss": 3.440354823100094,
      "tokens_seen": 1931870208
    },
    {
      "epoch": 0.16,
      "objective/train/aligned_token_count": 0.4935239553451538,
      "objective/train/avg_token_score": 0.006382493767887354,
      "objective/train/docs_used": 1092471,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.711350440979004,
      "objective/train/original_loss": 3.7253687381744385,
      "objective/train/theoretical_loss": 3.4403450259115465,
      "objective/train/tokens_used": 290794976,
      "objective/train/unlikelihood_loss": 0.1075059249997139,
      "theoretical_loss": 3.4403450259115465,
      "tokens_seen": 1931935744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042171717171717173,
      "loss": 1.5274,
      "theoretical_loss": 3.4403450259115465,
      "tokens_seen": 1931935744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004216969696969697,
      "loss": 1.6173,
      "theoretical_loss": 3.4403352291483906,
      "tokens_seen": 1932001280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042167676767676765,
      "loss": 1.55,
      "theoretical_loss": 3.440325432810594,
      "tokens_seen": 1932066816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042165656565656566,
      "loss": 1.4703,
      "theoretical_loss": 3.4403156368981236,
      "tokens_seen": 1932132352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042163636363636367,
      "loss": 1.3505,
      "theoretical_loss": 3.4403058414109458,
      "tokens_seen": 1932197888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004216161616161616,
      "loss": 1.3519,
      "theoretical_loss": 3.4402960463490286,
      "tokens_seen": 1932263424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215959595959596,
      "loss": 1.5252,
      "theoretical_loss": 3.4402862517123394,
      "tokens_seen": 1932328960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042157575757575754,
      "loss": 1.5055,
      "theoretical_loss": 3.440276457500844,
      "tokens_seen": 1932394496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215555555555556,
      "loss": 1.6192,
      "theoretical_loss": 3.440266663714511,
      "tokens_seen": 1932460032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042153535353535356,
      "loss": 1.56,
      "theoretical_loss": 3.440256870353307,
      "tokens_seen": 1932525568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004215151515151515,
      "loss": 1.5159,
      "theoretical_loss": 3.4402470774171987,
      "tokens_seen": 1932591104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214949494949495,
      "loss": 1.418,
      "theoretical_loss": 3.4402372849061535,
      "tokens_seen": 1932656640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214747474747475,
      "loss": 1.3176,
      "theoretical_loss": 3.4402274928201386,
      "tokens_seen": 1932722176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214545454545455,
      "loss": 1.3761,
      "theoretical_loss": 3.4402177011591206,
      "tokens_seen": 1932787712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042143434343434345,
      "loss": 1.3156,
      "theoretical_loss": 3.4402079099230676,
      "tokens_seen": 1932853248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004214141414141414,
      "loss": 1.4477,
      "theoretical_loss": 3.4401981191119466,
      "tokens_seen": 1932918784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042139393939393937,
      "loss": 1.3894,
      "theoretical_loss": 3.440188328725724,
      "tokens_seen": 1932984320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042137373737373743,
      "loss": 1.4207,
      "theoretical_loss": 3.4401785387643677,
      "tokens_seen": 1933049856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213535353535354,
      "loss": 1.3732,
      "theoretical_loss": 3.440168749227844,
      "tokens_seen": 1933115392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042133333333333335,
      "loss": 1.4545,
      "theoretical_loss": 3.440158960116121,
      "tokens_seen": 1933180928
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004213131313131313,
      "loss": 1.4779,
      "theoretical_loss": 3.4401491714291654,
      "tokens_seen": 1933246464
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042129292929292926,
      "loss": 1.3513,
      "theoretical_loss": 3.4401393831669447,
      "tokens_seen": 1933312000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004212727272727273,
      "loss": 1.4636,
      "theoretical_loss": 3.440129595329425,
      "tokens_seen": 1933377536
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004212525252525253,
      "loss": 1.3884,
      "theoretical_loss": 3.440119807916575,
      "tokens_seen": 1933443072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042123232323232324,
      "loss": 1.3359,
      "theoretical_loss": 3.440110020928361,
      "tokens_seen": 1933508608
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.2790200412273407,
      "objective/train/avg_token_score": 0.020743899047374725,
      "objective/train/docs_used": 1092955,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.856033444404602,
      "objective/train/original_loss": 4.6658196449279785,
      "objective/train/theoretical_loss": 3.440100234364751,
      "objective/train/tokens_used": 292433376,
      "objective/train/unlikelihood_loss": 0.06594061106443405,
      "theoretical_loss": 3.440100234364751,
      "tokens_seen": 1933574144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004212121212121212,
      "loss": 1.2959,
      "theoretical_loss": 3.440100234364751,
      "tokens_seen": 1933574144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211919191919192,
      "loss": 1.3046,
      "theoretical_loss": 3.4400904482257104,
      "tokens_seen": 1933639680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211717171717172,
      "loss": 1.4219,
      "theoretical_loss": 3.4400806625112086,
      "tokens_seen": 1933705216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211515151515152,
      "loss": 1.3372,
      "theoretical_loss": 3.440070877221211,
      "tokens_seen": 1933770752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042113131313131313,
      "loss": 1.5524,
      "theoretical_loss": 3.440061092355686,
      "tokens_seen": 1933836288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004211111111111111,
      "loss": 1.4611,
      "theoretical_loss": 3.4400513079146,
      "tokens_seen": 1933901824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004210909090909091,
      "loss": 1.4109,
      "theoretical_loss": 3.440041523897921,
      "tokens_seen": 1933967360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004210707070707071,
      "loss": 1.488,
      "theoretical_loss": 3.4400317403056153,
      "tokens_seen": 1934032896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042105050505050507,
      "loss": 1.3147,
      "theoretical_loss": 3.4400219571376507,
      "tokens_seen": 1934098432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000421030303030303,
      "loss": 1.5583,
      "theoretical_loss": 3.4400121743939946,
      "tokens_seen": 1934163968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042101010101010104,
      "loss": 1.6352,
      "theoretical_loss": 3.4400023920746134,
      "tokens_seen": 1934229504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420989898989899,
      "loss": 1.4111,
      "theoretical_loss": 3.439992610179475,
      "tokens_seen": 1934295040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420969696969697,
      "loss": 1.7752,
      "theoretical_loss": 3.4399828287085468,
      "tokens_seen": 1934360576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042094949494949496,
      "loss": 1.4621,
      "theoretical_loss": 3.4399730476617956,
      "tokens_seen": 1934426112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004209292929292929,
      "loss": 1.3782,
      "theoretical_loss": 3.4399632670391886,
      "tokens_seen": 1934491648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042090909090909093,
      "loss": 1.3861,
      "theoretical_loss": 3.439953486840693,
      "tokens_seen": 1934557184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208888888888889,
      "loss": 1.51,
      "theoretical_loss": 3.4399437070662766,
      "tokens_seen": 1934622720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208686868686869,
      "loss": 1.2562,
      "theoretical_loss": 3.439933927715906,
      "tokens_seen": 1934688256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042084848484848485,
      "loss": 1.3318,
      "theoretical_loss": 3.439924148789549,
      "tokens_seen": 1934753792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042082828282828286,
      "loss": 1.4035,
      "theoretical_loss": 3.4399143702871724,
      "tokens_seen": 1934819328
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004208080808080808,
      "loss": 1.5151,
      "theoretical_loss": 3.4399045922087437,
      "tokens_seen": 1934884864
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207878787878788,
      "loss": 1.4422,
      "theoretical_loss": 3.4398948145542296,
      "tokens_seen": 1934950400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207676767676768,
      "loss": 1.4189,
      "theoretical_loss": 3.4398850373235987,
      "tokens_seen": 1935015936
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042074747474747475,
      "loss": 1.4982,
      "theoretical_loss": 3.439875260516817,
      "tokens_seen": 1935081472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042072727272727276,
      "loss": 1.2191,
      "theoretical_loss": 3.439865484133852,
      "tokens_seen": 1935147008
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.492118775844574,
      "objective/train/avg_token_score": 0.013941711746156216,
      "objective/train/docs_used": 1093966,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5089529752731323,
      "objective/train/original_loss": 4.055489540100098,
      "objective/train/theoretical_loss": 3.4398557081746715,
      "objective/train/tokens_used": 294071776,
      "objective/train/unlikelihood_loss": 0.046310193836688995,
      "theoretical_loss": 3.4398557081746715,
      "tokens_seen": 1935212544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004207070707070707,
      "loss": 1.4559,
      "theoretical_loss": 3.4398557081746715,
      "tokens_seen": 1935212544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042068686868686867,
      "loss": 1.6231,
      "theoretical_loss": 3.439845932639242,
      "tokens_seen": 1935278080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004206666666666667,
      "loss": 1.3089,
      "theoretical_loss": 3.4398361575275316,
      "tokens_seen": 1935343616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042064646464646464,
      "loss": 1.6297,
      "theoretical_loss": 3.439826382839507,
      "tokens_seen": 1935409152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042062626262626265,
      "loss": 1.5352,
      "theoretical_loss": 3.4398166085751356,
      "tokens_seen": 1935474688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004206060606060606,
      "loss": 1.6264,
      "theoretical_loss": 3.4398068347343855,
      "tokens_seen": 1935540224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042058585858585856,
      "loss": 1.6027,
      "theoretical_loss": 3.439797061317223,
      "tokens_seen": 1935605760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205656565656566,
      "loss": 1.5475,
      "theoretical_loss": 3.439787288323615,
      "tokens_seen": 1935671296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205454545454546,
      "loss": 1.3246,
      "theoretical_loss": 3.43977751575353,
      "tokens_seen": 1935736832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042052525252525254,
      "loss": 1.4094,
      "theoretical_loss": 3.439767743606935,
      "tokens_seen": 1935802368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004205050505050505,
      "loss": 1.2892,
      "theoretical_loss": 3.439757971883797,
      "tokens_seen": 1935867904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042048484848484846,
      "loss": 1.4243,
      "theoretical_loss": 3.4397482005840834,
      "tokens_seen": 1935933440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042046464646464647,
      "loss": 1.6329,
      "theoretical_loss": 3.439738429707762,
      "tokens_seen": 1935998976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004204444444444445,
      "loss": 1.4389,
      "theoretical_loss": 3.439728659254799,
      "tokens_seen": 1936064512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042042424242424244,
      "loss": 1.3894,
      "theoretical_loss": 3.439718889225163,
      "tokens_seen": 1936130048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004204040404040404,
      "loss": 1.3502,
      "theoretical_loss": 3.4397091196188203,
      "tokens_seen": 1936195584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042038383838383835,
      "loss": 1.6292,
      "theoretical_loss": 3.4396993504357387,
      "tokens_seen": 1936261120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004203636363636364,
      "loss": 1.271,
      "theoretical_loss": 3.4396895816758857,
      "tokens_seen": 1936326656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042034343434343437,
      "loss": 1.6302,
      "theoretical_loss": 3.439679813339229,
      "tokens_seen": 1936392192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042032323232323233,
      "loss": 1.3143,
      "theoretical_loss": 3.439670045425734,
      "tokens_seen": 1936457728
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004203030303030303,
      "loss": 1.597,
      "theoretical_loss": 3.4396602779353707,
      "tokens_seen": 1936523264
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042028282828282824,
      "loss": 1.5688,
      "theoretical_loss": 3.4396505108681046,
      "tokens_seen": 1936588800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004202626262626263,
      "loss": 1.5527,
      "theoretical_loss": 3.4396407442239036,
      "tokens_seen": 1936654336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042024242424242426,
      "loss": 1.5146,
      "theoretical_loss": 3.4396309780027354,
      "tokens_seen": 1936719872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004202222222222222,
      "loss": 1.5646,
      "theoretical_loss": 3.4396212122045666,
      "tokens_seen": 1936785408
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.5265762209892273,
      "objective/train/avg_token_score": 0.011445840820670128,
      "objective/train/docs_used": 1094637,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9047353267669678,
      "objective/train/original_loss": 3.971524477005005,
      "objective/train/theoretical_loss": 3.4396114468293657,
      "objective/train/tokens_used": 295710176,
      "objective/train/unlikelihood_loss": 0.05321503430604935,
      "theoretical_loss": 3.4396114468293657,
      "tokens_seen": 1936850944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004202020202020202,
      "loss": 1.4187,
      "theoretical_loss": 3.4396114468293657,
      "tokens_seen": 1936850944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004201818181818182,
      "loss": 1.3045,
      "theoretical_loss": 3.4396016818770994,
      "tokens_seen": 1936916480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004201616161616162,
      "loss": 1.5315,
      "theoretical_loss": 3.4395919173477347,
      "tokens_seen": 1936982016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042014141414141416,
      "loss": 1.3532,
      "theoretical_loss": 3.43958215324124,
      "tokens_seen": 1937047552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004201212121212121,
      "loss": 1.4987,
      "theoretical_loss": 3.439572389557581,
      "tokens_seen": 1937113088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042010101010101007,
      "loss": 1.5557,
      "theoretical_loss": 3.439562626296727,
      "tokens_seen": 1937178624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200808080808081,
      "loss": 1.4468,
      "theoretical_loss": 3.439552863458644,
      "tokens_seen": 1937244160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004200606060606061,
      "loss": 1.6046,
      "theoretical_loss": 3.4395431010433,
      "tokens_seen": 1937309696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042004040404040405,
      "loss": 1.3819,
      "theoretical_loss": 3.439533339050662,
      "tokens_seen": 1937375232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000420020202020202,
      "loss": 1.4179,
      "theoretical_loss": 3.4395235774806983,
      "tokens_seen": 1937440768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00042,
      "loss": 1.4089,
      "theoretical_loss": 3.439513816333376,
      "tokens_seen": 1937506304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000419979797979798,
      "loss": 1.379,
      "theoretical_loss": 3.439504055608661,
      "tokens_seen": 1937571840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000419959595959596,
      "loss": 1.4083,
      "theoretical_loss": 3.439494295306523,
      "tokens_seen": 1937637376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041993939393939394,
      "loss": 1.2468,
      "theoretical_loss": 3.439484535426928,
      "tokens_seen": 1937702912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004199191919191919,
      "loss": 1.345,
      "theoretical_loss": 3.4394747759698436,
      "tokens_seen": 1937768448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198989898989899,
      "loss": 1.2737,
      "theoretical_loss": 3.439465016935238,
      "tokens_seen": 1937833984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041987878787878787,
      "loss": 1.4496,
      "theoretical_loss": 3.4394552583230773,
      "tokens_seen": 1937899520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004198585858585859,
      "loss": 1.3009,
      "theoretical_loss": 3.4394455001333295,
      "tokens_seen": 1937965056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041983838383838383,
      "loss": 1.6033,
      "theoretical_loss": 3.439435742365963,
      "tokens_seen": 1938030592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041981818181818185,
      "loss": 1.4372,
      "theoretical_loss": 3.439425985020944,
      "tokens_seen": 1938096128
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004197979797979798,
      "loss": 1.5184,
      "theoretical_loss": 3.43941622809824,
      "tokens_seen": 1938161664
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004197777777777778,
      "loss": 1.6132,
      "theoretical_loss": 3.439406471597819,
      "tokens_seen": 1938227200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041975757575757577,
      "loss": 1.4743,
      "theoretical_loss": 3.4393967155196483,
      "tokens_seen": 1938292736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041973737373737373,
      "loss": 1.5898,
      "theoretical_loss": 3.439386959863695,
      "tokens_seen": 1938358272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041971717171717174,
      "loss": 1.5563,
      "theoretical_loss": 3.439377204629927,
      "tokens_seen": 1938423808
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.22458454966545105,
      "objective/train/avg_token_score": 0.036178138107061386,
      "objective/train/docs_used": 1095318,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.6959583759307861,
      "objective/train/original_loss": 4.687398910522461,
      "objective/train/theoretical_loss": 3.4393674498183118,
      "objective/train/tokens_used": 297348576,
      "objective/train/unlikelihood_loss": 0.08731012046337128,
      "theoretical_loss": 3.4393674498183118,
      "tokens_seen": 1938489344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196969696969697,
      "loss": 1.2169,
      "theoretical_loss": 3.4393674498183118,
      "tokens_seen": 1938489344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196767676767677,
      "loss": 1.2217,
      "theoretical_loss": 3.439357695428816,
      "tokens_seen": 1938554880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041965656565656566,
      "loss": 1.437,
      "theoretical_loss": 3.4393479414614085,
      "tokens_seen": 1938620416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004196363636363637,
      "loss": 1.5371,
      "theoretical_loss": 3.4393381879160554,
      "tokens_seen": 1938685952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041961616161616163,
      "loss": 1.417,
      "theoretical_loss": 3.439328434792725,
      "tokens_seen": 1938751488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195959595959596,
      "loss": 1.4628,
      "theoretical_loss": 3.439318682091385,
      "tokens_seen": 1938817024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195757575757576,
      "loss": 1.4443,
      "theoretical_loss": 3.4393089298120016,
      "tokens_seen": 1938882560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041955555555555556,
      "loss": 1.3587,
      "theoretical_loss": 3.4392991779545437,
      "tokens_seen": 1938948096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041953535353535357,
      "loss": 1.497,
      "theoretical_loss": 3.439289426518978,
      "tokens_seen": 1939013632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004195151515151515,
      "loss": 1.6095,
      "theoretical_loss": 3.4392796755052712,
      "tokens_seen": 1939079168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194949494949495,
      "loss": 1.4658,
      "theoretical_loss": 3.439269924913393,
      "tokens_seen": 1939144704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194747474747475,
      "loss": 1.5622,
      "theoretical_loss": 3.439260174743309,
      "tokens_seen": 1939210240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194545454545455,
      "loss": 1.5161,
      "theoretical_loss": 3.439250424994988,
      "tokens_seen": 1939275776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041943434343434346,
      "loss": 1.4962,
      "theoretical_loss": 3.4392406756683958,
      "tokens_seen": 1939341312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004194141414141414,
      "loss": 1.4209,
      "theoretical_loss": 3.4392309267635017,
      "tokens_seen": 1939406848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193939393939394,
      "loss": 1.2259,
      "theoretical_loss": 3.439221178280272,
      "tokens_seen": 1939472384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193737373737374,
      "loss": 1.4304,
      "theoretical_loss": 3.439211430218675,
      "tokens_seen": 1939537920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193535353535354,
      "loss": 1.4592,
      "theoretical_loss": 3.439201682578678,
      "tokens_seen": 1939603456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041933333333333335,
      "loss": 1.6319,
      "theoretical_loss": 3.439191935360249,
      "tokens_seen": 1939668992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004193131313131313,
      "loss": 1.351,
      "theoretical_loss": 3.439182188563354,
      "tokens_seen": 1939734528
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041929292929292927,
      "loss": 1.6761,
      "theoretical_loss": 3.4391724421879615,
      "tokens_seen": 1939800064
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041927272727272733,
      "loss": 1.5499,
      "theoretical_loss": 3.4391626962340394,
      "tokens_seen": 1939865600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004192525252525253,
      "loss": 1.5205,
      "theoretical_loss": 3.439152950701555,
      "tokens_seen": 1939931136
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041923232323232325,
      "loss": 1.4779,
      "theoretical_loss": 3.439143205590475,
      "tokens_seen": 1939996672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004192121212121212,
      "loss": 1.4338,
      "theoretical_loss": 3.439133460900768,
      "tokens_seen": 1940062208
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.47330155968666077,
      "objective/train/avg_token_score": 0.009856240823864937,
      "objective/train/docs_used": 1096386,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5892785787582397,
      "objective/train/original_loss": 4.016402721405029,
      "objective/train/theoretical_loss": 3.439123716632401,
      "objective/train/tokens_used": 298986976,
      "objective/train/unlikelihood_loss": 0.07871534675359726,
      "theoretical_loss": 3.439123716632401,
      "tokens_seen": 1940127744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041919191919191916,
      "loss": 1.7263,
      "theoretical_loss": 3.439123716632401,
      "tokens_seen": 1940127744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191717171717172,
      "loss": 1.4301,
      "theoretical_loss": 3.439113972785342,
      "tokens_seen": 1940193280
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191515151515152,
      "loss": 1.4758,
      "theoretical_loss": 3.4391042293595584,
      "tokens_seen": 1940258816
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041913131313131314,
      "loss": 1.4596,
      "theoretical_loss": 3.4390944863550175,
      "tokens_seen": 1940324352
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004191111111111111,
      "loss": 1.4683,
      "theoretical_loss": 3.4390847437716863,
      "tokens_seen": 1940389888
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041909090909090905,
      "loss": 1.3133,
      "theoretical_loss": 3.4390750016095337,
      "tokens_seen": 1940455424
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190707070707071,
      "loss": 1.6132,
      "theoretical_loss": 3.439065259868527,
      "tokens_seen": 1940520960
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004190505050505051,
      "loss": 1.4538,
      "theoretical_loss": 3.4390555185486327,
      "tokens_seen": 1940586496
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041903030303030303,
      "loss": 1.4229,
      "theoretical_loss": 3.4390457776498193,
      "tokens_seen": 1940652032
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000419010101010101,
      "loss": 1.363,
      "theoretical_loss": 3.4390360371720545,
      "tokens_seen": 1940717568
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418989898989899,
      "loss": 1.4503,
      "theoretical_loss": 3.439026297115305,
      "tokens_seen": 1940783104
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418969696969697,
      "loss": 1.3993,
      "theoretical_loss": 3.4390165574795395,
      "tokens_seen": 1940848640
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041894949494949497,
      "loss": 1.4746,
      "theoretical_loss": 3.4390068182647244,
      "tokens_seen": 1940914176
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189292929292929,
      "loss": 1.5665,
      "theoretical_loss": 3.4389970794708287,
      "tokens_seen": 1940979712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004189090909090909,
      "loss": 1.6576,
      "theoretical_loss": 3.4389873410978185,
      "tokens_seen": 1941045248
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004188888888888889,
      "loss": 1.4832,
      "theoretical_loss": 3.4389776031456623,
      "tokens_seen": 1941110784
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004188686868686869,
      "loss": 1.1606,
      "theoretical_loss": 3.4389678656143277,
      "tokens_seen": 1941176320
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041884848484848486,
      "loss": 1.5457,
      "theoretical_loss": 3.4389581285037822,
      "tokens_seen": 1941241856
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004188282828282828,
      "loss": 1.2376,
      "theoretical_loss": 3.4389483918139927,
      "tokens_seen": 1941307392
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041880808080808083,
      "loss": 1.5873,
      "theoretical_loss": 3.438938655544928,
      "tokens_seen": 1941372928
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187878787878788,
      "loss": 1.4951,
      "theoretical_loss": 3.438928919696555,
      "tokens_seen": 1941438464
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187676767676768,
      "loss": 1.655,
      "theoretical_loss": 3.4389191842688414,
      "tokens_seen": 1941504000
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041874747474747475,
      "loss": 1.3632,
      "theoretical_loss": 3.438909449261755,
      "tokens_seen": 1941569536
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187272727272727,
      "loss": 1.4523,
      "theoretical_loss": 3.4388997146752636,
      "tokens_seen": 1941635072
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004187070707070707,
      "loss": 1.3786,
      "theoretical_loss": 3.4388899805093343,
      "tokens_seen": 1941700608
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.43194037675857544,
      "objective/train/avg_token_score": 0.020467616617679596,
      "objective/train/docs_used": 1097742,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5595964193344116,
      "objective/train/original_loss": 4.220158576965332,
      "objective/train/theoretical_loss": 3.438880246763935,
      "objective/train/tokens_used": 300625376,
      "objective/train/unlikelihood_loss": 0.06359304487705231,
      "theoretical_loss": 3.438880246763935,
      "tokens_seen": 1941766144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186868686868687,
      "loss": 1.4585,
      "theoretical_loss": 3.438880246763935,
      "tokens_seen": 1941766144
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186666666666667,
      "loss": 1.3775,
      "theoretical_loss": 3.438870513439033,
      "tokens_seen": 1941831680
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041864646464646464,
      "loss": 1.4134,
      "theoretical_loss": 3.438860780534597,
      "tokens_seen": 1941897216
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041862626262626266,
      "loss": 1.2068,
      "theoretical_loss": 3.438851048050594,
      "tokens_seen": 1941962752
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004186060606060606,
      "loss": 1.6119,
      "theoretical_loss": 3.4388413159869913,
      "tokens_seen": 1942028288
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041858585858585857,
      "loss": 1.4467,
      "theoretical_loss": 3.438831584343757,
      "tokens_seen": 1942093824
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004185656565656566,
      "loss": 1.4658,
      "theoretical_loss": 3.4388218531208583,
      "tokens_seen": 1942159360
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041854545454545454,
      "loss": 1.5447,
      "theoretical_loss": 3.438812122318263,
      "tokens_seen": 1942224896
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041852525252525255,
      "loss": 1.5643,
      "theoretical_loss": 3.4388023919359396,
      "tokens_seen": 1942290432
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004185050505050505,
      "loss": 1.305,
      "theoretical_loss": 3.4387926619738547,
      "tokens_seen": 1942355968
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041848484848484846,
      "loss": 1.4774,
      "theoretical_loss": 3.438782932431976,
      "tokens_seen": 1942421504
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184646464646465,
      "loss": 1.4883,
      "theoretical_loss": 3.438773203310272,
      "tokens_seen": 1942487040
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184444444444445,
      "loss": 1.7094,
      "theoretical_loss": 3.4387634746087103,
      "tokens_seen": 1942552576
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041842424242424244,
      "loss": 1.3692,
      "theoretical_loss": 3.4387537463272575,
      "tokens_seen": 1942618112
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004184040404040404,
      "loss": 1.5318,
      "theoretical_loss": 3.438744018465882,
      "tokens_seen": 1942683648
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041838383838383835,
      "loss": 1.2879,
      "theoretical_loss": 3.438734291024552,
      "tokens_seen": 1942749184
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041836363636363637,
      "loss": 1.6327,
      "theoretical_loss": 3.438724564003235,
      "tokens_seen": 1942814720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183434343434344,
      "loss": 1.2888,
      "theoretical_loss": 3.438714837401897,
      "tokens_seen": 1942880256
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041832323232323233,
      "loss": 1.5383,
      "theoretical_loss": 3.4387051112205085,
      "tokens_seen": 1942945792
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004183030303030303,
      "loss": 1.4541,
      "theoretical_loss": 3.4386953854590354,
      "tokens_seen": 1943011328
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041828282828282825,
      "loss": 1.4903,
      "theoretical_loss": 3.4386856601174456,
      "tokens_seen": 1943076864
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182626262626263,
      "loss": 1.3577,
      "theoretical_loss": 3.4386759351957066,
      "tokens_seen": 1943142400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041824242424242427,
      "loss": 1.4549,
      "theoretical_loss": 3.4386662106937873,
      "tokens_seen": 1943207936
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182222222222222,
      "loss": 1.5026,
      "theoretical_loss": 3.438656486611654,
      "tokens_seen": 1943273472
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004182020202020202,
      "loss": 1.4573,
      "theoretical_loss": 3.4386467629492756,
      "tokens_seen": 1943339008
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.38086509704589844,
      "objective/train/avg_token_score": 0.017717378214001656,
      "objective/train/docs_used": 1098461,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2602291107177734,
      "objective/train/original_loss": 4.616687297821045,
      "objective/train/theoretical_loss": 3.4386370397066193,
      "objective/train/tokens_used": 302263776,
      "objective/train/unlikelihood_loss": 0.05086660385131836,
      "theoretical_loss": 3.4386370397066193,
      "tokens_seen": 1943404544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041818181818181814,
      "loss": 1.4643,
      "theoretical_loss": 3.4386370397066193,
      "tokens_seen": 1943404544
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181616161616162,
      "loss": 1.4196,
      "theoretical_loss": 3.4386273168836525,
      "tokens_seen": 1943470080
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041814141414141416,
      "loss": 1.3719,
      "theoretical_loss": 3.4386175944803434,
      "tokens_seen": 1943535616
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181212121212121,
      "loss": 1.6352,
      "theoretical_loss": 3.4386078724966596,
      "tokens_seen": 1943601152
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004181010101010101,
      "loss": 1.4791,
      "theoretical_loss": 3.438598150932569,
      "tokens_seen": 1943666688
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041808080808080814,
      "loss": 1.2838,
      "theoretical_loss": 3.4385884297880387,
      "tokens_seen": 1943732224
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004180606060606061,
      "loss": 1.6374,
      "theoretical_loss": 3.4385787090630373,
      "tokens_seen": 1943797760
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041804040404040406,
      "loss": 1.5597,
      "theoretical_loss": 3.438568988757532,
      "tokens_seen": 1943863296
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000418020202020202,
      "loss": 1.3555,
      "theoretical_loss": 3.438559268871491,
      "tokens_seen": 1943928832
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041799999999999997,
      "loss": 1.5381,
      "theoretical_loss": 3.4385495494048817,
      "tokens_seen": 1943994368
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041797979797979803,
      "loss": 1.6316,
      "theoretical_loss": 3.4385398303576715,
      "tokens_seen": 1944059904
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417959595959596,
      "loss": 1.5205,
      "theoretical_loss": 3.438530111729829,
      "tokens_seen": 1944125440
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041793939393939395,
      "loss": 1.4276,
      "theoretical_loss": 3.4385203935213218,
      "tokens_seen": 1944190976
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004179191919191919,
      "loss": 1.4871,
      "theoretical_loss": 3.4385106757321173,
      "tokens_seen": 1944256512
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178989898989899,
      "loss": 1.406,
      "theoretical_loss": 3.438500958362183,
      "tokens_seen": 1944322048
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178787878787879,
      "loss": 1.6157,
      "theoretical_loss": 3.438491241411487,
      "tokens_seen": 1944387584
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178585858585859,
      "loss": 1.3643,
      "theoretical_loss": 3.438481524879998,
      "tokens_seen": 1944453120
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041783838383838384,
      "loss": 1.5163,
      "theoretical_loss": 3.438471808767683,
      "tokens_seen": 1944518656
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004178181818181818,
      "loss": 1.5139,
      "theoretical_loss": 3.4384620930745093,
      "tokens_seen": 1944584192
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177979797979798,
      "loss": 1.6269,
      "theoretical_loss": 3.438452377800445,
      "tokens_seen": 1944649728
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177777777777778,
      "loss": 1.3211,
      "theoretical_loss": 3.4384426629454587,
      "tokens_seen": 1944715264
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004177575757575758,
      "loss": 1.4293,
      "theoretical_loss": 3.438432948509517,
      "tokens_seen": 1944780800
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041773737373737373,
      "loss": 1.4411,
      "theoretical_loss": 3.4384232344925882,
      "tokens_seen": 1944846336
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041771717171717174,
      "loss": 1.4302,
      "theoretical_loss": 3.438413520894641,
      "tokens_seen": 1944911872
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176969696969697,
      "loss": 1.5839,
      "theoretical_loss": 3.438403807715641,
      "tokens_seen": 1944977408
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.33687683939933777,
      "objective/train/avg_token_score": 0.05385596305131912,
      "objective/train/docs_used": 1099474,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2468277215957642,
      "objective/train/original_loss": 4.72367000579834,
      "objective/train/theoretical_loss": 3.4383940949555587,
      "objective/train/tokens_used": 303902176,
      "objective/train/unlikelihood_loss": 0.06228525936603546,
      "theoretical_loss": 3.4383940949555587,
      "tokens_seen": 1945042944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176767676767677,
      "loss": 1.3905,
      "theoretical_loss": 3.4383940949555587,
      "tokens_seen": 1945042944
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041765656565656567,
      "loss": 1.418,
      "theoretical_loss": 3.43838438261436,
      "tokens_seen": 1945108480
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004176363636363636,
      "loss": 1.3343,
      "theoretical_loss": 3.4383746706920135,
      "tokens_seen": 1945174016
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041761616161616164,
      "loss": 1.3567,
      "theoretical_loss": 3.438364959188487,
      "tokens_seen": 1945239552
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175959595959596,
      "loss": 1.467,
      "theoretical_loss": 3.438355248103748,
      "tokens_seen": 1945305088
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004175757575757576,
      "loss": 1.5035,
      "theoretical_loss": 3.4383455374377645,
      "tokens_seen": 1945370624
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041755555555555556,
      "loss": 1.5008,
      "theoretical_loss": 3.4383358271905045,
      "tokens_seen": 1945436160
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041753535353535357,
      "loss": 1.2581,
      "theoretical_loss": 3.4383261173619353,
      "tokens_seen": 1945501696
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041751515151515153,
      "loss": 1.6451,
      "theoretical_loss": 3.438316407952026,
      "tokens_seen": 1945567232
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174949494949495,
      "loss": 1.4594,
      "theoretical_loss": 3.438306698960743,
      "tokens_seen": 1945632768
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174747474747475,
      "loss": 1.5545,
      "theoretical_loss": 3.4382969903880545,
      "tokens_seen": 1945698304
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041745454545454545,
      "loss": 1.6054,
      "theoretical_loss": 3.4382872822339294,
      "tokens_seen": 1945763840
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041743434343434347,
      "loss": 1.6339,
      "theoretical_loss": 3.4382775744983345,
      "tokens_seen": 1945829376
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004174141414141414,
      "loss": 1.3249,
      "theoretical_loss": 3.4382678671812377,
      "tokens_seen": 1945894912
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173939393939394,
      "loss": 1.4215,
      "theoretical_loss": 3.438258160282607,
      "tokens_seen": 1945960448
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173737373737374,
      "loss": 1.6104,
      "theoretical_loss": 3.4382484538024105,
      "tokens_seen": 1946025984
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041735353535353535,
      "loss": 1.3356,
      "theoretical_loss": 3.438238747740616,
      "tokens_seen": 1946091520
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041733333333333336,
      "loss": 1.1932,
      "theoretical_loss": 3.4382290420971913,
      "tokens_seen": 1946157056
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004173131313131313,
      "loss": 1.4409,
      "theoretical_loss": 3.438219336872104,
      "tokens_seen": 1946222592
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041729292929292927,
      "loss": 1.5784,
      "theoretical_loss": 3.4382096320653224,
      "tokens_seen": 1946288128
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004172727272727273,
      "loss": 1.7057,
      "theoretical_loss": 3.4381999276768145,
      "tokens_seen": 1946353664
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004172525252525253,
      "loss": 1.2891,
      "theoretical_loss": 3.438190223706548,
      "tokens_seen": 1946419200
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041723232323232325,
      "loss": 1.4872,
      "theoretical_loss": 3.43818052015449,
      "tokens_seen": 1946484736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004172121212121212,
      "loss": 1.4779,
      "theoretical_loss": 3.43817081702061,
      "tokens_seen": 1946550272
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041719191919191916,
      "loss": 1.3737,
      "theoretical_loss": 3.4381611143048745,
      "tokens_seen": 1946615808
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.4748900234699249,
      "objective/train/avg_token_score": 0.007229257840663195,
      "objective/train/docs_used": 1099984,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5736761093139648,
      "objective/train/original_loss": 4.259389400482178,
      "objective/train/theoretical_loss": 3.4381514120072523,
      "objective/train/tokens_used": 305540576,
      "objective/train/unlikelihood_loss": 0.05011461675167084,
      "theoretical_loss": 3.4381514120072523,
      "tokens_seen": 1946681344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171717171717172,
      "loss": 1.6191,
      "theoretical_loss": 3.4381514120072523,
      "tokens_seen": 1946681344
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171515151515152,
      "loss": 1.5331,
      "theoretical_loss": 3.4381417101277107,
      "tokens_seen": 1946746880
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041713131313131314,
      "loss": 1.5493,
      "theoretical_loss": 3.438132008666218,
      "tokens_seen": 1946812416
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004171111111111111,
      "loss": 1.3735,
      "theoretical_loss": 3.438122307622742,
      "tokens_seen": 1946877952
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041709090909090906,
      "loss": 1.5348,
      "theoretical_loss": 3.43811260699725,
      "tokens_seen": 1946943488
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170707070707071,
      "loss": 1.5874,
      "theoretical_loss": 3.4381029067897115,
      "tokens_seen": 1947009024
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004170505050505051,
      "loss": 1.4854,
      "theoretical_loss": 3.438093207000093,
      "tokens_seen": 1947074560
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041703030303030304,
      "loss": 1.6018,
      "theoretical_loss": 3.4380835076283622,
      "tokens_seen": 1947140096
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000417010101010101,
      "loss": 1.3937,
      "theoretical_loss": 3.4380738086744884,
      "tokens_seen": 1947205632
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041698989898989895,
      "loss": 1.3372,
      "theoretical_loss": 3.4380641101384386,
      "tokens_seen": 1947271168
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.000416969696969697,
      "loss": 1.4951,
      "theoretical_loss": 3.4380544120201812,
      "tokens_seen": 1947336704
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041694949494949497,
      "loss": 1.2962,
      "theoretical_loss": 3.438044714319684,
      "tokens_seen": 1947402240
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041692929292929293,
      "loss": 1.4141,
      "theoretical_loss": 3.4380350170369143,
      "tokens_seen": 1947467776
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004169090909090909,
      "loss": 1.3446,
      "theoretical_loss": 3.438025320171841,
      "tokens_seen": 1947533312
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168888888888889,
      "loss": 1.469,
      "theoretical_loss": 3.438015623724432,
      "tokens_seen": 1947598848
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168686868686869,
      "loss": 1.516,
      "theoretical_loss": 3.4380059276946544,
      "tokens_seen": 1947664384
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041684848484848486,
      "loss": 1.4418,
      "theoretical_loss": 3.437996232082477,
      "tokens_seen": 1947729920
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168282828282828,
      "loss": 1.5376,
      "theoretical_loss": 3.437986536887867,
      "tokens_seen": 1947795456
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004168080808080808,
      "loss": 1.396,
      "theoretical_loss": 3.437976842110793,
      "tokens_seen": 1947860992
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167878787878788,
      "loss": 1.2724,
      "theoretical_loss": 3.4379671477512233,
      "tokens_seen": 1947926528
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167676767676768,
      "loss": 1.306,
      "theoretical_loss": 3.4379574538091244,
      "tokens_seen": 1947992064
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.00041674747474747476,
      "loss": 1.656,
      "theoretical_loss": 3.437947760284466,
      "tokens_seen": 1948057600
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167272727272727,
      "loss": 1.4577,
      "theoretical_loss": 3.437938067177215,
      "tokens_seen": 1948123136
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004167070707070707,
      "loss": 1.3489,
      "theoretical_loss": 3.43792837448734,
      "tokens_seen": 1948188672
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004166868686868687,
      "loss": 1.574,
      "theoretical_loss": 3.437918682214808,
      "tokens_seen": 1948254208
    },
    {
      "epoch": 0.17,
      "objective/train/aligned_token_count": 0.3907013535499573,
      "objective/train/avg_token_score": 0.006235840264707804,
      "objective/train/docs_used": 1101410,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.310508370399475,
      "objective/train/original_loss": 4.164927005767822,
      "objective/train/theoretical_loss": 3.4379089903595883,
      "objective/train/tokens_used": 307178976,
      "objective/train/unlikelihood_loss": 0.08828544616699219,
      "theoretical_loss": 3.4379089903595883,
      "tokens_seen": 1948319744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0004166666666666667,
      "loss": 1.4446,
      "theoretical_loss": 3.4379089903595883,
      "tokens_seen": 1948319744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041664646464646465,
      "loss": 1.258,
      "theoretical_loss": 3.437899298921648,
      "tokens_seen": 1948385280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004166262626262626,
      "loss": 1.2467,
      "theoretical_loss": 3.4378896079009555,
      "tokens_seen": 1948450816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004166060606060606,
      "loss": 1.2485,
      "theoretical_loss": 3.437879917297479,
      "tokens_seen": 1948516352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165858585858586,
      "loss": 1.4547,
      "theoretical_loss": 3.437870227111186,
      "tokens_seen": 1948581888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165656565656566,
      "loss": 1.5381,
      "theoretical_loss": 3.4378605373420443,
      "tokens_seen": 1948647424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041654545454545454,
      "loss": 1.4898,
      "theoretical_loss": 3.4378508479900227,
      "tokens_seen": 1948712960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041652525252525255,
      "loss": 1.6897,
      "theoretical_loss": 3.437841159055089,
      "tokens_seen": 1948778496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004165050505050505,
      "loss": 1.5336,
      "theoretical_loss": 3.4378314705372106,
      "tokens_seen": 1948844032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041648484848484847,
      "loss": 1.4052,
      "theoretical_loss": 3.437821782436356,
      "tokens_seen": 1948909568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164646464646465,
      "loss": 1.5426,
      "theoretical_loss": 3.4378120947524935,
      "tokens_seen": 1948975104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041644444444444444,
      "loss": 1.5688,
      "theoretical_loss": 3.4378024074855906,
      "tokens_seen": 1949040640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041642424242424245,
      "loss": 1.5614,
      "theoretical_loss": 3.437792720635616,
      "tokens_seen": 1949106176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004164040404040404,
      "loss": 1.4658,
      "theoretical_loss": 3.437783034202537,
      "tokens_seen": 1949171712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041638383838383836,
      "loss": 1.407,
      "theoretical_loss": 3.4377733481863215,
      "tokens_seen": 1949237248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041636363636363637,
      "loss": 1.343,
      "theoretical_loss": 3.4377636625869386,
      "tokens_seen": 1949302784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163434343434344,
      "loss": 1.5019,
      "theoretical_loss": 3.4377539774043555,
      "tokens_seen": 1949368320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041632323232323234,
      "loss": 1.4887,
      "theoretical_loss": 3.4377442926385404,
      "tokens_seen": 1949433856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004163030303030303,
      "loss": 1.5129,
      "theoretical_loss": 3.4377346082894618,
      "tokens_seen": 1949499392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162828282828283,
      "loss": 1.3991,
      "theoretical_loss": 3.437724924357087,
      "tokens_seen": 1949564928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041626262626262626,
      "loss": 1.6064,
      "theoretical_loss": 3.437715240841385,
      "tokens_seen": 1949630464
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162424242424243,
      "loss": 1.5792,
      "theoretical_loss": 3.437705557742323,
      "tokens_seen": 1949696000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041622222222222223,
      "loss": 1.6014,
      "theoretical_loss": 3.4376958750598696,
      "tokens_seen": 1949761536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004162020202020202,
      "loss": 1.515,
      "theoretical_loss": 3.437686192793992,
      "tokens_seen": 1949827072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161818181818182,
      "loss": 1.3131,
      "theoretical_loss": 3.4376765109446596,
      "tokens_seen": 1949892608
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.24596774578094482,
      "objective/train/avg_token_score": 0.017320265993475914,
      "objective/train/docs_used": 1102181,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8094052076339722,
      "objective/train/original_loss": 4.91034460067749,
      "objective/train/theoretical_loss": 3.4376668295118398,
      "objective/train/tokens_used": 308817376,
      "objective/train/unlikelihood_loss": 0.06164025887846947,
      "theoretical_loss": 3.4376668295118398,
      "tokens_seen": 1949958144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161616161616162,
      "loss": 1.3284,
      "theoretical_loss": 3.4376668295118398,
      "tokens_seen": 1949958144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041614141414141417,
      "loss": 1.476,
      "theoretical_loss": 3.4376571484955005,
      "tokens_seen": 1950023680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161212121212121,
      "loss": 1.4625,
      "theoretical_loss": 3.4376474678956104,
      "tokens_seen": 1950089216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004161010101010101,
      "loss": 1.2883,
      "theoretical_loss": 3.437637787712137,
      "tokens_seen": 1950154752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004160808080808081,
      "loss": 1.4984,
      "theoretical_loss": 3.4376281079450486,
      "tokens_seen": 1950220288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004160606060606061,
      "loss": 1.6164,
      "theoretical_loss": 3.437618428594313,
      "tokens_seen": 1950285824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041604040404040406,
      "loss": 1.4649,
      "theoretical_loss": 3.4376087496598986,
      "tokens_seen": 1950351360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000416020202020202,
      "loss": 1.5006,
      "theoretical_loss": 3.437599071141774,
      "tokens_seen": 1950416896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000416,
      "loss": 1.4269,
      "theoretical_loss": 3.4375893930399064,
      "tokens_seen": 1950482432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041597979797979804,
      "loss": 1.3195,
      "theoretical_loss": 3.4375797153542647,
      "tokens_seen": 1950547968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415959595959596,
      "loss": 1.5313,
      "theoretical_loss": 3.437570038084816,
      "tokens_seen": 1950613504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041593939393939395,
      "loss": 1.3833,
      "theoretical_loss": 3.4375603612315295,
      "tokens_seen": 1950679040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004159191919191919,
      "loss": 1.5523,
      "theoretical_loss": 3.437550684794373,
      "tokens_seen": 1950744576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041589898989898987,
      "loss": 1.3026,
      "theoretical_loss": 3.4375410087733145,
      "tokens_seen": 1950810112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041587878787878793,
      "loss": 1.1543,
      "theoretical_loss": 3.437531333168322,
      "tokens_seen": 1950875648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004158585858585859,
      "loss": 1.4297,
      "theoretical_loss": 3.4375216579793637,
      "tokens_seen": 1950941184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041583838383838385,
      "loss": 1.3946,
      "theoretical_loss": 3.4375119832064076,
      "tokens_seen": 1951006720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004158181818181818,
      "loss": 1.637,
      "theoretical_loss": 3.4375023088494228,
      "tokens_seen": 1951072256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041579797979797976,
      "loss": 1.4739,
      "theoretical_loss": 3.4374926349083763,
      "tokens_seen": 1951137792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157777777777778,
      "loss": 1.1972,
      "theoretical_loss": 3.4374829613832363,
      "tokens_seen": 1951203328
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157575757575758,
      "loss": 1.2674,
      "theoretical_loss": 3.4374732882739716,
      "tokens_seen": 1951268864
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041573737373737374,
      "loss": 1.2435,
      "theoretical_loss": 3.43746361558055,
      "tokens_seen": 1951334400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004157171717171717,
      "loss": 1.283,
      "theoretical_loss": 3.4374539433029394,
      "tokens_seen": 1951399936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004156969696969697,
      "loss": 1.5095,
      "theoretical_loss": 3.437444271441109,
      "tokens_seen": 1951465472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004156767676767677,
      "loss": 1.5597,
      "theoretical_loss": 3.4374345999950253,
      "tokens_seen": 1951531008
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.5144794583320618,
      "objective/train/avg_token_score": 0.006237310823053122,
      "objective/train/docs_used": 1103575,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6771392822265625,
      "objective/train/original_loss": 3.8277225494384766,
      "objective/train/theoretical_loss": 3.437424928964658,
      "objective/train/tokens_used": 310455776,
      "objective/train/unlikelihood_loss": 0.050984472036361694,
      "theoretical_loss": 3.437424928964658,
      "tokens_seen": 1951596544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004156565656565657,
      "loss": 1.5589,
      "theoretical_loss": 3.437424928964658,
      "tokens_seen": 1951596544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041563636363636363,
      "loss": 1.3017,
      "theoretical_loss": 3.437415258349975,
      "tokens_seen": 1951662080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004156161616161616,
      "loss": 1.5017,
      "theoretical_loss": 3.437405588150943,
      "tokens_seen": 1951727616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004155959595959596,
      "loss": 1.3667,
      "theoretical_loss": 3.437395918367532,
      "tokens_seen": 1951793152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004155757575757576,
      "loss": 1.6398,
      "theoretical_loss": 3.43738624899971,
      "tokens_seen": 1951858688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041555555555555557,
      "loss": 1.5077,
      "theoretical_loss": 3.437376580047444,
      "tokens_seen": 1951924224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004155353535353535,
      "loss": 1.6662,
      "theoretical_loss": 3.437366911510703,
      "tokens_seen": 1951989760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041551515151515154,
      "loss": 1.524,
      "theoretical_loss": 3.437357243389455,
      "tokens_seen": 1952055296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154949494949495,
      "loss": 1.2516,
      "theoretical_loss": 3.437347575683668,
      "tokens_seen": 1952120832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154747474747475,
      "loss": 1.4622,
      "theoretical_loss": 3.437337908393311,
      "tokens_seen": 1952186368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041545454545454546,
      "loss": 1.5639,
      "theoretical_loss": 3.437328241518352,
      "tokens_seen": 1952251904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004154343434343434,
      "loss": 1.4582,
      "theoretical_loss": 3.4373185750587583,
      "tokens_seen": 1952317440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041541414141414143,
      "loss": 1.2993,
      "theoretical_loss": 3.4373089090144986,
      "tokens_seen": 1952382976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153939393939394,
      "loss": 1.689,
      "theoretical_loss": 3.4372992433855414,
      "tokens_seen": 1952448512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153737373737374,
      "loss": 1.5325,
      "theoretical_loss": 3.4372895781718547,
      "tokens_seen": 1952514048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041535353535353535,
      "loss": 1.5222,
      "theoretical_loss": 3.437279913373407,
      "tokens_seen": 1952579584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041533333333333336,
      "loss": 1.2857,
      "theoretical_loss": 3.437270248990166,
      "tokens_seen": 1952645120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004153131313131313,
      "loss": 1.4812,
      "theoretical_loss": 3.4372605850221003,
      "tokens_seen": 1952710656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152929292929293,
      "loss": 1.4163,
      "theoretical_loss": 3.437250921469178,
      "tokens_seen": 1952776192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152727272727273,
      "loss": 1.58,
      "theoretical_loss": 3.437241258331367,
      "tokens_seen": 1952841728
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041525252525252525,
      "loss": 1.5932,
      "theoretical_loss": 3.437231595608636,
      "tokens_seen": 1952907264
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041523232323232326,
      "loss": 1.3163,
      "theoretical_loss": 3.4372219333009535,
      "tokens_seen": 1952972800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004152121212121212,
      "loss": 1.4865,
      "theoretical_loss": 3.437212271408287,
      "tokens_seen": 1953038336
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041519191919191917,
      "loss": 1.3725,
      "theoretical_loss": 3.4372026099306057,
      "tokens_seen": 1953103872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151717171717172,
      "loss": 1.4658,
      "theoretical_loss": 3.4371929488678767,
      "tokens_seen": 1953169408
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.3589320480823517,
      "objective/train/avg_token_score": 0.013197287917137146,
      "objective/train/docs_used": 1104173,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2983180284500122,
      "objective/train/original_loss": 4.760884761810303,
      "objective/train/theoretical_loss": 3.4371832882200692,
      "objective/train/tokens_used": 312094176,
      "objective/train/unlikelihood_loss": 0.052073884755373,
      "theoretical_loss": 3.4371832882200692,
      "tokens_seen": 1953234944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151515151515152,
      "loss": 1.3511,
      "theoretical_loss": 3.4371832882200692,
      "tokens_seen": 1953234944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041513131313131315,
      "loss": 1.3083,
      "theoretical_loss": 3.4371736279871508,
      "tokens_seen": 1953300480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004151111111111111,
      "loss": 1.6078,
      "theoretical_loss": 3.43716396816909,
      "tokens_seen": 1953366016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041509090909090906,
      "loss": 1.5163,
      "theoretical_loss": 3.4371543087658556,
      "tokens_seen": 1953431552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004150707070707071,
      "loss": 1.3672,
      "theoretical_loss": 3.437144649777415,
      "tokens_seen": 1953497088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004150505050505051,
      "loss": 1.3603,
      "theoretical_loss": 3.4371349912037372,
      "tokens_seen": 1953562624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041503030303030304,
      "loss": 1.472,
      "theoretical_loss": 3.43712533304479,
      "tokens_seen": 1953628160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000415010101010101,
      "loss": 1.4868,
      "theoretical_loss": 3.437115675300542,
      "tokens_seen": 1953693696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041498989898989896,
      "loss": 1.3721,
      "theoretical_loss": 3.4371060179709607,
      "tokens_seen": 1953759232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414969696969697,
      "loss": 1.5007,
      "theoretical_loss": 3.4370963610560157,
      "tokens_seen": 1953824768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414949494949495,
      "loss": 1.3869,
      "theoretical_loss": 3.437086704555674,
      "tokens_seen": 1953890304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041492929292929293,
      "loss": 1.5926,
      "theoretical_loss": 3.437077048469905,
      "tokens_seen": 1953955840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004149090909090909,
      "loss": 1.5715,
      "theoretical_loss": 3.437067392798676,
      "tokens_seen": 1954021376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041488888888888885,
      "loss": 1.688,
      "theoretical_loss": 3.4370577375419566,
      "tokens_seen": 1954086912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148686868686869,
      "loss": 1.3722,
      "theoretical_loss": 3.4370480826997136,
      "tokens_seen": 1954152448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041484848484848487,
      "loss": 1.4423,
      "theoretical_loss": 3.437038428271916,
      "tokens_seen": 1954217984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041482828282828283,
      "loss": 1.4912,
      "theoretical_loss": 3.4370287742585326,
      "tokens_seen": 1954283520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004148080808080808,
      "loss": 1.453,
      "theoretical_loss": 3.4370191206595306,
      "tokens_seen": 1954349056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147878787878788,
      "loss": 1.473,
      "theoretical_loss": 3.437009467474879,
      "tokens_seen": 1954414592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147676767676768,
      "loss": 1.533,
      "theoretical_loss": 3.4369998147045466,
      "tokens_seen": 1954480128
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041474747474747476,
      "loss": 1.1881,
      "theoretical_loss": 3.4369901623485006,
      "tokens_seen": 1954545664
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147272727272727,
      "loss": 1.4963,
      "theoretical_loss": 3.43698051040671,
      "tokens_seen": 1954611200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004147070707070707,
      "loss": 1.501,
      "theoretical_loss": 3.4369708588791434,
      "tokens_seen": 1954676736
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146868686868687,
      "loss": 1.4789,
      "theoretical_loss": 3.4369612077657687,
      "tokens_seen": 1954742272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146666666666667,
      "loss": 1.4487,
      "theoretical_loss": 3.4369515570665543,
      "tokens_seen": 1954807808
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.42155423760414124,
      "objective/train/avg_token_score": 0.007324489299207926,
      "objective/train/docs_used": 1105519,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2957619428634644,
      "objective/train/original_loss": 4.201072692871094,
      "objective/train/theoretical_loss": 3.4369419067814686,
      "objective/train/tokens_used": 313732576,
      "objective/train/unlikelihood_loss": 0.06063661724328995,
      "theoretical_loss": 3.4369419067814686,
      "tokens_seen": 1954873344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041464646464646466,
      "loss": 1.3553,
      "theoretical_loss": 3.4369419067814686,
      "tokens_seen": 1954873344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146262626262626,
      "loss": 1.3954,
      "theoretical_loss": 3.4369322569104797,
      "tokens_seen": 1954938880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004146060606060606,
      "loss": 1.4729,
      "theoretical_loss": 3.4369226074535564,
      "tokens_seen": 1955004416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041458585858585863,
      "loss": 1.5528,
      "theoretical_loss": 3.436912958410667,
      "tokens_seen": 1955069952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145656565656566,
      "loss": 1.2495,
      "theoretical_loss": 3.436903309781779,
      "tokens_seen": 1955135488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041454545454545455,
      "loss": 1.6581,
      "theoretical_loss": 3.436893661566862,
      "tokens_seen": 1955201024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145252525252525,
      "loss": 1.3339,
      "theoretical_loss": 3.4368840137658836,
      "tokens_seen": 1955266560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004145050505050505,
      "loss": 1.4273,
      "theoretical_loss": 3.436874366378812,
      "tokens_seen": 1955332096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041448484848484853,
      "loss": 1.1655,
      "theoretical_loss": 3.4368647194056168,
      "tokens_seen": 1955397632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004144646464646465,
      "loss": 1.4596,
      "theoretical_loss": 3.436855072846265,
      "tokens_seen": 1955463168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041444444444444444,
      "loss": 1.295,
      "theoretical_loss": 3.4368454267007253,
      "tokens_seen": 1955528704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041442424242424245,
      "loss": 1.5501,
      "theoretical_loss": 3.4368357809689667,
      "tokens_seen": 1955594240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004144040404040404,
      "loss": 1.4875,
      "theoretical_loss": 3.436826135650957,
      "tokens_seen": 1955659776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143838383838384,
      "loss": 1.4391,
      "theoretical_loss": 3.436816490746665,
      "tokens_seen": 1955725312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143636363636364,
      "loss": 1.436,
      "theoretical_loss": 3.436806846256058,
      "tokens_seen": 1955790848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041434343434343433,
      "loss": 1.6272,
      "theoretical_loss": 3.436797202179106,
      "tokens_seen": 1955856384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041432323232323235,
      "loss": 1.5892,
      "theoretical_loss": 3.4367875585157766,
      "tokens_seen": 1955921920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004143030303030303,
      "loss": 1.5377,
      "theoretical_loss": 3.4367779152660383,
      "tokens_seen": 1955987456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142828282828283,
      "loss": 1.1497,
      "theoretical_loss": 3.436768272429859,
      "tokens_seen": 1956052992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041426262626262627,
      "loss": 1.3424,
      "theoretical_loss": 3.4367586300072075,
      "tokens_seen": 1956118528
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142424242424243,
      "loss": 1.6651,
      "theoretical_loss": 3.4367489879980524,
      "tokens_seen": 1956184064
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041422222222222224,
      "loss": 1.3298,
      "theoretical_loss": 3.436739346402362,
      "tokens_seen": 1956249600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004142020202020202,
      "loss": 1.48,
      "theoretical_loss": 3.436729705220105,
      "tokens_seen": 1956315136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141818181818182,
      "loss": 1.5618,
      "theoretical_loss": 3.436720064451249,
      "tokens_seen": 1956380672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041416161616161616,
      "loss": 1.265,
      "theoretical_loss": 3.4367104240957635,
      "tokens_seen": 1956446208
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.32251954078674316,
      "objective/train/avg_token_score": 0.009885277599096298,
      "objective/train/docs_used": 1106102,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.99889075756073,
      "objective/train/original_loss": 4.139228820800781,
      "objective/train/theoretical_loss": 3.436700784153616,
      "objective/train/tokens_used": 315370976,
      "objective/train/unlikelihood_loss": 0.07105885446071625,
      "theoretical_loss": 3.436700784153616,
      "tokens_seen": 1956511744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141414141414142,
      "loss": 1.3508,
      "theoretical_loss": 3.436700784153616,
      "tokens_seen": 1956511744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041412121212121213,
      "loss": 1.4683,
      "theoretical_loss": 3.4366911446247754,
      "tokens_seen": 1956577280
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004141010101010101,
      "loss": 1.5079,
      "theoretical_loss": 3.43668150550921,
      "tokens_seen": 1956642816
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004140808080808081,
      "loss": 1.5571,
      "theoretical_loss": 3.4366718668068885,
      "tokens_seen": 1956708352
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041406060606060606,
      "loss": 1.4941,
      "theoretical_loss": 3.436662228517779,
      "tokens_seen": 1956773888
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041404040404040407,
      "loss": 1.5562,
      "theoretical_loss": 3.4366525906418497,
      "tokens_seen": 1956839424
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414020202020202,
      "loss": 1.7608,
      "theoretical_loss": 3.43664295317907,
      "tokens_seen": 1956904960
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000414,
      "loss": 1.1626,
      "theoretical_loss": 3.4366333161294076,
      "tokens_seen": 1956970496
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000413979797979798,
      "loss": 1.2521,
      "theoretical_loss": 3.436623679492831,
      "tokens_seen": 1957036032
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000413959595959596,
      "loss": 1.5362,
      "theoretical_loss": 3.436614043269309,
      "tokens_seen": 1957101568
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041393939393939396,
      "loss": 1.3286,
      "theoretical_loss": 3.43660440745881,
      "tokens_seen": 1957167104
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004139191919191919,
      "loss": 1.4399,
      "theoretical_loss": 3.4365947720613015,
      "tokens_seen": 1957232640
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041389898989898987,
      "loss": 1.478,
      "theoretical_loss": 3.4365851370767535,
      "tokens_seen": 1957298176
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004138787878787879,
      "loss": 1.4755,
      "theoretical_loss": 3.436575502505134,
      "tokens_seen": 1957363712
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004138585858585859,
      "loss": 1.505,
      "theoretical_loss": 3.4365658683464106,
      "tokens_seen": 1957429248
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041383838383838385,
      "loss": 1.4277,
      "theoretical_loss": 3.4365562346005527,
      "tokens_seen": 1957494784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004138181818181818,
      "loss": 1.2607,
      "theoretical_loss": 3.436546601267529,
      "tokens_seen": 1957560320
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041379797979797977,
      "loss": 1.4686,
      "theoretical_loss": 3.4365369683473066,
      "tokens_seen": 1957625856
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041377777777777783,
      "loss": 1.1683,
      "theoretical_loss": 3.4365273358398554,
      "tokens_seen": 1957691392
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004137575757575758,
      "loss": 1.3763,
      "theoretical_loss": 3.4365177037451433,
      "tokens_seen": 1957756928
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041373737373737374,
      "loss": 1.1928,
      "theoretical_loss": 3.4365080720631394,
      "tokens_seen": 1957822464
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004137171717171717,
      "loss": 1.5801,
      "theoretical_loss": 3.4364984407938115,
      "tokens_seen": 1957888000
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041369696969696966,
      "loss": 1.2228,
      "theoretical_loss": 3.436488809937128,
      "tokens_seen": 1957953536
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004136767676767677,
      "loss": 1.3379,
      "theoretical_loss": 3.4364791794930576,
      "tokens_seen": 1958019072
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004136565656565657,
      "loss": 1.3664,
      "theoretical_loss": 3.4364695494615694,
      "tokens_seen": 1958084608
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.4745234549045563,
      "objective/train/avg_token_score": 0.011926386505365372,
      "objective/train/docs_used": 1106840,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.578378438949585,
      "objective/train/original_loss": 4.3692216873168945,
      "objective/train/theoretical_loss": 3.4364599198426307,
      "objective/train/tokens_used": 317009376,
      "objective/train/unlikelihood_loss": 0.04125186428427696,
      "theoretical_loss": 3.4364599198426307,
      "tokens_seen": 1958150144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041363636363636364,
      "loss": 1.6483,
      "theoretical_loss": 3.4364599198426307,
      "tokens_seen": 1958150144
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004136161616161616,
      "loss": 1.5616,
      "theoretical_loss": 3.4364502906362118,
      "tokens_seen": 1958215680
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004135959595959596,
      "loss": 1.3994,
      "theoretical_loss": 3.4364406618422794,
      "tokens_seen": 1958281216
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004135757575757576,
      "loss": 1.3587,
      "theoretical_loss": 3.436431033460803,
      "tokens_seen": 1958346752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004135555555555556,
      "loss": 1.526,
      "theoretical_loss": 3.4364214054917515,
      "tokens_seen": 1958412288
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041353535353535353,
      "loss": 1.2816,
      "theoretical_loss": 3.436411777935092,
      "tokens_seen": 1958477824
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004135151515151515,
      "loss": 1.3146,
      "theoretical_loss": 3.4364021507907943,
      "tokens_seen": 1958543360
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134949494949495,
      "loss": 1.3607,
      "theoretical_loss": 3.4363925240588267,
      "tokens_seen": 1958608896
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134747474747475,
      "loss": 1.4933,
      "theoretical_loss": 3.4363828977391577,
      "tokens_seen": 1958674432
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041345454545454547,
      "loss": 1.4624,
      "theoretical_loss": 3.436373271831755,
      "tokens_seen": 1958739968
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004134343434343434,
      "loss": 1.4388,
      "theoretical_loss": 3.4363636463365888,
      "tokens_seen": 1958805504
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041341414141414143,
      "loss": 1.3727,
      "theoretical_loss": 3.436354021253626,
      "tokens_seen": 1958871040
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133939393939394,
      "loss": 1.3955,
      "theoretical_loss": 3.436344396582837,
      "tokens_seen": 1958936576
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133737373737374,
      "loss": 1.3918,
      "theoretical_loss": 3.436334772324188,
      "tokens_seen": 1959002112
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041335353535353536,
      "loss": 1.4894,
      "theoretical_loss": 3.43632514847765,
      "tokens_seen": 1959067648
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133333333333333,
      "loss": 1.4914,
      "theoretical_loss": 3.4363155250431894,
      "tokens_seen": 1959133184
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004133131313131313,
      "loss": 1.487,
      "theoretical_loss": 3.4363059020207762,
      "tokens_seen": 1959198720
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132929292929293,
      "loss": 1.4438,
      "theoretical_loss": 3.436296279410379,
      "tokens_seen": 1959264256
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132727272727273,
      "loss": 1.4592,
      "theoretical_loss": 3.436286657211965,
      "tokens_seen": 1959329792
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041325252525252525,
      "loss": 1.5674,
      "theoretical_loss": 3.4362770354255043,
      "tokens_seen": 1959395328
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041323232323232326,
      "loss": 1.5201,
      "theoretical_loss": 3.436267414050965,
      "tokens_seen": 1959460864
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004132121212121212,
      "loss": 1.3334,
      "theoretical_loss": 3.4362577930883154,
      "tokens_seen": 1959526400
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131919191919192,
      "loss": 1.6545,
      "theoretical_loss": 3.436248172537524,
      "tokens_seen": 1959591936
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131717171717172,
      "loss": 1.5059,
      "theoretical_loss": 3.43623855239856,
      "tokens_seen": 1959657472
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041315151515151514,
      "loss": 1.4031,
      "theoretical_loss": 3.4362289326713915,
      "tokens_seen": 1959723008
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.34634652733802795,
      "objective/train/avg_token_score": 0.02731332927942276,
      "objective/train/docs_used": 1108280,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.078272819519043,
      "objective/train/original_loss": 4.436617851257324,
      "objective/train/theoretical_loss": 3.4362193133559877,
      "objective/train/tokens_used": 318647776,
      "objective/train/unlikelihood_loss": 0.09181275963783264,
      "theoretical_loss": 3.4362193133559877,
      "tokens_seen": 1959788544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041313131313131315,
      "loss": 1.3301,
      "theoretical_loss": 3.4362193133559877,
      "tokens_seen": 1959788544
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004131111111111111,
      "loss": 1.2653,
      "theoretical_loss": 3.4362096944523164,
      "tokens_seen": 1959854080
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041309090909090907,
      "loss": 1.5523,
      "theoretical_loss": 3.436200075960347,
      "tokens_seen": 1959919616
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130707070707071,
      "loss": 1.5079,
      "theoretical_loss": 3.4361904578800466,
      "tokens_seen": 1959985152
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004130505050505051,
      "loss": 1.3166,
      "theoretical_loss": 3.436180840211386,
      "tokens_seen": 1960050688
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041303030303030305,
      "loss": 1.5395,
      "theoretical_loss": 3.436171222954332,
      "tokens_seen": 1960116224
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000413010101010101,
      "loss": 1.2939,
      "theoretical_loss": 3.436161606108855,
      "tokens_seen": 1960181760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041298989898989896,
      "loss": 1.4107,
      "theoretical_loss": 3.436151989674922,
      "tokens_seen": 1960247296
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041296969696969697,
      "loss": 1.3592,
      "theoretical_loss": 3.436142373652502,
      "tokens_seen": 1960312832
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000412949494949495,
      "loss": 1.5549,
      "theoretical_loss": 3.4361327580415644,
      "tokens_seen": 1960378368
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041292929292929294,
      "loss": 1.5716,
      "theoretical_loss": 3.436123142842077,
      "tokens_seen": 1960443904
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004129090909090909,
      "loss": 1.3391,
      "theoretical_loss": 3.4361135280540087,
      "tokens_seen": 1960509440
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128888888888889,
      "loss": 1.4607,
      "theoretical_loss": 3.436103913677328,
      "tokens_seen": 1960574976
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128686868686869,
      "loss": 1.35,
      "theoretical_loss": 3.4360942997120043,
      "tokens_seen": 1960640512
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128484848484849,
      "loss": 1.1904,
      "theoretical_loss": 3.436084686158005,
      "tokens_seen": 1960706048
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041282828282828283,
      "loss": 1.216,
      "theoretical_loss": 3.4360750730153002,
      "tokens_seen": 1960771584
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004128080808080808,
      "loss": 1.3311,
      "theoretical_loss": 3.4360654602838574,
      "tokens_seen": 1960837120
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004127878787878788,
      "loss": 1.4853,
      "theoretical_loss": 3.4360558479636456,
      "tokens_seen": 1960902656
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004127676767676768,
      "loss": 1.6454,
      "theoretical_loss": 3.4360462360546333,
      "tokens_seen": 1960968192
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041274747474747477,
      "loss": 1.2993,
      "theoretical_loss": 3.43603662455679,
      "tokens_seen": 1961033728
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004127272727272727,
      "loss": 1.2265,
      "theoretical_loss": 3.436027013470083,
      "tokens_seen": 1961099264
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004127070707070707,
      "loss": 1.3192,
      "theoretical_loss": 3.436017402794482,
      "tokens_seen": 1961164800
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041268686868686875,
      "loss": 1.6334,
      "theoretical_loss": 3.4360077925299555,
      "tokens_seen": 1961230336
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126666666666667,
      "loss": 1.3333,
      "theoretical_loss": 3.435998182676472,
      "tokens_seen": 1961295872
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041264646464646466,
      "loss": 1.3261,
      "theoretical_loss": 3.4359885732340008,
      "tokens_seen": 1961361408
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.43071848154067993,
      "objective/train/avg_token_score": 0.014302323572337627,
      "objective/train/docs_used": 1108943,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3354324102401733,
      "objective/train/original_loss": 4.234708309173584,
      "objective/train/theoretical_loss": 3.4359789642025094,
      "objective/train/tokens_used": 320286176,
      "objective/train/unlikelihood_loss": 0.055439990013837814,
      "theoretical_loss": 3.4359789642025094,
      "tokens_seen": 1961426944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126262626262626,
      "loss": 1.5046,
      "theoretical_loss": 3.4359789642025094,
      "tokens_seen": 1961426944
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004126060606060606,
      "loss": 1.2966,
      "theoretical_loss": 3.4359693555819675,
      "tokens_seen": 1961492480
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041258585858585864,
      "loss": 1.4952,
      "theoretical_loss": 3.435959747372343,
      "tokens_seen": 1961558016
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125656565656566,
      "loss": 1.3245,
      "theoretical_loss": 3.4359501395736056,
      "tokens_seen": 1961623552
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041254545454545455,
      "loss": 1.3205,
      "theoretical_loss": 3.4359405321857235,
      "tokens_seen": 1961689088
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004125252525252525,
      "loss": 1.2586,
      "theoretical_loss": 3.4359309252086647,
      "tokens_seen": 1961754624
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041250505050505047,
      "loss": 1.124,
      "theoretical_loss": 3.435921318642399,
      "tokens_seen": 1961820160
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041248484848484853,
      "loss": 1.2376,
      "theoretical_loss": 3.4359117124868948,
      "tokens_seen": 1961885696
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004124646464646465,
      "loss": 1.5335,
      "theoretical_loss": 3.4359021067421205,
      "tokens_seen": 1961951232
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041244444444444445,
      "loss": 1.5446,
      "theoretical_loss": 3.435892501408045,
      "tokens_seen": 1962016768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004124242424242424,
      "loss": 1.2378,
      "theoretical_loss": 3.435882896484637,
      "tokens_seen": 1962082304
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004124040404040404,
      "loss": 1.3928,
      "theoretical_loss": 3.4358732919718653,
      "tokens_seen": 1962147840
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123838383838384,
      "loss": 1.5574,
      "theoretical_loss": 3.435863687869699,
      "tokens_seen": 1962213376
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123636363636364,
      "loss": 1.2501,
      "theoretical_loss": 3.4358540841781062,
      "tokens_seen": 1962278912
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041234343434343434,
      "loss": 1.3118,
      "theoretical_loss": 3.435844480897056,
      "tokens_seen": 1962344448
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123232323232323,
      "loss": 1.3476,
      "theoretical_loss": 3.435834878026516,
      "tokens_seen": 1962409984
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004123030303030303,
      "loss": 1.2193,
      "theoretical_loss": 3.435825275566457,
      "tokens_seen": 1962475520
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004122828282828283,
      "loss": 1.7092,
      "theoretical_loss": 3.4358156735168466,
      "tokens_seen": 1962541056
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004122626262626263,
      "loss": 1.3348,
      "theoretical_loss": 3.4358060718776535,
      "tokens_seen": 1962606592
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041224242424242423,
      "loss": 1.4579,
      "theoretical_loss": 3.435796470648846,
      "tokens_seen": 1962672128
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041222222222222224,
      "loss": 1.4916,
      "theoretical_loss": 3.4357868698303946,
      "tokens_seen": 1962737664
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004122020202020202,
      "loss": 1.5911,
      "theoretical_loss": 3.435777269422266,
      "tokens_seen": 1962803200
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121818181818182,
      "loss": 1.3839,
      "theoretical_loss": 3.4357676694244303,
      "tokens_seen": 1962868736
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041216161616161617,
      "loss": 1.4081,
      "theoretical_loss": 3.4357580698368557,
      "tokens_seen": 1962934272
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121414141414141,
      "loss": 1.2923,
      "theoretical_loss": 3.435748470659511,
      "tokens_seen": 1962999808
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.3285679519176483,
      "objective/train/avg_token_score": 0.03589082136750221,
      "objective/train/docs_used": 1110285,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1867942810058594,
      "objective/train/original_loss": 5.2249436378479,
      "objective/train/theoretical_loss": 3.4357388718923656,
      "objective/train/tokens_used": 321924576,
      "objective/train/unlikelihood_loss": 0.03941744193434715,
      "theoretical_loss": 3.4357388718923656,
      "tokens_seen": 1963065344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041212121212121214,
      "loss": 1.4527,
      "theoretical_loss": 3.4357388718923656,
      "tokens_seen": 1963065344
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004121010101010101,
      "loss": 1.3501,
      "theoretical_loss": 3.4357292735353875,
      "tokens_seen": 1963130880
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004120808080808081,
      "loss": 1.4519,
      "theoretical_loss": 3.4357196755885457,
      "tokens_seen": 1963196416
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041206060606060606,
      "loss": 1.4366,
      "theoretical_loss": 3.435710078051809,
      "tokens_seen": 1963261952
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041204040404040407,
      "loss": 1.3558,
      "theoretical_loss": 3.435700480925146,
      "tokens_seen": 1963327488
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041202020202020203,
      "loss": 1.6353,
      "theoretical_loss": 3.435690884208526,
      "tokens_seen": 1963393024
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000412,
      "loss": 1.477,
      "theoretical_loss": 3.4356812879019176,
      "tokens_seen": 1963458560
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.000411979797979798,
      "loss": 1.6616,
      "theoretical_loss": 3.4356716920052897,
      "tokens_seen": 1963524096
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041195959595959595,
      "loss": 1.2516,
      "theoretical_loss": 3.4356620965186107,
      "tokens_seen": 1963589632
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041193939393939396,
      "loss": 1.4148,
      "theoretical_loss": 3.435652501441849,
      "tokens_seen": 1963655168
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004119191919191919,
      "loss": 1.4648,
      "theoretical_loss": 3.4356429067749747,
      "tokens_seen": 1963720704
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004118989898989899,
      "loss": 1.4111,
      "theoretical_loss": 3.4356333125179557,
      "tokens_seen": 1963786240
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004118787878787879,
      "loss": 1.4834,
      "theoretical_loss": 3.435623718670761,
      "tokens_seen": 1963851776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004118585858585859,
      "loss": 1.4329,
      "theoretical_loss": 3.43561412523336,
      "tokens_seen": 1963917312
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041183838383838386,
      "loss": 1.4315,
      "theoretical_loss": 3.4356045322057205,
      "tokens_seen": 1963982848
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004118181818181818,
      "loss": 1.1834,
      "theoretical_loss": 3.435594939587812,
      "tokens_seen": 1964048384
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041179797979797977,
      "loss": 1.1604,
      "theoretical_loss": 3.435585347379603,
      "tokens_seen": 1964113920
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004117777777777778,
      "loss": 1.3149,
      "theoretical_loss": 3.4355757555810627,
      "tokens_seen": 1964179456
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004117575757575758,
      "loss": 1.4457,
      "theoretical_loss": 3.435566164192159,
      "tokens_seen": 1964244992
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041173737373737375,
      "loss": 1.4566,
      "theoretical_loss": 3.435556573212862,
      "tokens_seen": 1964310528
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004117171717171717,
      "loss": 1.4909,
      "theoretical_loss": 3.4355469826431406,
      "tokens_seen": 1964376064
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041169696969696966,
      "loss": 1.3193,
      "theoretical_loss": 3.435537392482962,
      "tokens_seen": 1964441600
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041167676767676773,
      "loss": 1.5391,
      "theoretical_loss": 3.4355278027322966,
      "tokens_seen": 1964507136
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116565656565657,
      "loss": 1.4511,
      "theoretical_loss": 3.435518213391113,
      "tokens_seen": 1964572672
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.00041163636363636364,
      "loss": 1.4846,
      "theoretical_loss": 3.4355086244593793,
      "tokens_seen": 1964638208
    },
    {
      "epoch": 0.18,
      "objective/train/aligned_token_count": 0.3200146555900574,
      "objective/train/avg_token_score": 0.009661064483225346,
      "objective/train/docs_used": 1110737,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0394518375396729,
      "objective/train/original_loss": 3.6994714736938477,
      "objective/train/theoretical_loss": 3.435499035937065,
      "objective/train/tokens_used": 323562976,
      "objective/train/unlikelihood_loss": 0.08720573782920837,
      "theoretical_loss": 3.435499035937065,
      "tokens_seen": 1964703744
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0004116161616161616,
      "loss": 1.2846,
      "theoretical_loss": 3.435499035937065,
      "tokens_seen": 1964703744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041159595959595956,
      "loss": 1.4118,
      "theoretical_loss": 3.435489447824139,
      "tokens_seen": 1964769280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115757575757576,
      "loss": 1.3332,
      "theoretical_loss": 3.43547986012057,
      "tokens_seen": 1964834816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115555555555556,
      "loss": 1.371,
      "theoretical_loss": 3.4354702728263264,
      "tokens_seen": 1964900352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041153535353535354,
      "loss": 1.5411,
      "theoretical_loss": 3.435460685941378,
      "tokens_seen": 1964965888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004115151515151515,
      "loss": 1.4533,
      "theoretical_loss": 3.435451099465693,
      "tokens_seen": 1965031424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114949494949495,
      "loss": 1.2664,
      "theoretical_loss": 3.4354415133992404,
      "tokens_seen": 1965096960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114747474747475,
      "loss": 1.4433,
      "theoretical_loss": 3.4354319277419894,
      "tokens_seen": 1965162496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041145454545454547,
      "loss": 1.4658,
      "theoretical_loss": 3.4354223424939088,
      "tokens_seen": 1965228032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041143434343434343,
      "loss": 1.382,
      "theoretical_loss": 3.435412757654967,
      "tokens_seen": 1965293568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004114141414141414,
      "loss": 1.3177,
      "theoretical_loss": 3.4354031732251333,
      "tokens_seen": 1965359104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004113939393939394,
      "loss": 1.2478,
      "theoretical_loss": 3.4353935892043768,
      "tokens_seen": 1965424640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004113737373737374,
      "loss": 1.7692,
      "theoretical_loss": 3.435384005592666,
      "tokens_seen": 1965490176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041135353535353536,
      "loss": 1.5251,
      "theoretical_loss": 3.4353744223899696,
      "tokens_seen": 1965555712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004113333333333333,
      "loss": 1.441,
      "theoretical_loss": 3.4353648395962573,
      "tokens_seen": 1965621248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041131313131313133,
      "loss": 1.3334,
      "theoretical_loss": 3.435355257211498,
      "tokens_seen": 1965686784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112929292929293,
      "loss": 1.4239,
      "theoretical_loss": 3.4353456752356593,
      "tokens_seen": 1965752320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112727272727273,
      "loss": 1.3418,
      "theoretical_loss": 3.4353360936687114,
      "tokens_seen": 1965817856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041125252525252526,
      "loss": 1.1314,
      "theoretical_loss": 3.4353265125106223,
      "tokens_seen": 1965883392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112323232323232,
      "loss": 1.4341,
      "theoretical_loss": 3.4353169317613617,
      "tokens_seen": 1965948928
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004112121212121212,
      "loss": 1.5507,
      "theoretical_loss": 3.4353073514208985,
      "tokens_seen": 1966014464
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041119191919191924,
      "loss": 1.3749,
      "theoretical_loss": 3.4352977714892017,
      "tokens_seen": 1966080000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111717171717172,
      "loss": 1.5129,
      "theoretical_loss": 3.4352881919662392,
      "tokens_seen": 1966145536
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041115151515151515,
      "loss": 1.382,
      "theoretical_loss": 3.435278612851981,
      "tokens_seen": 1966211072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041113131313131316,
      "loss": 1.6181,
      "theoretical_loss": 3.435269034146396,
      "tokens_seen": 1966276608
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.4528347849845886,
      "objective/train/avg_token_score": 0.006985062267631292,
      "objective/train/docs_used": 1112038,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5920250415802002,
      "objective/train/original_loss": 4.1393818855285645,
      "objective/train/theoretical_loss": 3.4352594558494527,
      "objective/train/tokens_used": 325201376,
      "objective/train/unlikelihood_loss": 0.0694524422287941,
      "theoretical_loss": 3.4352594558494527,
      "tokens_seen": 1966342144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004111111111111111,
      "loss": 1.3822,
      "theoretical_loss": 3.4352594558494527,
      "tokens_seen": 1966342144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041109090909090913,
      "loss": 1.7697,
      "theoretical_loss": 3.4352498779611196,
      "tokens_seen": 1966407680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004110707070707071,
      "loss": 1.4781,
      "theoretical_loss": 3.435240300481367,
      "tokens_seen": 1966473216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041105050505050504,
      "loss": 1.381,
      "theoretical_loss": 3.4352307234101627,
      "tokens_seen": 1966538752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041103030303030305,
      "loss": 1.4222,
      "theoretical_loss": 3.4352211467474767,
      "tokens_seen": 1966604288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000411010101010101,
      "loss": 1.5048,
      "theoretical_loss": 3.4352115704932764,
      "tokens_seen": 1966669824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410989898989899,
      "loss": 1.3564,
      "theoretical_loss": 3.4352019946475325,
      "tokens_seen": 1966735360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410969696969697,
      "loss": 1.1223,
      "theoretical_loss": 3.435192419210213,
      "tokens_seen": 1966800896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410949494949495,
      "loss": 1.5934,
      "theoretical_loss": 3.435182844181287,
      "tokens_seen": 1966866432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041092929292929295,
      "loss": 1.2154,
      "theoretical_loss": 3.435173269560723,
      "tokens_seen": 1966931968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004109090909090909,
      "loss": 1.4263,
      "theoretical_loss": 3.435163695348491,
      "tokens_seen": 1966997504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108888888888889,
      "loss": 1.4043,
      "theoretical_loss": 3.435154121544559,
      "tokens_seen": 1967063040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041086868686868687,
      "loss": 1.263,
      "theoretical_loss": 3.4351445481488967,
      "tokens_seen": 1967128576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108484848484849,
      "loss": 1.5556,
      "theoretical_loss": 3.435134975161473,
      "tokens_seen": 1967194112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041082828282828284,
      "loss": 1.6561,
      "theoretical_loss": 3.4351254025822566,
      "tokens_seen": 1967259648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004108080808080808,
      "loss": 1.3542,
      "theoretical_loss": 3.435115830411217,
      "tokens_seen": 1967325184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107878787878788,
      "loss": 1.328,
      "theoretical_loss": 3.4351062586483225,
      "tokens_seen": 1967390720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041076767676767676,
      "loss": 1.6009,
      "theoretical_loss": 3.4350966872935422,
      "tokens_seen": 1967456256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107474747474748,
      "loss": 1.3577,
      "theoretical_loss": 3.4350871163468457,
      "tokens_seen": 1967521792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041072727272727273,
      "loss": 1.3894,
      "theoretical_loss": 3.435077545808202,
      "tokens_seen": 1967587328
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004107070707070707,
      "loss": 1.3641,
      "theoretical_loss": 3.435067975677579,
      "tokens_seen": 1967652864
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106868686868687,
      "loss": 1.4913,
      "theoretical_loss": 3.4350584059549467,
      "tokens_seen": 1967718400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106666666666667,
      "loss": 1.6145,
      "theoretical_loss": 3.435048836640274,
      "tokens_seen": 1967783936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041064646464646467,
      "loss": 1.5496,
      "theoretical_loss": 3.4350392677335297,
      "tokens_seen": 1967849472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106262626262626,
      "loss": 1.4591,
      "theoretical_loss": 3.435029699234683,
      "tokens_seen": 1967915008
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.3913734257221222,
      "objective/train/avg_token_score": 0.011882202699780464,
      "objective/train/docs_used": 1112624,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2495681047439575,
      "objective/train/original_loss": 4.346351623535156,
      "objective/train/theoretical_loss": 3.435020131143703,
      "objective/train/tokens_used": 326839776,
      "objective/train/unlikelihood_loss": 0.07031120359897614,
      "theoretical_loss": 3.435020131143703,
      "tokens_seen": 1967980544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004106060606060606,
      "loss": 1.2495,
      "theoretical_loss": 3.435020131143703,
      "tokens_seen": 1967980544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105858585858586,
      "loss": 1.5182,
      "theoretical_loss": 3.4350105634605583,
      "tokens_seen": 1968046080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105656565656566,
      "loss": 1.4065,
      "theoretical_loss": 3.435000996185219,
      "tokens_seen": 1968111616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041054545454545456,
      "loss": 1.4003,
      "theoretical_loss": 3.4349914293176527,
      "tokens_seen": 1968177152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105252525252525,
      "loss": 1.4241,
      "theoretical_loss": 3.4349818628578292,
      "tokens_seen": 1968242688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004105050505050505,
      "loss": 1.5984,
      "theoretical_loss": 3.4349722968057175,
      "tokens_seen": 1968308224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041048484848484854,
      "loss": 1.6424,
      "theoretical_loss": 3.4349627311612867,
      "tokens_seen": 1968373760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004104646464646465,
      "loss": 1.5381,
      "theoretical_loss": 3.434953165924506,
      "tokens_seen": 1968439296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041044444444444445,
      "loss": 1.2385,
      "theoretical_loss": 3.434943601095344,
      "tokens_seen": 1968504832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004104242424242424,
      "loss": 1.2404,
      "theoretical_loss": 3.43493403667377,
      "tokens_seen": 1968570368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041040404040404037,
      "loss": 1.4081,
      "theoretical_loss": 3.4349244726597528,
      "tokens_seen": 1968635904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041038383838383843,
      "loss": 1.4257,
      "theoretical_loss": 3.434914909053262,
      "tokens_seen": 1968701440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103636363636364,
      "loss": 1.2265,
      "theoretical_loss": 3.434905345854266,
      "tokens_seen": 1968766976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041034343434343435,
      "loss": 1.258,
      "theoretical_loss": 3.4348957830627347,
      "tokens_seen": 1968832512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103232323232323,
      "loss": 1.4968,
      "theoretical_loss": 3.434886220678637,
      "tokens_seen": 1968898048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004103030303030303,
      "loss": 1.4694,
      "theoretical_loss": 3.434876658701941,
      "tokens_seen": 1968963584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102828282828283,
      "loss": 1.5733,
      "theoretical_loss": 3.434867097132617,
      "tokens_seen": 1969029120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102626262626263,
      "loss": 1.5411,
      "theoretical_loss": 3.4348575359706333,
      "tokens_seen": 1969094656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041024242424242424,
      "loss": 1.4478,
      "theoretical_loss": 3.43484797521596,
      "tokens_seen": 1969160192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102222222222222,
      "loss": 1.3074,
      "theoretical_loss": 3.4348384148685644,
      "tokens_seen": 1969225728
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004102020202020202,
      "loss": 1.2974,
      "theoretical_loss": 3.4348288549284174,
      "tokens_seen": 1969291264
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101818181818182,
      "loss": 1.3546,
      "theoretical_loss": 3.434819295395487,
      "tokens_seen": 1969356800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101616161616162,
      "loss": 1.1625,
      "theoretical_loss": 3.434809736269743,
      "tokens_seen": 1969422336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041014141414141413,
      "loss": 1.5824,
      "theoretical_loss": 3.4348001775511543,
      "tokens_seen": 1969487872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041012121212121214,
      "loss": 1.3634,
      "theoretical_loss": 3.4347906192396893,
      "tokens_seen": 1969553408
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.27834799885749817,
      "objective/train/avg_token_score": 0.007226386573165655,
      "objective/train/docs_used": 1113989,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9873694777488708,
      "objective/train/original_loss": 4.861571788787842,
      "objective/train/theoretical_loss": 3.4347810613353182,
      "objective/train/tokens_used": 328478176,
      "objective/train/unlikelihood_loss": 0.06607561558485031,
      "theoretical_loss": 3.4347810613353182,
      "tokens_seen": 1969618944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004101010101010101,
      "loss": 1.2623,
      "theoretical_loss": 3.4347810613353182,
      "tokens_seen": 1969618944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004100808080808081,
      "loss": 1.5619,
      "theoretical_loss": 3.434771503838009,
      "tokens_seen": 1969684480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041006060606060607,
      "loss": 1.1316,
      "theoretical_loss": 3.434761946747732,
      "tokens_seen": 1969750016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000410040404040404,
      "loss": 1.38,
      "theoretical_loss": 3.434752390064456,
      "tokens_seen": 1969815552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041002020202020203,
      "loss": 1.4082,
      "theoretical_loss": 3.4347428337881496,
      "tokens_seen": 1969881088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00041,
      "loss": 1.2224,
      "theoretical_loss": 3.434733277918782,
      "tokens_seen": 1969946624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000409979797979798,
      "loss": 1.3039,
      "theoretical_loss": 3.4347237224563223,
      "tokens_seen": 1970012160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040995959595959596,
      "loss": 1.3388,
      "theoretical_loss": 3.4347141674007404,
      "tokens_seen": 1970077696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040993939393939397,
      "loss": 1.3502,
      "theoretical_loss": 3.4347046127520047,
      "tokens_seen": 1970143232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004099191919191919,
      "loss": 1.3016,
      "theoretical_loss": 3.4346950585100853,
      "tokens_seen": 1970208768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098989898989899,
      "loss": 1.3919,
      "theoretical_loss": 3.43468550467495,
      "tokens_seen": 1970274304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098787878787879,
      "loss": 1.5056,
      "theoretical_loss": 3.434675951246568,
      "tokens_seen": 1970339840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040985858585858585,
      "loss": 1.2892,
      "theoretical_loss": 3.43466639822491,
      "tokens_seen": 1970405376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040983838383838386,
      "loss": 1.4575,
      "theoretical_loss": 3.434656845609944,
      "tokens_seen": 1970470912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004098181818181818,
      "loss": 1.4339,
      "theoretical_loss": 3.434647293401639,
      "tokens_seen": 1970536448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097979797979798,
      "loss": 1.1932,
      "theoretical_loss": 3.434637741599965,
      "tokens_seen": 1970601984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097777777777778,
      "loss": 1.4734,
      "theoretical_loss": 3.4346281902048905,
      "tokens_seen": 1970667520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097575757575758,
      "loss": 1.4094,
      "theoretical_loss": 3.4346186392163847,
      "tokens_seen": 1970733056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040973737373737376,
      "loss": 1.5061,
      "theoretical_loss": 3.434609088634417,
      "tokens_seen": 1970798592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004097171717171717,
      "loss": 1.7504,
      "theoretical_loss": 3.4345995384589565,
      "tokens_seen": 1970864128
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040969696969696967,
      "loss": 1.6252,
      "theoretical_loss": 3.4345899886899725,
      "tokens_seen": 1970929664
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096767676767677,
      "loss": 1.6023,
      "theoretical_loss": 3.434580439327434,
      "tokens_seen": 1970995200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096565656565657,
      "loss": 1.4679,
      "theoretical_loss": 3.4345708903713104,
      "tokens_seen": 1971060736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040963636363636365,
      "loss": 1.3235,
      "theoretical_loss": 3.4345613418215706,
      "tokens_seen": 1971126272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004096161616161616,
      "loss": 1.1655,
      "theoretical_loss": 3.4345517936781835,
      "tokens_seen": 1971191808
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.5766128897666931,
      "objective/train/avg_token_score": 0.006864483468234539,
      "objective/train/docs_used": 1114737,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9049957990646362,
      "objective/train/original_loss": 4.060426235198975,
      "objective/train/theoretical_loss": 3.4345422459411195,
      "objective/train/tokens_used": 330116576,
      "objective/train/unlikelihood_loss": 0.039708707481622696,
      "theoretical_loss": 3.4345422459411195,
      "tokens_seen": 1971257344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040959595959595956,
      "loss": 1.6513,
      "theoretical_loss": 3.4345422459411195,
      "tokens_seen": 1971257344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040957575757575763,
      "loss": 1.5602,
      "theoretical_loss": 3.434532698610347,
      "tokens_seen": 1971322880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095555555555556,
      "loss": 1.3076,
      "theoretical_loss": 3.434523151685835,
      "tokens_seen": 1971388416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040953535353535354,
      "loss": 1.1315,
      "theoretical_loss": 3.4345136051675524,
      "tokens_seen": 1971453952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004095151515151515,
      "loss": 1.3409,
      "theoretical_loss": 3.43450405905547,
      "tokens_seen": 1971519488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040949494949494945,
      "loss": 1.1566,
      "theoretical_loss": 3.4344945133495552,
      "tokens_seen": 1971585024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094747474747475,
      "loss": 1.4211,
      "theoretical_loss": 3.4344849680497784,
      "tokens_seen": 1971650560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094545454545455,
      "loss": 1.323,
      "theoretical_loss": 3.4344754231561083,
      "tokens_seen": 1971716096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040943434343434343,
      "loss": 1.5279,
      "theoretical_loss": 3.4344658786685143,
      "tokens_seen": 1971781632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004094141414141414,
      "loss": 1.3783,
      "theoretical_loss": 3.4344563345869656,
      "tokens_seen": 1971847168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040939393939393946,
      "loss": 1.3743,
      "theoretical_loss": 3.4344467909114313,
      "tokens_seen": 1971912704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093737373737374,
      "loss": 1.3636,
      "theoretical_loss": 3.4344372476418807,
      "tokens_seen": 1971978240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040935353535353537,
      "loss": 1.3911,
      "theoretical_loss": 3.434427704778283,
      "tokens_seen": 1972043776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093333333333333,
      "loss": 1.4582,
      "theoretical_loss": 3.434418162320608,
      "tokens_seen": 1972109312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004093131313131313,
      "loss": 1.1357,
      "theoretical_loss": 3.434408620268824,
      "tokens_seen": 1972174848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040929292929292935,
      "loss": 1.5361,
      "theoretical_loss": 3.434399078622901,
      "tokens_seen": 1972240384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092727272727273,
      "loss": 1.7598,
      "theoretical_loss": 3.434389537382808,
      "tokens_seen": 1972305920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040925252525252526,
      "loss": 1.365,
      "theoretical_loss": 3.434379996548514,
      "tokens_seen": 1972371456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092323232323232,
      "loss": 1.533,
      "theoretical_loss": 3.4343704561199884,
      "tokens_seen": 1972436992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004092121212121212,
      "loss": 1.4524,
      "theoretical_loss": 3.4343609160972006,
      "tokens_seen": 1972502528
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040919191919191924,
      "loss": 1.4198,
      "theoretical_loss": 3.4343513764801203,
      "tokens_seen": 1972568064
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004091717171717172,
      "loss": 1.3793,
      "theoretical_loss": 3.4343418372687156,
      "tokens_seen": 1972633600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040915151515151515,
      "loss": 1.3158,
      "theoretical_loss": 3.434332298462957,
      "tokens_seen": 1972699136
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004091313131313131,
      "loss": 1.382,
      "theoretical_loss": 3.434322760062813,
      "tokens_seen": 1972764672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004091111111111111,
      "loss": 1.3222,
      "theoretical_loss": 3.434313222068253,
      "tokens_seen": 1972830208
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.38007086515426636,
      "objective/train/avg_token_score": 0.014573216438293457,
      "objective/train/docs_used": 1116152,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2298966646194458,
      "objective/train/original_loss": 4.61524772644043,
      "objective/train/theoretical_loss": 3.4343036844792465,
      "objective/train/tokens_used": 331754976,
      "objective/train/unlikelihood_loss": 0.04789040982723236,
      "theoretical_loss": 3.4343036844792465,
      "tokens_seen": 1972895744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040909090909090913,
      "loss": 1.2967,
      "theoretical_loss": 3.4343036844792465,
      "tokens_seen": 1972895744
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004090707070707071,
      "loss": 1.4073,
      "theoretical_loss": 3.4342941472957627,
      "tokens_seen": 1972961280
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040905050505050505,
      "loss": 1.4403,
      "theoretical_loss": 3.4342846105177705,
      "tokens_seen": 1973026816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000409030303030303,
      "loss": 1.3333,
      "theoretical_loss": 3.4342750741452397,
      "tokens_seen": 1973092352
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000409010101010101,
      "loss": 1.5288,
      "theoretical_loss": 3.43426553817814,
      "tokens_seen": 1973157888
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408989898989899,
      "loss": 1.3896,
      "theoretical_loss": 3.4342560026164395,
      "tokens_seen": 1973223424
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408969696969697,
      "loss": 1.3534,
      "theoretical_loss": 3.4342464674601083,
      "tokens_seen": 1973288960
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040894949494949494,
      "loss": 1.3614,
      "theoretical_loss": 3.4342369327091156,
      "tokens_seen": 1973354496
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040892929292929295,
      "loss": 1.6497,
      "theoretical_loss": 3.434227398363431,
      "tokens_seen": 1973420032
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004089090909090909,
      "loss": 1.577,
      "theoretical_loss": 3.4342178644230232,
      "tokens_seen": 1973485568
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088888888888889,
      "loss": 1.2795,
      "theoretical_loss": 3.4342083308878615,
      "tokens_seen": 1973551104
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088686868686869,
      "loss": 1.4452,
      "theoretical_loss": 3.4341987977579156,
      "tokens_seen": 1973616640
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040884848484848483,
      "loss": 1.4619,
      "theoretical_loss": 3.434189265033155,
      "tokens_seen": 1973682176
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040882828282828284,
      "loss": 1.4424,
      "theoretical_loss": 3.4341797327135484,
      "tokens_seen": 1973747712
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004088080808080808,
      "loss": 1.1546,
      "theoretical_loss": 3.4341702007990658,
      "tokens_seen": 1973813248
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087878787878788,
      "loss": 1.4294,
      "theoretical_loss": 3.434160669289676,
      "tokens_seen": 1973878784
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040876767676767677,
      "loss": 0.9521,
      "theoretical_loss": 3.4341511381853485,
      "tokens_seen": 1973944320
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087474747474748,
      "loss": 1.5766,
      "theoretical_loss": 3.434141607486053,
      "tokens_seen": 1974009856
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040872727272727274,
      "loss": 1.4657,
      "theoretical_loss": 3.4341320771917583,
      "tokens_seen": 1974075392
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004087070707070707,
      "loss": 1.4426,
      "theoretical_loss": 3.434122547302434,
      "tokens_seen": 1974140928
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004086868686868687,
      "loss": 1.5492,
      "theoretical_loss": 3.4341130178180492,
      "tokens_seen": 1974206464
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040866666666666666,
      "loss": 1.3706,
      "theoretical_loss": 3.4341034887385735,
      "tokens_seen": 1974272000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040864646464646467,
      "loss": 1.5491,
      "theoretical_loss": 3.4340939600639766,
      "tokens_seen": 1974337536
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040862626262626263,
      "loss": 1.3823,
      "theoretical_loss": 3.4340844317942274,
      "tokens_seen": 1974403072
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004086060606060606,
      "loss": 1.3105,
      "theoretical_loss": 3.4340749039292957,
      "tokens_seen": 1974468608
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.46144917607307434,
      "objective/train/avg_token_score": 0.03224779665470123,
      "objective/train/docs_used": 1116806,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.653957486152649,
      "objective/train/original_loss": 4.292056560516357,
      "objective/train/theoretical_loss": 3.4340653764691496,
      "objective/train/tokens_used": 333393376,
      "objective/train/unlikelihood_loss": 0.05980272591114044,
      "theoretical_loss": 3.4340653764691496,
      "tokens_seen": 1974534144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085858585858586,
      "loss": 1.4107,
      "theoretical_loss": 3.4340653764691496,
      "tokens_seen": 1974534144
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085656565656566,
      "loss": 1.4226,
      "theoretical_loss": 3.43405584941376,
      "tokens_seen": 1974599680
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040854545454545457,
      "loss": 1.3297,
      "theoretical_loss": 3.4340463227630957,
      "tokens_seen": 1974665216
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085252525252525,
      "loss": 1.4029,
      "theoretical_loss": 3.434036796517126,
      "tokens_seen": 1974730752
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004085050505050505,
      "loss": 1.3152,
      "theoretical_loss": 3.43402727067582,
      "tokens_seen": 1974796288
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084848484848485,
      "loss": 1.6887,
      "theoretical_loss": 3.4340177452391476,
      "tokens_seen": 1974861824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084646464646465,
      "loss": 1.3779,
      "theoretical_loss": 3.4340082202070774,
      "tokens_seen": 1974927360
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040844444444444446,
      "loss": 1.3655,
      "theoretical_loss": 3.43399869557958,
      "tokens_seen": 1974992896
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004084242424242424,
      "loss": 1.5568,
      "theoretical_loss": 3.4339891713566244,
      "tokens_seen": 1975058432
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040840404040404037,
      "loss": 1.4729,
      "theoretical_loss": 3.433979647538179,
      "tokens_seen": 1975123968
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040838383838383844,
      "loss": 1.3622,
      "theoretical_loss": 3.4339701241242144,
      "tokens_seen": 1975189504
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083636363636364,
      "loss": 1.2509,
      "theoretical_loss": 3.4339606011146993,
      "tokens_seen": 1975255040
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040834343434343435,
      "loss": 1.2185,
      "theoretical_loss": 3.4339510785096037,
      "tokens_seen": 1975320576
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004083232323232323,
      "loss": 1.4309,
      "theoretical_loss": 3.4339415563088966,
      "tokens_seen": 1975386112
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040830303030303026,
      "loss": 1.4844,
      "theoretical_loss": 3.433932034512547,
      "tokens_seen": 1975451648
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040828282828282833,
      "loss": 1.2712,
      "theoretical_loss": 3.4339225131205255,
      "tokens_seen": 1975517184
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082626262626263,
      "loss": 1.4414,
      "theoretical_loss": 3.4339129921328,
      "tokens_seen": 1975582720
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040824242424242424,
      "loss": 1.5263,
      "theoretical_loss": 3.4339034715493417,
      "tokens_seen": 1975648256
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082222222222222,
      "loss": 1.1537,
      "theoretical_loss": 3.433893951370118,
      "tokens_seen": 1975713792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004082020202020202,
      "loss": 1.4704,
      "theoretical_loss": 3.4338844315951,
      "tokens_seen": 1975779328
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004081818181818182,
      "loss": 1.41,
      "theoretical_loss": 3.4338749122242564,
      "tokens_seen": 1975844864
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004081616161616162,
      "loss": 1.5023,
      "theoretical_loss": 3.433865393257557,
      "tokens_seen": 1975910400
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040814141414141414,
      "loss": 1.5095,
      "theoretical_loss": 3.433855874694971,
      "tokens_seen": 1975975936
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004081212121212121,
      "loss": 1.5097,
      "theoretical_loss": 3.433846356536468,
      "tokens_seen": 1976041472
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004081010101010101,
      "loss": 1.1897,
      "theoretical_loss": 3.4338368387820166,
      "tokens_seen": 1976107008
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.38624143600463867,
      "objective/train/avg_token_score": 0.0034812039230018854,
      "objective/train/docs_used": 1117380,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2491487264633179,
      "objective/train/original_loss": 4.108737945556641,
      "objective/train/theoretical_loss": 3.433827321431587,
      "objective/train/tokens_used": 335031776,
      "objective/train/unlikelihood_loss": 0.0697765201330185,
      "theoretical_loss": 3.433827321431587,
      "tokens_seen": 1976172544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004080808080808081,
      "loss": 1.4446,
      "theoretical_loss": 3.433827321431587,
      "tokens_seen": 1976172544
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040806060606060607,
      "loss": 1.293,
      "theoretical_loss": 3.433817804485149,
      "tokens_seen": 1976238080
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040804040404040403,
      "loss": 1.415,
      "theoretical_loss": 3.4338082879426715,
      "tokens_seen": 1976303616
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040802020202020204,
      "loss": 1.3463,
      "theoretical_loss": 3.433798771804124,
      "tokens_seen": 1976369152
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000408,
      "loss": 1.6793,
      "theoretical_loss": 3.433789256069476,
      "tokens_seen": 1976434688
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407979797979798,
      "loss": 1.3098,
      "theoretical_loss": 3.433779740738697,
      "tokens_seen": 1976500224
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040795959595959596,
      "loss": 1.4687,
      "theoretical_loss": 3.433770225811757,
      "tokens_seen": 1976565760
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004079393939393939,
      "loss": 1.6195,
      "theoretical_loss": 3.433760711288624,
      "tokens_seen": 1976631296
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040791919191919193,
      "loss": 1.4593,
      "theoretical_loss": 3.4337511971692694,
      "tokens_seen": 1976696832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078989898989899,
      "loss": 1.4784,
      "theoretical_loss": 3.433741683453661,
      "tokens_seen": 1976762368
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078787878787879,
      "loss": 1.4241,
      "theoretical_loss": 3.4337321701417696,
      "tokens_seen": 1976827904
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040785858585858586,
      "loss": 1.6483,
      "theoretical_loss": 3.433722657233564,
      "tokens_seen": 1976893440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040783838383838387,
      "loss": 1.3813,
      "theoretical_loss": 3.4337131447290132,
      "tokens_seen": 1976958976
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004078181818181818,
      "loss": 1.5208,
      "theoretical_loss": 3.4337036326280876,
      "tokens_seen": 1977024512
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004077979797979798,
      "loss": 1.3666,
      "theoretical_loss": 3.433694120930756,
      "tokens_seen": 1977090048
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004077777777777778,
      "loss": 1.5639,
      "theoretical_loss": 3.4336846096369893,
      "tokens_seen": 1977155584
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040775757575757575,
      "loss": 1.3248,
      "theoretical_loss": 3.433675098746755,
      "tokens_seen": 1977221120
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040773737373737376,
      "loss": 1.2001,
      "theoretical_loss": 3.4336655882600238,
      "tokens_seen": 1977286656
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004077171717171717,
      "loss": 1.3965,
      "theoretical_loss": 3.433656078176765,
      "tokens_seen": 1977352192
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040769696969696973,
      "loss": 1.3558,
      "theoretical_loss": 3.4336465684969477,
      "tokens_seen": 1977417728
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076767676767677,
      "loss": 1.4859,
      "theoretical_loss": 3.4336370592205423,
      "tokens_seen": 1977483264
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076565656565657,
      "loss": 1.3785,
      "theoretical_loss": 3.4336275503475178,
      "tokens_seen": 1977548800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040763636363636365,
      "loss": 1.6469,
      "theoretical_loss": 3.4336180418778435,
      "tokens_seen": 1977614336
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004076161616161616,
      "loss": 1.3878,
      "theoretical_loss": 3.4336085338114897,
      "tokens_seen": 1977679872
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075959595959596,
      "loss": 1.2246,
      "theoretical_loss": 3.433599026148425,
      "tokens_seen": 1977745408
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.2834188640117645,
      "objective/train/avg_token_score": 0.007966935634613037,
      "objective/train/docs_used": 1118619,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9379705190658569,
      "objective/train/original_loss": 4.463184356689453,
      "objective/train/theoretical_loss": 3.4335895188886187,
      "objective/train/tokens_used": 336670176,
      "objective/train/unlikelihood_loss": 0.07986027747392654,
      "theoretical_loss": 3.4335895188886187,
      "tokens_seen": 1977810944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075757575757576,
      "loss": 1.4292,
      "theoretical_loss": 3.4335895188886187,
      "tokens_seen": 1977810944
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075555555555556,
      "loss": 1.4095,
      "theoretical_loss": 3.433580012032042,
      "tokens_seen": 1977876480
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040753535353535355,
      "loss": 1.4814,
      "theoretical_loss": 3.433570505578663,
      "tokens_seen": 1977942016
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004075151515151515,
      "loss": 1.5789,
      "theoretical_loss": 3.433560999528451,
      "tokens_seen": 1978007552
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074949494949495,
      "loss": 1.5362,
      "theoretical_loss": 3.433551493881377,
      "tokens_seen": 1978073088
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040747474747474747,
      "loss": 1.3804,
      "theoretical_loss": 3.4335419886374092,
      "tokens_seen": 1978138624
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074545454545455,
      "loss": 1.6188,
      "theoretical_loss": 3.433532483796518,
      "tokens_seen": 1978204160
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040743434343434344,
      "loss": 1.1727,
      "theoretical_loss": 3.433522979358673,
      "tokens_seen": 1978269696
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004074141414141414,
      "loss": 1.5376,
      "theoretical_loss": 3.433513475323843,
      "tokens_seen": 1978335232
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073939393939394,
      "loss": 1.2815,
      "theoretical_loss": 3.433503971691998,
      "tokens_seen": 1978400768
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073737373737374,
      "loss": 1.4291,
      "theoretical_loss": 3.433494468463107,
      "tokens_seen": 1978466304
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073535353535354,
      "loss": 1.3798,
      "theoretical_loss": 3.4334849656371405,
      "tokens_seen": 1978531840
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040733333333333333,
      "loss": 1.3716,
      "theoretical_loss": 3.4334754632140676,
      "tokens_seen": 1978597376
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004073131313131313,
      "loss": 1.327,
      "theoretical_loss": 3.433465961193858,
      "tokens_seen": 1978662912
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072929292929293,
      "loss": 1.3915,
      "theoretical_loss": 3.4334564595764814,
      "tokens_seen": 1978728448
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072727272727273,
      "loss": 1.543,
      "theoretical_loss": 3.4334469583619067,
      "tokens_seen": 1978793984
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040725252525252527,
      "loss": 1.3711,
      "theoretical_loss": 3.433437457550104,
      "tokens_seen": 1978859520
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072323232323232,
      "loss": 1.2373,
      "theoretical_loss": 3.4334279571410438,
      "tokens_seen": 1978925056
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004072121212121212,
      "loss": 1.4841,
      "theoretical_loss": 3.4334184571346933,
      "tokens_seen": 1978990592
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040719191919191925,
      "loss": 1.5336,
      "theoretical_loss": 3.4334089575310247,
      "tokens_seen": 1979056128
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004071717171717172,
      "loss": 1.4586,
      "theoretical_loss": 3.433399458330006,
      "tokens_seen": 1979121664
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040715151515151516,
      "loss": 1.2878,
      "theoretical_loss": 3.433389959531607,
      "tokens_seen": 1979187200
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004071313131313131,
      "loss": 1.5874,
      "theoretical_loss": 3.433380461135798,
      "tokens_seen": 1979252736
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004071111111111111,
      "loss": 1.5125,
      "theoretical_loss": 3.433370963142548,
      "tokens_seen": 1979318272
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040709090909090914,
      "loss": 1.7091,
      "theoretical_loss": 3.433361465551827,
      "tokens_seen": 1979383808
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.34341397881507874,
      "objective/train/avg_token_score": 0.009266973473131657,
      "objective/train/docs_used": 1119354,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2784351110458374,
      "objective/train/original_loss": 4.967345714569092,
      "objective/train/theoretical_loss": 3.4333519683636036,
      "objective/train/tokens_used": 338308576,
      "objective/train/unlikelihood_loss": 0.047723278403282166,
      "theoretical_loss": 3.4333519683636036,
      "tokens_seen": 1979449344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004070707070707071,
      "loss": 1.5983,
      "theoretical_loss": 3.4333519683636036,
      "tokens_seen": 1979449344
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040705050505050505,
      "loss": 1.3467,
      "theoretical_loss": 3.433342471577849,
      "tokens_seen": 1979514880
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407030303030303,
      "loss": 1.4703,
      "theoretical_loss": 3.4333329751945314,
      "tokens_seen": 1979580416
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000407010101010101,
      "loss": 1.607,
      "theoretical_loss": 3.433323479213622,
      "tokens_seen": 1979645952
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040698989898989903,
      "loss": 1.3784,
      "theoretical_loss": 3.4333139836350886,
      "tokens_seen": 1979711488
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000406969696969697,
      "loss": 1.3995,
      "theoretical_loss": 3.433304488458902,
      "tokens_seen": 1979777024
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040694949494949495,
      "loss": 1.6541,
      "theoretical_loss": 3.433294993685031,
      "tokens_seen": 1979842560
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004069292929292929,
      "loss": 1.2676,
      "theoretical_loss": 3.4332854993134463,
      "tokens_seen": 1979908096
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004069090909090909,
      "loss": 1.3127,
      "theoretical_loss": 3.4332760053441174,
      "tokens_seen": 1979973632
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068888888888889,
      "loss": 1.4902,
      "theoretical_loss": 3.4332665117770125,
      "tokens_seen": 1980039168
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068686868686869,
      "loss": 1.1425,
      "theoretical_loss": 3.433257018612103,
      "tokens_seen": 1980104704
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040684848484848484,
      "loss": 1.5914,
      "theoretical_loss": 3.433247525849358,
      "tokens_seen": 1980170240
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040682828282828285,
      "loss": 1.4876,
      "theoretical_loss": 3.433238033488747,
      "tokens_seen": 1980235776
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004068080808080808,
      "loss": 1.5058,
      "theoretical_loss": 3.433228541530239,
      "tokens_seen": 1980301312
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067878787878788,
      "loss": 1.3863,
      "theoretical_loss": 3.433219049973805,
      "tokens_seen": 1980366848
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067676767676768,
      "loss": 1.5363,
      "theoretical_loss": 3.4332095588194136,
      "tokens_seen": 1980432384
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040674747474747473,
      "loss": 1.1681,
      "theoretical_loss": 3.433200068067035,
      "tokens_seen": 1980497920
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040672727272727274,
      "loss": 1.3775,
      "theoretical_loss": 3.4331905777166387,
      "tokens_seen": 1980563456
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004067070707070707,
      "loss": 1.5538,
      "theoretical_loss": 3.4331810877681948,
      "tokens_seen": 1980628992
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004066868686868687,
      "loss": 1.3664,
      "theoretical_loss": 3.433171598221672,
      "tokens_seen": 1980694528
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040666666666666667,
      "loss": 1.2929,
      "theoretical_loss": 3.43316210907704,
      "tokens_seen": 1980760064
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004066464646464647,
      "loss": 1.6327,
      "theoretical_loss": 3.43315262033427,
      "tokens_seen": 1980825600
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00040662626262626264,
      "loss": 1.3788,
      "theoretical_loss": 3.4331431319933303,
      "tokens_seen": 1980891136
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004066060606060606,
      "loss": 1.3966,
      "theoretical_loss": 3.4331336440541915,
      "tokens_seen": 1980956672
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0004065858585858586,
      "loss": 1.5779,
      "theoretical_loss": 3.433124156516823,
      "tokens_seen": 1981022208
    },
    {
      "epoch": 0.19,
      "objective/train/aligned_token_count": 0.41037389636039734,
      "objective/train/avg_token_score": 0.023919859901070595,
      "objective/train/docs_used": 1120511,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3339953422546387,
      "objective/train/original_loss": 4.790691375732422,
      "objective/train/theoretical_loss": 3.4331146693811934,
      "objective/train/tokens_used": 339946976,
      "objective/train/unlikelihood_loss": 0.049351904541254044,
      "theoretical_loss": 3.4331146693811934,
      "tokens_seen": 1981087744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040656565656565656,
      "loss": 1.422,
      "theoretical_loss": 3.4331146693811934,
      "tokens_seen": 1981087744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040654545454545457,
      "loss": 1.6359,
      "theoretical_loss": 3.4331051826472736,
      "tokens_seen": 1981153280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040652525252525253,
      "loss": 1.3955,
      "theoretical_loss": 3.433095696315033,
      "tokens_seen": 1981218816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004065050505050505,
      "loss": 1.5511,
      "theoretical_loss": 3.4330862103844417,
      "tokens_seen": 1981284352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064848484848485,
      "loss": 1.4746,
      "theoretical_loss": 3.4330767248554688,
      "tokens_seen": 1981349888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064646464646465,
      "loss": 1.6425,
      "theoretical_loss": 3.433067239728084,
      "tokens_seen": 1981415424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040644444444444446,
      "loss": 1.1774,
      "theoretical_loss": 3.4330577550022574,
      "tokens_seen": 1981480960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064242424242424,
      "loss": 1.5055,
      "theoretical_loss": 3.4330482706779586,
      "tokens_seen": 1981546496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004064040404040404,
      "loss": 1.3855,
      "theoretical_loss": 3.4330387867551573,
      "tokens_seen": 1981612032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004063838383838384,
      "loss": 1.3515,
      "theoretical_loss": 3.4330293032338233,
      "tokens_seen": 1981677568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004063636363636364,
      "loss": 1.4944,
      "theoretical_loss": 3.433019820113926,
      "tokens_seen": 1981743104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040634343434343436,
      "loss": 1.4746,
      "theoretical_loss": 3.4330103373954355,
      "tokens_seen": 1981808640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004063232323232323,
      "loss": 1.2959,
      "theoretical_loss": 3.4330008550783213,
      "tokens_seen": 1981874176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040630303030303027,
      "loss": 1.5334,
      "theoretical_loss": 3.4329913731625536,
      "tokens_seen": 1981939712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040628282828282834,
      "loss": 1.2631,
      "theoretical_loss": 3.4329818916481014,
      "tokens_seen": 1982005248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062626262626263,
      "loss": 1.6203,
      "theoretical_loss": 3.4329724105349353,
      "tokens_seen": 1982070784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040624242424242425,
      "loss": 1.4313,
      "theoretical_loss": 3.432962929823024,
      "tokens_seen": 1982136320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004062222222222222,
      "loss": 1.415,
      "theoretical_loss": 3.432953449512338,
      "tokens_seen": 1982201856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040620202020202016,
      "loss": 1.4721,
      "theoretical_loss": 3.432943969602847,
      "tokens_seen": 1982267392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040618181818181823,
      "loss": 1.4335,
      "theoretical_loss": 3.432934490094521,
      "tokens_seen": 1982332928
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004061616161616162,
      "loss": 1.4509,
      "theoretical_loss": 3.4329250109873293,
      "tokens_seen": 1982398464
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040614141414141414,
      "loss": 1.4384,
      "theoretical_loss": 3.4329155322812417,
      "tokens_seen": 1982464000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004061212121212121,
      "loss": 1.4729,
      "theoretical_loss": 3.432906053976228,
      "tokens_seen": 1982529536
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040610101010101006,
      "loss": 1.4978,
      "theoretical_loss": 3.432896576072258,
      "tokens_seen": 1982595072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004060808080808081,
      "loss": 1.5901,
      "theoretical_loss": 3.432887098569301,
      "tokens_seen": 1982660608
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.5727028250694275,
      "objective/train/avg_token_score": 0.004548849072307348,
      "objective/train/docs_used": 1121670,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.048727512359619,
      "objective/train/original_loss": 4.066925525665283,
      "objective/train/theoretical_loss": 3.4328776214673278,
      "objective/train/tokens_used": 341585376,
      "objective/train/unlikelihood_loss": 0.07542819529771805,
      "theoretical_loss": 3.4328776214673278,
      "tokens_seen": 1982726144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004060606060606061,
      "loss": 1.7648,
      "theoretical_loss": 3.4328776214673278,
      "tokens_seen": 1982726144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040604040404040403,
      "loss": 1.4492,
      "theoretical_loss": 3.432868144766308,
      "tokens_seen": 1982791680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000406020202020202,
      "loss": 1.4899,
      "theoretical_loss": 3.432858668466211,
      "tokens_seen": 1982857216
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040600000000000006,
      "loss": 1.5625,
      "theoretical_loss": 3.432849192567006,
      "tokens_seen": 1982922752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405979797979798,
      "loss": 1.49,
      "theoretical_loss": 3.4328397170686635,
      "tokens_seen": 1982988288
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040595959595959597,
      "loss": 1.6991,
      "theoretical_loss": 3.4328302419711534,
      "tokens_seen": 1983053824
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040593939393939393,
      "loss": 1.5363,
      "theoretical_loss": 3.4328207672744453,
      "tokens_seen": 1983119360
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004059191919191919,
      "loss": 1.3888,
      "theoretical_loss": 3.432811292978509,
      "tokens_seen": 1983184896
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040589898989898995,
      "loss": 1.4623,
      "theoretical_loss": 3.432801819083315,
      "tokens_seen": 1983250432
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058787878787879,
      "loss": 1.2032,
      "theoretical_loss": 3.4327923455888314,
      "tokens_seen": 1983315968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040585858585858586,
      "loss": 1.5196,
      "theoretical_loss": 3.4327828724950296,
      "tokens_seen": 1983381504
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004058383838383838,
      "loss": 1.3848,
      "theoretical_loss": 3.432773399801879,
      "tokens_seen": 1983447040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040581818181818183,
      "loss": 1.2752,
      "theoretical_loss": 3.432763927509349,
      "tokens_seen": 1983512576
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040579797979797984,
      "loss": 1.4994,
      "theoretical_loss": 3.4327544556174097,
      "tokens_seen": 1983578112
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057777777777778,
      "loss": 1.5548,
      "theoretical_loss": 3.432744984126031,
      "tokens_seen": 1983643648
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040575757575757576,
      "loss": 1.3647,
      "theoretical_loss": 3.4327355130351824,
      "tokens_seen": 1983709184
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057373737373737,
      "loss": 1.3819,
      "theoretical_loss": 3.4327260423448345,
      "tokens_seen": 1983774720
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004057171717171717,
      "loss": 1.1727,
      "theoretical_loss": 3.4327165720549564,
      "tokens_seen": 1983840256
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040569696969696973,
      "loss": 1.442,
      "theoretical_loss": 3.4327071021655176,
      "tokens_seen": 1983905792
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004056767676767677,
      "loss": 1.3211,
      "theoretical_loss": 3.432697632676489,
      "tokens_seen": 1983971328
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040565656565656565,
      "loss": 1.4399,
      "theoretical_loss": 3.43268816358784,
      "tokens_seen": 1984036864
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040563636363636366,
      "loss": 1.401,
      "theoretical_loss": 3.4326786948995407,
      "tokens_seen": 1984102400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004056161616161616,
      "loss": 1.7314,
      "theoretical_loss": 3.43266922661156,
      "tokens_seen": 1984167936
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040559595959595963,
      "loss": 1.4885,
      "theoretical_loss": 3.4326597587238687,
      "tokens_seen": 1984233472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055757575757576,
      "loss": 1.193,
      "theoretical_loss": 3.432650291236436,
      "tokens_seen": 1984299008
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.46981915831565857,
      "objective/train/avg_token_score": 0.007897254079580307,
      "objective/train/docs_used": 1122413,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6054514646530151,
      "objective/train/original_loss": 4.072227954864502,
      "objective/train/theoretical_loss": 3.4326408241492326,
      "objective/train/tokens_used": 343223776,
      "objective/train/unlikelihood_loss": 0.05646788701415062,
      "theoretical_loss": 3.4326408241492326,
      "tokens_seen": 1984364544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040555555555555554,
      "loss": 1.4734,
      "theoretical_loss": 3.4326408241492326,
      "tokens_seen": 1984364544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040553535353535355,
      "loss": 1.5419,
      "theoretical_loss": 3.4326313574622276,
      "tokens_seen": 1984430080
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004055151515151515,
      "loss": 1.4651,
      "theoretical_loss": 3.432621891175391,
      "tokens_seen": 1984495616
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004054949494949495,
      "loss": 1.5656,
      "theoretical_loss": 3.432612425288693,
      "tokens_seen": 1984561152
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004054747474747475,
      "loss": 1.3567,
      "theoretical_loss": 3.4326029598021033,
      "tokens_seen": 1984626688
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004054545454545455,
      "loss": 1.5375,
      "theoretical_loss": 3.432593494715592,
      "tokens_seen": 1984692224
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040543434343434344,
      "loss": 1.2787,
      "theoretical_loss": 3.432584030029128,
      "tokens_seen": 1984757760
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004054141414141414,
      "loss": 1.5412,
      "theoretical_loss": 3.4325745657426827,
      "tokens_seen": 1984823296
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053939393939394,
      "loss": 1.6001,
      "theoretical_loss": 3.432565101856224,
      "tokens_seen": 1984888832
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040537373737373737,
      "loss": 1.4601,
      "theoretical_loss": 3.4325556383697244,
      "tokens_seen": 1984954368
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053535353535354,
      "loss": 1.4659,
      "theoretical_loss": 3.432546175283152,
      "tokens_seen": 1985019904
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040533333333333334,
      "loss": 1.4451,
      "theoretical_loss": 3.4325367125964767,
      "tokens_seen": 1985085440
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004053131313131313,
      "loss": 1.476,
      "theoretical_loss": 3.4325272503096693,
      "tokens_seen": 1985150976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052929292929293,
      "loss": 1.4559,
      "theoretical_loss": 3.4325177884226985,
      "tokens_seen": 1985216512
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052727272727273,
      "loss": 1.4112,
      "theoretical_loss": 3.4325083269355354,
      "tokens_seen": 1985282048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052525252525253,
      "loss": 1.6486,
      "theoretical_loss": 3.4324988658481494,
      "tokens_seen": 1985347584
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040523232323232323,
      "loss": 1.5827,
      "theoretical_loss": 3.43248940516051,
      "tokens_seen": 1985413120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004052121212121212,
      "loss": 1.5105,
      "theoretical_loss": 3.432479944872588,
      "tokens_seen": 1985478656
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051919191919192,
      "loss": 1.4755,
      "theoretical_loss": 3.4324704849843526,
      "tokens_seen": 1985544192
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051717171717172,
      "loss": 1.5322,
      "theoretical_loss": 3.4324610254957744,
      "tokens_seen": 1985609728
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040515151515151517,
      "loss": 1.359,
      "theoretical_loss": 3.4324515664068223,
      "tokens_seen": 1985675264
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051313131313131,
      "loss": 1.3518,
      "theoretical_loss": 3.432442107717467,
      "tokens_seen": 1985740800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004051111111111111,
      "loss": 1.5694,
      "theoretical_loss": 3.4324326494276782,
      "tokens_seen": 1985806336
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040509090909090915,
      "loss": 1.2736,
      "theoretical_loss": 3.432423191537426,
      "tokens_seen": 1985871872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004050707070707071,
      "loss": 1.4792,
      "theoretical_loss": 3.4324137340466807,
      "tokens_seen": 1985937408
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.2547653913497925,
      "objective/train/avg_token_score": 0.016683774068951607,
      "objective/train/docs_used": 1123893,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.856845498085022,
      "objective/train/original_loss": 5.153318881988525,
      "objective/train/theoretical_loss": 3.432404276955411,
      "objective/train/tokens_used": 344862176,
      "objective/train/unlikelihood_loss": 0.05154005065560341,
      "theoretical_loss": 3.432404276955411,
      "tokens_seen": 1986002944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040505050505050506,
      "loss": 1.2664,
      "theoretical_loss": 3.432404276955411,
      "tokens_seen": 1986002944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000405030303030303,
      "loss": 1.6352,
      "theoretical_loss": 3.432394820263588,
      "tokens_seen": 1986068480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040501010101010097,
      "loss": 1.4826,
      "theoretical_loss": 3.432385363971181,
      "tokens_seen": 1986134016
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040498989898989904,
      "loss": 1.3789,
      "theoretical_loss": 3.4323759080781606,
      "tokens_seen": 1986199552
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404969696969697,
      "loss": 1.4896,
      "theoretical_loss": 3.432366452584496,
      "tokens_seen": 1986265088
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040494949494949495,
      "loss": 1.5823,
      "theoretical_loss": 3.4323569974901575,
      "tokens_seen": 1986330624
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004049292929292929,
      "loss": 1.5142,
      "theoretical_loss": 3.4323475427951156,
      "tokens_seen": 1986396160
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004049090909090909,
      "loss": 1.4073,
      "theoretical_loss": 3.4323380884993395,
      "tokens_seen": 1986461696
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040488888888888893,
      "loss": 1.4661,
      "theoretical_loss": 3.4323286346027997,
      "tokens_seen": 1986527232
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004048686868686869,
      "loss": 1.5341,
      "theoretical_loss": 3.4323191811054654,
      "tokens_seen": 1986592768
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040484848484848484,
      "loss": 1.314,
      "theoretical_loss": 3.432309728007307,
      "tokens_seen": 1986658304
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004048282828282828,
      "loss": 1.2981,
      "theoretical_loss": 3.432300275308295,
      "tokens_seen": 1986723840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004048080808080808,
      "loss": 1.372,
      "theoretical_loss": 3.4322908230083984,
      "tokens_seen": 1986789376
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047878787878788,
      "loss": 1.4414,
      "theoretical_loss": 3.432281371107588,
      "tokens_seen": 1986854912
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047676767676768,
      "loss": 1.6455,
      "theoretical_loss": 3.4322719196058338,
      "tokens_seen": 1986920448
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040474747474747474,
      "loss": 1.3235,
      "theoretical_loss": 3.4322624685031053,
      "tokens_seen": 1986985984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040472727272727275,
      "loss": 1.3656,
      "theoretical_loss": 3.4322530177993724,
      "tokens_seen": 1987051520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004047070707070707,
      "loss": 1.3143,
      "theoretical_loss": 3.4322435674946057,
      "tokens_seen": 1987117056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004046868686868687,
      "loss": 1.6096,
      "theoretical_loss": 3.4322341175887745,
      "tokens_seen": 1987182592
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040466666666666667,
      "loss": 1.2566,
      "theoretical_loss": 3.4322246680818496,
      "tokens_seen": 1987248128
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040464646464646463,
      "loss": 1.2232,
      "theoretical_loss": 3.4322152189738,
      "tokens_seen": 1987313664
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040462626262626264,
      "loss": 1.5018,
      "theoretical_loss": 3.4322057702645967,
      "tokens_seen": 1987379200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004046060606060606,
      "loss": 1.5121,
      "theoretical_loss": 3.4321963219542098,
      "tokens_seen": 1987444736
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004045858585858586,
      "loss": 1.3607,
      "theoretical_loss": 3.4321868740426082,
      "tokens_seen": 1987510272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040456565656565657,
      "loss": 1.818,
      "theoretical_loss": 3.432177426529763,
      "tokens_seen": 1987575808
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.5406891703605652,
      "objective/train/avg_token_score": 0.005704171024262905,
      "objective/train/docs_used": 1124511,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8316115140914917,
      "objective/train/original_loss": 3.5460124015808105,
      "objective/train/theoretical_loss": 3.4321679794156434,
      "objective/train/tokens_used": 346500576,
      "objective/train/unlikelihood_loss": 0.07432541251182556,
      "theoretical_loss": 3.4321679794156434,
      "tokens_seen": 1987641344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004045454545454546,
      "loss": 1.5832,
      "theoretical_loss": 3.4321679794156434,
      "tokens_seen": 1987641344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040452525252525253,
      "loss": 1.3207,
      "theoretical_loss": 3.4321585327002198,
      "tokens_seen": 1987706880
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004045050505050505,
      "loss": 1.3359,
      "theoretical_loss": 3.432149086383462,
      "tokens_seen": 1987772416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004044848484848485,
      "loss": 1.4143,
      "theoretical_loss": 3.4321396404653406,
      "tokens_seen": 1987837952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040446464646464646,
      "loss": 1.3399,
      "theoretical_loss": 3.4321301949458256,
      "tokens_seen": 1987903488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040444444444444447,
      "loss": 1.4967,
      "theoretical_loss": 3.432120749824886,
      "tokens_seen": 1987969024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004044242424242424,
      "loss": 1.6424,
      "theoretical_loss": 3.4321113051024934,
      "tokens_seen": 1988034560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004044040404040404,
      "loss": 1.656,
      "theoretical_loss": 3.4321018607786167,
      "tokens_seen": 1988100096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004043838383838384,
      "loss": 1.3635,
      "theoretical_loss": 3.4320924168532256,
      "tokens_seen": 1988165632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040436363636363635,
      "loss": 1.4125,
      "theoretical_loss": 3.432082973326292,
      "tokens_seen": 1988231168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040434343434343436,
      "loss": 1.5523,
      "theoretical_loss": 3.432073530197784,
      "tokens_seen": 1988296704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004043232323232323,
      "loss": 1.4857,
      "theoretical_loss": 3.4320640874676727,
      "tokens_seen": 1988362240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040430303030303033,
      "loss": 1.3104,
      "theoretical_loss": 3.4320546451359277,
      "tokens_seen": 1988427776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042828282828283,
      "loss": 1.4525,
      "theoretical_loss": 3.4320452032025193,
      "tokens_seen": 1988493312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042626262626263,
      "loss": 1.4042,
      "theoretical_loss": 3.4320357616674175,
      "tokens_seen": 1988558848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040424242424242425,
      "loss": 1.387,
      "theoretical_loss": 3.4320263205305928,
      "tokens_seen": 1988624384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042222222222222,
      "loss": 1.398,
      "theoretical_loss": 3.4320168797920143,
      "tokens_seen": 1988689920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004042020202020202,
      "loss": 1.3234,
      "theoretical_loss": 3.432007439451653,
      "tokens_seen": 1988755456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041818181818182,
      "loss": 1.6412,
      "theoretical_loss": 3.4319979995094787,
      "tokens_seen": 1988820992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041616161616162,
      "loss": 1.5261,
      "theoretical_loss": 3.431988559965461,
      "tokens_seen": 1988886528
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040414141414141415,
      "loss": 1.5544,
      "theoretical_loss": 3.431979120819571,
      "tokens_seen": 1988952064
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041212121212121,
      "loss": 1.4711,
      "theoretical_loss": 3.4319696820717778,
      "tokens_seen": 1989017600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004041010101010101,
      "loss": 1.2866,
      "theoretical_loss": 3.431960243722052,
      "tokens_seen": 1989083136
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004040808080808081,
      "loss": 1.435,
      "theoretical_loss": 3.4319508057703634,
      "tokens_seen": 1989148672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004040606060606061,
      "loss": 1.4636,
      "theoretical_loss": 3.4319413682166826,
      "tokens_seen": 1989214208
    },
    {
      "debugging/Self-BLEU-5": 0.4683363022964717,
      "debugging/distinct-1-grams": 0.8034092501484725,
      "debugging/distinct-2-grams": 0.9641002767459668,
      "debugging/entropy-1-grams": 6.198537273223117,
      "debugging/entropy-2-grams": 7.019141508423205,
      "debugging/length": 468.3529411764706,
      "debugging/num_segments": 17,
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.21560361981391907,
      "objective/train/avg_token_score": 0.04828590154647827,
      "objective/train/docs_used": 1125297,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8591703772544861,
      "objective/train/original_loss": 5.175046443939209,
      "objective/train/theoretical_loss": 3.431931931060979,
      "objective/train/tokens_used": 348138976,
      "objective/train/unlikelihood_loss": 0.05250702425837517,
      "theoretical_loss": 3.431931931060979,
      "tokens_seen": 1989279744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040404040404040404,
      "loss": 1.3473,
      "theoretical_loss": 3.431931931060979,
      "tokens_seen": 1989279744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404020202020202,
      "loss": 1.4626,
      "theoretical_loss": 3.4319224943032234,
      "tokens_seen": 1989345280
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000404,
      "loss": 1.278,
      "theoretical_loss": 3.431913057943386,
      "tokens_seen": 1989410816
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403979797979798,
      "loss": 1.5107,
      "theoretical_loss": 3.4319036219814363,
      "tokens_seen": 1989476352
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403959595959596,
      "loss": 1.5861,
      "theoretical_loss": 3.4318941864173445,
      "tokens_seen": 1989541888
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040393939393939393,
      "loss": 1.2866,
      "theoretical_loss": 3.4318847512510806,
      "tokens_seen": 1989607424
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004039191919191919,
      "loss": 1.5729,
      "theoretical_loss": 3.431875316482615,
      "tokens_seen": 1989672960
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040389898989898995,
      "loss": 1.3685,
      "theoretical_loss": 3.4318658821119183,
      "tokens_seen": 1989738496
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004038787878787879,
      "loss": 1.5106,
      "theoretical_loss": 3.43185644813896,
      "tokens_seen": 1989804032
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040385858585858587,
      "loss": 1.2137,
      "theoretical_loss": 3.43184701456371,
      "tokens_seen": 1989869568
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004038383838383838,
      "loss": 1.449,
      "theoretical_loss": 3.431837581386139,
      "tokens_seen": 1989935104
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004038181818181818,
      "loss": 1.2781,
      "theoretical_loss": 3.4318281486062174,
      "tokens_seen": 1990000640
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040379797979797985,
      "loss": 1.7174,
      "theoretical_loss": 3.4318187162239147,
      "tokens_seen": 1990066176
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037777777777778,
      "loss": 1.3277,
      "theoretical_loss": 3.4318092842392005,
      "tokens_seen": 1990131712
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040375757575757576,
      "loss": 1.4206,
      "theoretical_loss": 3.4317998526520466,
      "tokens_seen": 1990197248
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004037373737373737,
      "loss": 1.425,
      "theoretical_loss": 3.4317904214624217,
      "tokens_seen": 1990262784
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040371717171717173,
      "loss": 1.4692,
      "theoretical_loss": 3.4317809906702967,
      "tokens_seen": 1990328320
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040369696969696974,
      "loss": 1.345,
      "theoretical_loss": 3.4317715602756413,
      "tokens_seen": 1990393856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036767676767677,
      "loss": 1.3247,
      "theoretical_loss": 3.4317621302784262,
      "tokens_seen": 1990459392
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040365656565656565,
      "loss": 1.5162,
      "theoretical_loss": 3.4317527006786217,
      "tokens_seen": 1990524928
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036363636363636,
      "loss": 1.4397,
      "theoretical_loss": 3.431743271476197,
      "tokens_seen": 1990590464
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004036161616161616,
      "loss": 1.5959,
      "theoretical_loss": 3.4317338426711226,
      "tokens_seen": 1990656000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040359595959595963,
      "loss": 1.3966,
      "theoretical_loss": 3.4317244142633694,
      "tokens_seen": 1990721536
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004035757575757576,
      "loss": 1.3737,
      "theoretical_loss": 3.4317149862529073,
      "tokens_seen": 1990787072
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040355555555555555,
      "loss": 1.4397,
      "theoretical_loss": 3.4317055586397056,
      "tokens_seen": 1990852608
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.2753543555736542,
      "objective/train/avg_token_score": 0.040420789271593094,
      "objective/train/docs_used": 1126404,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8407698273658752,
      "objective/train/original_loss": 5.04019832611084,
      "objective/train/theoretical_loss": 3.4316961314237355,
      "objective/train/tokens_used": 349777376,
      "objective/train/unlikelihood_loss": 0.060525793582201004,
      "theoretical_loss": 3.4316961314237355,
      "tokens_seen": 1990918144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040353535353535356,
      "loss": 1.468,
      "theoretical_loss": 3.4316961314237355,
      "tokens_seen": 1990918144
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004035151515151515,
      "loss": 1.349,
      "theoretical_loss": 3.431686704604967,
      "tokens_seen": 1990983680
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034949494949495,
      "loss": 1.2932,
      "theoretical_loss": 3.4316772781833693,
      "tokens_seen": 1991049216
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034747474747475,
      "loss": 1.5111,
      "theoretical_loss": 3.431667852158914,
      "tokens_seen": 1991114752
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040345454545454544,
      "loss": 1.5132,
      "theoretical_loss": 3.4316584265315706,
      "tokens_seen": 1991180288
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040343434343434345,
      "loss": 1.6366,
      "theoretical_loss": 3.43164900130131,
      "tokens_seen": 1991245824
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004034141414141414,
      "loss": 1.3234,
      "theoretical_loss": 3.4316395764681014,
      "tokens_seen": 1991311360
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033939393939394,
      "loss": 1.3436,
      "theoretical_loss": 3.4316301520319152,
      "tokens_seen": 1991376896
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033737373737374,
      "loss": 1.3677,
      "theoretical_loss": 3.4316207279927218,
      "tokens_seen": 1991442432
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033535353535354,
      "loss": 1.3794,
      "theoretical_loss": 3.4316113043504917,
      "tokens_seen": 1991507968
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040333333333333334,
      "loss": 1.1624,
      "theoretical_loss": 3.4316018811051947,
      "tokens_seen": 1991573504
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004033131313131313,
      "loss": 1.4437,
      "theoretical_loss": 3.4315924582568016,
      "tokens_seen": 1991639040
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004032929292929293,
      "loss": 1.0514,
      "theoretical_loss": 3.4315830358052817,
      "tokens_seen": 1991704576
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040327272727272727,
      "loss": 1.433,
      "theoretical_loss": 3.431573613750606,
      "tokens_seen": 1991770112
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004032525252525253,
      "loss": 1.3241,
      "theoretical_loss": 3.431564192092744,
      "tokens_seen": 1991835648
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040323232323232324,
      "loss": 1.4644,
      "theoretical_loss": 3.431554770831667,
      "tokens_seen": 1991901184
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004032121212121212,
      "loss": 1.4245,
      "theoretical_loss": 3.4315453499673443,
      "tokens_seen": 1991966720
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004031919191919192,
      "loss": 1.5533,
      "theoretical_loss": 3.4315359294997467,
      "tokens_seen": 1992032256
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004031717171717172,
      "loss": 1.4905,
      "theoretical_loss": 3.431526509428844,
      "tokens_seen": 1992097792
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040315151515151517,
      "loss": 1.3922,
      "theoretical_loss": 3.431517089754607,
      "tokens_seen": 1992163328
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040313131313131313,
      "loss": 1.4556,
      "theoretical_loss": 3.4315076704770053,
      "tokens_seen": 1992228864
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004031111111111111,
      "loss": 1.4792,
      "theoretical_loss": 3.4314982515960093,
      "tokens_seen": 1992294400
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004030909090909091,
      "loss": 1.7444,
      "theoretical_loss": 3.4314888331115894,
      "tokens_seen": 1992359936
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004030707070707071,
      "loss": 1.3701,
      "theoretical_loss": 3.4314794150237162,
      "tokens_seen": 1992425472
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040305050505050506,
      "loss": 1.2929,
      "theoretical_loss": 3.4314699973323597,
      "tokens_seen": 1992491008
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.40243157744407654,
      "objective/train/avg_token_score": 0.0219747181981802,
      "objective/train/docs_used": 1126980,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4165253639221191,
      "objective/train/original_loss": 4.742091655731201,
      "objective/train/theoretical_loss": 3.4314605800374895,
      "objective/train/tokens_used": 351415776,
      "objective/train/unlikelihood_loss": 0.05318494141101837,
      "theoretical_loss": 3.4314605800374895,
      "tokens_seen": 1992556544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403030303030303,
      "loss": 1.4698,
      "theoretical_loss": 3.4314605800374895,
      "tokens_seen": 1992556544
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000403010101010101,
      "loss": 1.3088,
      "theoretical_loss": 3.4314511631390774,
      "tokens_seen": 1992622080
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040298989898989904,
      "loss": 1.3889,
      "theoretical_loss": 3.431441746637092,
      "tokens_seen": 1992687616
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402969696969697,
      "loss": 1.4465,
      "theoretical_loss": 3.4314323305315044,
      "tokens_seen": 1992753152
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040294949494949496,
      "loss": 1.3891,
      "theoretical_loss": 3.431422914822285,
      "tokens_seen": 1992818688
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004029292929292929,
      "loss": 1.5381,
      "theoretical_loss": 3.431413499509404,
      "tokens_seen": 1992884224
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040290909090909087,
      "loss": 1.5889,
      "theoretical_loss": 3.431404084592831,
      "tokens_seen": 1992949760
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040288888888888894,
      "loss": 1.3384,
      "theoretical_loss": 3.4313946700725375,
      "tokens_seen": 1993015296
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004028686868686869,
      "loss": 1.1691,
      "theoretical_loss": 3.4313852559484928,
      "tokens_seen": 1993080832
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040284848484848485,
      "loss": 1.2463,
      "theoretical_loss": 3.4313758422206675,
      "tokens_seen": 1993146368
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004028282828282828,
      "loss": 1.4076,
      "theoretical_loss": 3.4313664288890324,
      "tokens_seen": 1993211904
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040280808080808076,
      "loss": 1.5663,
      "theoretical_loss": 3.431357015953557,
      "tokens_seen": 1993277440
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040278787878787883,
      "loss": 1.505,
      "theoretical_loss": 3.4313476034142116,
      "tokens_seen": 1993342976
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004027676767676768,
      "loss": 1.3878,
      "theoretical_loss": 3.431338191270967,
      "tokens_seen": 1993408512
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040274747474747474,
      "loss": 1.7233,
      "theoretical_loss": 3.4313287795237932,
      "tokens_seen": 1993474048
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004027272727272727,
      "loss": 1.4578,
      "theoretical_loss": 3.4313193681726606,
      "tokens_seen": 1993539584
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004027070707070707,
      "loss": 1.3499,
      "theoretical_loss": 3.4313099572175405,
      "tokens_seen": 1993605120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026868686868687,
      "loss": 1.5321,
      "theoretical_loss": 3.4313005466584015,
      "tokens_seen": 1993670656
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026666666666667,
      "loss": 1.6351,
      "theoretical_loss": 3.431291136495215,
      "tokens_seen": 1993736192
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040264646464646464,
      "loss": 1.1717,
      "theoretical_loss": 3.4312817267279505,
      "tokens_seen": 1993801728
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026262626262626,
      "loss": 1.3782,
      "theoretical_loss": 3.431272317356579,
      "tokens_seen": 1993867264
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004026060606060606,
      "loss": 1.5529,
      "theoretical_loss": 3.431262908381071,
      "tokens_seen": 1993932800
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004025858585858586,
      "loss": 1.255,
      "theoretical_loss": 3.4312534998013966,
      "tokens_seen": 1993998336
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040256565656565657,
      "loss": 1.4759,
      "theoretical_loss": 3.431244091617526,
      "tokens_seen": 1994063872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040254545454545453,
      "loss": 1.3079,
      "theoretical_loss": 3.4312346838294294,
      "tokens_seen": 1994129408
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.42699170112609863,
      "objective/train/avg_token_score": 0.011511247605085373,
      "objective/train/docs_used": 1128284,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.334111213684082,
      "objective/train/original_loss": 3.9049336910247803,
      "objective/train/theoretical_loss": 3.431225276437078,
      "objective/train/tokens_used": 353054176,
      "objective/train/unlikelihood_loss": 0.11700768768787384,
      "theoretical_loss": 3.431225276437078,
      "tokens_seen": 1994194944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040252525252525254,
      "loss": 1.3111,
      "theoretical_loss": 3.431225276437078,
      "tokens_seen": 1994194944
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040250505050505055,
      "loss": 1.3331,
      "theoretical_loss": 3.431215869440441,
      "tokens_seen": 1994260480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004024848484848485,
      "loss": 1.3455,
      "theoretical_loss": 3.431206462839489,
      "tokens_seen": 1994326016
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040246464646464646,
      "loss": 1.5316,
      "theoretical_loss": 3.4311970566341934,
      "tokens_seen": 1994391552
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004024444444444444,
      "loss": 1.3737,
      "theoretical_loss": 3.4311876508245227,
      "tokens_seen": 1994457088
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040242424242424243,
      "loss": 1.1763,
      "theoretical_loss": 3.4311782454104494,
      "tokens_seen": 1994522624
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040240404040404044,
      "loss": 1.65,
      "theoretical_loss": 3.4311688403919423,
      "tokens_seen": 1994588160
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004023838383838384,
      "loss": 1.3469,
      "theoretical_loss": 3.431159435768972,
      "tokens_seen": 1994653696
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040236363636363636,
      "loss": 1.2375,
      "theoretical_loss": 3.4311500315415095,
      "tokens_seen": 1994719232
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040234343434343437,
      "loss": 1.7511,
      "theoretical_loss": 3.431140627709525,
      "tokens_seen": 1994784768
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004023232323232323,
      "loss": 1.3529,
      "theoretical_loss": 3.4311312242729883,
      "tokens_seen": 1994850304
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040230303030303034,
      "loss": 1.8496,
      "theoretical_loss": 3.43112182123187,
      "tokens_seen": 1994915840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004022828282828283,
      "loss": 1.4838,
      "theoretical_loss": 3.431112418586141,
      "tokens_seen": 1994981376
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040226262626262625,
      "loss": 1.6903,
      "theoretical_loss": 3.431103016335771,
      "tokens_seen": 1995046912
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040224242424242426,
      "loss": 1.2681,
      "theoretical_loss": 3.4310936144807314,
      "tokens_seen": 1995112448
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004022222222222222,
      "loss": 1.1243,
      "theoretical_loss": 3.4310842130209913,
      "tokens_seen": 1995177984
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040220202020202023,
      "loss": 1.6698,
      "theoretical_loss": 3.431074811956522,
      "tokens_seen": 1995243520
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021818181818182,
      "loss": 1.5313,
      "theoretical_loss": 3.4310654112872934,
      "tokens_seen": 1995309056
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021616161616162,
      "loss": 1.3351,
      "theoretical_loss": 3.4310560110132764,
      "tokens_seen": 1995374592
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040214141414141415,
      "loss": 1.5461,
      "theoretical_loss": 3.431046611134441,
      "tokens_seen": 1995440128
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021212121212121,
      "loss": 1.3952,
      "theoretical_loss": 3.4310372116507573,
      "tokens_seen": 1995505664
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004021010101010101,
      "loss": 1.34,
      "theoretical_loss": 3.4310278125621965,
      "tokens_seen": 1995571200
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004020808080808081,
      "loss": 1.3774,
      "theoretical_loss": 3.431018413868729,
      "tokens_seen": 1995636736
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004020606060606061,
      "loss": 1.5431,
      "theoretical_loss": 3.4310090155703246,
      "tokens_seen": 1995702272
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040204040404040405,
      "loss": 1.535,
      "theoretical_loss": 3.430999617666954,
      "tokens_seen": 1995767808
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.4018206298351288,
      "objective/train/avg_token_score": 0.02627016417682171,
      "objective/train/docs_used": 1128921,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3620922565460205,
      "objective/train/original_loss": 4.719934463500977,
      "objective/train/theoretical_loss": 3.430990220158588,
      "objective/train/tokens_used": 354692576,
      "objective/train/unlikelihood_loss": 0.050740428268909454,
      "theoretical_loss": 3.430990220158588,
      "tokens_seen": 1995833344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402020202020202,
      "loss": 1.5693,
      "theoretical_loss": 3.430990220158588,
      "tokens_seen": 1995833344
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000402,
      "loss": 1.3973,
      "theoretical_loss": 3.4309808230451955,
      "tokens_seen": 1995898880
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000401979797979798,
      "loss": 1.3402,
      "theoretical_loss": 3.430971426326749,
      "tokens_seen": 1995964416
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000401959595959596,
      "loss": 1.4646,
      "theoretical_loss": 3.4309620300032178,
      "tokens_seen": 1996029952
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040193939393939394,
      "loss": 1.5222,
      "theoretical_loss": 3.430952634074573,
      "tokens_seen": 1996095488
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004019191919191919,
      "loss": 1.4229,
      "theoretical_loss": 3.4309432385407836,
      "tokens_seen": 1996161024
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018989898989899,
      "loss": 1.3307,
      "theoretical_loss": 3.4309338434018217,
      "tokens_seen": 1996226560
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018787878787879,
      "loss": 1.6634,
      "theoretical_loss": 3.430924448657657,
      "tokens_seen": 1996292096
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018585858585859,
      "loss": 1.394,
      "theoretical_loss": 3.43091505430826,
      "tokens_seen": 1996357632
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040183838383838383,
      "loss": 1.5225,
      "theoretical_loss": 3.4309056603536012,
      "tokens_seen": 1996423168
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004018181818181818,
      "loss": 1.4662,
      "theoretical_loss": 3.430896266793651,
      "tokens_seen": 1996488704
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040179797979797985,
      "loss": 1.4994,
      "theoretical_loss": 3.43088687362838,
      "tokens_seen": 1996554240
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004017777777777778,
      "loss": 1.3143,
      "theoretical_loss": 3.4308774808577587,
      "tokens_seen": 1996619776
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040175757575757577,
      "loss": 1.3971,
      "theoretical_loss": 3.4308680884817573,
      "tokens_seen": 1996685312
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004017373737373737,
      "loss": 1.4626,
      "theoretical_loss": 3.430858696500346,
      "tokens_seen": 1996750848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004017171717171717,
      "loss": 1.5061,
      "theoretical_loss": 3.430849304913496,
      "tokens_seen": 1996816384
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040169696969696975,
      "loss": 1.4394,
      "theoretical_loss": 3.4308399137211776,
      "tokens_seen": 1996881920
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004016767676767677,
      "loss": 1.6052,
      "theoretical_loss": 3.430830522923361,
      "tokens_seen": 1996947456
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040165656565656566,
      "loss": 1.4002,
      "theoretical_loss": 3.4308211325200166,
      "tokens_seen": 1997012992
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004016363636363636,
      "loss": 1.5824,
      "theoretical_loss": 3.430811742511115,
      "tokens_seen": 1997078528
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040161616161616163,
      "loss": 1.5992,
      "theoretical_loss": 3.4308023528966274,
      "tokens_seen": 1997144064
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040159595959595964,
      "loss": 1.5798,
      "theoretical_loss": 3.4307929636765233,
      "tokens_seen": 1997209600
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004015757575757576,
      "loss": 1.6261,
      "theoretical_loss": 3.4307835748507736,
      "tokens_seen": 1997275136
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00040155555555555555,
      "loss": 1.4311,
      "theoretical_loss": 3.430774186419349,
      "tokens_seen": 1997340672
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004015353535353535,
      "loss": 1.3639,
      "theoretical_loss": 3.4307647983822194,
      "tokens_seen": 1997406208
    },
    {
      "epoch": 0.2,
      "objective/train/aligned_token_count": 0.4560728371143341,
      "objective/train/avg_token_score": 0.014928154647350311,
      "objective/train/docs_used": 1129917,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6200169324874878,
      "objective/train/original_loss": 4.3188934326171875,
      "objective/train/theoretical_loss": 3.430755410739356,
      "objective/train/tokens_used": 356330976,
      "objective/train/unlikelihood_loss": 0.06730559468269348,
      "theoretical_loss": 3.430755410739356,
      "tokens_seen": 1997471744
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0004015151515151515,
      "loss": 1.5374,
      "theoretical_loss": 3.430755410739356,
      "tokens_seen": 1997471744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040149494949494953,
      "loss": 1.5942,
      "theoretical_loss": 3.4307460234907285,
      "tokens_seen": 1997537280
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004014747474747475,
      "loss": 1.2555,
      "theoretical_loss": 3.430736636636308,
      "tokens_seen": 1997602816
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040145454545454545,
      "loss": 1.2842,
      "theoretical_loss": 3.430727250176065,
      "tokens_seen": 1997668352
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040143434343434346,
      "loss": 1.2268,
      "theoretical_loss": 3.4307178641099703,
      "tokens_seen": 1997733888
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004014141414141414,
      "loss": 1.4628,
      "theoretical_loss": 3.4307084784379933,
      "tokens_seen": 1997799424
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004013939393939394,
      "loss": 1.5295,
      "theoretical_loss": 3.430699093160106,
      "tokens_seen": 1997864960
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004013737373737374,
      "loss": 1.3422,
      "theoretical_loss": 3.4306897082762777,
      "tokens_seen": 1997930496
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040135353535353534,
      "loss": 1.3854,
      "theoretical_loss": 3.4306803237864796,
      "tokens_seen": 1997996032
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040133333333333335,
      "loss": 1.3874,
      "theoretical_loss": 3.4306709396906823,
      "tokens_seen": 1998061568
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004013131313131313,
      "loss": 1.5439,
      "theoretical_loss": 3.4306615559888556,
      "tokens_seen": 1998127104
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004012929292929293,
      "loss": 1.3111,
      "theoretical_loss": 3.4306521726809707,
      "tokens_seen": 1998192640
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004012727272727273,
      "loss": 1.514,
      "theoretical_loss": 3.430642789766998,
      "tokens_seen": 1998258176
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004012525252525253,
      "loss": 1.4821,
      "theoretical_loss": 3.4306334072469085,
      "tokens_seen": 1998323712
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040123232323232324,
      "loss": 1.4354,
      "theoretical_loss": 3.4306240251206717,
      "tokens_seen": 1998389248
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004012121212121212,
      "loss": 1.5244,
      "theoretical_loss": 3.4306146433882585,
      "tokens_seen": 1998454784
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004011919191919192,
      "loss": 1.6676,
      "theoretical_loss": 3.43060526204964,
      "tokens_seen": 1998520320
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040117171717171717,
      "loss": 1.4527,
      "theoretical_loss": 3.4305958811047863,
      "tokens_seen": 1998585856
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004011515151515152,
      "loss": 1.7453,
      "theoretical_loss": 3.4305865005536686,
      "tokens_seen": 1998651392
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040113131313131313,
      "loss": 1.3844,
      "theoretical_loss": 3.430577120396256,
      "tokens_seen": 1998716928
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004011111111111111,
      "loss": 1.4724,
      "theoretical_loss": 3.4305677406325206,
      "tokens_seen": 1998782464
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004010909090909091,
      "loss": 1.5013,
      "theoretical_loss": 3.430558361262433,
      "tokens_seen": 1998848000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040107070707070706,
      "loss": 1.6508,
      "theoretical_loss": 3.430548982285962,
      "tokens_seen": 1998913536
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040105050505050507,
      "loss": 1.5671,
      "theoretical_loss": 3.43053960370308,
      "tokens_seen": 1998979072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000401030303030303,
      "loss": 1.3459,
      "theoretical_loss": 3.4305302255137566,
      "tokens_seen": 1999044608
    },
    {
      "epoch": 0.21,
      "objective/train/aligned_token_count": 0.5745356678962708,
      "objective/train/avg_token_score": 0.00940377451479435,
      "objective/train/docs_used": 1130465,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 2.279651403427124,
      "objective/train/original_loss": 3.9225895404815674,
      "objective/train/theoretical_loss": 3.430520847717963,
      "objective/train/tokens_used": 357969376,
      "objective/train/unlikelihood_loss": 0.0918278694152832,
      "theoretical_loss": 3.430520847717963,
      "tokens_seen": 1999110144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000401010101010101,
      "loss": 1.5423,
      "theoretical_loss": 3.430520847717963,
      "tokens_seen": 1999110144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400989898989899,
      "loss": 1.7275,
      "theoretical_loss": 3.430511470315669,
      "tokens_seen": 1999175680
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400969696969697,
      "loss": 1.2737,
      "theoretical_loss": 3.4305020933068464,
      "tokens_seen": 1999241216
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040094949494949496,
      "loss": 1.4778,
      "theoretical_loss": 3.4304927166914645,
      "tokens_seen": 1999306752
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004009292929292929,
      "loss": 1.6099,
      "theoretical_loss": 3.430483340469495,
      "tokens_seen": 1999372288
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004009090909090909,
      "loss": 1.5138,
      "theoretical_loss": 3.4304739646409077,
      "tokens_seen": 1999437824
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008888888888889,
      "loss": 1.3048,
      "theoretical_loss": 3.4304645892056733,
      "tokens_seen": 1999503360
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008686868686869,
      "loss": 1.2571,
      "theoretical_loss": 3.4304552141637625,
      "tokens_seen": 1999568896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040084848484848486,
      "loss": 1.2765,
      "theoretical_loss": 3.4304458395151465,
      "tokens_seen": 1999634432
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008282828282828,
      "loss": 1.223,
      "theoretical_loss": 3.430436465259795,
      "tokens_seen": 1999699968
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004008080808080808,
      "loss": 1.5195,
      "theoretical_loss": 3.4304270913976795,
      "tokens_seen": 1999765504
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040078787878787883,
      "loss": 1.2751,
      "theoretical_loss": 3.430417717928769,
      "tokens_seen": 1999831040
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004007676767676768,
      "loss": 1.3775,
      "theoretical_loss": 3.4304083448530367,
      "tokens_seen": 1999896576
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040074747474747475,
      "loss": 1.3148,
      "theoretical_loss": 3.4303989721704506,
      "tokens_seen": 1999962112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004007272727272727,
      "loss": 1.3953,
      "theoretical_loss": 3.430389599880983,
      "tokens_seen": 2000027648
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004007070707070707,
      "loss": 1.2631,
      "theoretical_loss": 3.430380227984604,
      "tokens_seen": 2000093184
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040068686868686873,
      "loss": 1.2196,
      "theoretical_loss": 3.4303708564812845,
      "tokens_seen": 2000158720
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004006666666666667,
      "loss": 1.3836,
      "theoretical_loss": 3.4303614853709945,
      "tokens_seen": 2000224256
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040064646464646464,
      "loss": 1.2242,
      "theoretical_loss": 3.4303521146537053,
      "tokens_seen": 2000289792
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004006262626262626,
      "loss": 1.28,
      "theoretical_loss": 3.430342744329387,
      "tokens_seen": 2000355328
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040060606060606066,
      "loss": 1.3343,
      "theoretical_loss": 3.430333374398011,
      "tokens_seen": 2000420864
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004005858585858586,
      "loss": 1.4141,
      "theoretical_loss": 3.430324004859547,
      "tokens_seen": 2000486400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004005656565656566,
      "loss": 1.0115,
      "theoretical_loss": 3.430314635713967,
      "tokens_seen": 2000551936
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040054545454545453,
      "loss": 1.3156,
      "theoretical_loss": 3.4303052669612395,
      "tokens_seen": 2000617472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004005252525252525,
      "loss": 1.7144,
      "theoretical_loss": 3.4302958986013374,
      "tokens_seen": 2000683008
    },
    {
      "epoch": 0.21,
      "objective/train/aligned_token_count": 0.3841642141342163,
      "objective/train/avg_token_score": 0.048149384558200836,
      "objective/train/docs_used": 1131148,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.318234920501709,
      "objective/train/original_loss": 5.058935165405273,
      "objective/train/theoretical_loss": 3.43028653063423,
      "objective/train/tokens_used": 359607776,
      "objective/train/unlikelihood_loss": 0.04659086838364601,
      "theoretical_loss": 3.43028653063423,
      "tokens_seen": 2000748544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040050505050505056,
      "loss": 1.4759,
      "theoretical_loss": 3.43028653063423,
      "tokens_seen": 2000748544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004004848484848485,
      "loss": 1.1273,
      "theoretical_loss": 3.4302771630598885,
      "tokens_seen": 2000814080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040046464646464647,
      "loss": 1.4563,
      "theoretical_loss": 3.4302677958782835,
      "tokens_seen": 2000879616
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004004444444444444,
      "loss": 1.294,
      "theoretical_loss": 3.4302584290893856,
      "tokens_seen": 2000945152
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040042424242424244,
      "loss": 1.3029,
      "theoretical_loss": 3.430249062693165,
      "tokens_seen": 2001010688
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040040404040404045,
      "loss": 1.4544,
      "theoretical_loss": 3.430239696689594,
      "tokens_seen": 2001076224
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004003838383838384,
      "loss": 1.4593,
      "theoretical_loss": 3.4302303310786417,
      "tokens_seen": 2001141760
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040036363636363636,
      "loss": 1.2595,
      "theoretical_loss": 3.4302209658602787,
      "tokens_seen": 2001207296
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004003434343434343,
      "loss": 1.395,
      "theoretical_loss": 3.4302116010344763,
      "tokens_seen": 2001272832
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040032323232323233,
      "loss": 1.2054,
      "theoretical_loss": 3.4302022366012057,
      "tokens_seen": 2001338368
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040030303030303034,
      "loss": 1.3016,
      "theoretical_loss": 3.4301928725604367,
      "tokens_seen": 2001403904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004002828282828283,
      "loss": 1.7202,
      "theoretical_loss": 3.4301835089121404,
      "tokens_seen": 2001469440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040026262626262625,
      "loss": 1.542,
      "theoretical_loss": 3.430174145656287,
      "tokens_seen": 2001534976
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040024242424242427,
      "loss": 1.2283,
      "theoretical_loss": 3.4301647827928483,
      "tokens_seen": 2001600512
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004002222222222222,
      "loss": 1.3033,
      "theoretical_loss": 3.430155420321794,
      "tokens_seen": 2001666048
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040020202020202023,
      "loss": 1.4857,
      "theoretical_loss": 3.430146058243095,
      "tokens_seen": 2001731584
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004001818181818182,
      "loss": 1.4668,
      "theoretical_loss": 3.430136696556722,
      "tokens_seen": 2001797120
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040016161616161615,
      "loss": 1.0117,
      "theoretical_loss": 3.4301273352626462,
      "tokens_seen": 2001862656
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040014141414141416,
      "loss": 1.3209,
      "theoretical_loss": 3.430117974360838,
      "tokens_seen": 2001928192
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004001212121212121,
      "loss": 1.4727,
      "theoretical_loss": 3.4301086138512678,
      "tokens_seen": 2001993728
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004001010101010101,
      "loss": 1.6293,
      "theoretical_loss": 3.430099253733907,
      "tokens_seen": 2002059264
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004000808080808081,
      "loss": 1.6588,
      "theoretical_loss": 3.4300898940087254,
      "tokens_seen": 2002124800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004000606060606061,
      "loss": 1.2346,
      "theoretical_loss": 3.430080534675695,
      "tokens_seen": 2002190336
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00040004040404040405,
      "loss": 1.2629,
      "theoretical_loss": 3.4300711757347853,
      "tokens_seen": 2002255872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000400020202020202,
      "loss": 1.2379,
      "theoretical_loss": 3.4300618171859676,
      "tokens_seen": 2002321408
    },
    {
      "epoch": 0.21,
      "objective/train/aligned_token_count": 0.39876589179039,
      "objective/train/avg_token_score": 0.012741667218506336,
      "objective/train/docs_used": 1132271,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2703096866607666,
      "objective/train/original_loss": 3.7862753868103027,
      "objective/train/theoretical_loss": 3.4300524590292127,
      "objective/train/tokens_used": 361246176,
      "objective/train/unlikelihood_loss": 0.11513296514749527,
      "theoretical_loss": 3.4300524590292127,
      "tokens_seen": 2002386944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0004,
      "loss": 1.4336,
      "theoretical_loss": 3.4300524590292127,
      "tokens_seen": 2002386944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000399979797979798,
      "loss": 1.1594,
      "theoretical_loss": 3.4300431012644914,
      "tokens_seen": 2002452480
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000399959595959596,
      "loss": 1.5808,
      "theoretical_loss": 3.430033743891774,
      "tokens_seen": 2002518016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039993939393939394,
      "loss": 1.4235,
      "theoretical_loss": 3.4300243869110316,
      "tokens_seen": 2002583552
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003999191919191919,
      "loss": 1.4838,
      "theoretical_loss": 3.430015030322235,
      "tokens_seen": 2002649088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998989898989899,
      "loss": 1.3642,
      "theoretical_loss": 3.4300056741253546,
      "tokens_seen": 2002714624
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998787878787879,
      "loss": 1.4876,
      "theoretical_loss": 3.4299963183203617,
      "tokens_seen": 2002780160
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998585858585859,
      "loss": 1.5863,
      "theoretical_loss": 3.4299869629072264,
      "tokens_seen": 2002845696
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039983838383838384,
      "loss": 1.5689,
      "theoretical_loss": 3.4299776078859203,
      "tokens_seen": 2002911232
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003998181818181818,
      "loss": 1.1708,
      "theoretical_loss": 3.4299682532564137,
      "tokens_seen": 2002976768
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003997979797979798,
      "loss": 1.1919,
      "theoretical_loss": 3.4299588990186773,
      "tokens_seen": 2003042304
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003997777777777778,
      "loss": 1.4953,
      "theoretical_loss": 3.429949545172682,
      "tokens_seen": 2003107840
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039975757575757577,
      "loss": 1.3504,
      "theoretical_loss": 3.429940191718398,
      "tokens_seen": 2003173376
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039973737373737373,
      "loss": 1.5487,
      "theoretical_loss": 3.429930838655797,
      "tokens_seen": 2003238912
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003997171717171717,
      "loss": 1.5136,
      "theoretical_loss": 3.4299214859848495,
      "tokens_seen": 2003304448
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039969696969696975,
      "loss": 1.2122,
      "theoretical_loss": 3.429912133705526,
      "tokens_seen": 2003369984
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003996767676767677,
      "loss": 1.3716,
      "theoretical_loss": 3.4299027818177974,
      "tokens_seen": 2003435520
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039965656565656567,
      "loss": 1.6886,
      "theoretical_loss": 3.429893430321635,
      "tokens_seen": 2003501056
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003996363636363636,
      "loss": 0.969,
      "theoretical_loss": 3.4298840792170084,
      "tokens_seen": 2003566592
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003996161616161616,
      "loss": 1.4569,
      "theoretical_loss": 3.42987472850389,
      "tokens_seen": 2003632128
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039959595959595964,
      "loss": 1.4078,
      "theoretical_loss": 3.4298653781822495,
      "tokens_seen": 2003697664
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003995757575757576,
      "loss": 1.3704,
      "theoretical_loss": 3.429856028252058,
      "tokens_seen": 2003763200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039955555555555556,
      "loss": 1.3195,
      "theoretical_loss": 3.429846678713286,
      "tokens_seen": 2003828736
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003995353535353535,
      "loss": 1.4617,
      "theoretical_loss": 3.4298373295659053,
      "tokens_seen": 2003894272
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039951515151515147,
      "loss": 1.4341,
      "theoretical_loss": 3.4298279808098853,
      "tokens_seen": 2003959808
    },
    {
      "epoch": 0.21,
      "objective/train/aligned_token_count": 0.291177898645401,
      "objective/train/avg_token_score": 0.03287602216005325,
      "objective/train/docs_used": 1132794,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9526001214981079,
      "objective/train/original_loss": 5.209911823272705,
      "objective/train/theoretical_loss": 3.4298186324451976,
      "objective/train/tokens_used": 362884576,
      "objective/train/unlikelihood_loss": 0.04816959425806999,
      "theoretical_loss": 3.4298186324451976,
      "tokens_seen": 2004025344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039949494949494954,
      "loss": 1.1675,
      "theoretical_loss": 3.4298186324451976,
      "tokens_seen": 2004025344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994747474747475,
      "loss": 1.4201,
      "theoretical_loss": 3.4298092844718133,
      "tokens_seen": 2004090880
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039945454545454545,
      "loss": 1.2932,
      "theoretical_loss": 3.4297999368897027,
      "tokens_seen": 2004156416
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994343434343434,
      "loss": 1.3336,
      "theoretical_loss": 3.429790589698837,
      "tokens_seen": 2004221952
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003994141414141414,
      "loss": 1.3756,
      "theoretical_loss": 3.4297812428991867,
      "tokens_seen": 2004287488
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039939393939393943,
      "loss": 1.5859,
      "theoretical_loss": 3.429771896490723,
      "tokens_seen": 2004353024
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003993737373737374,
      "loss": 1.6585,
      "theoretical_loss": 3.4297625504734164,
      "tokens_seen": 2004418560
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039935353535353534,
      "loss": 1.1558,
      "theoretical_loss": 3.4297532048472377,
      "tokens_seen": 2004484096
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003993333333333333,
      "loss": 1.3409,
      "theoretical_loss": 3.4297438596121577,
      "tokens_seen": 2004549632
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003993131313131313,
      "loss": 1.4537,
      "theoretical_loss": 3.429734514768148,
      "tokens_seen": 2004615168
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003992929292929293,
      "loss": 1.3233,
      "theoretical_loss": 3.4297251703151783,
      "tokens_seen": 2004680704
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003992727272727273,
      "loss": 1.3178,
      "theoretical_loss": 3.4297158262532204,
      "tokens_seen": 2004746240
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039925252525252524,
      "loss": 1.46,
      "theoretical_loss": 3.4297064825822448,
      "tokens_seen": 2004811776
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039923232323232325,
      "loss": 1.1032,
      "theoretical_loss": 3.4296971393022226,
      "tokens_seen": 2004877312
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003992121212121212,
      "loss": 1.4979,
      "theoretical_loss": 3.4296877964131243,
      "tokens_seen": 2004942848
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003991919191919192,
      "loss": 1.4713,
      "theoretical_loss": 3.4296784539149208,
      "tokens_seen": 2005008384
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039917171717171717,
      "loss": 1.525,
      "theoretical_loss": 3.4296691118075833,
      "tokens_seen": 2005073920
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039915151515151513,
      "loss": 1.1359,
      "theoretical_loss": 3.4296597700910825,
      "tokens_seen": 2005139456
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039913131313131314,
      "loss": 1.364,
      "theoretical_loss": 3.429650428765389,
      "tokens_seen": 2005204992
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039911111111111115,
      "loss": 1.0891,
      "theoretical_loss": 3.4296410878304737,
      "tokens_seen": 2005270528
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003990909090909091,
      "loss": 1.1563,
      "theoretical_loss": 3.429631747286308,
      "tokens_seen": 2005336064
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039907070707070706,
      "loss": 1.3294,
      "theoretical_loss": 3.4296224071328623,
      "tokens_seen": 2005401600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003990505050505051,
      "loss": 1.498,
      "theoretical_loss": 3.4296130673701075,
      "tokens_seen": 2005467136
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039903030303030303,
      "loss": 1.3699,
      "theoretical_loss": 3.429603727998015,
      "tokens_seen": 2005532672
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039901010101010104,
      "loss": 1.3808,
      "theoretical_loss": 3.429594389016555,
      "tokens_seen": 2005598208
    },
    {
      "epoch": 0.21,
      "objective/train/aligned_token_count": 0.48179373145103455,
      "objective/train/avg_token_score": 0.01661764457821846,
      "objective/train/docs_used": 1134045,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4383306503295898,
      "objective/train/original_loss": 4.333360195159912,
      "objective/train/theoretical_loss": 3.429585050425699,
      "objective/train/tokens_used": 364522976,
      "objective/train/unlikelihood_loss": 0.05595013126730919,
      "theoretical_loss": 3.429585050425699,
      "tokens_seen": 2005663744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000398989898989899,
      "loss": 1.3411,
      "theoretical_loss": 3.429585050425699,
      "tokens_seen": 2005663744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039896969696969696,
      "loss": 1.3543,
      "theoretical_loss": 3.429575712225418,
      "tokens_seen": 2005729280
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039894949494949497,
      "loss": 1.3326,
      "theoretical_loss": 3.429566374415682,
      "tokens_seen": 2005794816
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003989292929292929,
      "loss": 1.6406,
      "theoretical_loss": 3.4295570369964623,
      "tokens_seen": 2005860352
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039890909090909094,
      "loss": 1.3855,
      "theoretical_loss": 3.4295476999677303,
      "tokens_seen": 2005925888
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003988888888888889,
      "loss": 1.3958,
      "theoretical_loss": 3.4295383633294563,
      "tokens_seen": 2005991424
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003988686868686869,
      "loss": 1.4905,
      "theoretical_loss": 3.429529027081612,
      "tokens_seen": 2006056960
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039884848484848486,
      "loss": 1.5166,
      "theoretical_loss": 3.4295196912241677,
      "tokens_seen": 2006122496
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003988282828282828,
      "loss": 1.6309,
      "theoretical_loss": 3.429510355757094,
      "tokens_seen": 2006188032
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039880808080808083,
      "loss": 1.585,
      "theoretical_loss": 3.4295010206803624,
      "tokens_seen": 2006253568
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987878787878788,
      "loss": 1.4573,
      "theoretical_loss": 3.429491685993944,
      "tokens_seen": 2006319104
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987676767676768,
      "loss": 1.6665,
      "theoretical_loss": 3.4294823516978092,
      "tokens_seen": 2006384640
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039874747474747475,
      "loss": 1.2809,
      "theoretical_loss": 3.4294730177919295,
      "tokens_seen": 2006450176
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987272727272727,
      "loss": 1.6269,
      "theoretical_loss": 3.429463684276275,
      "tokens_seen": 2006515712
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003987070707070707,
      "loss": 1.7012,
      "theoretical_loss": 3.429454351150817,
      "tokens_seen": 2006581248
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039868686868686873,
      "loss": 1.489,
      "theoretical_loss": 3.429445018415527,
      "tokens_seen": 2006646784
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003986666666666667,
      "loss": 1.5224,
      "theoretical_loss": 3.4294356860703754,
      "tokens_seen": 2006712320
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039864646464646465,
      "loss": 1.4711,
      "theoretical_loss": 3.429426354115334,
      "tokens_seen": 2006777856
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003986262626262626,
      "loss": 1.2435,
      "theoretical_loss": 3.4294170225503717,
      "tokens_seen": 2006843392
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003986060606060606,
      "loss": 1.3795,
      "theoretical_loss": 3.429407691375461,
      "tokens_seen": 2006908928
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003985858585858586,
      "loss": 1.4629,
      "theoretical_loss": 3.4293983605905733,
      "tokens_seen": 2006974464
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003985656565656566,
      "loss": 1.569,
      "theoretical_loss": 3.429389030195679,
      "tokens_seen": 2007040000
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039854545454545454,
      "loss": 1.3673,
      "theoretical_loss": 3.4293797001907484,
      "tokens_seen": 2007105536
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003985252525252525,
      "loss": 1.4482,
      "theoretical_loss": 3.4293703705757537,
      "tokens_seen": 2007171072
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039850505050505056,
      "loss": 1.1585,
      "theoretical_loss": 3.4293610413506643,
      "tokens_seen": 2007236608
    },
    {
      "epoch": 0.21,
      "objective/train/aligned_token_count": 0.4596163332462311,
      "objective/train/avg_token_score": 0.01717677153646946,
      "objective/train/docs_used": 1134700,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4232748746871948,
      "objective/train/original_loss": 4.048245429992676,
      "objective/train/theoretical_loss": 3.4293517125154525,
      "objective/train/tokens_used": 366161376,
      "objective/train/unlikelihood_loss": 0.062166180461645126,
      "theoretical_loss": 3.4293517125154525,
      "tokens_seen": 2007302144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003984848484848485,
      "loss": 1.3018,
      "theoretical_loss": 3.4293517125154525,
      "tokens_seen": 2007302144
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003984646464646465,
      "loss": 1.0955,
      "theoretical_loss": 3.429342384070089,
      "tokens_seen": 2007367680
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039844444444444443,
      "loss": 1.2975,
      "theoretical_loss": 3.4293330560145447,
      "tokens_seen": 2007433216
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003984242424242424,
      "loss": 1.5909,
      "theoretical_loss": 3.4293237283487903,
      "tokens_seen": 2007498752
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039840404040404045,
      "loss": 1.2435,
      "theoretical_loss": 3.4293144010727974,
      "tokens_seen": 2007564288
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003983838383838384,
      "loss": 1.2224,
      "theoretical_loss": 3.4293050741865363,
      "tokens_seen": 2007629824
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039836363636363637,
      "loss": 1.4889,
      "theoretical_loss": 3.429295747689978,
      "tokens_seen": 2007695360
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003983434343434343,
      "loss": 1.6156,
      "theoretical_loss": 3.4292864215830945,
      "tokens_seen": 2007760896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039832323232323234,
      "loss": 1.2967,
      "theoretical_loss": 3.4292770958658556,
      "tokens_seen": 2007826432
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039830303030303035,
      "loss": 1.4694,
      "theoretical_loss": 3.4292677705382335,
      "tokens_seen": 2007891968
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003982828282828283,
      "loss": 1.5067,
      "theoretical_loss": 3.4292584456001975,
      "tokens_seen": 2007957504
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039826262626262626,
      "loss": 1.389,
      "theoretical_loss": 3.42924912105172,
      "tokens_seen": 2008023040
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003982424242424242,
      "loss": 1.2721,
      "theoretical_loss": 3.429239796892772,
      "tokens_seen": 2008088576
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039822222222222223,
      "loss": 1.4431,
      "theoretical_loss": 3.429230473123324,
      "tokens_seen": 2008154112
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039820202020202024,
      "loss": 1.4681,
      "theoretical_loss": 3.429221149743347,
      "tokens_seen": 2008219648
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003981818181818182,
      "loss": 1.6894,
      "theoretical_loss": 3.4292118267528124,
      "tokens_seen": 2008285184
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039816161616161615,
      "loss": 1.1353,
      "theoretical_loss": 3.429202504151691,
      "tokens_seen": 2008350720
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039814141414141416,
      "loss": 1.6095,
      "theoretical_loss": 3.4291931819399535,
      "tokens_seen": 2008416256
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003981212121212121,
      "loss": 1.6248,
      "theoretical_loss": 3.429183860117572,
      "tokens_seen": 2008481792
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039810101010101013,
      "loss": 1.436,
      "theoretical_loss": 3.4291745386845163,
      "tokens_seen": 2008547328
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003980808080808081,
      "loss": 1.2587,
      "theoretical_loss": 3.429165217640758,
      "tokens_seen": 2008612864
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039806060606060605,
      "loss": 1.6218,
      "theoretical_loss": 3.429155896986268,
      "tokens_seen": 2008678400
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039804040404040406,
      "loss": 1.6518,
      "theoretical_loss": 3.4291465767210174,
      "tokens_seen": 2008743936
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000398020202020202,
      "loss": 1.5358,
      "theoretical_loss": 3.429137256844977,
      "tokens_seen": 2008809472
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000398,
      "loss": 1.4281,
      "theoretical_loss": 3.429127937358119,
      "tokens_seen": 2008875008
    },
    {
      "epoch": 0.21,
      "objective/train/aligned_token_count": 0.4087854325771332,
      "objective/train/avg_token_score": 0.03126665577292442,
      "objective/train/docs_used": 1135755,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2452770471572876,
      "objective/train/original_loss": 4.195167541503906,
      "objective/train/theoretical_loss": 3.429118618260413,
      "objective/train/tokens_used": 367799776,
      "objective/train/unlikelihood_loss": 0.037951596081256866,
      "theoretical_loss": 3.429118618260413,
      "tokens_seen": 2008940544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397979797979798,
      "loss": 1.351,
      "theoretical_loss": 3.429118618260413,
      "tokens_seen": 2008940544
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397959595959596,
      "loss": 1.3136,
      "theoretical_loss": 3.429109299551831,
      "tokens_seen": 2009006080
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039793939393939395,
      "loss": 1.1987,
      "theoretical_loss": 3.4290999812323433,
      "tokens_seen": 2009071616
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003979191919191919,
      "loss": 1.4176,
      "theoretical_loss": 3.4290906633019214,
      "tokens_seen": 2009137152
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978989898989899,
      "loss": 1.6675,
      "theoretical_loss": 3.4290813457605362,
      "tokens_seen": 2009202688
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978787878787879,
      "loss": 1.4206,
      "theoretical_loss": 3.429072028608159,
      "tokens_seen": 2009268224
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978585858585859,
      "loss": 1.5396,
      "theoretical_loss": 3.429062711844761,
      "tokens_seen": 2009333760
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039783838383838384,
      "loss": 1.4403,
      "theoretical_loss": 3.429053395470313,
      "tokens_seen": 2009399296
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003978181818181818,
      "loss": 1.2628,
      "theoretical_loss": 3.429044079484786,
      "tokens_seen": 2009464832
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003977979797979798,
      "loss": 1.4167,
      "theoretical_loss": 3.4290347638881515,
      "tokens_seen": 2009530368
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039777777777777777,
      "loss": 1.5803,
      "theoretical_loss": 3.42902544868038,
      "tokens_seen": 2009595904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003977575757575758,
      "loss": 1.5627,
      "theoretical_loss": 3.4290161338614427,
      "tokens_seen": 2009661440
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039773737373737373,
      "loss": 1.4839,
      "theoretical_loss": 3.4290068194313115,
      "tokens_seen": 2009726976
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003977171717171717,
      "loss": 1.6212,
      "theoretical_loss": 3.428997505389956,
      "tokens_seen": 2009792512
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003976969696969697,
      "loss": 1.5033,
      "theoretical_loss": 3.4289881917373486,
      "tokens_seen": 2009858048
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003976767676767677,
      "loss": 1.3601,
      "theoretical_loss": 3.42897887847346,
      "tokens_seen": 2009923584
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039765656565656567,
      "loss": 1.4848,
      "theoretical_loss": 3.428969565598261,
      "tokens_seen": 2009989120
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039763636363636363,
      "loss": 1.5622,
      "theoretical_loss": 3.428960253111723,
      "tokens_seen": 2010054656
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003976161616161616,
      "loss": 1.3449,
      "theoretical_loss": 3.4289509410138175,
      "tokens_seen": 2010120192
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003975959595959596,
      "loss": 1.3104,
      "theoretical_loss": 3.4289416293045147,
      "tokens_seen": 2010185728
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003975757575757576,
      "loss": 1.4158,
      "theoretical_loss": 3.4289323179837865,
      "tokens_seen": 2010251264
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039755555555555556,
      "loss": 1.485,
      "theoretical_loss": 3.4289230070516035,
      "tokens_seen": 2010316800
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003975353535353535,
      "loss": 1.5817,
      "theoretical_loss": 3.428913696507937,
      "tokens_seen": 2010382336
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003975151515151515,
      "loss": 1.3679,
      "theoretical_loss": 3.428904386352758,
      "tokens_seen": 2010447872
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039749494949494954,
      "loss": 1.4348,
      "theoretical_loss": 3.428895076586038,
      "tokens_seen": 2010513408
    },
    {
      "epoch": 0.21,
      "objective/train/aligned_token_count": 0.3502565920352936,
      "objective/train/avg_token_score": 0.01835888810455799,
      "objective/train/docs_used": 1136767,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1449387073516846,
      "objective/train/original_loss": 4.458366870880127,
      "objective/train/theoretical_loss": 3.428885767207748,
      "objective/train/tokens_used": 369438176,
      "objective/train/unlikelihood_loss": 0.04602991044521332,
      "theoretical_loss": 3.428885767207748,
      "tokens_seen": 2010578944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003974747474747475,
      "loss": 1.2614,
      "theoretical_loss": 3.428885767207748,
      "tokens_seen": 2010578944
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039745454545454546,
      "loss": 1.568,
      "theoretical_loss": 3.428876458217859,
      "tokens_seen": 2010644480
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003974343434343434,
      "loss": 1.4913,
      "theoretical_loss": 3.4288671496163423,
      "tokens_seen": 2010710016
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039741414141414137,
      "loss": 1.5301,
      "theoretical_loss": 3.4288578414031687,
      "tokens_seen": 2010775552
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039739393939393944,
      "loss": 1.4495,
      "theoretical_loss": 3.42884853357831,
      "tokens_seen": 2010841088
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003973737373737374,
      "loss": 1.4207,
      "theoretical_loss": 3.4288392261417364,
      "tokens_seen": 2010906624
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039735353535353535,
      "loss": 1.2996,
      "theoretical_loss": 3.4288299190934195,
      "tokens_seen": 2010972160
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003973333333333333,
      "loss": 1.4799,
      "theoretical_loss": 3.4288206124333307,
      "tokens_seen": 2011037696
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039731313131313137,
      "loss": 1.6606,
      "theoretical_loss": 3.4288113061614407,
      "tokens_seen": 2011103232
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039729292929292933,
      "loss": 1.4349,
      "theoretical_loss": 3.4288020002777215,
      "tokens_seen": 2011168768
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003972727272727273,
      "loss": 1.3954,
      "theoretical_loss": 3.4287926947821434,
      "tokens_seen": 2011234304
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039725252525252524,
      "loss": 1.3447,
      "theoretical_loss": 3.4287833896746776,
      "tokens_seen": 2011299840
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003972323232323232,
      "loss": 1.491,
      "theoretical_loss": 3.428774084955296,
      "tokens_seen": 2011365376
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039721212121212126,
      "loss": 1.6107,
      "theoretical_loss": 3.428764780623969,
      "tokens_seen": 2011430912
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003971919191919192,
      "loss": 1.4626,
      "theoretical_loss": 3.428755476680668,
      "tokens_seen": 2011496448
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003971717171717172,
      "loss": 1.6105,
      "theoretical_loss": 3.428746173125364,
      "tokens_seen": 2011561984
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039715151515151513,
      "loss": 1.5182,
      "theoretical_loss": 3.428736869958029,
      "tokens_seen": 2011627520
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039713131313131315,
      "loss": 1.391,
      "theoretical_loss": 3.4287275671786333,
      "tokens_seen": 2011693056
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039711111111111116,
      "loss": 1.3069,
      "theoretical_loss": 3.4287182647871486,
      "tokens_seen": 2011758592
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003970909090909091,
      "loss": 1.3114,
      "theoretical_loss": 3.4287089627835456,
      "tokens_seen": 2011824128
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039707070707070707,
      "loss": 1.6529,
      "theoretical_loss": 3.428699661167796,
      "tokens_seen": 2011889664
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000397050505050505,
      "loss": 1.5254,
      "theoretical_loss": 3.4286903599398704,
      "tokens_seen": 2011955200
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039703030303030304,
      "loss": 1.3794,
      "theoretical_loss": 3.4286810590997403,
      "tokens_seen": 2012020736
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039701010101010105,
      "loss": 1.3395,
      "theoretical_loss": 3.4286717586473774,
      "tokens_seen": 2012086272
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000396989898989899,
      "loss": 1.4321,
      "theoretical_loss": 3.4286624585827523,
      "tokens_seen": 2012151808
    },
    {
      "epoch": 0.21,
      "objective/train/aligned_token_count": 0.4283357858657837,
      "objective/train/avg_token_score": 0.013622998259961605,
      "objective/train/docs_used": 1137326,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4791117906570435,
      "objective/train/original_loss": 4.347757816314697,
      "objective/train/theoretical_loss": 3.428653158905836,
      "objective/train/tokens_used": 371076576,
      "objective/train/unlikelihood_loss": 0.06590590626001358,
      "theoretical_loss": 3.428653158905836,
      "tokens_seen": 2012217344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039696969696969696,
      "loss": 1.5796,
      "theoretical_loss": 3.428653158905836,
      "tokens_seen": 2012217344
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.000396949494949495,
      "loss": 1.5627,
      "theoretical_loss": 3.4286438596166007,
      "tokens_seen": 2012282880
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039692929292929293,
      "loss": 1.6311,
      "theoretical_loss": 3.428634560715017,
      "tokens_seen": 2012348416
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039690909090909094,
      "loss": 1.4296,
      "theoretical_loss": 3.428625262201056,
      "tokens_seen": 2012413952
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003968888888888889,
      "loss": 1.5677,
      "theoretical_loss": 3.4286159640746883,
      "tokens_seen": 2012479488
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039686868686868686,
      "loss": 1.3217,
      "theoretical_loss": 3.4286066663358863,
      "tokens_seen": 2012545024
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039684848484848487,
      "loss": 1.4378,
      "theoretical_loss": 3.428597368984621,
      "tokens_seen": 2012610560
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003968282828282828,
      "loss": 1.1096,
      "theoretical_loss": 3.4285880720208635,
      "tokens_seen": 2012676096
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039680808080808083,
      "loss": 1.3569,
      "theoretical_loss": 3.428578775444585,
      "tokens_seen": 2012741632
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967878787878788,
      "loss": 1.5794,
      "theoretical_loss": 3.428569479255756,
      "tokens_seen": 2012807168
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967676767676768,
      "loss": 1.4991,
      "theoretical_loss": 3.4285601834543487,
      "tokens_seen": 2012872704
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039674747474747476,
      "loss": 1.4536,
      "theoretical_loss": 3.428550888040334,
      "tokens_seen": 2012938240
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003967272727272727,
      "loss": 1.4292,
      "theoretical_loss": 3.4285415930136836,
      "tokens_seen": 2013003776
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039670707070707073,
      "loss": 1.4315,
      "theoretical_loss": 3.4285322983743676,
      "tokens_seen": 2013069312
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966868686868687,
      "loss": 1.3883,
      "theoretical_loss": 3.4285230041223587,
      "tokens_seen": 2013134848
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966666666666667,
      "loss": 1.5132,
      "theoretical_loss": 3.4285137102576275,
      "tokens_seen": 2013200384
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039664646464646465,
      "loss": 1.4263,
      "theoretical_loss": 3.428504416780145,
      "tokens_seen": 2013265920
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966262626262626,
      "loss": 1.6979,
      "theoretical_loss": 3.428495123689882,
      "tokens_seen": 2013331456
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003966060606060606,
      "loss": 1.3121,
      "theoretical_loss": 3.428485830986811,
      "tokens_seen": 2013396992
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039658585858585863,
      "loss": 1.2552,
      "theoretical_loss": 3.4284765386709033,
      "tokens_seen": 2013462528
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003965656565656566,
      "loss": 1.4098,
      "theoretical_loss": 3.4284672467421284,
      "tokens_seen": 2013528064
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00039654545454545454,
      "loss": 1.5877,
      "theoretical_loss": 3.4284579552004595,
      "tokens_seen": 2013593600
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003965252525252525,
      "loss": 1.471,
      "theoretical_loss": 3.4284486640458667,
      "tokens_seen": 2013659136
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003965050505050505,
      "loss": 1.5591,
      "theoretical_loss": 3.428439373278322,
      "tokens_seen": 2013724672
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003964848484848485,
      "loss": 1.3222,
      "theoretical_loss": 3.4284300828977967,
      "tokens_seen": 2013790208
    },
    {
      "epoch": 0.21,
      "objective/train/aligned_token_count": 0.3340053856372833,
      "objective/train/avg_token_score": 0.0038737088907510042,
      "objective/train/docs_used": 1138753,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0530929565429688,
      "objective/train/original_loss": 4.1840667724609375,
      "objective/train/theoretical_loss": 3.428420792904261,
      "objective/train/tokens_used": 372714976,
      "objective/train/unlikelihood_loss": 0.06946754455566406,
      "theoretical_loss": 3.428420792904261,
      "tokens_seen": 2013855744
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0003964646464646465,
      "loss": 1.5015,
      "theoretical_loss": 3.428420792904261,
      "tokens_seen": 2013855744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039644444444444444,
      "loss": 1.3655,
      "theoretical_loss": 3.428411503297687,
      "tokens_seen": 2013921280
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003964242424242424,
      "loss": 1.8818,
      "theoretical_loss": 3.428402214078046,
      "tokens_seen": 2013986816
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039640404040404046,
      "loss": 1.3193,
      "theoretical_loss": 3.4283929252453094,
      "tokens_seen": 2014052352
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003963838383838384,
      "loss": 1.3773,
      "theoretical_loss": 3.4283836367994485,
      "tokens_seen": 2014117888
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003963636363636364,
      "loss": 1.659,
      "theoretical_loss": 3.428374348740434,
      "tokens_seen": 2014183424
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039634343434343433,
      "loss": 1.3025,
      "theoretical_loss": 3.428365061068238,
      "tokens_seen": 2014248960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003963232323232323,
      "loss": 1.2988,
      "theoretical_loss": 3.4283557737828314,
      "tokens_seen": 2014314496
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039630303030303035,
      "loss": 1.7744,
      "theoretical_loss": 3.4283464868841853,
      "tokens_seen": 2014380032
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962828282828283,
      "loss": 1.4166,
      "theoretical_loss": 3.428337200372271,
      "tokens_seen": 2014445568
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039626262626262627,
      "loss": 1.6612,
      "theoretical_loss": 3.4283279142470606,
      "tokens_seen": 2014511104
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962424242424242,
      "loss": 1.5514,
      "theoretical_loss": 3.4283186285085243,
      "tokens_seen": 2014576640
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003962222222222222,
      "loss": 1.451,
      "theoretical_loss": 3.4283093431566343,
      "tokens_seen": 2014642176
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039620202020202024,
      "loss": 1.3729,
      "theoretical_loss": 3.428300058191362,
      "tokens_seen": 2014707712
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003961818181818182,
      "loss": 1.6455,
      "theoretical_loss": 3.428290773612678,
      "tokens_seen": 2014773248
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039616161616161616,
      "loss": 1.3161,
      "theoretical_loss": 3.428281489420554,
      "tokens_seen": 2014838784
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003961414141414141,
      "loss": 1.6431,
      "theoretical_loss": 3.4282722056149613,
      "tokens_seen": 2014904320
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003961212121212121,
      "loss": 1.6435,
      "theoretical_loss": 3.428262922195871,
      "tokens_seen": 2014969856
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039610101010101014,
      "loss": 1.5589,
      "theoretical_loss": 3.428253639163255,
      "tokens_seen": 2015035392
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003960808080808081,
      "loss": 1.7296,
      "theoretical_loss": 3.428244356517084,
      "tokens_seen": 2015100928
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039606060606060605,
      "loss": 1.4722,
      "theoretical_loss": 3.42823507425733,
      "tokens_seen": 2015166464
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000396040404040404,
      "loss": 1.4072,
      "theoretical_loss": 3.428225792383964,
      "tokens_seen": 2015232000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000396020202020202,
      "loss": 1.3557,
      "theoretical_loss": 3.4282165108969576,
      "tokens_seen": 2015297536
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039600000000000003,
      "loss": 1.5808,
      "theoretical_loss": 3.428207229796281,
      "tokens_seen": 2015363072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000395979797979798,
      "loss": 1.3163,
      "theoretical_loss": 3.4281979490819072,
      "tokens_seen": 2015428608
    },
    {
      "epoch": 0.22,
      "objective/train/aligned_token_count": 0.4355449676513672,
      "objective/train/avg_token_score": 0.022654006257653236,
      "objective/train/docs_used": 1139429,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4072726964950562,
      "objective/train/original_loss": 4.453670024871826,
      "objective/train/theoretical_loss": 3.4281886687538066,
      "objective/train/tokens_used": 374353376,
      "objective/train/unlikelihood_loss": 0.06321793794631958,
      "theoretical_loss": 3.4281886687538066,
      "tokens_seen": 2015494144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039595959595959594,
      "loss": 1.3024,
      "theoretical_loss": 3.4281886687538066,
      "tokens_seen": 2015494144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039593939393939396,
      "loss": 1.6393,
      "theoretical_loss": 3.428179388811951,
      "tokens_seen": 2015559680
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003959191919191919,
      "loss": 1.2718,
      "theoretical_loss": 3.4281701092563117,
      "tokens_seen": 2015625216
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003958989898989899,
      "loss": 1.3737,
      "theoretical_loss": 3.42816083008686,
      "tokens_seen": 2015690752
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003958787878787879,
      "loss": 1.6082,
      "theoretical_loss": 3.428151551303567,
      "tokens_seen": 2015756288
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039585858585858584,
      "loss": 1.4045,
      "theoretical_loss": 3.428142272906404,
      "tokens_seen": 2015821824
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039583838383838385,
      "loss": 1.1955,
      "theoretical_loss": 3.4281329948953427,
      "tokens_seen": 2015887360
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003958181818181818,
      "loss": 1.4748,
      "theoretical_loss": 3.428123717270355,
      "tokens_seen": 2015952896
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003957979797979798,
      "loss": 1.2493,
      "theoretical_loss": 3.428114440031411,
      "tokens_seen": 2016018432
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039577777777777777,
      "loss": 1.3435,
      "theoretical_loss": 3.4281051631784836,
      "tokens_seen": 2016083968
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003957575757575758,
      "loss": 1.4313,
      "theoretical_loss": 3.428095886711543,
      "tokens_seen": 2016149504
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039573737373737374,
      "loss": 1.479,
      "theoretical_loss": 3.4280866106305607,
      "tokens_seen": 2016215040
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003957171717171717,
      "loss": 1.5382,
      "theoretical_loss": 3.4280773349355087,
      "tokens_seen": 2016280576
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003956969696969697,
      "loss": 1.2953,
      "theoretical_loss": 3.4280680596263586,
      "tokens_seen": 2016346112
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039567676767676767,
      "loss": 1.2436,
      "theoretical_loss": 3.4280587847030812,
      "tokens_seen": 2016411648
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003956565656565657,
      "loss": 1.4693,
      "theoretical_loss": 3.428049510165647,
      "tokens_seen": 2016477184
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039563636363636363,
      "loss": 1.5161,
      "theoretical_loss": 3.4280402360140294,
      "tokens_seen": 2016542720
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039561616161616164,
      "loss": 1.5267,
      "theoretical_loss": 3.4280309622481986,
      "tokens_seen": 2016608256
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003955959595959596,
      "loss": 1.3441,
      "theoretical_loss": 3.4280216888681263,
      "tokens_seen": 2016673792
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003955757575757576,
      "loss": 1.5796,
      "theoretical_loss": 3.4280124158737837,
      "tokens_seen": 2016739328
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039555555555555557,
      "loss": 1.4962,
      "theoretical_loss": 3.4280031432651423,
      "tokens_seen": 2016804864
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003955353535353535,
      "loss": 1.3655,
      "theoretical_loss": 3.4279938710421742,
      "tokens_seen": 2016870400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039551515151515154,
      "loss": 1.5749,
      "theoretical_loss": 3.4279845992048497,
      "tokens_seen": 2016935936
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003954949494949495,
      "loss": 1.5773,
      "theoretical_loss": 3.427975327753141,
      "tokens_seen": 2017001472
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003954747474747475,
      "loss": 1.3477,
      "theoretical_loss": 3.427966056687019,
      "tokens_seen": 2017067008
    },
    {
      "epoch": 0.22,
      "objective/train/aligned_token_count": 0.41849949955940247,
      "objective/train/avg_token_score": 0.04628181830048561,
      "objective/train/docs_used": 1140637,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.280837893486023,
      "objective/train/original_loss": 4.614250659942627,
      "objective/train/theoretical_loss": 3.427956786006456,
      "objective/train/tokens_used": 375991776,
      "objective/train/unlikelihood_loss": 0.0635131448507309,
      "theoretical_loss": 3.427956786006456,
      "tokens_seen": 2017132544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039545454545454546,
      "loss": 1.4989,
      "theoretical_loss": 3.427956786006456,
      "tokens_seen": 2017132544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003954343434343434,
      "loss": 1.3419,
      "theoretical_loss": 3.4279475157114225,
      "tokens_seen": 2017198080
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039541414141414143,
      "loss": 1.5053,
      "theoretical_loss": 3.42793824580189,
      "tokens_seen": 2017263616
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039539393939393944,
      "loss": 1.671,
      "theoretical_loss": 3.427928976277831,
      "tokens_seen": 2017329152
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953737373737374,
      "loss": 1.5888,
      "theoretical_loss": 3.427919707139216,
      "tokens_seen": 2017394688
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039535353535353535,
      "loss": 1.3773,
      "theoretical_loss": 3.4279104383860166,
      "tokens_seen": 2017460224
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953333333333333,
      "loss": 1.6882,
      "theoretical_loss": 3.4279011700182043,
      "tokens_seen": 2017525760
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003953131313131313,
      "loss": 1.3838,
      "theoretical_loss": 3.427891902035751,
      "tokens_seen": 2017591296
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039529292929292933,
      "loss": 1.406,
      "theoretical_loss": 3.427882634438627,
      "tokens_seen": 2017656832
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003952727272727273,
      "loss": 1.5952,
      "theoretical_loss": 3.4278733672268045,
      "tokens_seen": 2017722368
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039525252525252525,
      "loss": 1.3994,
      "theoretical_loss": 3.4278641004002557,
      "tokens_seen": 2017787904
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003952323232323232,
      "loss": 1.6201,
      "theoretical_loss": 3.4278548339589507,
      "tokens_seen": 2017853440
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039521212121212127,
      "loss": 1.5805,
      "theoretical_loss": 3.427845567902862,
      "tokens_seen": 2017918976
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003951919191919192,
      "loss": 1.4173,
      "theoretical_loss": 3.427836302231961,
      "tokens_seen": 2017984512
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003951717171717172,
      "loss": 1.4336,
      "theoretical_loss": 3.4278270369462183,
      "tokens_seen": 2018050048
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039515151515151514,
      "loss": 1.3806,
      "theoretical_loss": 3.4278177720456062,
      "tokens_seen": 2018115584
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003951313131313131,
      "loss": 1.2122,
      "theoretical_loss": 3.427808507530096,
      "tokens_seen": 2018181120
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039511111111111116,
      "loss": 1.4988,
      "theoretical_loss": 3.427799243399659,
      "tokens_seen": 2018246656
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950909090909091,
      "loss": 1.2608,
      "theoretical_loss": 3.427789979654267,
      "tokens_seen": 2018312192
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003950707070707071,
      "loss": 1.5234,
      "theoretical_loss": 3.427780716293891,
      "tokens_seen": 2018377728
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039505050505050503,
      "loss": 1.6553,
      "theoretical_loss": 3.427771453318503,
      "tokens_seen": 2018443264
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039503030303030304,
      "loss": 1.4045,
      "theoretical_loss": 3.4277621907280746,
      "tokens_seen": 2018508800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039501010101010105,
      "loss": 1.4315,
      "theoretical_loss": 3.4277529285225765,
      "tokens_seen": 2018574336
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394989898989899,
      "loss": 1.5395,
      "theoretical_loss": 3.427743666701981,
      "tokens_seen": 2018639872
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039496969696969697,
      "loss": 1.3235,
      "theoretical_loss": 3.4277344052662597,
      "tokens_seen": 2018705408
    },
    {
      "epoch": 0.22,
      "objective/train/aligned_token_count": 0.4572947323322296,
      "objective/train/avg_token_score": 0.014024609699845314,
      "objective/train/docs_used": 1141300,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4917529821395874,
      "objective/train/original_loss": 4.333985328674316,
      "objective/train/theoretical_loss": 3.4277251442153833,
      "objective/train/tokens_used": 377630176,
      "objective/train/unlikelihood_loss": 0.04426649212837219,
      "theoretical_loss": 3.4277251442153833,
      "tokens_seen": 2018770944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003949494949494949,
      "loss": 1.4542,
      "theoretical_loss": 3.4277251442153833,
      "tokens_seen": 2018770944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039492929292929294,
      "loss": 1.5937,
      "theoretical_loss": 3.4277158835493236,
      "tokens_seen": 2018836480
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039490909090909095,
      "loss": 1.561,
      "theoretical_loss": 3.427706623268053,
      "tokens_seen": 2018902016
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003948888888888889,
      "loss": 1.3872,
      "theoretical_loss": 3.4276973633715415,
      "tokens_seen": 2018967552
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039486868686868686,
      "loss": 1.5779,
      "theoretical_loss": 3.4276881038597615,
      "tokens_seen": 2019033088
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039484848484848487,
      "loss": 1.5723,
      "theoretical_loss": 3.427678844732685,
      "tokens_seen": 2019098624
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039482828282828283,
      "loss": 1.4318,
      "theoretical_loss": 3.427669585990283,
      "tokens_seen": 2019164160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039480808080808084,
      "loss": 1.4772,
      "theoretical_loss": 3.427660327632527,
      "tokens_seen": 2019229696
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003947878787878788,
      "loss": 1.6609,
      "theoretical_loss": 3.4276510696593876,
      "tokens_seen": 2019295232
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039476767676767675,
      "loss": 1.3307,
      "theoretical_loss": 3.4276418120708385,
      "tokens_seen": 2019360768
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039474747474747476,
      "loss": 1.3786,
      "theoretical_loss": 3.4276325548668494,
      "tokens_seen": 2019426304
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003947272727272727,
      "loss": 1.2832,
      "theoretical_loss": 3.4276232980473926,
      "tokens_seen": 2019491840
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039470707070707073,
      "loss": 1.254,
      "theoretical_loss": 3.42761404161244,
      "tokens_seen": 2019557376
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003946868686868687,
      "loss": 1.6662,
      "theoretical_loss": 3.4276047855619622,
      "tokens_seen": 2019622912
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003946666666666667,
      "loss": 1.4569,
      "theoretical_loss": 3.4275955298959313,
      "tokens_seen": 2019688448
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039464646464646466,
      "loss": 1.2263,
      "theoretical_loss": 3.4275862746143186,
      "tokens_seen": 2019753984
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003946262626262626,
      "loss": 1.6766,
      "theoretical_loss": 3.4275770197170963,
      "tokens_seen": 2019819520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003946060606060606,
      "loss": 1.4078,
      "theoretical_loss": 3.427567765204236,
      "tokens_seen": 2019885056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945858585858586,
      "loss": 1.7381,
      "theoretical_loss": 3.4275585110757074,
      "tokens_seen": 2019950592
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945656565656566,
      "loss": 1.3675,
      "theoretical_loss": 3.4275492573314845,
      "tokens_seen": 2020016128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039454545454545455,
      "loss": 1.5744,
      "theoretical_loss": 3.4275400039715374,
      "tokens_seen": 2020081664
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945252525252525,
      "loss": 1.5906,
      "theoretical_loss": 3.427530750995839,
      "tokens_seen": 2020147200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003945050505050505,
      "loss": 1.5845,
      "theoretical_loss": 3.427521498404359,
      "tokens_seen": 2020212736
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944848484848485,
      "loss": 1.49,
      "theoretical_loss": 3.42751224619707,
      "tokens_seen": 2020278272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944646464646465,
      "loss": 1.3986,
      "theoretical_loss": 3.427502994373944,
      "tokens_seen": 2020343808
    },
    {
      "epoch": 0.22,
      "objective/train/aligned_token_count": 0.5370234847068787,
      "objective/train/avg_token_score": 0.015117557719349861,
      "objective/train/docs_used": 1142665,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6904584169387817,
      "objective/train/original_loss": 4.142788410186768,
      "objective/train/theoretical_loss": 3.427493742934952,
      "objective/train/tokens_used": 379268576,
      "objective/train/unlikelihood_loss": 0.05455305427312851,
      "theoretical_loss": 3.427493742934952,
      "tokens_seen": 2020409344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039444444444444444,
      "loss": 1.4157,
      "theoretical_loss": 3.427493742934952,
      "tokens_seen": 2020409344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944242424242424,
      "loss": 1.4263,
      "theoretical_loss": 3.4274844918800658,
      "tokens_seen": 2020474880
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003944040404040404,
      "loss": 1.516,
      "theoretical_loss": 3.4274752412092573,
      "tokens_seen": 2020540416
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003943838383838384,
      "loss": 1.6332,
      "theoretical_loss": 3.427465990922497,
      "tokens_seen": 2020605952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003943636363636364,
      "loss": 1.5494,
      "theoretical_loss": 3.427456741019758,
      "tokens_seen": 2020671488
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039434343434343434,
      "loss": 1.2278,
      "theoretical_loss": 3.4274474915010105,
      "tokens_seen": 2020737024
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003943232323232323,
      "loss": 1.4388,
      "theoretical_loss": 3.427438242366227,
      "tokens_seen": 2020802560
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003943030303030303,
      "loss": 1.4224,
      "theoretical_loss": 3.427428993615379,
      "tokens_seen": 2020868096
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003942828282828283,
      "loss": 1.5674,
      "theoretical_loss": 3.4274197452484376,
      "tokens_seen": 2020933632
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039426262626262627,
      "loss": 1.5341,
      "theoretical_loss": 3.4274104972653747,
      "tokens_seen": 2020999168
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039424242424242423,
      "loss": 1.4315,
      "theoretical_loss": 3.4274012496661626,
      "tokens_seen": 2021064704
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003942222222222222,
      "loss": 1.3652,
      "theoretical_loss": 3.4273920024507722,
      "tokens_seen": 2021130240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039420202020202025,
      "loss": 1.5896,
      "theoretical_loss": 3.427382755619175,
      "tokens_seen": 2021195776
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003941818181818182,
      "loss": 1.3944,
      "theoretical_loss": 3.427373509171343,
      "tokens_seen": 2021261312
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039416161616161616,
      "loss": 1.4165,
      "theoretical_loss": 3.427364263107248,
      "tokens_seen": 2021326848
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003941414141414141,
      "loss": 1.5089,
      "theoretical_loss": 3.4273550174268603,
      "tokens_seen": 2021392384
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003941212121212121,
      "loss": 1.4739,
      "theoretical_loss": 3.4273457721301535,
      "tokens_seen": 2021457920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039410101010101014,
      "loss": 1.5756,
      "theoretical_loss": 3.4273365272170984,
      "tokens_seen": 2021523456
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003940808080808081,
      "loss": 1.4581,
      "theoretical_loss": 3.427327282687666,
      "tokens_seen": 2021588992
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039406060606060606,
      "loss": 1.5966,
      "theoretical_loss": 3.427318038541829,
      "tokens_seen": 2021654528
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394040404040404,
      "loss": 1.586,
      "theoretical_loss": 3.4273087947795586,
      "tokens_seen": 2021720064
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000394020202020202,
      "loss": 1.4665,
      "theoretical_loss": 3.427299551400826,
      "tokens_seen": 2021785600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039400000000000004,
      "loss": 1.4767,
      "theoretical_loss": 3.4272903084056034,
      "tokens_seen": 2021851136
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000393979797979798,
      "loss": 1.4185,
      "theoretical_loss": 3.4272810657938626,
      "tokens_seen": 2021916672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039395959595959595,
      "loss": 1.6319,
      "theoretical_loss": 3.427271823565574,
      "tokens_seen": 2021982208
    },
    {
      "epoch": 0.22,
      "objective/train/aligned_token_count": 0.5471652150154114,
      "objective/train/avg_token_score": 0.012184202671051025,
      "objective/train/docs_used": 1143239,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7339967489242554,
      "objective/train/original_loss": 4.035579204559326,
      "objective/train/theoretical_loss": 3.4272625817207114,
      "objective/train/tokens_used": 380906976,
      "objective/train/unlikelihood_loss": 0.04034143686294556,
      "theoretical_loss": 3.4272625817207114,
      "tokens_seen": 2022047744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003939393939393939,
      "loss": 1.4043,
      "theoretical_loss": 3.4272625817207114,
      "tokens_seen": 2022047744
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039391919191919197,
      "loss": 1.6556,
      "theoretical_loss": 3.427253340259245,
      "tokens_seen": 2022113280
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039389898989898993,
      "loss": 1.3459,
      "theoretical_loss": 3.427244099181147,
      "tokens_seen": 2022178816
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003938787878787879,
      "loss": 1.3638,
      "theoretical_loss": 3.427234858486388,
      "tokens_seen": 2022244352
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039385858585858584,
      "loss": 1.5388,
      "theoretical_loss": 3.4272256181749414,
      "tokens_seen": 2022309888
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039383838383838385,
      "loss": 1.326,
      "theoretical_loss": 3.4272163782467775,
      "tokens_seen": 2022375424
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039381818181818186,
      "loss": 1.3037,
      "theoretical_loss": 3.4272071387018688,
      "tokens_seen": 2022440960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003937979797979798,
      "loss": 1.4877,
      "theoretical_loss": 3.427197899540187,
      "tokens_seen": 2022506496
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003937777777777778,
      "loss": 1.5708,
      "theoretical_loss": 3.4271886607617024,
      "tokens_seen": 2022572032
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039375757575757574,
      "loss": 1.4995,
      "theoretical_loss": 3.4271794223663887,
      "tokens_seen": 2022637568
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039373737373737375,
      "loss": 1.5048,
      "theoretical_loss": 3.427170184354216,
      "tokens_seen": 2022703104
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039371717171717176,
      "loss": 1.2581,
      "theoretical_loss": 3.427160946725157,
      "tokens_seen": 2022768640
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003936969696969697,
      "loss": 1.5276,
      "theoretical_loss": 3.4271517094791832,
      "tokens_seen": 2022834176
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039367676767676767,
      "loss": 1.7464,
      "theoretical_loss": 3.427142472616266,
      "tokens_seen": 2022899712
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003936565656565657,
      "loss": 1.2784,
      "theoretical_loss": 3.4271332361363775,
      "tokens_seen": 2022965248
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039363636363636364,
      "loss": 1.6781,
      "theoretical_loss": 3.4271240000394885,
      "tokens_seen": 2023030784
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039361616161616165,
      "loss": 1.5917,
      "theoretical_loss": 3.427114764325572,
      "tokens_seen": 2023096320
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003935959595959596,
      "loss": 1.4526,
      "theoretical_loss": 3.427105528994599,
      "tokens_seen": 2023161856
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039357575757575756,
      "loss": 1.4812,
      "theoretical_loss": 3.427096294046541,
      "tokens_seen": 2023227392
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003935555555555556,
      "loss": 1.5517,
      "theoretical_loss": 3.4270870594813703,
      "tokens_seen": 2023292928
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039353535353535353,
      "loss": 1.4609,
      "theoretical_loss": 3.427077825299058,
      "tokens_seen": 2023358464
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039351515151515154,
      "loss": 1.3124,
      "theoretical_loss": 3.4270685914995767,
      "tokens_seen": 2023424000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934949494949495,
      "loss": 1.5729,
      "theoretical_loss": 3.4270593580828974,
      "tokens_seen": 2023489536
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934747474747475,
      "loss": 1.3827,
      "theoretical_loss": 3.427050125048992,
      "tokens_seen": 2023555072
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039345454545454547,
      "loss": 1.6446,
      "theoretical_loss": 3.4270408923978324,
      "tokens_seen": 2023620608
    },
    {
      "epoch": 0.22,
      "objective/train/aligned_token_count": 0.27614858746528625,
      "objective/train/avg_token_score": 0.07473522424697876,
      "objective/train/docs_used": 1144661,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9461042284965515,
      "objective/train/original_loss": 4.836662292480469,
      "objective/train/theoretical_loss": 3.42703166012939,
      "objective/train/tokens_used": 382545376,
      "objective/train/unlikelihood_loss": 0.0774540901184082,
      "theoretical_loss": 3.42703166012939,
      "tokens_seen": 2023686144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003934343434343434,
      "loss": 1.4616,
      "theoretical_loss": 3.42703166012939,
      "tokens_seen": 2023686144
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039341414141414144,
      "loss": 1.5512,
      "theoretical_loss": 3.427022428243637,
      "tokens_seen": 2023751680
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003933939393939394,
      "loss": 1.5079,
      "theoretical_loss": 3.427013196740545,
      "tokens_seen": 2023817216
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003933737373737374,
      "loss": 1.5972,
      "theoretical_loss": 3.4270039656200852,
      "tokens_seen": 2023882752
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039335353535353536,
      "loss": 1.4106,
      "theoretical_loss": 3.4269947348822303,
      "tokens_seen": 2023948288
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003933333333333333,
      "loss": 1.5354,
      "theoretical_loss": 3.426985504526951,
      "tokens_seen": 2024013824
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039331313131313133,
      "loss": 1.3946,
      "theoretical_loss": 3.42697627455422,
      "tokens_seen": 2024079360
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039329292929292934,
      "loss": 1.3868,
      "theoretical_loss": 3.426967044964009,
      "tokens_seen": 2024144896
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003932727272727273,
      "loss": 1.4791,
      "theoretical_loss": 3.426957815756289,
      "tokens_seen": 2024210432
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039325252525252525,
      "loss": 1.2664,
      "theoretical_loss": 3.426948586931032,
      "tokens_seen": 2024275968
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003932323232323232,
      "loss": 1.4863,
      "theoretical_loss": 3.4269393584882106,
      "tokens_seen": 2024341504
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003932121212121212,
      "loss": 1.5515,
      "theoretical_loss": 3.4269301304277953,
      "tokens_seen": 2024407040
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039319191919191923,
      "loss": 1.4376,
      "theoretical_loss": 3.4269209027497594,
      "tokens_seen": 2024472576
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003931717171717172,
      "loss": 1.3563,
      "theoretical_loss": 3.4269116754540727,
      "tokens_seen": 2024538112
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039315151515151515,
      "loss": 1.6737,
      "theoretical_loss": 3.4269024485407087,
      "tokens_seen": 2024603648
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003931313131313131,
      "loss": 1.6364,
      "theoretical_loss": 3.426893222009639,
      "tokens_seen": 2024669184
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039311111111111117,
      "loss": 1.478,
      "theoretical_loss": 3.426883995860834,
      "tokens_seen": 2024734720
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003930909090909091,
      "loss": 1.49,
      "theoretical_loss": 3.4268747700942668,
      "tokens_seen": 2024800256
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003930707070707071,
      "loss": 1.2528,
      "theoretical_loss": 3.426865544709909,
      "tokens_seen": 2024865792
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039305050505050504,
      "loss": 1.4622,
      "theoretical_loss": 3.426856319707732,
      "tokens_seen": 2024931328
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000393030303030303,
      "loss": 1.5658,
      "theoretical_loss": 3.426847095087708,
      "tokens_seen": 2024996864
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039301010101010106,
      "loss": 1.5705,
      "theoretical_loss": 3.4268378708498086,
      "tokens_seen": 2025062400
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000392989898989899,
      "loss": 1.3251,
      "theoretical_loss": 3.426828646994006,
      "tokens_seen": 2025127936
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000392969696969697,
      "loss": 1.2437,
      "theoretical_loss": 3.4268194235202714,
      "tokens_seen": 2025193472
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039294949494949493,
      "loss": 1.4621,
      "theoretical_loss": 3.4268102004285765,
      "tokens_seen": 2025259008
    },
    {
      "epoch": 0.22,
      "objective/train/aligned_token_count": 0.309506356716156,
      "objective/train/avg_token_score": 0.01996792107820511,
      "objective/train/docs_used": 1145363,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8851459622383118,
      "objective/train/original_loss": 4.123068809509277,
      "objective/train/theoretical_loss": 3.4268009777188935,
      "objective/train/tokens_used": 384183776,
      "objective/train/unlikelihood_loss": 0.08348941057920456,
      "theoretical_loss": 3.4268009777188935,
      "tokens_seen": 2025324544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003929292929292929,
      "loss": 1.2802,
      "theoretical_loss": 3.4268009777188935,
      "tokens_seen": 2025324544
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039290909090909095,
      "loss": 1.3641,
      "theoretical_loss": 3.4267917553911946,
      "tokens_seen": 2025390080
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003928888888888889,
      "loss": 1.2474,
      "theoretical_loss": 3.426782533445451,
      "tokens_seen": 2025455616
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039286868686868687,
      "loss": 1.346,
      "theoretical_loss": 3.426773311881635,
      "tokens_seen": 2025521152
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003928484848484848,
      "loss": 1.5347,
      "theoretical_loss": 3.426764090699718,
      "tokens_seen": 2025586688
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039282828282828283,
      "loss": 1.3794,
      "theoretical_loss": 3.4267548698996717,
      "tokens_seen": 2025652224
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039280808080808085,
      "loss": 1.4771,
      "theoretical_loss": 3.4267456494814685,
      "tokens_seen": 2025717760
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003927878787878788,
      "loss": 1.3016,
      "theoretical_loss": 3.4267364294450795,
      "tokens_seen": 2025783296
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039276767676767676,
      "loss": 1.3826,
      "theoretical_loss": 3.426727209790478,
      "tokens_seen": 2025848832
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003927474747474747,
      "loss": 1.1181,
      "theoretical_loss": 3.426717990517634,
      "tokens_seen": 2025914368
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039272727272727273,
      "loss": 1.4132,
      "theoretical_loss": 3.4267087716265205,
      "tokens_seen": 2025979904
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039270707070707074,
      "loss": 1.4424,
      "theoretical_loss": 3.4266995531171087,
      "tokens_seen": 2026045440
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003926868686868687,
      "loss": 1.2487,
      "theoretical_loss": 3.426690334989371,
      "tokens_seen": 2026110976
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039266666666666665,
      "loss": 1.4395,
      "theoretical_loss": 3.426681117243279,
      "tokens_seen": 2026176512
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039264646464646466,
      "loss": 1.5357,
      "theoretical_loss": 3.426671899878804,
      "tokens_seen": 2026242048
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003926262626262626,
      "loss": 1.3508,
      "theoretical_loss": 3.4266626828959197,
      "tokens_seen": 2026307584
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039260606060606063,
      "loss": 1.5377,
      "theoretical_loss": 3.4266534662945958,
      "tokens_seen": 2026373120
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003925858585858586,
      "loss": 1.6388,
      "theoretical_loss": 3.4266442500748053,
      "tokens_seen": 2026438656
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039256565656565654,
      "loss": 1.5414,
      "theoretical_loss": 3.42663503423652,
      "tokens_seen": 2026504192
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039254545454545456,
      "loss": 1.3003,
      "theoretical_loss": 3.426625818779711,
      "tokens_seen": 2026569728
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003925252525252525,
      "loss": 1.6056,
      "theoretical_loss": 3.4266166037043515,
      "tokens_seen": 2026635264
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003925050505050505,
      "loss": 1.5774,
      "theoretical_loss": 3.4266073890104125,
      "tokens_seen": 2026700800
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924848484848485,
      "loss": 1.582,
      "theoretical_loss": 3.426598174697866,
      "tokens_seen": 2026766336
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924646464646465,
      "loss": 1.5685,
      "theoretical_loss": 3.4265889607666837,
      "tokens_seen": 2026831872
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039244444444444445,
      "loss": 1.5274,
      "theoretical_loss": 3.426579747216838,
      "tokens_seen": 2026897408
    },
    {
      "epoch": 0.22,
      "objective/train/aligned_token_count": 0.39002931118011475,
      "objective/train/avg_token_score": 0.004569432698190212,
      "objective/train/docs_used": 1146594,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3302489519119263,
      "objective/train/original_loss": 4.367794513702393,
      "objective/train/theoretical_loss": 3.4265705340483,
      "objective/train/tokens_used": 385822176,
      "objective/train/unlikelihood_loss": 0.05916104093194008,
      "theoretical_loss": 3.4265705340483,
      "tokens_seen": 2026962944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924242424242424,
      "loss": 1.3711,
      "theoretical_loss": 3.4265705340483,
      "tokens_seen": 2026962944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003924040404040404,
      "loss": 1.3188,
      "theoretical_loss": 3.426561321261042,
      "tokens_seen": 2027028480
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003923838383838384,
      "loss": 1.6713,
      "theoretical_loss": 3.426552108855037,
      "tokens_seen": 2027094016
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003923636363636364,
      "loss": 1.4991,
      "theoretical_loss": 3.426542896830255,
      "tokens_seen": 2027159552
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039234343434343434,
      "loss": 1.4604,
      "theoretical_loss": 3.4265336851866692,
      "tokens_seen": 2027225088
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003923232323232323,
      "loss": 1.3867,
      "theoretical_loss": 3.4265244739242506,
      "tokens_seen": 2027290624
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003923030303030303,
      "loss": 1.165,
      "theoretical_loss": 3.4265152630429725,
      "tokens_seen": 2027356160
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003922828282828283,
      "loss": 1.4778,
      "theoretical_loss": 3.426506052542805,
      "tokens_seen": 2027421696
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003922626262626263,
      "loss": 1.6086,
      "theoretical_loss": 3.426496842423721,
      "tokens_seen": 2027487232
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039224242424242423,
      "loss": 1.6731,
      "theoretical_loss": 3.4264876326856926,
      "tokens_seen": 2027552768
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039222222222222225,
      "loss": 1.4806,
      "theoretical_loss": 3.4264784233286916,
      "tokens_seen": 2027618304
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003922020202020202,
      "loss": 1.4064,
      "theoretical_loss": 3.4264692143526894,
      "tokens_seen": 2027683840
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003921818181818182,
      "loss": 1.5279,
      "theoretical_loss": 3.4264600057576584,
      "tokens_seen": 2027749376
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039216161616161617,
      "loss": 1.4998,
      "theoretical_loss": 3.4264507975435707,
      "tokens_seen": 2027814912
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003921414141414141,
      "loss": 1.6423,
      "theoretical_loss": 3.4264415897103975,
      "tokens_seen": 2027880448
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039212121212121214,
      "loss": 1.4058,
      "theoretical_loss": 3.4264323822581115,
      "tokens_seen": 2027945984
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039210101010101015,
      "loss": 1.4748,
      "theoretical_loss": 3.426423175186684,
      "tokens_seen": 2028011520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003920808080808081,
      "loss": 1.5133,
      "theoretical_loss": 3.4264139684960875,
      "tokens_seen": 2028077056
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039206060606060606,
      "loss": 1.608,
      "theoretical_loss": 3.4264047621862934,
      "tokens_seen": 2028142592
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000392040404040404,
      "loss": 1.489,
      "theoretical_loss": 3.4263955562572743,
      "tokens_seen": 2028208128
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039202020202020203,
      "loss": 1.4776,
      "theoretical_loss": 3.4263863507090018,
      "tokens_seen": 2028273664
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039200000000000004,
      "loss": 1.6193,
      "theoretical_loss": 3.426377145541448,
      "tokens_seen": 2028339200
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000391979797979798,
      "loss": 1.4625,
      "theoretical_loss": 3.426367940754584,
      "tokens_seen": 2028404736
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039195959595959596,
      "loss": 1.5303,
      "theoretical_loss": 3.4263587363483827,
      "tokens_seen": 2028470272
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003919393939393939,
      "loss": 1.3432,
      "theoretical_loss": 3.426349532322816,
      "tokens_seen": 2028535808
    },
    {
      "epoch": 0.22,
      "objective/train/aligned_token_count": 0.28635141253471375,
      "objective/train/avg_token_score": 0.03568592295050621,
      "objective/train/docs_used": 1147311,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8363280892372131,
      "objective/train/original_loss": 5.367769241333008,
      "objective/train/theoretical_loss": 3.4263403286778558,
      "objective/train/tokens_used": 387460576,
      "objective/train/unlikelihood_loss": 0.08712463080883026,
      "theoretical_loss": 3.4263403286778558,
      "tokens_seen": 2028601344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.000391919191919192,
      "loss": 1.2285,
      "theoretical_loss": 3.4263403286778558,
      "tokens_seen": 2028601344
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039189898989898993,
      "loss": 1.4714,
      "theoretical_loss": 3.4263311254134736,
      "tokens_seen": 2028666880
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003918787878787879,
      "loss": 1.5263,
      "theoretical_loss": 3.4263219225296417,
      "tokens_seen": 2028732416
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039185858585858585,
      "loss": 1.4768,
      "theoretical_loss": 3.4263127200263326,
      "tokens_seen": 2028797952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003918383838383838,
      "loss": 1.4421,
      "theoretical_loss": 3.4263035179035173,
      "tokens_seen": 2028863488
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039181818181818187,
      "loss": 1.337,
      "theoretical_loss": 3.4262943161611688,
      "tokens_seen": 2028929024
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003917979797979798,
      "loss": 1.546,
      "theoretical_loss": 3.4262851147992577,
      "tokens_seen": 2028994560
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003917777777777778,
      "loss": 1.4646,
      "theoretical_loss": 3.426275913817757,
      "tokens_seen": 2029060096
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039175757575757574,
      "loss": 1.3797,
      "theoretical_loss": 3.4262667132166387,
      "tokens_seen": 2029125632
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039173737373737375,
      "loss": 1.5535,
      "theoretical_loss": 3.4262575129958748,
      "tokens_seen": 2029191168
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039171717171717176,
      "loss": 1.4743,
      "theoretical_loss": 3.426248313155437,
      "tokens_seen": 2029256704
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003916969696969697,
      "loss": 1.4379,
      "theoretical_loss": 3.426239113695297,
      "tokens_seen": 2029322240
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003916767676767677,
      "loss": 1.6946,
      "theoretical_loss": 3.426229914615427,
      "tokens_seen": 2029387776
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039165656565656563,
      "loss": 1.5947,
      "theoretical_loss": 3.4262207159158,
      "tokens_seen": 2029453312
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039163636363636364,
      "loss": 1.344,
      "theoretical_loss": 3.4262115175963865,
      "tokens_seen": 2029518848
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039161616161616166,
      "loss": 1.5337,
      "theoretical_loss": 3.4262023196571594,
      "tokens_seen": 2029584384
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003915959595959596,
      "loss": 1.3476,
      "theoretical_loss": 3.4261931220980903,
      "tokens_seen": 2029649920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039157575757575757,
      "loss": 1.284,
      "theoretical_loss": 3.426183924919152,
      "tokens_seen": 2029715456
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003915555555555556,
      "loss": 1.4791,
      "theoretical_loss": 3.4261747281203148,
      "tokens_seen": 2029780992
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039153535353535354,
      "loss": 1.6132,
      "theoretical_loss": 3.4261655317015527,
      "tokens_seen": 2029846528
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039151515151515155,
      "loss": 1.2482,
      "theoretical_loss": 3.426156335662837,
      "tokens_seen": 2029912064
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003914949494949495,
      "loss": 1.5782,
      "theoretical_loss": 3.426147140004139,
      "tokens_seen": 2029977600
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039147474747474746,
      "loss": 1.3025,
      "theoretical_loss": 3.4261379447254314,
      "tokens_seen": 2030043136
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0003914545454545455,
      "loss": 1.3747,
      "theoretical_loss": 3.4261287498266864,
      "tokens_seen": 2030108672
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00039143434343434343,
      "loss": 1.4792,
      "theoretical_loss": 3.4261195553078756,
      "tokens_seen": 2030174208
    },
    {
      "epoch": 0.22,
      "objective/train/aligned_token_count": 0.4472140669822693,
      "objective/train/avg_token_score": 0.004431017208844423,
      "objective/train/docs_used": 1148399,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5445789098739624,
      "objective/train/original_loss": 4.722729682922363,
      "objective/train/theoretical_loss": 3.4261103611689716,
      "objective/train/tokens_used": 389098976,
      "objective/train/unlikelihood_loss": 0.04807444289326668,
      "theoretical_loss": 3.4261103611689716,
      "tokens_seen": 2030239744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039141414141414144,
      "loss": 1.2979,
      "theoretical_loss": 3.4261103611689716,
      "tokens_seen": 2030239744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913939393939394,
      "loss": 1.5433,
      "theoretical_loss": 3.4261011674099455,
      "tokens_seen": 2030305280
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913737373737374,
      "loss": 1.2657,
      "theoretical_loss": 3.4260919740307703,
      "tokens_seen": 2030370816
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039135353535353537,
      "loss": 1.4345,
      "theoretical_loss": 3.4260827810314174,
      "tokens_seen": 2030436352
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003913333333333333,
      "loss": 1.3582,
      "theoretical_loss": 3.426073588411859,
      "tokens_seen": 2030501888
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039131313131313133,
      "loss": 1.4711,
      "theoretical_loss": 3.426064396172068,
      "tokens_seen": 2030567424
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912929292929293,
      "loss": 1.4337,
      "theoretical_loss": 3.4260552043120147,
      "tokens_seen": 2030632960
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912727272727273,
      "loss": 1.6487,
      "theoretical_loss": 3.4260460128316725,
      "tokens_seen": 2030698496
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039125252525252526,
      "loss": 1.5703,
      "theoretical_loss": 3.4260368217310133,
      "tokens_seen": 2030764032
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912323232323232,
      "loss": 1.7008,
      "theoretical_loss": 3.426027631010009,
      "tokens_seen": 2030829568
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003912121212121212,
      "loss": 1.5387,
      "theoretical_loss": 3.4260184406686314,
      "tokens_seen": 2030895104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911919191919192,
      "loss": 1.5542,
      "theoretical_loss": 3.426009250706853,
      "tokens_seen": 2030960640
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911717171717172,
      "loss": 1.4692,
      "theoretical_loss": 3.4260000611246455,
      "tokens_seen": 2031026176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039115151515151515,
      "loss": 1.3309,
      "theoretical_loss": 3.4259908719219814,
      "tokens_seen": 2031091712
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911313131313131,
      "loss": 1.4352,
      "theoretical_loss": 3.4259816830988328,
      "tokens_seen": 2031157248
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003911111111111111,
      "loss": 1.4223,
      "theoretical_loss": 3.425972494655171,
      "tokens_seen": 2031222784
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039109090909090913,
      "loss": 1.5841,
      "theoretical_loss": 3.425963306590969,
      "tokens_seen": 2031288320
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003910707070707071,
      "loss": 1.4516,
      "theoretical_loss": 3.4259541189061977,
      "tokens_seen": 2031353856
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039105050505050504,
      "loss": 1.3569,
      "theoretical_loss": 3.4259449316008306,
      "tokens_seen": 2031419392
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000391030303030303,
      "loss": 1.2101,
      "theoretical_loss": 3.4259357446748395,
      "tokens_seen": 2031484928
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000391010101010101,
      "loss": 1.5249,
      "theoretical_loss": 3.4259265581281957,
      "tokens_seen": 2031550464
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000390989898989899,
      "loss": 1.4189,
      "theoretical_loss": 3.425917371960872,
      "tokens_seen": 2031616000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000390969696969697,
      "loss": 1.7349,
      "theoretical_loss": 3.4259081861728404,
      "tokens_seen": 2031681536
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039094949494949494,
      "loss": 1.5265,
      "theoretical_loss": 3.4258990007640726,
      "tokens_seen": 2031747072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003909292929292929,
      "loss": 1.501,
      "theoretical_loss": 3.425889815734541,
      "tokens_seen": 2031812608
    },
    {
      "epoch": 0.23,
      "objective/train/aligned_token_count": 0.4509408473968506,
      "objective/train/avg_token_score": 0.008637008257210255,
      "objective/train/docs_used": 1149081,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.402267336845398,
      "objective/train/original_loss": 3.668806552886963,
      "objective/train/theoretical_loss": 3.425880631084218,
      "objective/train/tokens_used": 390737376,
      "objective/train/unlikelihood_loss": 0.07837577164173126,
      "theoretical_loss": 3.425880631084218,
      "tokens_seen": 2031878144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039090909090909096,
      "loss": 1.4988,
      "theoretical_loss": 3.425880631084218,
      "tokens_seen": 2031878144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003908888888888889,
      "loss": 1.5695,
      "theoretical_loss": 3.4258714468130753,
      "tokens_seen": 2031943680
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039086868686868687,
      "loss": 1.2279,
      "theoretical_loss": 3.425862262921085,
      "tokens_seen": 2032009216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039084848484848483,
      "loss": 1.5898,
      "theoretical_loss": 3.425853079408219,
      "tokens_seen": 2032074752
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003908282828282828,
      "loss": 1.5135,
      "theoretical_loss": 3.4258438962744506,
      "tokens_seen": 2032140288
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039080808080808085,
      "loss": 1.5636,
      "theoretical_loss": 3.4258347135197504,
      "tokens_seen": 2032205824
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003907878787878788,
      "loss": 1.3192,
      "theoretical_loss": 3.425825531144092,
      "tokens_seen": 2032271360
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039076767676767677,
      "loss": 1.5681,
      "theoretical_loss": 3.4258163491474463,
      "tokens_seen": 2032336896
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003907474747474747,
      "loss": 1.3838,
      "theoretical_loss": 3.4258071675297854,
      "tokens_seen": 2032402432
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039072727272727273,
      "loss": 1.2697,
      "theoretical_loss": 3.425797986291083,
      "tokens_seen": 2032467968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039070707070707074,
      "loss": 1.4194,
      "theoretical_loss": 3.425788805431309,
      "tokens_seen": 2032533504
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003906868686868687,
      "loss": 1.327,
      "theoretical_loss": 3.4257796249504375,
      "tokens_seen": 2032599040
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039066666666666666,
      "loss": 1.5645,
      "theoretical_loss": 3.42577044484844,
      "tokens_seen": 2032664576
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003906464646464646,
      "loss": 1.5047,
      "theoretical_loss": 3.425761265125288,
      "tokens_seen": 2032730112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003906262626262626,
      "loss": 1.464,
      "theoretical_loss": 3.4257520857809545,
      "tokens_seen": 2032795648
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039060606060606064,
      "loss": 1.3326,
      "theoretical_loss": 3.4257429068154113,
      "tokens_seen": 2032861184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003905858585858586,
      "loss": 1.5455,
      "theoretical_loss": 3.4257337282286304,
      "tokens_seen": 2032926720
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039056565656565655,
      "loss": 1.317,
      "theoretical_loss": 3.4257245500205844,
      "tokens_seen": 2032992256
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039054545454545456,
      "loss": 1.5326,
      "theoretical_loss": 3.4257153721912452,
      "tokens_seen": 2033057792
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003905252525252525,
      "loss": 1.5445,
      "theoretical_loss": 3.4257061947405845,
      "tokens_seen": 2033123328
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039050505050505053,
      "loss": 1.182,
      "theoretical_loss": 3.4256970176685755,
      "tokens_seen": 2033188864
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003904848484848485,
      "loss": 1.4214,
      "theoretical_loss": 3.42568784097519,
      "tokens_seen": 2033254400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039046464646464644,
      "loss": 1.3142,
      "theoretical_loss": 3.425678664660399,
      "tokens_seen": 2033319936
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039044444444444445,
      "loss": 1.386,
      "theoretical_loss": 3.4256694887241768,
      "tokens_seen": 2033385472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039042424242424247,
      "loss": 1.426,
      "theoretical_loss": 3.4256603131664933,
      "tokens_seen": 2033451008
    },
    {
      "epoch": 0.23,
      "objective/train/aligned_token_count": 0.3975439965724945,
      "objective/train/avg_token_score": 0.041283704340457916,
      "objective/train/docs_used": 1149676,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4140899181365967,
      "objective/train/original_loss": 4.860000133514404,
      "objective/train/theoretical_loss": 3.4256511379873227,
      "objective/train/tokens_used": 392375776,
      "objective/train/unlikelihood_loss": 0.06886297464370728,
      "theoretical_loss": 3.4256511379873227,
      "tokens_seen": 2033516544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003904040404040404,
      "loss": 1.3443,
      "theoretical_loss": 3.4256511379873227,
      "tokens_seen": 2033516544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903838383838384,
      "loss": 1.3114,
      "theoretical_loss": 3.425641963186636,
      "tokens_seen": 2033582080
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903636363636364,
      "loss": 1.5314,
      "theoretical_loss": 3.425632788764406,
      "tokens_seen": 2033647616
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039034343434343435,
      "loss": 1.2772,
      "theoretical_loss": 3.425623614720604,
      "tokens_seen": 2033713152
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039032323232323236,
      "loss": 1.3318,
      "theoretical_loss": 3.4256144410552034,
      "tokens_seen": 2033778688
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003903030303030303,
      "loss": 1.412,
      "theoretical_loss": 3.4256052677681756,
      "tokens_seen": 2033844224
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039028282828282827,
      "loss": 1.3957,
      "theoretical_loss": 3.4255960948594932,
      "tokens_seen": 2033909760
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003902626262626263,
      "loss": 1.5125,
      "theoretical_loss": 3.425586922329128,
      "tokens_seen": 2033975296
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039024242424242424,
      "loss": 1.5286,
      "theoretical_loss": 3.4255777501770526,
      "tokens_seen": 2034040832
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039022222222222225,
      "loss": 1.5622,
      "theoretical_loss": 3.4255685784032393,
      "tokens_seen": 2034106368
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003902020202020202,
      "loss": 1.3586,
      "theoretical_loss": 3.4255594070076594,
      "tokens_seen": 2034171904
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003901818181818182,
      "loss": 1.3422,
      "theoretical_loss": 3.4255502359902863,
      "tokens_seen": 2034237440
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003901616161616162,
      "loss": 1.3987,
      "theoretical_loss": 3.4255410653510916,
      "tokens_seen": 2034302976
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039014141414141413,
      "loss": 1.2199,
      "theoretical_loss": 3.425531895090047,
      "tokens_seen": 2034368512
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039012121212121214,
      "loss": 1.3739,
      "theoretical_loss": 3.425522725207126,
      "tokens_seen": 2034434048
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003901010101010101,
      "loss": 1.475,
      "theoretical_loss": 3.4255135557022998,
      "tokens_seen": 2034499584
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003900808080808081,
      "loss": 1.256,
      "theoretical_loss": 3.4255043865755415,
      "tokens_seen": 2034565120
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039006060606060607,
      "loss": 1.4305,
      "theoretical_loss": 3.425495217826822,
      "tokens_seen": 2034630656
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000390040404040404,
      "loss": 1.4553,
      "theoretical_loss": 3.425486049456115,
      "tokens_seen": 2034696192
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039002020202020204,
      "loss": 1.5309,
      "theoretical_loss": 3.425476881463392,
      "tokens_seen": 2034761728
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00039000000000000005,
      "loss": 1.3208,
      "theoretical_loss": 3.425467713848625,
      "tokens_seen": 2034827264
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389979797979798,
      "loss": 1.4335,
      "theoretical_loss": 3.425458546611787,
      "tokens_seen": 2034892800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038995959595959596,
      "loss": 1.6035,
      "theoretical_loss": 3.42544937975285,
      "tokens_seen": 2034958336
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003899393939393939,
      "loss": 1.303,
      "theoretical_loss": 3.4254402132717856,
      "tokens_seen": 2035023872
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038991919191919193,
      "loss": 1.5508,
      "theoretical_loss": 3.4254310471685665,
      "tokens_seen": 2035089408
    },
    {
      "epoch": 0.23,
      "objective/train/aligned_token_count": 0.40835776925086975,
      "objective/train/avg_token_score": 0.025650721043348312,
      "objective/train/docs_used": 1150523,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4754818677902222,
      "objective/train/original_loss": 4.815614223480225,
      "objective/train/theoretical_loss": 3.4254218814431656,
      "objective/train/tokens_used": 394014176,
      "objective/train/unlikelihood_loss": 0.04576752707362175,
      "theoretical_loss": 3.4254218814431656,
      "tokens_seen": 2035154944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038989898989898994,
      "loss": 1.3807,
      "theoretical_loss": 3.4254218814431656,
      "tokens_seen": 2035154944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898787878787879,
      "loss": 1.6743,
      "theoretical_loss": 3.425412716095554,
      "tokens_seen": 2035220480
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038985858585858585,
      "loss": 1.4136,
      "theoretical_loss": 3.4254035511257053,
      "tokens_seen": 2035286016
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898383838383838,
      "loss": 1.4809,
      "theoretical_loss": 3.4253943865335903,
      "tokens_seen": 2035351552
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003898181818181819,
      "loss": 1.3785,
      "theoretical_loss": 3.425385222319182,
      "tokens_seen": 2035417088
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038979797979797983,
      "loss": 1.5739,
      "theoretical_loss": 3.425376058482453,
      "tokens_seen": 2035482624
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897777777777778,
      "loss": 1.1614,
      "theoretical_loss": 3.425366895023375,
      "tokens_seen": 2035548160
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038975757575757575,
      "loss": 1.3595,
      "theoretical_loss": 3.425357731941921,
      "tokens_seen": 2035613696
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003897373737373737,
      "loss": 1.4393,
      "theoretical_loss": 3.425348569238062,
      "tokens_seen": 2035679232
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038971717171717177,
      "loss": 1.2865,
      "theoretical_loss": 3.4253394069117715,
      "tokens_seen": 2035744768
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003896969696969697,
      "loss": 1.499,
      "theoretical_loss": 3.425330244963021,
      "tokens_seen": 2035810304
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003896767676767677,
      "loss": 1.568,
      "theoretical_loss": 3.4253210833917835,
      "tokens_seen": 2035875840
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038965656565656564,
      "loss": 1.4883,
      "theoretical_loss": 3.425311922198031,
      "tokens_seen": 2035941376
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003896363636363636,
      "loss": 1.4195,
      "theoretical_loss": 3.4253027613817357,
      "tokens_seen": 2036006912
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038961616161616166,
      "loss": 1.4323,
      "theoretical_loss": 3.4252936009428696,
      "tokens_seen": 2036072448
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003895959595959596,
      "loss": 1.3823,
      "theoretical_loss": 3.4252844408814056,
      "tokens_seen": 2036137984
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003895757575757576,
      "loss": 1.3188,
      "theoretical_loss": 3.4252752811973157,
      "tokens_seen": 2036203520
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038955555555555553,
      "loss": 1.4645,
      "theoretical_loss": 3.4252661218905724,
      "tokens_seen": 2036269056
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038953535353535354,
      "loss": 1.3862,
      "theoretical_loss": 3.4252569629611473,
      "tokens_seen": 2036334592
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038951515151515155,
      "loss": 1.5763,
      "theoretical_loss": 3.4252478044090138,
      "tokens_seen": 2036400128
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003894949494949495,
      "loss": 1.1362,
      "theoretical_loss": 3.4252386462341438,
      "tokens_seen": 2036465664
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038947474747474747,
      "loss": 1.5676,
      "theoretical_loss": 3.4252294884365093,
      "tokens_seen": 2036531200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003894545454545454,
      "loss": 1.6026,
      "theoretical_loss": 3.425220331016083,
      "tokens_seen": 2036596736
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038943434343434344,
      "loss": 1.3879,
      "theoretical_loss": 3.4252111739728366,
      "tokens_seen": 2036662272
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038941414141414145,
      "loss": 1.5851,
      "theoretical_loss": 3.4252020173067432,
      "tokens_seen": 2036727808
    },
    {
      "epoch": 0.23,
      "objective/train/aligned_token_count": 0.3683406710624695,
      "objective/train/avg_token_score": 0.011941329576075077,
      "objective/train/docs_used": 1151732,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1528146266937256,
      "objective/train/original_loss": 4.464666843414307,
      "objective/train/theoretical_loss": 3.4251928610177744,
      "objective/train/tokens_used": 395652576,
      "objective/train/unlikelihood_loss": 0.0723705142736435,
      "theoretical_loss": 3.4251928610177744,
      "tokens_seen": 2036793344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003893939393939394,
      "loss": 1.3575,
      "theoretical_loss": 3.4251928610177744,
      "tokens_seen": 2036793344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038937373737373736,
      "loss": 1.1522,
      "theoretical_loss": 3.4251837051059035,
      "tokens_seen": 2036858880
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038935353535353537,
      "loss": 1.2316,
      "theoretical_loss": 3.425174549571102,
      "tokens_seen": 2036924416
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038933333333333333,
      "loss": 1.2988,
      "theoretical_loss": 3.4251653944133427,
      "tokens_seen": 2036989952
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038931313131313134,
      "loss": 1.4529,
      "theoretical_loss": 3.4251562396325976,
      "tokens_seen": 2037055488
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003892929292929293,
      "loss": 1.6243,
      "theoretical_loss": 3.4251470852288395,
      "tokens_seen": 2037121024
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038927272727272725,
      "loss": 1.3962,
      "theoretical_loss": 3.4251379312020402,
      "tokens_seen": 2037186560
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038925252525252526,
      "loss": 1.6076,
      "theoretical_loss": 3.4251287775521724,
      "tokens_seen": 2037252096
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003892323232323232,
      "loss": 1.4508,
      "theoretical_loss": 3.4251196242792084,
      "tokens_seen": 2037317632
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038921212121212123,
      "loss": 1.3475,
      "theoretical_loss": 3.42511047138312,
      "tokens_seen": 2037383168
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891919191919192,
      "loss": 1.4607,
      "theoretical_loss": 3.425101318863881,
      "tokens_seen": 2037448704
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891717171717172,
      "loss": 1.4235,
      "theoretical_loss": 3.4250921667214627,
      "tokens_seen": 2037514240
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038915151515151516,
      "loss": 1.3172,
      "theoretical_loss": 3.4250830149558373,
      "tokens_seen": 2037579776
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891313131313131,
      "loss": 1.2515,
      "theoretical_loss": 3.4250738635669773,
      "tokens_seen": 2037645312
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003891111111111111,
      "loss": 1.2808,
      "theoretical_loss": 3.425064712554856,
      "tokens_seen": 2037710848
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003890909090909091,
      "loss": 1.0805,
      "theoretical_loss": 3.425055561919444,
      "tokens_seen": 2037776384
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003890707070707071,
      "loss": 1.4789,
      "theoretical_loss": 3.4250464116607153,
      "tokens_seen": 2037841920
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038905050505050505,
      "loss": 1.2794,
      "theoretical_loss": 3.4250372617786415,
      "tokens_seen": 2037907456
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389030303030303,
      "loss": 1.4194,
      "theoretical_loss": 3.4250281122731954,
      "tokens_seen": 2037972992
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000389010101010101,
      "loss": 1.433,
      "theoretical_loss": 3.4250189631443493,
      "tokens_seen": 2038038528
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038898989898989903,
      "loss": 1.4705,
      "theoretical_loss": 3.425009814392075,
      "tokens_seen": 2038104064
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000388969696969697,
      "loss": 1.3959,
      "theoretical_loss": 3.425000666016346,
      "tokens_seen": 2038169600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038894949494949494,
      "loss": 1.4009,
      "theoretical_loss": 3.4249915180171335,
      "tokens_seen": 2038235136
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003889292929292929,
      "loss": 1.2109,
      "theoretical_loss": 3.4249823703944102,
      "tokens_seen": 2038300672
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003889090909090909,
      "loss": 1.6845,
      "theoretical_loss": 3.4249732231481493,
      "tokens_seen": 2038366208
    },
    {
      "epoch": 0.23,
      "objective/train/aligned_token_count": 0.2941104471683502,
      "objective/train/avg_token_score": 0.020645203068852425,
      "objective/train/docs_used": 1152471,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9252520203590393,
      "objective/train/original_loss": 4.827944278717041,
      "objective/train/theoretical_loss": 3.4249640762783224,
      "objective/train/tokens_used": 397290976,
      "objective/train/unlikelihood_loss": 0.03434165567159653,
      "theoretical_loss": 3.4249640762783224,
      "tokens_seen": 2038431744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888888888888889,
      "loss": 1.2632,
      "theoretical_loss": 3.4249640762783224,
      "tokens_seen": 2038431744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888686868686869,
      "loss": 1.6081,
      "theoretical_loss": 3.424954929784902,
      "tokens_seen": 2038497280
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038884848484848483,
      "loss": 1.3837,
      "theoretical_loss": 3.424945783667861,
      "tokens_seen": 2038562816
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003888282828282828,
      "loss": 1.5471,
      "theoretical_loss": 3.4249366379271713,
      "tokens_seen": 2038628352
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038880808080808086,
      "loss": 1.455,
      "theoretical_loss": 3.4249274925628055,
      "tokens_seen": 2038693888
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003887878787878788,
      "loss": 1.5115,
      "theoretical_loss": 3.424918347574736,
      "tokens_seen": 2038759424
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038876767676767677,
      "loss": 1.6937,
      "theoretical_loss": 3.4249092029629353,
      "tokens_seen": 2038824960
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038874747474747473,
      "loss": 1.3687,
      "theoretical_loss": 3.424900058727376,
      "tokens_seen": 2038890496
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038872727272727274,
      "loss": 1.7146,
      "theoretical_loss": 3.4248909148680298,
      "tokens_seen": 2038956032
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038870707070707075,
      "loss": 1.5177,
      "theoretical_loss": 3.42488177138487,
      "tokens_seen": 2039021568
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886868686868687,
      "loss": 1.4942,
      "theoretical_loss": 3.424872628277868,
      "tokens_seen": 2039087104
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038866666666666666,
      "loss": 1.5334,
      "theoretical_loss": 3.4248634855469975,
      "tokens_seen": 2039152640
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886464646464646,
      "loss": 1.3483,
      "theoretical_loss": 3.42485434319223,
      "tokens_seen": 2039218176
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003886262626262627,
      "loss": 1.3849,
      "theoretical_loss": 3.4248452012135386,
      "tokens_seen": 2039283712
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038860606060606064,
      "loss": 1.4504,
      "theoretical_loss": 3.424836059610895,
      "tokens_seen": 2039349248
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003885858585858586,
      "loss": 1.4475,
      "theoretical_loss": 3.424826918384272,
      "tokens_seen": 2039414784
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038856565656565656,
      "loss": 1.5788,
      "theoretical_loss": 3.424817777533643,
      "tokens_seen": 2039480320
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003885454545454545,
      "loss": 1.3884,
      "theoretical_loss": 3.4248086370589785,
      "tokens_seen": 2039545856
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003885252525252526,
      "loss": 1.6731,
      "theoretical_loss": 3.4247994969602527,
      "tokens_seen": 2039611392
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038850505050505054,
      "loss": 1.3777,
      "theoretical_loss": 3.4247903572374367,
      "tokens_seen": 2039676928
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003884848484848485,
      "loss": 1.5438,
      "theoretical_loss": 3.424781217890504,
      "tokens_seen": 2039742464
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038846464646464645,
      "loss": 1.052,
      "theoretical_loss": 3.4247720789194274,
      "tokens_seen": 2039808000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038844444444444446,
      "loss": 1.3154,
      "theoretical_loss": 3.424762940324178,
      "tokens_seen": 2039873536
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038842424242424247,
      "loss": 1.2718,
      "theoretical_loss": 3.424753802104729,
      "tokens_seen": 2039939072
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038840404040404043,
      "loss": 1.4838,
      "theoretical_loss": 3.4247446642610524,
      "tokens_seen": 2040004608
    },
    {
      "epoch": 0.23,
      "objective/train/aligned_token_count": 0.39326736330986023,
      "objective/train/avg_token_score": 0.01161976158618927,
      "objective/train/docs_used": 1153876,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2003021240234375,
      "objective/train/original_loss": 4.390503883361816,
      "objective/train/theoretical_loss": 3.4247355267931217,
      "objective/train/tokens_used": 398929376,
      "objective/train/unlikelihood_loss": 0.05784899368882179,
      "theoretical_loss": 3.4247355267931217,
      "tokens_seen": 2040070144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003883838383838384,
      "loss": 1.3673,
      "theoretical_loss": 3.4247355267931217,
      "tokens_seen": 2040070144
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038836363636363634,
      "loss": 1.2824,
      "theoretical_loss": 3.4247263897009086,
      "tokens_seen": 2040135680
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038834343434343435,
      "loss": 1.4723,
      "theoretical_loss": 3.4247172529843857,
      "tokens_seen": 2040201216
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038832323232323236,
      "loss": 1.3064,
      "theoretical_loss": 3.4247081166435254,
      "tokens_seen": 2040266752
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003883030303030303,
      "loss": 1.5967,
      "theoretical_loss": 3.4246989806783006,
      "tokens_seen": 2040332288
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003882828282828283,
      "loss": 1.4823,
      "theoretical_loss": 3.4246898450886833,
      "tokens_seen": 2040397824
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003882626262626263,
      "loss": 1.1834,
      "theoretical_loss": 3.4246807098746466,
      "tokens_seen": 2040463360
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038824242424242425,
      "loss": 1.4359,
      "theoretical_loss": 3.424671575036162,
      "tokens_seen": 2040528896
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038822222222222226,
      "loss": 1.3799,
      "theoretical_loss": 3.424662440573203,
      "tokens_seen": 2040594432
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003882020202020202,
      "loss": 1.1812,
      "theoretical_loss": 3.4246533064857414,
      "tokens_seen": 2040659968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038818181818181817,
      "loss": 1.6693,
      "theoretical_loss": 3.4246441727737507,
      "tokens_seen": 2040725504
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003881616161616162,
      "loss": 1.5425,
      "theoretical_loss": 3.424635039437202,
      "tokens_seen": 2040791040
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038814141414141414,
      "loss": 1.3501,
      "theoretical_loss": 3.4246259064760687,
      "tokens_seen": 2040856576
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038812121212121215,
      "loss": 1.4213,
      "theoretical_loss": 3.4246167738903233,
      "tokens_seen": 2040922112
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003881010101010101,
      "loss": 1.2594,
      "theoretical_loss": 3.424607641679938,
      "tokens_seen": 2040987648
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003880808080808081,
      "loss": 1.407,
      "theoretical_loss": 3.424598509844886,
      "tokens_seen": 2041053184
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003880606060606061,
      "loss": 1.5835,
      "theoretical_loss": 3.4245893783851384,
      "tokens_seen": 2041118720
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038804040404040403,
      "loss": 1.484,
      "theoretical_loss": 3.4245802473006695,
      "tokens_seen": 2041184256
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038802020202020204,
      "loss": 1.5434,
      "theoretical_loss": 3.4245711165914505,
      "tokens_seen": 2041249792
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000388,
      "loss": 1.5105,
      "theoretical_loss": 3.4245619862574546,
      "tokens_seen": 2041315328
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000387979797979798,
      "loss": 1.4721,
      "theoretical_loss": 3.424552856298654,
      "tokens_seen": 2041380864
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038795959595959597,
      "loss": 1.338,
      "theoretical_loss": 3.4245437267150214,
      "tokens_seen": 2041446400
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003879393939393939,
      "loss": 1.5508,
      "theoretical_loss": 3.4245345975065296,
      "tokens_seen": 2041511936
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038791919191919193,
      "loss": 1.3857,
      "theoretical_loss": 3.4245254686731506,
      "tokens_seen": 2041577472
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878989898989899,
      "loss": 1.6434,
      "theoretical_loss": 3.424516340214857,
      "tokens_seen": 2041643008
    },
    {
      "epoch": 0.23,
      "objective/train/aligned_token_count": 0.4324902296066284,
      "objective/train/avg_token_score": 0.014002914540469646,
      "objective/train/docs_used": 1154501,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5224765539169312,
      "objective/train/original_loss": 4.884289264678955,
      "objective/train/theoretical_loss": 3.424507212131622,
      "objective/train/tokens_used": 400567776,
      "objective/train/unlikelihood_loss": 0.04442446678876877,
      "theoretical_loss": 3.424507212131622,
      "tokens_seen": 2041708544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878787878787879,
      "loss": 1.4539,
      "theoretical_loss": 3.424507212131622,
      "tokens_seen": 2041708544
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038785858585858586,
      "loss": 1.3583,
      "theoretical_loss": 3.4244980844234174,
      "tokens_seen": 2041774080
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878383838383838,
      "loss": 1.3776,
      "theoretical_loss": 3.424488957090216,
      "tokens_seen": 2041839616
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003878181818181818,
      "loss": 1.4747,
      "theoretical_loss": 3.4244798301319905,
      "tokens_seen": 2041905152
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038779797979797984,
      "loss": 1.2291,
      "theoretical_loss": 3.424470703548714,
      "tokens_seen": 2041970688
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003877777777777778,
      "loss": 1.2546,
      "theoretical_loss": 3.4244615773403577,
      "tokens_seen": 2042036224
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038775757575757575,
      "loss": 1.2885,
      "theoretical_loss": 3.4244524515068946,
      "tokens_seen": 2042101760
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003877373737373737,
      "loss": 1.3927,
      "theoretical_loss": 3.4244433260482983,
      "tokens_seen": 2042167296
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003877171717171717,
      "loss": 1.5276,
      "theoretical_loss": 3.4244342009645403,
      "tokens_seen": 2042232832
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038769696969696973,
      "loss": 1.3749,
      "theoretical_loss": 3.4244250762555937,
      "tokens_seen": 2042298368
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003876767676767677,
      "loss": 1.3225,
      "theoretical_loss": 3.424415951921431,
      "tokens_seen": 2042363904
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038765656565656564,
      "loss": 1.3656,
      "theoretical_loss": 3.4244068279620246,
      "tokens_seen": 2042429440
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003876363636363636,
      "loss": 1.4351,
      "theoretical_loss": 3.4243977043773466,
      "tokens_seen": 2042494976
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038761616161616167,
      "loss": 1.51,
      "theoretical_loss": 3.4243885811673707,
      "tokens_seen": 2042560512
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875959595959596,
      "loss": 1.5127,
      "theoretical_loss": 3.424379458332069,
      "tokens_seen": 2042626048
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875757575757576,
      "loss": 1.4925,
      "theoretical_loss": 3.424370335871414,
      "tokens_seen": 2042691584
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038755555555555554,
      "loss": 1.6616,
      "theoretical_loss": 3.424361213785378,
      "tokens_seen": 2042757120
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003875353535353535,
      "loss": 1.4192,
      "theoretical_loss": 3.424352092073934,
      "tokens_seen": 2042822656
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038751515151515156,
      "loss": 1.5003,
      "theoretical_loss": 3.424342970737055,
      "tokens_seen": 2042888192
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003874949494949495,
      "loss": 1.5405,
      "theoretical_loss": 3.4243338497747127,
      "tokens_seen": 2042953728
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003874747474747475,
      "loss": 1.5091,
      "theoretical_loss": 3.42432472918688,
      "tokens_seen": 2043019264
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038745454545454543,
      "loss": 1.5408,
      "theoretical_loss": 3.4243156089735303,
      "tokens_seen": 2043084800
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038743434343434344,
      "loss": 1.5459,
      "theoretical_loss": 3.424306489134635,
      "tokens_seen": 2043150336
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038741414141414145,
      "loss": 1.3214,
      "theoretical_loss": 3.4242973696701675,
      "tokens_seen": 2043215872
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003873939393939394,
      "loss": 1.3561,
      "theoretical_loss": 3.4242882505801,
      "tokens_seen": 2043281408
    },
    {
      "epoch": 0.23,
      "objective/train/aligned_token_count": 0.3992546498775482,
      "objective/train/avg_token_score": 0.018309487029910088,
      "objective/train/docs_used": 1155127,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4903119802474976,
      "objective/train/original_loss": 4.612830638885498,
      "objective/train/theoretical_loss": 3.424279131864406,
      "objective/train/tokens_used": 402206176,
      "objective/train/unlikelihood_loss": 0.05610045790672302,
      "theoretical_loss": 3.424279131864406,
      "tokens_seen": 2043346944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038737373737373737,
      "loss": 1.5274,
      "theoretical_loss": 3.424279131864406,
      "tokens_seen": 2043346944
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003873535353535353,
      "loss": 1.4273,
      "theoretical_loss": 3.4242700135230564,
      "tokens_seen": 2043412480
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038733333333333333,
      "loss": 1.3918,
      "theoretical_loss": 3.424260895556025,
      "tokens_seen": 2043478016
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038731313131313134,
      "loss": 1.3255,
      "theoretical_loss": 3.424251777963285,
      "tokens_seen": 2043543552
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003872929292929293,
      "loss": 1.4575,
      "theoretical_loss": 3.4242426607448078,
      "tokens_seen": 2043609088
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038727272727272726,
      "loss": 1.4562,
      "theoretical_loss": 3.424233543900567,
      "tokens_seen": 2043674624
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038725252525252527,
      "loss": 1.4065,
      "theoretical_loss": 3.4242244274305347,
      "tokens_seen": 2043740160
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003872323232323232,
      "loss": 1.3534,
      "theoretical_loss": 3.4242153113346836,
      "tokens_seen": 2043805696
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038721212121212124,
      "loss": 1.382,
      "theoretical_loss": 3.424206195612986,
      "tokens_seen": 2043871232
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003871919191919192,
      "loss": 1.176,
      "theoretical_loss": 3.424197080265415,
      "tokens_seen": 2043936768
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038717171717171715,
      "loss": 1.4214,
      "theoretical_loss": 3.424187965291944,
      "tokens_seen": 2044002304
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038715151515151516,
      "loss": 1.4899,
      "theoretical_loss": 3.4241788506925444,
      "tokens_seen": 2044067840
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003871313131313131,
      "loss": 1.5427,
      "theoretical_loss": 3.4241697364671886,
      "tokens_seen": 2044133376
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038711111111111113,
      "loss": 1.4991,
      "theoretical_loss": 3.424160622615851,
      "tokens_seen": 2044198912
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870909090909091,
      "loss": 1.3178,
      "theoretical_loss": 3.424151509138502,
      "tokens_seen": 2044264448
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003870707070707071,
      "loss": 1.4449,
      "theoretical_loss": 3.4241423960351165,
      "tokens_seen": 2044329984
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038705050505050505,
      "loss": 1.1632,
      "theoretical_loss": 3.4241332833056664,
      "tokens_seen": 2044395520
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038703030303030307,
      "loss": 1.251,
      "theoretical_loss": 3.4241241709501233,
      "tokens_seen": 2044461056
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000387010101010101,
      "loss": 1.3892,
      "theoretical_loss": 3.424115058968461,
      "tokens_seen": 2044526592
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000386989898989899,
      "loss": 1.2911,
      "theoretical_loss": 3.424105947360652,
      "tokens_seen": 2044592128
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.000386969696969697,
      "loss": 1.5039,
      "theoretical_loss": 3.424096836126669,
      "tokens_seen": 2044657664
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038694949494949495,
      "loss": 1.4078,
      "theoretical_loss": 3.4240877252664834,
      "tokens_seen": 2044723200
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038692929292929296,
      "loss": 1.5386,
      "theoretical_loss": 3.42407861478007,
      "tokens_seen": 2044788736
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003869090909090909,
      "loss": 1.6039,
      "theoretical_loss": 3.4240695046674006,
      "tokens_seen": 2044854272
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003868888888888889,
      "loss": 1.4234,
      "theoretical_loss": 3.424060394928447,
      "tokens_seen": 2044919808
    },
    {
      "epoch": 0.23,
      "objective/train/aligned_token_count": 0.4493523836135864,
      "objective/train/avg_token_score": 0.012142237275838852,
      "objective/train/docs_used": 1156711,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4567104578018188,
      "objective/train/original_loss": 3.7896244525909424,
      "objective/train/theoretical_loss": 3.4240512855631833,
      "objective/train/tokens_used": 403844576,
      "objective/train/unlikelihood_loss": 0.09484424442052841,
      "theoretical_loss": 3.4240512855631833,
      "tokens_seen": 2044985344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003868686868686869,
      "loss": 1.4923,
      "theoretical_loss": 3.4240512855631833,
      "tokens_seen": 2044985344
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038684848484848484,
      "loss": 1.4911,
      "theoretical_loss": 3.424042176571582,
      "tokens_seen": 2045050880
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038682828282828285,
      "loss": 1.5525,
      "theoretical_loss": 3.424033067953615,
      "tokens_seen": 2045116416
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003868080808080808,
      "loss": 1.598,
      "theoretical_loss": 3.4240239597092548,
      "tokens_seen": 2045181952
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003867878787878788,
      "loss": 1.5819,
      "theoretical_loss": 3.424014851838475,
      "tokens_seen": 2045247488
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003867676767676768,
      "loss": 1.427,
      "theoretical_loss": 3.4240057443412484,
      "tokens_seen": 2045313024
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038674747474747473,
      "loss": 1.41,
      "theoretical_loss": 3.423996637217547,
      "tokens_seen": 2045378560
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038672727272727274,
      "loss": 1.4549,
      "theoretical_loss": 3.4239875304673437,
      "tokens_seen": 2045444096
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038670707070707076,
      "loss": 1.4496,
      "theoretical_loss": 3.4239784240906115,
      "tokens_seen": 2045509632
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003866868686868687,
      "loss": 1.304,
      "theoretical_loss": 3.423969318087323,
      "tokens_seen": 2045575168
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038666666666666667,
      "loss": 1.2372,
      "theoretical_loss": 3.423960212457451,
      "tokens_seen": 2045640704
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003866464646464646,
      "loss": 1.5265,
      "theoretical_loss": 3.423951107200968,
      "tokens_seen": 2045706240
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038662626262626264,
      "loss": 1.1508,
      "theoretical_loss": 3.423942002317846,
      "tokens_seen": 2045771776
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038660606060606065,
      "loss": 1.5112,
      "theoretical_loss": 3.4239328978080596,
      "tokens_seen": 2045837312
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865858585858586,
      "loss": 1.3431,
      "theoretical_loss": 3.4239237936715803,
      "tokens_seen": 2045902848
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038656565656565656,
      "loss": 1.434,
      "theoretical_loss": 3.4239146899083805,
      "tokens_seen": 2045968384
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865454545454545,
      "loss": 1.4307,
      "theoretical_loss": 3.4239055865184342,
      "tokens_seen": 2046033920
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003865252525252526,
      "loss": 1.2533,
      "theoretical_loss": 3.423896483501713,
      "tokens_seen": 2046099456
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038650505050505054,
      "loss": 1.2292,
      "theoretical_loss": 3.42388738085819,
      "tokens_seen": 2046164992
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003864848484848485,
      "loss": 1.329,
      "theoretical_loss": 3.423878278587838,
      "tokens_seen": 2046230528
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038646464646464645,
      "loss": 1.4768,
      "theoretical_loss": 3.42386917669063,
      "tokens_seen": 2046296064
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003864444444444444,
      "loss": 1.4991,
      "theoretical_loss": 3.4238600751665382,
      "tokens_seen": 2046361600
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003864242424242425,
      "loss": 1.2756,
      "theoretical_loss": 3.4238509740155356,
      "tokens_seen": 2046427136
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038640404040404043,
      "loss": 1.3897,
      "theoretical_loss": 3.4238418732375955,
      "tokens_seen": 2046492672
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0003863838383838384,
      "loss": 1.475,
      "theoretical_loss": 3.42383277283269,
      "tokens_seen": 2046558208
    },
    {
      "epoch": 0.23,
      "objective/train/aligned_token_count": 0.402309387922287,
      "objective/train/avg_token_score": 0.005781781394034624,
      "objective/train/docs_used": 1157531,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2796791791915894,
      "objective/train/original_loss": 3.6973581314086914,
      "objective/train/theoretical_loss": 3.4238236728007916,
      "objective/train/tokens_used": 405482976,
      "objective/train/unlikelihood_loss": 0.09952288120985031,
      "theoretical_loss": 3.4238236728007916,
      "tokens_seen": 2046623744
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00038636363636363635,
      "loss": 1.3517,
      "theoretical_loss": 3.4238236728007916,
      "tokens_seen": 2046623744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003863434343434343,
      "loss": 1.473,
      "theoretical_loss": 3.4238145731418737,
      "tokens_seen": 2046689280
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038632323232323237,
      "loss": 1.2185,
      "theoretical_loss": 3.4238054738559094,
      "tokens_seen": 2046754816
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003863030303030303,
      "loss": 1.4437,
      "theoretical_loss": 3.4237963749428704,
      "tokens_seen": 2046820352
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003862828282828283,
      "loss": 1.5039,
      "theoretical_loss": 3.4237872764027304,
      "tokens_seen": 2046885888
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038626262626262624,
      "loss": 1.4342,
      "theoretical_loss": 3.4237781782354615,
      "tokens_seen": 2046951424
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038624242424242425,
      "loss": 1.4443,
      "theoretical_loss": 3.423769080441037,
      "tokens_seen": 2047016960
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038622222222222226,
      "loss": 1.4982,
      "theoretical_loss": 3.423759983019429,
      "tokens_seen": 2047082496
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003862020202020202,
      "loss": 1.4878,
      "theoretical_loss": 3.4237508859706116,
      "tokens_seen": 2047148032
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003861818181818182,
      "loss": 1.4705,
      "theoretical_loss": 3.423741789294556,
      "tokens_seen": 2047213568
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038616161616161613,
      "loss": 1.358,
      "theoretical_loss": 3.423732692991236,
      "tokens_seen": 2047279104
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038614141414141414,
      "loss": 1.5082,
      "theoretical_loss": 3.4237235970606243,
      "tokens_seen": 2047344640
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038612121212121215,
      "loss": 1.3321,
      "theoretical_loss": 3.4237145015026935,
      "tokens_seen": 2047410176
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003861010101010101,
      "loss": 1.4646,
      "theoretical_loss": 3.4237054063174166,
      "tokens_seen": 2047475712
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038608080808080807,
      "loss": 1.5084,
      "theoretical_loss": 3.4236963115047656,
      "tokens_seen": 2047541248
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003860606060606061,
      "loss": 1.5789,
      "theoretical_loss": 3.423687217064715,
      "tokens_seen": 2047606784
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038604040404040404,
      "loss": 1.4218,
      "theoretical_loss": 3.4236781229972353,
      "tokens_seen": 2047672320
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038602020202020205,
      "loss": 1.2997,
      "theoretical_loss": 3.4236690293023018,
      "tokens_seen": 2047737856
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000386,
      "loss": 1.3044,
      "theoretical_loss": 3.4236599359798854,
      "tokens_seen": 2047803392
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038597979797979796,
      "loss": 1.3948,
      "theoretical_loss": 3.42365084302996,
      "tokens_seen": 2047868928
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038595959595959597,
      "loss": 1.5736,
      "theoretical_loss": 3.4236417504524974,
      "tokens_seen": 2047934464
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038593939393939393,
      "loss": 1.5154,
      "theoretical_loss": 3.4236326582474716,
      "tokens_seen": 2048000000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038591919191919194,
      "loss": 1.1755,
      "theoretical_loss": 3.4236235664148547,
      "tokens_seen": 2048065536
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003858989898989899,
      "loss": 1.4525,
      "theoretical_loss": 3.42361447495462,
      "tokens_seen": 2048131072
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003858787878787879,
      "loss": 1.2833,
      "theoretical_loss": 3.4236053838667395,
      "tokens_seen": 2048196608
    },
    {
      "epoch": 0.24,
      "objective/train/aligned_token_count": 0.46884164214134216,
      "objective/train/avg_token_score": 0.013134431093931198,
      "objective/train/docs_used": 1159006,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5853708982467651,
      "objective/train/original_loss": 4.141778469085693,
      "objective/train/theoretical_loss": 3.4235962931511867,
      "objective/train/tokens_used": 407121376,
      "objective/train/unlikelihood_loss": 0.04517946392297745,
      "theoretical_loss": 3.4235962931511867,
      "tokens_seen": 2048262144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038585858585858586,
      "loss": 1.5279,
      "theoretical_loss": 3.4235962931511867,
      "tokens_seen": 2048262144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003858383838383838,
      "loss": 1.3556,
      "theoretical_loss": 3.4235872028079353,
      "tokens_seen": 2048327680
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038581818181818183,
      "loss": 1.5147,
      "theoretical_loss": 3.423578112836956,
      "tokens_seen": 2048393216
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857979797979798,
      "loss": 1.3604,
      "theoretical_loss": 3.4235690232382234,
      "tokens_seen": 2048458752
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857777777777778,
      "loss": 1.6133,
      "theoretical_loss": 3.4235599340117098,
      "tokens_seen": 2048524288
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038575757575757576,
      "loss": 1.2753,
      "theoretical_loss": 3.4235508451573873,
      "tokens_seen": 2048589824
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857373737373737,
      "loss": 1.4364,
      "theoretical_loss": 3.42354175667523,
      "tokens_seen": 2048655360
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003857171717171717,
      "loss": 1.4124,
      "theoretical_loss": 3.42353266856521,
      "tokens_seen": 2048720896
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038569696969696974,
      "loss": 1.4019,
      "theoretical_loss": 3.4235235808273012,
      "tokens_seen": 2048786432
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003856767676767677,
      "loss": 1.3403,
      "theoretical_loss": 3.4235144934614747,
      "tokens_seen": 2048851968
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038565656565656565,
      "loss": 1.2462,
      "theoretical_loss": 3.423505406467705,
      "tokens_seen": 2048917504
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003856363636363636,
      "loss": 1.3504,
      "theoretical_loss": 3.4234963198459636,
      "tokens_seen": 2048983040
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003856161616161616,
      "loss": 1.4364,
      "theoretical_loss": 3.4234872335962248,
      "tokens_seen": 2049048576
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038559595959595963,
      "loss": 1.4967,
      "theoretical_loss": 3.42347814771846,
      "tokens_seen": 2049114112
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855757575757576,
      "loss": 1.3122,
      "theoretical_loss": 3.423469062212643,
      "tokens_seen": 2049179648
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038555555555555554,
      "loss": 1.4711,
      "theoretical_loss": 3.4234599770787466,
      "tokens_seen": 2049245184
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003855353535353535,
      "loss": 1.2197,
      "theoretical_loss": 3.4234508923167435,
      "tokens_seen": 2049310720
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038551515151515156,
      "loss": 1.3143,
      "theoretical_loss": 3.423441807926607,
      "tokens_seen": 2049376256
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003854949494949495,
      "loss": 1.1005,
      "theoretical_loss": 3.423432723908309,
      "tokens_seen": 2049441792
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003854747474747475,
      "loss": 1.5098,
      "theoretical_loss": 3.423423640261823,
      "tokens_seen": 2049507328
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038545454545454544,
      "loss": 1.4713,
      "theoretical_loss": 3.4234145569871224,
      "tokens_seen": 2049572864
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003854343434343434,
      "loss": 1.3695,
      "theoretical_loss": 3.4234054740841797,
      "tokens_seen": 2049638400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038541414141414146,
      "loss": 1.2553,
      "theoretical_loss": 3.4233963915529673,
      "tokens_seen": 2049703936
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003853939393939394,
      "loss": 1.6962,
      "theoretical_loss": 3.4233873093934584,
      "tokens_seen": 2049769472
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038537373737373737,
      "loss": 1.5217,
      "theoretical_loss": 3.4233782276056264,
      "tokens_seen": 2049835008
    },
    {
      "epoch": 0.24,
      "objective/train/aligned_token_count": 0.45326244831085205,
      "objective/train/avg_token_score": 0.009649938903748989,
      "objective/train/docs_used": 1159734,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5877481698989868,
      "objective/train/original_loss": 4.093586444854736,
      "objective/train/theoretical_loss": 3.4233691461894438,
      "objective/train/tokens_used": 408759776,
      "objective/train/unlikelihood_loss": 0.0731666311621666,
      "theoretical_loss": 3.4233691461894438,
      "tokens_seen": 2049900544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038535353535353533,
      "loss": 1.4922,
      "theoretical_loss": 3.4233691461894438,
      "tokens_seen": 2049900544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003853333333333334,
      "loss": 1.2207,
      "theoretical_loss": 3.4233600651448834,
      "tokens_seen": 2049966080
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038531313131313135,
      "loss": 1.1984,
      "theoretical_loss": 3.4233509844719183,
      "tokens_seen": 2050031616
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003852929292929293,
      "loss": 1.3338,
      "theoretical_loss": 3.4233419041705213,
      "tokens_seen": 2050097152
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038527272727272726,
      "loss": 1.3954,
      "theoretical_loss": 3.4233328242406653,
      "tokens_seen": 2050162688
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003852525252525252,
      "loss": 1.3318,
      "theoretical_loss": 3.4233237446823237,
      "tokens_seen": 2050228224
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003852323232323233,
      "loss": 1.2815,
      "theoretical_loss": 3.4233146654954685,
      "tokens_seen": 2050293760
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038521212121212124,
      "loss": 1.4702,
      "theoretical_loss": 3.4233055866800735,
      "tokens_seen": 2050359296
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003851919191919192,
      "loss": 1.3674,
      "theoretical_loss": 3.4232965082361106,
      "tokens_seen": 2050424832
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038517171717171716,
      "loss": 1.2718,
      "theoretical_loss": 3.4232874301635543,
      "tokens_seen": 2050490368
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038515151515151517,
      "loss": 1.579,
      "theoretical_loss": 3.4232783524623764,
      "tokens_seen": 2050555904
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003851313131313132,
      "loss": 1.3102,
      "theoretical_loss": 3.42326927513255,
      "tokens_seen": 2050621440
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038511111111111114,
      "loss": 1.4095,
      "theoretical_loss": 3.423260198174048,
      "tokens_seen": 2050686976
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003850909090909091,
      "loss": 1.6638,
      "theoretical_loss": 3.4232511215868433,
      "tokens_seen": 2050752512
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038507070707070705,
      "loss": 1.3702,
      "theoretical_loss": 3.42324204537091,
      "tokens_seen": 2050818048
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038505050505050506,
      "loss": 1.2757,
      "theoretical_loss": 3.423232969526219,
      "tokens_seen": 2050883584
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038503030303030307,
      "loss": 1.1061,
      "theoretical_loss": 3.423223894052745,
      "tokens_seen": 2050949120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038501010101010103,
      "loss": 1.4858,
      "theoretical_loss": 3.4232148189504596,
      "tokens_seen": 2051014656
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000384989898989899,
      "loss": 1.2475,
      "theoretical_loss": 3.423205744219337,
      "tokens_seen": 2051080192
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000384969696969697,
      "loss": 1.3413,
      "theoretical_loss": 3.4231966698593492,
      "tokens_seen": 2051145728
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038494949494949495,
      "loss": 1.525,
      "theoretical_loss": 3.42318759587047,
      "tokens_seen": 2051211264
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038492929292929296,
      "loss": 1.4001,
      "theoretical_loss": 3.423178522252672,
      "tokens_seen": 2051276800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003849090909090909,
      "loss": 1.5063,
      "theoretical_loss": 3.4231694490059272,
      "tokens_seen": 2051342336
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003848888888888889,
      "loss": 1.1284,
      "theoretical_loss": 3.4231603761302107,
      "tokens_seen": 2051407872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003848686868686869,
      "loss": 1.2998,
      "theoretical_loss": 3.4231513036254935,
      "tokens_seen": 2051473408
    },
    {
      "epoch": 0.24,
      "objective/train/aligned_token_count": 0.4117790758609772,
      "objective/train/avg_token_score": 0.015326161868870258,
      "objective/train/docs_used": 1160859,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1978123188018799,
      "objective/train/original_loss": 3.7221078872680664,
      "objective/train/theoretical_loss": 3.4231422314917497,
      "objective/train/tokens_used": 410398176,
      "objective/train/unlikelihood_loss": 0.08059471845626831,
      "theoretical_loss": 3.4231422314917497,
      "tokens_seen": 2051538944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038484848484848485,
      "loss": 1.3475,
      "theoretical_loss": 3.4231422314917497,
      "tokens_seen": 2051538944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038482828282828286,
      "loss": 1.3863,
      "theoretical_loss": 3.423133159728951,
      "tokens_seen": 2051604480
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003848080808080808,
      "loss": 1.4549,
      "theoretical_loss": 3.4231240883370724,
      "tokens_seen": 2051670016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038478787878787877,
      "loss": 1.1667,
      "theoretical_loss": 3.423115017316085,
      "tokens_seen": 2051735552
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003847676767676768,
      "loss": 1.3304,
      "theoretical_loss": 3.423105946665963,
      "tokens_seen": 2051801088
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038474747474747474,
      "loss": 1.4697,
      "theoretical_loss": 3.423096876386679,
      "tokens_seen": 2051866624
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038472727272727275,
      "loss": 1.3162,
      "theoretical_loss": 3.4230878064782058,
      "tokens_seen": 2051932160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003847070707070707,
      "loss": 1.4495,
      "theoretical_loss": 3.4230787369405165,
      "tokens_seen": 2051997696
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003846868686868687,
      "loss": 1.4223,
      "theoretical_loss": 3.4230696677735843,
      "tokens_seen": 2052063232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003846666666666667,
      "loss": 1.4094,
      "theoretical_loss": 3.423060598977382,
      "tokens_seen": 2052128768
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038464646464646463,
      "loss": 1.7047,
      "theoretical_loss": 3.4230515305518825,
      "tokens_seen": 2052194304
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038462626262626264,
      "loss": 1.4611,
      "theoretical_loss": 3.423042462497059,
      "tokens_seen": 2052259840
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003846060606060606,
      "loss": 1.3121,
      "theoretical_loss": 3.423033394812885,
      "tokens_seen": 2052325376
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003845858585858586,
      "loss": 1.2598,
      "theoretical_loss": 3.4230243274993324,
      "tokens_seen": 2052390912
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038456565656565657,
      "loss": 1.3826,
      "theoretical_loss": 3.423015260556375,
      "tokens_seen": 2052456448
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003845454545454545,
      "loss": 1.3184,
      "theoretical_loss": 3.4230061939839858,
      "tokens_seen": 2052521984
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038452525252525254,
      "loss": 1.37,
      "theoretical_loss": 3.4229971277821374,
      "tokens_seen": 2052587520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038450505050505055,
      "loss": 1.4403,
      "theoretical_loss": 3.4229880619508033,
      "tokens_seen": 2052653056
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003844848484848485,
      "loss": 1.4237,
      "theoretical_loss": 3.422978996489956,
      "tokens_seen": 2052718592
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038446464646464646,
      "loss": 1.4622,
      "theoretical_loss": 3.422969931399569,
      "tokens_seen": 2052784128
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003844444444444444,
      "loss": 1.273,
      "theoretical_loss": 3.422960866679616,
      "tokens_seen": 2052849664
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038442424242424243,
      "loss": 1.5379,
      "theoretical_loss": 3.422951802330068,
      "tokens_seen": 2052915200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038440404040404044,
      "loss": 1.2839,
      "theoretical_loss": 3.4229427383508995,
      "tokens_seen": 2052980736
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003843838383838384,
      "loss": 1.3775,
      "theoretical_loss": 3.422933674742084,
      "tokens_seen": 2053046272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038436363636363635,
      "loss": 1.4104,
      "theoretical_loss": 3.422924611503593,
      "tokens_seen": 2053111808
    },
    {
      "epoch": 0.24,
      "objective/train/aligned_token_count": 0.3802541494369507,
      "objective/train/avg_token_score": 0.012950729578733444,
      "objective/train/docs_used": 1161710,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1741505861282349,
      "objective/train/original_loss": 4.349769115447998,
      "objective/train/theoretical_loss": 3.422915548635401,
      "objective/train/tokens_used": 412036576,
      "objective/train/unlikelihood_loss": 0.06985551118850708,
      "theoretical_loss": 3.422915548635401,
      "tokens_seen": 2053177344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003843434343434343,
      "loss": 1.5293,
      "theoretical_loss": 3.422915548635401,
      "tokens_seen": 2053177344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003843232323232324,
      "loss": 1.3415,
      "theoretical_loss": 3.4229064861374803,
      "tokens_seen": 2053242880
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038430303030303033,
      "loss": 1.3631,
      "theoretical_loss": 3.422897424009804,
      "tokens_seen": 2053308416
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003842828282828283,
      "loss": 1.6595,
      "theoretical_loss": 3.422888362252345,
      "tokens_seen": 2053373952
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038426262626262625,
      "loss": 1.7031,
      "theoretical_loss": 3.4228793008650773,
      "tokens_seen": 2053439488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003842424242424242,
      "loss": 1.624,
      "theoretical_loss": 3.422870239847973,
      "tokens_seen": 2053505024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038422222222222227,
      "loss": 1.3284,
      "theoretical_loss": 3.4228611792010053,
      "tokens_seen": 2053570560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003842020202020202,
      "loss": 1.5913,
      "theoretical_loss": 3.4228521189241476,
      "tokens_seen": 2053636096
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003841818181818182,
      "loss": 1.5343,
      "theoretical_loss": 3.4228430590173726,
      "tokens_seen": 2053701632
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038416161616161614,
      "loss": 1.3534,
      "theoretical_loss": 3.4228339994806536,
      "tokens_seen": 2053767168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038414141414141415,
      "loss": 1.6541,
      "theoretical_loss": 3.422824940313964,
      "tokens_seen": 2053832704
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038412121212121216,
      "loss": 1.2919,
      "theoretical_loss": 3.422815881517276,
      "tokens_seen": 2053898240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003841010101010101,
      "loss": 1.2531,
      "theoretical_loss": 3.4228068230905633,
      "tokens_seen": 2053963776
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003840808080808081,
      "loss": 1.8086,
      "theoretical_loss": 3.422797765033799,
      "tokens_seen": 2054029312
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038406060606060603,
      "loss": 1.4963,
      "theoretical_loss": 3.4227887073469563,
      "tokens_seen": 2054094848
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038404040404040404,
      "loss": 1.3441,
      "theoretical_loss": 3.422779650030008,
      "tokens_seen": 2054160384
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038402020202020205,
      "loss": 1.4607,
      "theoretical_loss": 3.422770593082927,
      "tokens_seen": 2054225920
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000384,
      "loss": 1.5216,
      "theoretical_loss": 3.422761536505687,
      "tokens_seen": 2054291456
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038397979797979797,
      "loss": 1.3406,
      "theoretical_loss": 3.4227524802982607,
      "tokens_seen": 2054356992
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000383959595959596,
      "loss": 1.4556,
      "theoretical_loss": 3.422743424460621,
      "tokens_seen": 2054422528
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038393939393939393,
      "loss": 1.1927,
      "theoretical_loss": 3.4227343689927414,
      "tokens_seen": 2054488064
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038391919191919195,
      "loss": 1.3426,
      "theoretical_loss": 3.4227253138945946,
      "tokens_seen": 2054553600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003838989898989899,
      "loss": 1.2681,
      "theoretical_loss": 3.4227162591661546,
      "tokens_seen": 2054619136
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038387878787878786,
      "loss": 1.3177,
      "theoretical_loss": 3.422707204807393,
      "tokens_seen": 2054684672
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038385858585858587,
      "loss": 1.3171,
      "theoretical_loss": 3.4226981508182845,
      "tokens_seen": 2054750208
    },
    {
      "epoch": 0.24,
      "objective/train/aligned_token_count": 0.4054863154888153,
      "objective/train/avg_token_score": 0.023195790126919746,
      "objective/train/docs_used": 1163069,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4281989336013794,
      "objective/train/original_loss": 4.372689723968506,
      "objective/train/theoretical_loss": 3.422689097198802,
      "objective/train/tokens_used": 413674976,
      "objective/train/unlikelihood_loss": 0.06228712573647499,
      "theoretical_loss": 3.422689097198802,
      "tokens_seen": 2054815744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038383838383838383,
      "loss": 1.4037,
      "theoretical_loss": 3.422689097198802,
      "tokens_seen": 2054815744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038381818181818184,
      "loss": 1.5647,
      "theoretical_loss": 3.422680043948917,
      "tokens_seen": 2054881280
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003837979797979798,
      "loss": 1.5309,
      "theoretical_loss": 3.422670991068605,
      "tokens_seen": 2054946816
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003837777777777778,
      "loss": 1.4869,
      "theoretical_loss": 3.422661938557837,
      "tokens_seen": 2055012352
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038375757575757576,
      "loss": 1.5066,
      "theoretical_loss": 3.4226528864165875,
      "tokens_seen": 2055077888
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003837373737373737,
      "loss": 1.6479,
      "theoretical_loss": 3.4226438346448287,
      "tokens_seen": 2055143424
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038371717171717173,
      "loss": 1.2513,
      "theoretical_loss": 3.422634783242535,
      "tokens_seen": 2055208960
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836969696969697,
      "loss": 1.4289,
      "theoretical_loss": 3.422625732209678,
      "tokens_seen": 2055274496
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836767676767677,
      "loss": 1.3469,
      "theoretical_loss": 3.4226166815462316,
      "tokens_seen": 2055340032
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038365656565656566,
      "loss": 1.3066,
      "theoretical_loss": 3.42260763125217,
      "tokens_seen": 2055405568
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836363636363636,
      "loss": 1.4145,
      "theoretical_loss": 3.422598581327464,
      "tokens_seen": 2055471104
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003836161616161616,
      "loss": 1.4349,
      "theoretical_loss": 3.4225895317720885,
      "tokens_seen": 2055536640
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038359595959595963,
      "loss": 1.3257,
      "theoretical_loss": 3.4225804825860164,
      "tokens_seen": 2055602176
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003835757575757576,
      "loss": 1.2129,
      "theoretical_loss": 3.42257143376922,
      "tokens_seen": 2055667712
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038355555555555555,
      "loss": 1.409,
      "theoretical_loss": 3.422562385321674,
      "tokens_seen": 2055733248
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038353535353535356,
      "loss": 1.349,
      "theoretical_loss": 3.4225533372433503,
      "tokens_seen": 2055798784
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003835151515151515,
      "loss": 1.4454,
      "theoretical_loss": 3.422544289534222,
      "tokens_seen": 2055864320
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038349494949494953,
      "loss": 1.4329,
      "theoretical_loss": 3.422535242194263,
      "tokens_seen": 2055929856
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003834747474747475,
      "loss": 1.3337,
      "theoretical_loss": 3.4225261952234467,
      "tokens_seen": 2055995392
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038345454545454544,
      "loss": 1.6852,
      "theoretical_loss": 3.422517148621745,
      "tokens_seen": 2056060928
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038343434343434345,
      "loss": 1.2157,
      "theoretical_loss": 3.4225081023891324,
      "tokens_seen": 2056126464
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038341414141414146,
      "loss": 1.4933,
      "theoretical_loss": 3.422499056525581,
      "tokens_seen": 2056192000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003833939393939394,
      "loss": 1.2843,
      "theoretical_loss": 3.4224900110310648,
      "tokens_seen": 2056257536
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003833737373737374,
      "loss": 1.3396,
      "theoretical_loss": 3.422480965905556,
      "tokens_seen": 2056323072
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038335353535353533,
      "loss": 1.3149,
      "theoretical_loss": 3.4224719211490293,
      "tokens_seen": 2056388608
    },
    {
      "epoch": 0.24,
      "objective/train/aligned_token_count": 0.380987286567688,
      "objective/train/avg_token_score": 0.0735260620713234,
      "objective/train/docs_used": 1163822,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3353681564331055,
      "objective/train/original_loss": 4.655409812927246,
      "objective/train/theoretical_loss": 3.4224628767614567,
      "objective/train/tokens_used": 415313376,
      "objective/train/unlikelihood_loss": 0.04561014100909233,
      "theoretical_loss": 3.4224628767614567,
      "tokens_seen": 2056454144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038333333333333334,
      "loss": 1.5331,
      "theoretical_loss": 3.4224628767614567,
      "tokens_seen": 2056454144
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038331313131313136,
      "loss": 1.4982,
      "theoretical_loss": 3.4224538327428116,
      "tokens_seen": 2056519680
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003832929292929293,
      "loss": 1.1249,
      "theoretical_loss": 3.4224447890930674,
      "tokens_seen": 2056585216
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038327272727272727,
      "loss": 1.3073,
      "theoretical_loss": 3.422435745812197,
      "tokens_seen": 2056650752
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003832525252525252,
      "loss": 1.7152,
      "theoretical_loss": 3.4224267029001743,
      "tokens_seen": 2056716288
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003832323232323233,
      "loss": 1.682,
      "theoretical_loss": 3.422417660356972,
      "tokens_seen": 2056781824
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038321212121212125,
      "loss": 1.315,
      "theoretical_loss": 3.422408618182563,
      "tokens_seen": 2056847360
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003831919191919192,
      "loss": 1.2364,
      "theoretical_loss": 3.422399576376921,
      "tokens_seen": 2056912896
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038317171717171716,
      "loss": 1.3722,
      "theoretical_loss": 3.422390534940019,
      "tokens_seen": 2056978432
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003831515151515151,
      "loss": 1.4452,
      "theoretical_loss": 3.42238149387183,
      "tokens_seen": 2057043968
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003831313131313132,
      "loss": 1.7672,
      "theoretical_loss": 3.4223724531723283,
      "tokens_seen": 2057109504
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038311111111111114,
      "loss": 1.5932,
      "theoretical_loss": 3.4223634128414853,
      "tokens_seen": 2057175040
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003830909090909091,
      "loss": 1.4351,
      "theoretical_loss": 3.422354372879276,
      "tokens_seen": 2057240576
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038307070707070706,
      "loss": 1.3841,
      "theoretical_loss": 3.4223453332856724,
      "tokens_seen": 2057306112
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000383050505050505,
      "loss": 1.4025,
      "theoretical_loss": 3.4223362940606483,
      "tokens_seen": 2057371648
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003830303030303031,
      "loss": 1.4452,
      "theoretical_loss": 3.422327255204177,
      "tokens_seen": 2057437184
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038301010101010103,
      "loss": 1.4571,
      "theoretical_loss": 3.4223182167162314,
      "tokens_seen": 2057502720
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000382989898989899,
      "loss": 1.4462,
      "theoretical_loss": 3.4223091785967847,
      "tokens_seen": 2057568256
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038296969696969695,
      "loss": 1.4576,
      "theoretical_loss": 3.4223001408458105,
      "tokens_seen": 2057633792
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038294949494949496,
      "loss": 1.466,
      "theoretical_loss": 3.422291103463282,
      "tokens_seen": 2057699328
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038292929292929297,
      "loss": 1.4806,
      "theoretical_loss": 3.422282066449172,
      "tokens_seen": 2057764864
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003829090909090909,
      "loss": 1.5457,
      "theoretical_loss": 3.4222730298034545,
      "tokens_seen": 2057830400
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003828888888888889,
      "loss": 1.7115,
      "theoretical_loss": 3.4222639935261014,
      "tokens_seen": 2057895936
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038286868686868684,
      "loss": 1.4036,
      "theoretical_loss": 3.4222549576170875,
      "tokens_seen": 2057961472
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038284848484848485,
      "loss": 1.3656,
      "theoretical_loss": 3.422245922076386,
      "tokens_seen": 2058027008
    },
    {
      "epoch": 0.24,
      "objective/train/aligned_token_count": 0.24334067106246948,
      "objective/train/avg_token_score": 0.04306131601333618,
      "objective/train/docs_used": 1164494,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8037949800491333,
      "objective/train/original_loss": 5.774464130401611,
      "objective/train/theoretical_loss": 3.4222368869039688,
      "objective/train/tokens_used": 416951776,
      "objective/train/unlikelihood_loss": 0.0329064279794693,
      "theoretical_loss": 3.4222368869039688,
      "tokens_seen": 2058092544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038282828282828286,
      "loss": 1.0842,
      "theoretical_loss": 3.4222368869039688,
      "tokens_seen": 2058092544
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003828080808080808,
      "loss": 1.4003,
      "theoretical_loss": 3.42222785209981,
      "tokens_seen": 2058158080
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827878787878788,
      "loss": 1.5305,
      "theoretical_loss": 3.4222188176638833,
      "tokens_seen": 2058223616
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827676767676768,
      "loss": 1.5104,
      "theoretical_loss": 3.422209783596161,
      "tokens_seen": 2058289152
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038274747474747474,
      "loss": 1.6139,
      "theoretical_loss": 3.422200749896617,
      "tokens_seen": 2058354688
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038272727272727276,
      "loss": 1.2825,
      "theoretical_loss": 3.4221917165652243,
      "tokens_seen": 2058420224
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003827070707070707,
      "loss": 1.341,
      "theoretical_loss": 3.4221826836019567,
      "tokens_seen": 2058485760
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038268686868686867,
      "loss": 1.5737,
      "theoretical_loss": 3.4221736510067866,
      "tokens_seen": 2058551296
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003826666666666667,
      "loss": 1.5835,
      "theoretical_loss": 3.422164618779688,
      "tokens_seen": 2058616832
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038264646464646464,
      "loss": 1.2643,
      "theoretical_loss": 3.422155586920634,
      "tokens_seen": 2058682368
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038262626262626265,
      "loss": 1.3795,
      "theoretical_loss": 3.4221465554295976,
      "tokens_seen": 2058747904
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003826060606060606,
      "loss": 1.6555,
      "theoretical_loss": 3.422137524306552,
      "tokens_seen": 2058813440
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003825858585858586,
      "loss": 1.2681,
      "theoretical_loss": 3.4221284935514715,
      "tokens_seen": 2058878976
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038256565656565657,
      "loss": 1.2707,
      "theoretical_loss": 3.4221194631643286,
      "tokens_seen": 2058944512
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038254545454545453,
      "loss": 1.2383,
      "theoretical_loss": 3.4221104331450967,
      "tokens_seen": 2059010048
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038252525252525254,
      "loss": 1.2207,
      "theoretical_loss": 3.4221014034937487,
      "tokens_seen": 2059075584
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003825050505050505,
      "loss": 1.3845,
      "theoretical_loss": 3.4220923742102585,
      "tokens_seen": 2059141120
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003824848484848485,
      "loss": 1.4695,
      "theoretical_loss": 3.4220833452945993,
      "tokens_seen": 2059206656
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038246464646464647,
      "loss": 1.3729,
      "theoretical_loss": 3.4220743167467442,
      "tokens_seen": 2059272192
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003824444444444444,
      "loss": 1.438,
      "theoretical_loss": 3.422065288566667,
      "tokens_seen": 2059337728
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038242424242424243,
      "loss": 1.547,
      "theoretical_loss": 3.4220562607543403,
      "tokens_seen": 2059403264
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038240404040404044,
      "loss": 1.575,
      "theoretical_loss": 3.422047233309738,
      "tokens_seen": 2059468800
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823838383838384,
      "loss": 1.4872,
      "theoretical_loss": 3.422038206232833,
      "tokens_seen": 2059534336
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038236363636363636,
      "loss": 1.4038,
      "theoretical_loss": 3.422029179523599,
      "tokens_seen": 2059599872
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823434343434343,
      "loss": 1.4396,
      "theoretical_loss": 3.422020153182009,
      "tokens_seen": 2059665408
    },
    {
      "epoch": 0.24,
      "objective/train/aligned_token_count": 0.30645161867141724,
      "objective/train/avg_token_score": 0.02268202044069767,
      "objective/train/docs_used": 1165816,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.004839539527893,
      "objective/train/original_loss": 4.174016952514648,
      "objective/train/theoretical_loss": 3.4220111272080365,
      "objective/train/tokens_used": 418590176,
      "objective/train/unlikelihood_loss": 0.0708223283290863,
      "theoretical_loss": 3.4220111272080365,
      "tokens_seen": 2059730944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003823232323232323,
      "loss": 1.2435,
      "theoretical_loss": 3.4220111272080365,
      "tokens_seen": 2059730944
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038230303030303034,
      "loss": 1.5124,
      "theoretical_loss": 3.4220021016016546,
      "tokens_seen": 2059796480
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003822828282828283,
      "loss": 1.4852,
      "theoretical_loss": 3.4219930763628374,
      "tokens_seen": 2059862016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038226262626262625,
      "loss": 1.3858,
      "theoretical_loss": 3.421984051491557,
      "tokens_seen": 2059927552
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003822424242424242,
      "loss": 1.3737,
      "theoretical_loss": 3.4219750269877878,
      "tokens_seen": 2059993088
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003822222222222223,
      "loss": 1.4172,
      "theoretical_loss": 3.4219660028515024,
      "tokens_seen": 2060058624
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038220202020202023,
      "loss": 1.3067,
      "theoretical_loss": 3.421956979082675,
      "tokens_seen": 2060124160
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003821818181818182,
      "loss": 1.3377,
      "theoretical_loss": 3.4219479556812784,
      "tokens_seen": 2060189696
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038216161616161614,
      "loss": 1.5462,
      "theoretical_loss": 3.421938932647286,
      "tokens_seen": 2060255232
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003821414141414141,
      "loss": 1.5533,
      "theoretical_loss": 3.421929909980671,
      "tokens_seen": 2060320768
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038212121212121217,
      "loss": 1.5683,
      "theoretical_loss": 3.4219208876814067,
      "tokens_seen": 2060386304
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003821010101010101,
      "loss": 1.2615,
      "theoretical_loss": 3.4219118657494674,
      "tokens_seen": 2060451840
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003820808080808081,
      "loss": 1.5316,
      "theoretical_loss": 3.421902844184825,
      "tokens_seen": 2060517376
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038206060606060604,
      "loss": 1.3873,
      "theoretical_loss": 3.421893822987454,
      "tokens_seen": 2060582912
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038204040404040405,
      "loss": 1.4241,
      "theoretical_loss": 3.421884802157327,
      "tokens_seen": 2060648448
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038202020202020206,
      "loss": 1.4358,
      "theoretical_loss": 3.4218757816944185,
      "tokens_seen": 2060713984
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.000382,
      "loss": 1.3676,
      "theoretical_loss": 3.4218667615987,
      "tokens_seen": 2060779520
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038197979797979797,
      "loss": 1.3977,
      "theoretical_loss": 3.421857741870147,
      "tokens_seen": 2060845056
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038195959595959593,
      "loss": 1.4248,
      "theoretical_loss": 3.4218487225087317,
      "tokens_seen": 2060910592
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038193939393939394,
      "loss": 1.4304,
      "theoretical_loss": 3.4218397035144275,
      "tokens_seen": 2060976128
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038191919191919195,
      "loss": 1.5699,
      "theoretical_loss": 3.4218306848872073,
      "tokens_seen": 2061041664
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818989898989899,
      "loss": 1.3153,
      "theoretical_loss": 3.421821666627046,
      "tokens_seen": 2061107200
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038187878787878786,
      "loss": 1.3453,
      "theoretical_loss": 3.4218126487339156,
      "tokens_seen": 2061172736
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818585858585859,
      "loss": 1.7463,
      "theoretical_loss": 3.4218036312077906,
      "tokens_seen": 2061238272
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003818383838383839,
      "loss": 1.4188,
      "theoretical_loss": 3.4217946140486433,
      "tokens_seen": 2061303808
    },
    {
      "epoch": 0.24,
      "objective/train/aligned_token_count": 0.4169110357761383,
      "objective/train/avg_token_score": 0.013650353066623211,
      "objective/train/docs_used": 1166525,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3619921207427979,
      "objective/train/original_loss": 4.5464186668396,
      "objective/train/theoretical_loss": 3.4217855972564477,
      "objective/train/tokens_used": 420228576,
      "objective/train/unlikelihood_loss": 0.058496687561273575,
      "theoretical_loss": 3.4217855972564477,
      "tokens_seen": 2061369344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038181818181818184,
      "loss": 1.3514,
      "theoretical_loss": 3.4217855972564477,
      "tokens_seen": 2061369344
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003817979797979798,
      "loss": 1.3813,
      "theoretical_loss": 3.4217765808311773,
      "tokens_seen": 2061434880
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038177777777777776,
      "loss": 1.5357,
      "theoretical_loss": 3.4217675647728054,
      "tokens_seen": 2061500416
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038175757575757577,
      "loss": 1.7312,
      "theoretical_loss": 3.4217585490813045,
      "tokens_seen": 2061565952
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003817373737373738,
      "loss": 1.4736,
      "theoretical_loss": 3.4217495337566497,
      "tokens_seen": 2061631488
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038171717171717174,
      "loss": 1.5037,
      "theoretical_loss": 3.4217405187988135,
      "tokens_seen": 2061697024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003816969696969697,
      "loss": 1.4296,
      "theoretical_loss": 3.421731504207769,
      "tokens_seen": 2061762560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003816767676767677,
      "loss": 1.511,
      "theoretical_loss": 3.42172248998349,
      "tokens_seen": 2061828096
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038165656565656566,
      "loss": 1.2112,
      "theoretical_loss": 3.42171347612595,
      "tokens_seen": 2061893632
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038163636363636367,
      "loss": 1.6706,
      "theoretical_loss": 3.4217044626351223,
      "tokens_seen": 2061959168
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038161616161616163,
      "loss": 1.3244,
      "theoretical_loss": 3.42169544951098,
      "tokens_seen": 2062024704
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003815959595959596,
      "loss": 1.4235,
      "theoretical_loss": 3.4216864367534976,
      "tokens_seen": 2062090240
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003815757575757576,
      "loss": 1.233,
      "theoretical_loss": 3.4216774243626475,
      "tokens_seen": 2062155776
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038155555555555555,
      "loss": 1.2825,
      "theoretical_loss": 3.4216684123384034,
      "tokens_seen": 2062221312
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038153535353535357,
      "loss": 1.395,
      "theoretical_loss": 3.421659400680739,
      "tokens_seen": 2062286848
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003815151515151515,
      "loss": 1.6011,
      "theoretical_loss": 3.421650389389627,
      "tokens_seen": 2062352384
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003814949494949495,
      "loss": 1.3961,
      "theoretical_loss": 3.4216413784650417,
      "tokens_seen": 2062417920
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003814747474747475,
      "loss": 1.3719,
      "theoretical_loss": 3.4216323679069562,
      "tokens_seen": 2062483456
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038145454545454545,
      "loss": 1.4384,
      "theoretical_loss": 3.421623357715344,
      "tokens_seen": 2062548992
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038143434343434346,
      "loss": 1.4021,
      "theoretical_loss": 3.421614347890178,
      "tokens_seen": 2062614528
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003814141414141414,
      "loss": 1.5168,
      "theoretical_loss": 3.4216053384314327,
      "tokens_seen": 2062680064
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003813939393939394,
      "loss": 1.4452,
      "theoretical_loss": 3.4215963293390805,
      "tokens_seen": 2062745600
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003813737373737374,
      "loss": 1.2376,
      "theoretical_loss": 3.4215873206130962,
      "tokens_seen": 2062811136
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038135353535353534,
      "loss": 1.3614,
      "theoretical_loss": 3.4215783122534518,
      "tokens_seen": 2062876672
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00038133333333333335,
      "loss": 1.3911,
      "theoretical_loss": 3.421569304260122,
      "tokens_seen": 2062942208
    },
    {
      "epoch": 0.24,
      "objective/train/aligned_token_count": 0.5072702765464783,
      "objective/train/avg_token_score": 0.02509508840739727,
      "objective/train/docs_used": 1167825,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6743861436843872,
      "objective/train/original_loss": 4.351998329162598,
      "objective/train/theoretical_loss": 3.421560296633079,
      "objective/train/tokens_used": 421866976,
      "objective/train/unlikelihood_loss": 0.04990775138139725,
      "theoretical_loss": 3.421560296633079,
      "tokens_seen": 2063007744
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0003813131313131313,
      "loss": 1.5035,
      "theoretical_loss": 3.421560296633079,
      "tokens_seen": 2063007744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003812929292929293,
      "loss": 1.2393,
      "theoretical_loss": 3.4215512893722972,
      "tokens_seen": 2063073280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003812727272727273,
      "loss": 1.5631,
      "theoretical_loss": 3.4215422824777497,
      "tokens_seen": 2063138816
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038125252525252523,
      "loss": 1.2478,
      "theoretical_loss": 3.42153327594941,
      "tokens_seen": 2063204352
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038123232323232324,
      "loss": 1.5025,
      "theoretical_loss": 3.421524269787252,
      "tokens_seen": 2063269888
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038121212121212125,
      "loss": 1.4308,
      "theoretical_loss": 3.4215152639912487,
      "tokens_seen": 2063335424
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003811919191919192,
      "loss": 1.4696,
      "theoretical_loss": 3.421506258561374,
      "tokens_seen": 2063400960
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038117171717171717,
      "loss": 1.3622,
      "theoretical_loss": 3.421497253497601,
      "tokens_seen": 2063466496
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003811515151515151,
      "loss": 1.4055,
      "theoretical_loss": 3.4214882487999034,
      "tokens_seen": 2063532032
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038113131313131314,
      "loss": 1.1969,
      "theoretical_loss": 3.4214792444682542,
      "tokens_seen": 2063597568
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038111111111111115,
      "loss": 1.3362,
      "theoretical_loss": 3.421470240502628,
      "tokens_seen": 2063663104
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003810909090909091,
      "loss": 1.4218,
      "theoretical_loss": 3.4214612369029966,
      "tokens_seen": 2063728640
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038107070707070706,
      "loss": 1.3588,
      "theoretical_loss": 3.421452233669335,
      "tokens_seen": 2063794176
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000381050505050505,
      "loss": 1.423,
      "theoretical_loss": 3.421443230801616,
      "tokens_seen": 2063859712
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003810303030303031,
      "loss": 1.4211,
      "theoretical_loss": 3.421434228299814,
      "tokens_seen": 2063925248
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038101010101010104,
      "loss": 1.3854,
      "theoretical_loss": 3.421425226163901,
      "tokens_seen": 2063990784
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380989898989899,
      "loss": 1.4755,
      "theoretical_loss": 3.4214162243938517,
      "tokens_seen": 2064056320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038096969696969695,
      "loss": 1.4164,
      "theoretical_loss": 3.4214072229896395,
      "tokens_seen": 2064121856
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003809494949494949,
      "loss": 1.4524,
      "theoretical_loss": 3.4213982219512373,
      "tokens_seen": 2064187392
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000380929292929293,
      "loss": 1.0907,
      "theoretical_loss": 3.421389221278619,
      "tokens_seen": 2064252928
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038090909090909093,
      "loss": 1.4692,
      "theoretical_loss": 3.4213802209717583,
      "tokens_seen": 2064318464
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003808888888888889,
      "loss": 1.345,
      "theoretical_loss": 3.4213712210306286,
      "tokens_seen": 2064384000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038086868686868685,
      "loss": 1.4035,
      "theoretical_loss": 3.421362221455203,
      "tokens_seen": 2064449536
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038084848484848486,
      "loss": 1.48,
      "theoretical_loss": 3.4213532222454557,
      "tokens_seen": 2064515072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038082828282828287,
      "loss": 1.3966,
      "theoretical_loss": 3.42134422340136,
      "tokens_seen": 2064580608
    },
    {
      "epoch": 0.25,
      "objective/train/aligned_token_count": 0.5281035900115967,
      "objective/train/avg_token_score": 0.018594948574900627,
      "objective/train/docs_used": 1168450,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7348947525024414,
      "objective/train/original_loss": 3.9402873516082764,
      "objective/train/theoretical_loss": 3.421335224922889,
      "objective/train/tokens_used": 423505376,
      "objective/train/unlikelihood_loss": 0.0670773833990097,
      "theoretical_loss": 3.421335224922889,
      "tokens_seen": 2064646144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003808080808080808,
      "loss": 1.7727,
      "theoretical_loss": 3.421335224922889,
      "tokens_seen": 2064646144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003807878787878788,
      "loss": 1.2763,
      "theoretical_loss": 3.4213262268100166,
      "tokens_seen": 2064711680
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038076767676767674,
      "loss": 1.3587,
      "theoretical_loss": 3.4213172290627165,
      "tokens_seen": 2064777216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038074747474747475,
      "loss": 1.6083,
      "theoretical_loss": 3.421308231680962,
      "tokens_seen": 2064842752
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038072727272727276,
      "loss": 1.5072,
      "theoretical_loss": 3.421299234664727,
      "tokens_seen": 2064908288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003807070707070707,
      "loss": 1.4557,
      "theoretical_loss": 3.421290238013985,
      "tokens_seen": 2064973824
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806868686868687,
      "loss": 1.4156,
      "theoretical_loss": 3.4212812417287086,
      "tokens_seen": 2065039360
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806666666666667,
      "loss": 1.4262,
      "theoretical_loss": 3.421272245808873,
      "tokens_seen": 2065104896
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038064646464646464,
      "loss": 1.2626,
      "theoretical_loss": 3.42126325025445,
      "tokens_seen": 2065170432
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038062626262626265,
      "loss": 1.5695,
      "theoretical_loss": 3.4212542550654144,
      "tokens_seen": 2065235968
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003806060606060606,
      "loss": 1.3867,
      "theoretical_loss": 3.4212452602417396,
      "tokens_seen": 2065301504
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038058585858585857,
      "loss": 1.5888,
      "theoretical_loss": 3.421236265783399,
      "tokens_seen": 2065367040
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003805656565656566,
      "loss": 1.3008,
      "theoretical_loss": 3.4212272716903653,
      "tokens_seen": 2065432576
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038054545454545454,
      "loss": 1.5071,
      "theoretical_loss": 3.421218277962614,
      "tokens_seen": 2065498112
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038052525252525255,
      "loss": 1.3812,
      "theoretical_loss": 3.421209284600117,
      "tokens_seen": 2065563648
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003805050505050505,
      "loss": 1.4338,
      "theoretical_loss": 3.4212002916028483,
      "tokens_seen": 2065629184
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003804848484848485,
      "loss": 1.358,
      "theoretical_loss": 3.421191298970782,
      "tokens_seen": 2065694720
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038046464646464647,
      "loss": 1.5437,
      "theoretical_loss": 3.4211823067038907,
      "tokens_seen": 2065760256
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038044444444444443,
      "loss": 1.3364,
      "theoretical_loss": 3.4211733148021493,
      "tokens_seen": 2065825792
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038042424242424244,
      "loss": 1.243,
      "theoretical_loss": 3.42116432326553,
      "tokens_seen": 2065891328
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038042424242424244,
      "loss": 1.4324,
      "theoretical_loss": 3.421155332094008,
      "tokens_seen": 2065956864
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003804040404040404,
      "loss": 1.3529,
      "theoretical_loss": 3.4211463412875553,
      "tokens_seen": 2066022400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003803838383838384,
      "loss": 1.3293,
      "theoretical_loss": 3.4211373508461462,
      "tokens_seen": 2066087936
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038036363636363636,
      "loss": 1.4924,
      "theoretical_loss": 3.4211283607697545,
      "tokens_seen": 2066153472
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003803434343434343,
      "loss": 1.1353,
      "theoretical_loss": 3.421119371058354,
      "tokens_seen": 2066219008
    },
    {
      "epoch": 0.25,
      "objective/train/aligned_token_count": 0.4422653913497925,
      "objective/train/avg_token_score": 0.037376537919044495,
      "objective/train/docs_used": 1169491,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3472920656204224,
      "objective/train/original_loss": 4.558820724487305,
      "objective/train/theoretical_loss": 3.421110381711917,
      "objective/train/tokens_used": 425143776,
      "objective/train/unlikelihood_loss": 0.06192596256732941,
      "theoretical_loss": 3.421110381711917,
      "tokens_seen": 2066284544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038032323232323233,
      "loss": 1.4788,
      "theoretical_loss": 3.421110381711917,
      "tokens_seen": 2066284544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038030303030303034,
      "loss": 1.3516,
      "theoretical_loss": 3.421101392730418,
      "tokens_seen": 2066350080
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802828282828283,
      "loss": 1.4592,
      "theoretical_loss": 3.421092404113831,
      "tokens_seen": 2066415616
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038026262626262626,
      "loss": 1.3995,
      "theoretical_loss": 3.421083415862129,
      "tokens_seen": 2066481152
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802424242424242,
      "loss": 1.5694,
      "theoretical_loss": 3.4210744279752863,
      "tokens_seen": 2066546688
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003802222222222222,
      "loss": 1.3938,
      "theoretical_loss": 3.421065440453276,
      "tokens_seen": 2066612224
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038020202020202024,
      "loss": 1.5165,
      "theoretical_loss": 3.421056453296071,
      "tokens_seen": 2066677760
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003801818181818182,
      "loss": 1.4164,
      "theoretical_loss": 3.4210474665036466,
      "tokens_seen": 2066743296
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038016161616161615,
      "loss": 1.5343,
      "theoretical_loss": 3.421038480075975,
      "tokens_seen": 2066808832
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038014141414141416,
      "loss": 1.3788,
      "theoretical_loss": 3.421029494013031,
      "tokens_seen": 2066874368
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038012121212121217,
      "loss": 1.5264,
      "theoretical_loss": 3.421020508314787,
      "tokens_seen": 2066939904
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038010101010101013,
      "loss": 1.3608,
      "theoretical_loss": 3.421011522981217,
      "tokens_seen": 2067005440
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003800808080808081,
      "loss": 1.5184,
      "theoretical_loss": 3.4210025380122953,
      "tokens_seen": 2067070976
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038006060606060604,
      "loss": 1.351,
      "theoretical_loss": 3.420993553407995,
      "tokens_seen": 2067136512
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038004040404040405,
      "loss": 1.5483,
      "theoretical_loss": 3.42098456916829,
      "tokens_seen": 2067202048
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038002020202020206,
      "loss": 1.4377,
      "theoretical_loss": 3.4209755852931534,
      "tokens_seen": 2067267584
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00038,
      "loss": 1.4813,
      "theoretical_loss": 3.4209666017825597,
      "tokens_seen": 2067333120
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379979797979798,
      "loss": 1.2604,
      "theoretical_loss": 3.4209576186364816,
      "tokens_seen": 2067398656
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037995959595959593,
      "loss": 1.2219,
      "theoretical_loss": 3.4209486358548933,
      "tokens_seen": 2067464192
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379939393939394,
      "loss": 1.2292,
      "theoretical_loss": 3.420939653437769,
      "tokens_seen": 2067529728
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037991919191919196,
      "loss": 1.241,
      "theoretical_loss": 3.420930671385081,
      "tokens_seen": 2067595264
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003798989898989899,
      "loss": 1.443,
      "theoretical_loss": 3.4209216896968044,
      "tokens_seen": 2067660800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037987878787878787,
      "loss": 1.5578,
      "theoretical_loss": 3.420912708372912,
      "tokens_seen": 2067726336
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037985858585858583,
      "loss": 1.6159,
      "theoretical_loss": 3.4209037274133776,
      "tokens_seen": 2067791872
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003798383838383839,
      "loss": 1.5248,
      "theoretical_loss": 3.420894746818175,
      "tokens_seen": 2067857408
    },
    {
      "epoch": 0.25,
      "objective/train/aligned_token_count": 0.42247068881988525,
      "objective/train/avg_token_score": 0.005036116577684879,
      "objective/train/docs_used": 1170265,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2853634357452393,
      "objective/train/original_loss": 3.96811842918396,
      "objective/train/theoretical_loss": 3.4208857665872774,
      "objective/train/tokens_used": 426782176,
      "objective/train/unlikelihood_loss": 0.09357985854148865,
      "theoretical_loss": 3.4208857665872774,
      "tokens_seen": 2067922944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037981818181818185,
      "loss": 1.4816,
      "theoretical_loss": 3.4208857665872774,
      "tokens_seen": 2067922944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003797979797979798,
      "loss": 1.4999,
      "theoretical_loss": 3.4208767867206595,
      "tokens_seen": 2067988480
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037977777777777776,
      "loss": 1.429,
      "theoretical_loss": 3.4208678072182943,
      "tokens_seen": 2068054016
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003797575757575757,
      "loss": 1.6531,
      "theoretical_loss": 3.4208588280801555,
      "tokens_seen": 2068119552
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003797373737373738,
      "loss": 1.3178,
      "theoretical_loss": 3.420849849306216,
      "tokens_seen": 2068185088
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037971717171717174,
      "loss": 1.3174,
      "theoretical_loss": 3.4208408708964515,
      "tokens_seen": 2068250624
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003796969696969697,
      "loss": 1.5182,
      "theoretical_loss": 3.420831892850834,
      "tokens_seen": 2068316160
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037967676767676766,
      "loss": 1.5146,
      "theoretical_loss": 3.420822915169338,
      "tokens_seen": 2068381696
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037965656565656567,
      "loss": 1.5209,
      "theoretical_loss": 3.4208139378519364,
      "tokens_seen": 2068447232
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003796363636363637,
      "loss": 1.2463,
      "theoretical_loss": 3.4208049608986038,
      "tokens_seen": 2068512768
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037961616161616163,
      "loss": 1.4292,
      "theoretical_loss": 3.420795984309313,
      "tokens_seen": 2068578304
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003795959595959596,
      "loss": 1.494,
      "theoretical_loss": 3.420787008084039,
      "tokens_seen": 2068643840
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037957575757575755,
      "loss": 1.5767,
      "theoretical_loss": 3.4207780322227537,
      "tokens_seen": 2068709376
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037955555555555556,
      "loss": 1.4401,
      "theoretical_loss": 3.4207690567254327,
      "tokens_seen": 2068774912
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037953535353535357,
      "loss": 1.7829,
      "theoretical_loss": 3.4207600815920487,
      "tokens_seen": 2068840448
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037951515151515153,
      "loss": 1.3836,
      "theoretical_loss": 3.4207511068225753,
      "tokens_seen": 2068905984
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003794949494949495,
      "loss": 1.5816,
      "theoretical_loss": 3.4207421324169864,
      "tokens_seen": 2068971520
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003794747474747475,
      "loss": 1.3778,
      "theoretical_loss": 3.4207331583752563,
      "tokens_seen": 2069037056
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037945454545454545,
      "loss": 1.3386,
      "theoretical_loss": 3.4207241846973577,
      "tokens_seen": 2069102592
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037943434343434346,
      "loss": 1.5548,
      "theoretical_loss": 3.420715211383265,
      "tokens_seen": 2069168128
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003794141414141414,
      "loss": 1.4954,
      "theoretical_loss": 3.4207062384329516,
      "tokens_seen": 2069233664
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003793939393939394,
      "loss": 1.2441,
      "theoretical_loss": 3.4206972658463917,
      "tokens_seen": 2069299200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003793737373737374,
      "loss": 1.5984,
      "theoretical_loss": 3.420688293623559,
      "tokens_seen": 2069364736
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037935353535353535,
      "loss": 1.3587,
      "theoretical_loss": 3.4206793217644265,
      "tokens_seen": 2069430272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037933333333333336,
      "loss": 1.3259,
      "theoretical_loss": 3.420670350268969,
      "tokens_seen": 2069495808
    },
    {
      "epoch": 0.25,
      "objective/train/aligned_token_count": 0.39571115374565125,
      "objective/train/avg_token_score": 0.015201877802610397,
      "objective/train/docs_used": 1170942,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3105769157409668,
      "objective/train/original_loss": 4.975223064422607,
      "objective/train/theoretical_loss": 3.420661379137159,
      "objective/train/tokens_used": 428420576,
      "objective/train/unlikelihood_loss": 0.05080878734588623,
      "theoretical_loss": 3.420661379137159,
      "tokens_seen": 2069561344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003793131313131313,
      "loss": 1.4032,
      "theoretical_loss": 3.420661379137159,
      "tokens_seen": 2069561344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003792929292929293,
      "loss": 1.3405,
      "theoretical_loss": 3.420652408368971,
      "tokens_seen": 2069626880
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003792727272727273,
      "loss": 1.4213,
      "theoretical_loss": 3.4206434379643786,
      "tokens_seen": 2069692416
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037925252525252524,
      "loss": 1.3388,
      "theoretical_loss": 3.420634467923356,
      "tokens_seen": 2069757952
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037923232323232325,
      "loss": 1.2898,
      "theoretical_loss": 3.420625498245876,
      "tokens_seen": 2069823488
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003792121212121212,
      "loss": 1.3475,
      "theoretical_loss": 3.4206165289319133,
      "tokens_seen": 2069889024
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003791919191919192,
      "loss": 1.4743,
      "theoretical_loss": 3.420607559981441,
      "tokens_seen": 2069954560
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003791717171717172,
      "loss": 1.279,
      "theoretical_loss": 3.420598591394433,
      "tokens_seen": 2070020096
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037915151515151513,
      "loss": 1.5811,
      "theoretical_loss": 3.420589623170864,
      "tokens_seen": 2070085632
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037913131313131314,
      "loss": 1.4274,
      "theoretical_loss": 3.4205806553107063,
      "tokens_seen": 2070151168
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037911111111111115,
      "loss": 1.2846,
      "theoretical_loss": 3.4205716878139345,
      "tokens_seen": 2070216704
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003790909090909091,
      "loss": 1.4081,
      "theoretical_loss": 3.420562720680522,
      "tokens_seen": 2070282240
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037907070707070707,
      "loss": 1.6356,
      "theoretical_loss": 3.420553753910443,
      "tokens_seen": 2070347776
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000379050505050505,
      "loss": 1.512,
      "theoretical_loss": 3.4205447875036707,
      "tokens_seen": 2070413312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037903030303030303,
      "loss": 1.4141,
      "theoretical_loss": 3.42053582146018,
      "tokens_seen": 2070478848
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037901010101010105,
      "loss": 1.393,
      "theoretical_loss": 3.4205268557799435,
      "tokens_seen": 2070544384
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378989898989899,
      "loss": 1.5733,
      "theoretical_loss": 3.420517890462935,
      "tokens_seen": 2070609920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037896969696969696,
      "loss": 1.1883,
      "theoretical_loss": 3.420508925509129,
      "tokens_seen": 2070675456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003789494949494949,
      "loss": 1.4328,
      "theoretical_loss": 3.420499960918499,
      "tokens_seen": 2070740992
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378929292929293,
      "loss": 1.3877,
      "theoretical_loss": 3.420490996691019,
      "tokens_seen": 2070806528
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037890909090909094,
      "loss": 1.5501,
      "theoretical_loss": 3.420482032826662,
      "tokens_seen": 2070872064
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003788888888888889,
      "loss": 1.4109,
      "theoretical_loss": 3.4204730693254035,
      "tokens_seen": 2070937600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037886868686868685,
      "loss": 1.3661,
      "theoretical_loss": 3.4204641061872154,
      "tokens_seen": 2071003136
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003788484848484848,
      "loss": 1.4302,
      "theoretical_loss": 3.420455143412072,
      "tokens_seen": 2071068672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003788282828282829,
      "loss": 1.4722,
      "theoretical_loss": 3.4204461809999476,
      "tokens_seen": 2071134208
    },
    {
      "epoch": 0.25,
      "objective/train/aligned_token_count": 0.4280914068222046,
      "objective/train/avg_token_score": 0.011312766931951046,
      "objective/train/docs_used": 1172086,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4044407606124878,
      "objective/train/original_loss": 3.66874361038208,
      "objective/train/theoretical_loss": 3.4204372189508163,
      "objective/train/tokens_used": 430058976,
      "objective/train/unlikelihood_loss": 0.0882360190153122,
      "theoretical_loss": 3.4204372189508163,
      "tokens_seen": 2071199744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037880808080808083,
      "loss": 1.3167,
      "theoretical_loss": 3.4204372189508163,
      "tokens_seen": 2071199744
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003787878787878788,
      "loss": 1.4364,
      "theoretical_loss": 3.4204282572646507,
      "tokens_seen": 2071265280
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037876767676767674,
      "loss": 1.5054,
      "theoretical_loss": 3.420419295941426,
      "tokens_seen": 2071330816
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037874747474747476,
      "loss": 1.5036,
      "theoretical_loss": 3.4204103349811152,
      "tokens_seen": 2071396352
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037872727272727277,
      "loss": 1.4788,
      "theoretical_loss": 3.420401374383692,
      "tokens_seen": 2071461888
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003787070707070707,
      "loss": 1.2752,
      "theoretical_loss": 3.4203924141491306,
      "tokens_seen": 2071527424
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786868686868687,
      "loss": 1.209,
      "theoretical_loss": 3.4203834542774043,
      "tokens_seen": 2071592960
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037866666666666664,
      "loss": 1.4583,
      "theoretical_loss": 3.420374494768488,
      "tokens_seen": 2071658496
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037864646464646465,
      "loss": 1.3893,
      "theoretical_loss": 3.420365535622355,
      "tokens_seen": 2071724032
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037862626262626266,
      "loss": 1.3295,
      "theoretical_loss": 3.4203565768389783,
      "tokens_seen": 2071789568
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003786060606060606,
      "loss": 1.4072,
      "theoretical_loss": 3.420347618418333,
      "tokens_seen": 2071855104
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003785858585858586,
      "loss": 1.4761,
      "theoretical_loss": 3.4203386603603922,
      "tokens_seen": 2071920640
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003785656565656566,
      "loss": 1.4877,
      "theoretical_loss": 3.4203297026651303,
      "tokens_seen": 2071986176
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037854545454545454,
      "loss": 1.5255,
      "theoretical_loss": 3.4203207453325204,
      "tokens_seen": 2072051712
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037852525252525255,
      "loss": 1.2721,
      "theoretical_loss": 3.4203117883625365,
      "tokens_seen": 2072117248
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003785050505050505,
      "loss": 1.3668,
      "theoretical_loss": 3.420302831755153,
      "tokens_seen": 2072182784
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037848484848484847,
      "loss": 1.4005,
      "theoretical_loss": 3.4202938755103434,
      "tokens_seen": 2072248320
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003784646464646465,
      "loss": 1.4546,
      "theoretical_loss": 3.420284919628082,
      "tokens_seen": 2072313856
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003784444444444445,
      "loss": 1.5243,
      "theoretical_loss": 3.4202759641083413,
      "tokens_seen": 2072379392
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037842424242424244,
      "loss": 1.0941,
      "theoretical_loss": 3.4202670089510967,
      "tokens_seen": 2072444928
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003784040404040404,
      "loss": 1.1452,
      "theoretical_loss": 3.4202580541563217,
      "tokens_seen": 2072510464
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003783838383838384,
      "loss": 1.349,
      "theoretical_loss": 3.4202490997239896,
      "tokens_seen": 2072576000
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037836363636363637,
      "loss": 1.4917,
      "theoretical_loss": 3.4202401456540747,
      "tokens_seen": 2072641536
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003783434343434344,
      "loss": 1.5066,
      "theoretical_loss": 3.420231191946551,
      "tokens_seen": 2072707072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037832323232323234,
      "loss": 1.5049,
      "theoretical_loss": 3.420222238601392,
      "tokens_seen": 2072772608
    },
    {
      "epoch": 0.25,
      "objective/train/aligned_token_count": 0.39668866991996765,
      "objective/train/avg_token_score": 0.006178613286465406,
      "objective/train/docs_used": 1172787,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4482109546661377,
      "objective/train/original_loss": 4.506319522857666,
      "objective/train/theoretical_loss": 3.420213285618572,
      "objective/train/tokens_used": 431697376,
      "objective/train/unlikelihood_loss": 0.05467967316508293,
      "theoretical_loss": 3.420213285618572,
      "tokens_seen": 2072838144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003783030303030303,
      "loss": 1.2255,
      "theoretical_loss": 3.420213285618572,
      "tokens_seen": 2072838144
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782828282828283,
      "loss": 1.4164,
      "theoretical_loss": 3.420204332998064,
      "tokens_seen": 2072903680
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037826262626262626,
      "loss": 1.386,
      "theoretical_loss": 3.420195380739843,
      "tokens_seen": 2072969216
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782424242424243,
      "loss": 1.3431,
      "theoretical_loss": 3.4201864288438824,
      "tokens_seen": 2073034752
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037822222222222223,
      "loss": 1.4041,
      "theoretical_loss": 3.4201774773101556,
      "tokens_seen": 2073100288
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003782020202020202,
      "loss": 1.277,
      "theoretical_loss": 3.4201685261386374,
      "tokens_seen": 2073165824
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003781818181818182,
      "loss": 1.3984,
      "theoretical_loss": 3.4201595753293015,
      "tokens_seen": 2073231360
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037816161616161615,
      "loss": 1.358,
      "theoretical_loss": 3.420150624882121,
      "tokens_seen": 2073296896
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037814141414141417,
      "loss": 1.3982,
      "theoretical_loss": 3.4201416747970708,
      "tokens_seen": 2073362432
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003781212121212121,
      "loss": 1.4702,
      "theoretical_loss": 3.4201327250741245,
      "tokens_seen": 2073427968
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037810101010101013,
      "loss": 1.4477,
      "theoretical_loss": 3.4201237757132557,
      "tokens_seen": 2073493504
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003780808080808081,
      "loss": 1.442,
      "theoretical_loss": 3.420114826714438,
      "tokens_seen": 2073559040
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037806060606060605,
      "loss": 1.3965,
      "theoretical_loss": 3.4201058780776465,
      "tokens_seen": 2073624576
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037804040404040406,
      "loss": 1.4009,
      "theoretical_loss": 3.4200969298028543,
      "tokens_seen": 2073690112
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378020202020202,
      "loss": 1.5525,
      "theoretical_loss": 3.4200879818900347,
      "tokens_seen": 2073755648
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000378,
      "loss": 1.5832,
      "theoretical_loss": 3.420079034339163,
      "tokens_seen": 2073821184
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000377979797979798,
      "loss": 1.3648,
      "theoretical_loss": 3.4200700871502123,
      "tokens_seen": 2073886720
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037795959595959594,
      "loss": 1.6466,
      "theoretical_loss": 3.420061140323157,
      "tokens_seen": 2073952256
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037793939393939395,
      "loss": 1.3599,
      "theoretical_loss": 3.4200521938579707,
      "tokens_seen": 2074017792
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037791919191919196,
      "loss": 1.2817,
      "theoretical_loss": 3.420043247754627,
      "tokens_seen": 2074083328
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003778989898989899,
      "loss": 1.3824,
      "theoretical_loss": 3.4200343020131,
      "tokens_seen": 2074148864
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003778787878787879,
      "loss": 1.3301,
      "theoretical_loss": 3.4200253566333645,
      "tokens_seen": 2074214400
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037785858585858583,
      "loss": 1.4708,
      "theoretical_loss": 3.4200164116153937,
      "tokens_seen": 2074279936
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037783838383838384,
      "loss": 1.2739,
      "theoretical_loss": 3.420007466959161,
      "tokens_seen": 2074345472
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037781818181818186,
      "loss": 1.3532,
      "theoretical_loss": 3.419998522664641,
      "tokens_seen": 2074411008
    },
    {
      "epoch": 0.25,
      "objective/train/aligned_token_count": 0.3930840790271759,
      "objective/train/avg_token_score": 0.019307734444737434,
      "objective/train/docs_used": 1174027,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3908665180206299,
      "objective/train/original_loss": 3.919964551925659,
      "objective/train/theoretical_loss": 3.4199895787318084,
      "objective/train/tokens_used": 433335776,
      "objective/train/unlikelihood_loss": 0.07689140737056732,
      "theoretical_loss": 3.4199895787318084,
      "tokens_seen": 2074476544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003777979797979798,
      "loss": 1.4928,
      "theoretical_loss": 3.4199895787318084,
      "tokens_seen": 2074476544
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037777777777777777,
      "loss": 1.4619,
      "theoretical_loss": 3.4199806351606354,
      "tokens_seen": 2074542080
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003777575757575757,
      "loss": 1.4482,
      "theoretical_loss": 3.4199716919510976,
      "tokens_seen": 2074607616
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003777373737373738,
      "loss": 1.2923,
      "theoretical_loss": 3.4199627491031674,
      "tokens_seen": 2074673152
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037771717171717175,
      "loss": 1.1326,
      "theoretical_loss": 3.41995380661682,
      "tokens_seen": 2074738688
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003776969696969697,
      "loss": 1.6064,
      "theoretical_loss": 3.419944864492029,
      "tokens_seen": 2074804224
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037767676767676766,
      "loss": 1.3603,
      "theoretical_loss": 3.4199359227287687,
      "tokens_seen": 2074869760
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003776565656565656,
      "loss": 1.453,
      "theoretical_loss": 3.4199269813270123,
      "tokens_seen": 2074935296
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003776363636363637,
      "loss": 1.5569,
      "theoretical_loss": 3.419918040286734,
      "tokens_seen": 2075000832
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037761616161616164,
      "loss": 1.8289,
      "theoretical_loss": 3.4199090996079082,
      "tokens_seen": 2075066368
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003775959595959596,
      "loss": 1.3843,
      "theoretical_loss": 3.419900159290509,
      "tokens_seen": 2075131904
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037757575757575755,
      "loss": 1.5317,
      "theoretical_loss": 3.419891219334509,
      "tokens_seen": 2075197440
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037755555555555557,
      "loss": 1.4498,
      "theoretical_loss": 3.419882279739884,
      "tokens_seen": 2075262976
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003775353535353536,
      "loss": 1.4489,
      "theoretical_loss": 3.4198733405066064,
      "tokens_seen": 2075328512
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037751515151515153,
      "loss": 1.501,
      "theoretical_loss": 3.4198644016346513,
      "tokens_seen": 2075394048
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003774949494949495,
      "loss": 1.3273,
      "theoretical_loss": 3.4198554631239926,
      "tokens_seen": 2075459584
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037747474747474745,
      "loss": 1.3192,
      "theoretical_loss": 3.4198465249746035,
      "tokens_seen": 2075525120
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037745454545454546,
      "loss": 1.4783,
      "theoretical_loss": 3.419837587186459,
      "tokens_seen": 2075590656
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037743434343434347,
      "loss": 1.3299,
      "theoretical_loss": 3.419828649759532,
      "tokens_seen": 2075656192
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003774141414141414,
      "loss": 1.6143,
      "theoretical_loss": 3.419819712693797,
      "tokens_seen": 2075721728
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003773939393939394,
      "loss": 1.337,
      "theoretical_loss": 3.4198107759892284,
      "tokens_seen": 2075787264
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003773737373737374,
      "loss": 1.3659,
      "theoretical_loss": 3.4198018396457996,
      "tokens_seen": 2075852800
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037735353535353535,
      "loss": 1.3633,
      "theoretical_loss": 3.4197929036634855,
      "tokens_seen": 2075918336
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037733333333333336,
      "loss": 1.5404,
      "theoretical_loss": 3.4197839680422586,
      "tokens_seen": 2075983872
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003773131313131313,
      "loss": 1.6373,
      "theoretical_loss": 3.419775032782095,
      "tokens_seen": 2076049408
    },
    {
      "epoch": 0.25,
      "objective/train/aligned_token_count": 0.4901026487350464,
      "objective/train/avg_token_score": 0.014933989383280277,
      "objective/train/docs_used": 1174812,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6007699966430664,
      "objective/train/original_loss": 4.3715410232543945,
      "objective/train/theoretical_loss": 3.4197660978829667,
      "objective/train/tokens_used": 434974176,
      "objective/train/unlikelihood_loss": 0.044615305960178375,
      "theoretical_loss": 3.4197660978829667,
      "tokens_seen": 2076114944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003772929292929293,
      "loss": 1.4418,
      "theoretical_loss": 3.4197660978829667,
      "tokens_seen": 2076114944
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003772727272727273,
      "loss": 1.5686,
      "theoretical_loss": 3.4197571633448485,
      "tokens_seen": 2076180480
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037725252525252524,
      "loss": 1.462,
      "theoretical_loss": 3.4197482291677144,
      "tokens_seen": 2076246016
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037723232323232325,
      "loss": 1.3599,
      "theoretical_loss": 3.4197392953515386,
      "tokens_seen": 2076311552
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003772121212121212,
      "loss": 1.2252,
      "theoretical_loss": 3.419730361896295,
      "tokens_seen": 2076377088
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003771919191919192,
      "loss": 1.352,
      "theoretical_loss": 3.4197214288019575,
      "tokens_seen": 2076442624
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003771717171717172,
      "loss": 1.349,
      "theoretical_loss": 3.4197124960685006,
      "tokens_seen": 2076508160
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037715151515151514,
      "loss": 1.5102,
      "theoretical_loss": 3.419703563695898,
      "tokens_seen": 2076573696
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037713131313131315,
      "loss": 1.2434,
      "theoretical_loss": 3.4196946316841235,
      "tokens_seen": 2076639232
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003771111111111111,
      "loss": 1.624,
      "theoretical_loss": 3.419685700033151,
      "tokens_seen": 2076704768
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003770909090909091,
      "loss": 1.3274,
      "theoretical_loss": 3.4196767687429555,
      "tokens_seen": 2076770304
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037707070707070707,
      "loss": 1.3486,
      "theoretical_loss": 3.41966783781351,
      "tokens_seen": 2076835840
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037705050505050503,
      "loss": 1.4912,
      "theoretical_loss": 3.4196589072447896,
      "tokens_seen": 2076901376
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037703030303030304,
      "loss": 1.5207,
      "theoretical_loss": 3.4196499770367668,
      "tokens_seen": 2076966912
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037701010101010105,
      "loss": 1.4682,
      "theoretical_loss": 3.419641047189417,
      "tokens_seen": 2077032448
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.000376989898989899,
      "loss": 1.3966,
      "theoretical_loss": 3.4196321177027142,
      "tokens_seen": 2077097984
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037696969696969696,
      "loss": 1.4683,
      "theoretical_loss": 3.4196231885766317,
      "tokens_seen": 2077163520
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003769494949494949,
      "loss": 1.3407,
      "theoretical_loss": 3.4196142598111443,
      "tokens_seen": 2077229056
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037692929292929293,
      "loss": 1.4778,
      "theoretical_loss": 3.4196053314062254,
      "tokens_seen": 2077294592
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037690909090909094,
      "loss": 1.3302,
      "theoretical_loss": 3.419596403361849,
      "tokens_seen": 2077360128
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768888888888889,
      "loss": 1.2834,
      "theoretical_loss": 3.4195874756779903,
      "tokens_seen": 2077425664
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037686868686868686,
      "loss": 1.4535,
      "theoretical_loss": 3.4195785483546226,
      "tokens_seen": 2077491200
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768484848484848,
      "loss": 1.4252,
      "theoretical_loss": 3.4195696213917195,
      "tokens_seen": 2077556736
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003768282828282829,
      "loss": 1.3688,
      "theoretical_loss": 3.419560694789256,
      "tokens_seen": 2077622272
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037680808080808084,
      "loss": 1.6246,
      "theoretical_loss": 3.419551768547205,
      "tokens_seen": 2077687808
    },
    {
      "epoch": 0.25,
      "objective/train/aligned_token_count": 0.33198925852775574,
      "objective/train/avg_token_score": 0.010006610304117203,
      "objective/train/docs_used": 1175928,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0175628662109375,
      "objective/train/original_loss": 4.5042290687561035,
      "objective/train/theoretical_loss": 3.419542842665542,
      "objective/train/tokens_used": 436612576,
      "objective/train/unlikelihood_loss": 0.040932752192020416,
      "theoretical_loss": 3.419542842665542,
      "tokens_seen": 2077753344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003767878787878788,
      "loss": 1.4852,
      "theoretical_loss": 3.419542842665542,
      "tokens_seen": 2077753344
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037676767676767675,
      "loss": 1.6318,
      "theoretical_loss": 3.41953391714424,
      "tokens_seen": 2077818880
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003767474747474747,
      "loss": 1.4043,
      "theoretical_loss": 3.4195249919832738,
      "tokens_seen": 2077884416
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037672727272727277,
      "loss": 1.6004,
      "theoretical_loss": 3.419516067182617,
      "tokens_seen": 2077949952
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037670707070707073,
      "loss": 1.3982,
      "theoretical_loss": 3.419507142742244,
      "tokens_seen": 2078015488
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003766868686868687,
      "loss": 1.2151,
      "theoretical_loss": 3.4194982186621288,
      "tokens_seen": 2078081024
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037666666666666664,
      "loss": 1.4042,
      "theoretical_loss": 3.4194892949422453,
      "tokens_seen": 2078146560
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003766464646464647,
      "loss": 1.372,
      "theoretical_loss": 3.419480371582568,
      "tokens_seen": 2078212096
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037662626262626266,
      "loss": 1.5022,
      "theoretical_loss": 3.419471448583071,
      "tokens_seen": 2078277632
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003766060606060606,
      "loss": 1.4701,
      "theoretical_loss": 3.419462525943727,
      "tokens_seen": 2078343168
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003765858585858586,
      "loss": 1.4062,
      "theoretical_loss": 3.419453603664512,
      "tokens_seen": 2078408704
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037656565656565654,
      "loss": 1.4994,
      "theoretical_loss": 3.4194446817453996,
      "tokens_seen": 2078474240
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003765454545454546,
      "loss": 1.456,
      "theoretical_loss": 3.4194357601863636,
      "tokens_seen": 2078539776
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037652525252525256,
      "loss": 1.5474,
      "theoretical_loss": 3.419426838987378,
      "tokens_seen": 2078605312
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003765050505050505,
      "loss": 1.5788,
      "theoretical_loss": 3.419417918148417,
      "tokens_seen": 2078670848
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037648484848484847,
      "loss": 1.3919,
      "theoretical_loss": 3.4194089976694553,
      "tokens_seen": 2078736384
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037646464646464643,
      "loss": 1.3453,
      "theoretical_loss": 3.419400077550466,
      "tokens_seen": 2078801920
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003764444444444445,
      "loss": 1.4328,
      "theoretical_loss": 3.4193911577914244,
      "tokens_seen": 2078867456
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037642424242424245,
      "loss": 1.3432,
      "theoretical_loss": 3.419382238392304,
      "tokens_seen": 2078932992
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003764040404040404,
      "loss": 1.4633,
      "theoretical_loss": 3.4193733193530784,
      "tokens_seen": 2078998528
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037638383838383836,
      "loss": 1.4656,
      "theoretical_loss": 3.419364400673723,
      "tokens_seen": 2079064064
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003763636363636364,
      "loss": 1.5146,
      "theoretical_loss": 3.4193554823542107,
      "tokens_seen": 2079129600
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003763434343434344,
      "loss": 1.6026,
      "theoretical_loss": 3.419346564394516,
      "tokens_seen": 2079195136
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00037632323232323234,
      "loss": 1.3136,
      "theoretical_loss": 3.419337646794614,
      "tokens_seen": 2079260672
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0003763030303030303,
      "loss": 1.374,
      "theoretical_loss": 3.4193287295544774,
      "tokens_seen": 2079326208
    },
    {
      "epoch": 0.25,
      "objective/train/aligned_token_count": 0.49841153621673584,
      "objective/train/avg_token_score": 0.009859995916485786,
      "objective/train/docs_used": 1176572,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5007386207580566,
      "objective/train/original_loss": 3.7215120792388916,
      "objective/train/theoretical_loss": 3.419319812674081,
      "objective/train/tokens_used": 438250976,
      "objective/train/unlikelihood_loss": 0.09313442558050156,
      "theoretical_loss": 3.419319812674081,
      "tokens_seen": 2079391744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037628282828282826,
      "loss": 1.6772,
      "theoretical_loss": 3.419319812674081,
      "tokens_seen": 2079391744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037626262626262627,
      "loss": 1.4377,
      "theoretical_loss": 3.4193108961533993,
      "tokens_seen": 2079457280
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003762424242424243,
      "loss": 1.5206,
      "theoretical_loss": 3.4193019799924063,
      "tokens_seen": 2079522816
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037622222222222224,
      "loss": 1.4173,
      "theoretical_loss": 3.4192930641910753,
      "tokens_seen": 2079588352
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003762020202020202,
      "loss": 1.2782,
      "theoretical_loss": 3.4192841487493815,
      "tokens_seen": 2079653888
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003761818181818182,
      "loss": 1.2694,
      "theoretical_loss": 3.4192752336672987,
      "tokens_seen": 2079719424
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037616161616161616,
      "loss": 1.5735,
      "theoretical_loss": 3.4192663189448007,
      "tokens_seen": 2079784960
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037614141414141417,
      "loss": 1.4436,
      "theoretical_loss": 3.4192574045818622,
      "tokens_seen": 2079850496
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037612121212121213,
      "loss": 1.3291,
      "theoretical_loss": 3.4192484905784575,
      "tokens_seen": 2079916032
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003761010101010101,
      "loss": 1.6037,
      "theoretical_loss": 3.41923957693456,
      "tokens_seen": 2079981568
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003760808080808081,
      "loss": 1.3677,
      "theoretical_loss": 3.4192306636501444,
      "tokens_seen": 2080047104
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037606060606060605,
      "loss": 1.4373,
      "theoretical_loss": 3.4192217507251854,
      "tokens_seen": 2080112640
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037604040404040406,
      "loss": 1.397,
      "theoretical_loss": 3.419212838159656,
      "tokens_seen": 2080178176
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000376020202020202,
      "loss": 1.429,
      "theoretical_loss": 3.4192039259535307,
      "tokens_seen": 2080243712
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037600000000000003,
      "loss": 1.2791,
      "theoretical_loss": 3.4191950141067844,
      "tokens_seen": 2080309248
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000375979797979798,
      "loss": 1.2927,
      "theoretical_loss": 3.419186102619391,
      "tokens_seen": 2080374784
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037595959595959595,
      "loss": 1.4066,
      "theoretical_loss": 3.419177191491324,
      "tokens_seen": 2080440320
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037593939393939396,
      "loss": 1.522,
      "theoretical_loss": 3.4191682807225585,
      "tokens_seen": 2080505856
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003759191919191919,
      "loss": 1.5626,
      "theoretical_loss": 3.419159370313068,
      "tokens_seen": 2080571392
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003758989898989899,
      "loss": 1.4265,
      "theoretical_loss": 3.419150460262827,
      "tokens_seen": 2080636928
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003758787878787879,
      "loss": 1.4347,
      "theoretical_loss": 3.41914155057181,
      "tokens_seen": 2080702464
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037585858585858584,
      "loss": 1.4526,
      "theoretical_loss": 3.419132641239991,
      "tokens_seen": 2080768000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037583838383838385,
      "loss": 1.5855,
      "theoretical_loss": 3.419123732267344,
      "tokens_seen": 2080833536
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037581818181818186,
      "loss": 1.2248,
      "theoretical_loss": 3.4191148236538433,
      "tokens_seen": 2080899072
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003757979797979798,
      "loss": 1.4762,
      "theoretical_loss": 3.419105915399463,
      "tokens_seen": 2080964608
    },
    {
      "epoch": 0.26,
      "objective/train/aligned_token_count": 0.43218475580215454,
      "objective/train/avg_token_score": 0.01707909256219864,
      "objective/train/docs_used": 1177438,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4051634073257446,
      "objective/train/original_loss": 4.354172229766846,
      "objective/train/theoretical_loss": 3.4190970075041776,
      "objective/train/tokens_used": 439889376,
      "objective/train/unlikelihood_loss": 0.06684666872024536,
      "theoretical_loss": 3.4190970075041776,
      "tokens_seen": 2081030144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003757777777777778,
      "loss": 1.4901,
      "theoretical_loss": 3.4190970075041776,
      "tokens_seen": 2081030144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037575757575757573,
      "loss": 1.3864,
      "theoretical_loss": 3.4190880999679614,
      "tokens_seen": 2081095680
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037573737373737374,
      "loss": 1.3233,
      "theoretical_loss": 3.4190791927907878,
      "tokens_seen": 2081161216
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037571717171717175,
      "loss": 1.5315,
      "theoretical_loss": 3.419070285972632,
      "tokens_seen": 2081226752
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003756969696969697,
      "loss": 1.3804,
      "theoretical_loss": 3.419061379513468,
      "tokens_seen": 2081292288
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037567676767676767,
      "loss": 1.3739,
      "theoretical_loss": 3.4190524734132692,
      "tokens_seen": 2081357824
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003756565656565656,
      "loss": 1.5024,
      "theoretical_loss": 3.419043567672011,
      "tokens_seen": 2081423360
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003756363636363637,
      "loss": 1.3044,
      "theoretical_loss": 3.4190346622896675,
      "tokens_seen": 2081488896
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037561616161616165,
      "loss": 1.5856,
      "theoretical_loss": 3.419025757266212,
      "tokens_seen": 2081554432
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755959595959596,
      "loss": 1.494,
      "theoretical_loss": 3.4190168526016196,
      "tokens_seen": 2081619968
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037557575757575756,
      "loss": 1.2753,
      "theoretical_loss": 3.4190079482958637,
      "tokens_seen": 2081685504
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755555555555555,
      "loss": 1.477,
      "theoretical_loss": 3.4189990443489195,
      "tokens_seen": 2081751040
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003755353535353536,
      "loss": 1.5872,
      "theoretical_loss": 3.4189901407607612,
      "tokens_seen": 2081816576
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037551515151515154,
      "loss": 1.6336,
      "theoretical_loss": 3.418981237531362,
      "tokens_seen": 2081882112
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754949494949495,
      "loss": 1.4524,
      "theoretical_loss": 3.4189723346606975,
      "tokens_seen": 2081947648
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037547474747474745,
      "loss": 1.4721,
      "theoretical_loss": 3.4189634321487405,
      "tokens_seen": 2082013184
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037545454545454546,
      "loss": 1.4003,
      "theoretical_loss": 3.4189545299954665,
      "tokens_seen": 2082078720
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003754343434343435,
      "loss": 1.5892,
      "theoretical_loss": 3.4189456282008495,
      "tokens_seen": 2082144256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037541414141414143,
      "loss": 1.4227,
      "theoretical_loss": 3.4189367267648634,
      "tokens_seen": 2082209792
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003753939393939394,
      "loss": 1.4975,
      "theoretical_loss": 3.4189278256874824,
      "tokens_seen": 2082275328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037537373737373735,
      "loss": 1.6269,
      "theoretical_loss": 3.4189189249686813,
      "tokens_seen": 2082340864
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037535353535353536,
      "loss": 1.3594,
      "theoretical_loss": 3.418910024608434,
      "tokens_seen": 2082406400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037533333333333337,
      "loss": 1.3369,
      "theoretical_loss": 3.4189011246067147,
      "tokens_seen": 2082471936
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003753131313131313,
      "loss": 1.4857,
      "theoretical_loss": 3.418892224963498,
      "tokens_seen": 2082537472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003752929292929293,
      "loss": 1.3987,
      "theoretical_loss": 3.418883325678758,
      "tokens_seen": 2082603008
    },
    {
      "epoch": 0.26,
      "objective/train/aligned_token_count": 0.34909579157829285,
      "objective/train/avg_token_score": 0.03110983781516552,
      "objective/train/docs_used": 1177725,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2431418895721436,
      "objective/train/original_loss": 4.56602668762207,
      "objective/train/theoretical_loss": 3.4188744267524687,
      "objective/train/tokens_used": 441527776,
      "objective/train/unlikelihood_loss": 0.06893003731966019,
      "theoretical_loss": 3.4188744267524687,
      "tokens_seen": 2082668544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003752727272727273,
      "loss": 1.4113,
      "theoretical_loss": 3.4188744267524687,
      "tokens_seen": 2082668544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037525252525252525,
      "loss": 1.5249,
      "theoretical_loss": 3.4188655281846048,
      "tokens_seen": 2082734080
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037523232323232326,
      "loss": 1.6509,
      "theoretical_loss": 3.4188566299751404,
      "tokens_seen": 2082799616
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003752121212121212,
      "loss": 1.4777,
      "theoretical_loss": 3.41884773212405,
      "tokens_seen": 2082865152
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003751919191919192,
      "loss": 1.463,
      "theoretical_loss": 3.418838834631307,
      "tokens_seen": 2082930688
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003751717171717172,
      "loss": 1.6285,
      "theoretical_loss": 3.4188299374968874,
      "tokens_seen": 2082996224
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037515151515151514,
      "loss": 1.4813,
      "theoretical_loss": 3.418821040720764,
      "tokens_seen": 2083061760
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037513131313131315,
      "loss": 1.7179,
      "theoretical_loss": 3.418812144302912,
      "tokens_seen": 2083127296
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003751111111111111,
      "loss": 1.5771,
      "theoretical_loss": 3.418803248243305,
      "tokens_seen": 2083192832
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003750909090909091,
      "loss": 1.6911,
      "theoretical_loss": 3.4187943525419175,
      "tokens_seen": 2083258368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003750707070707071,
      "loss": 1.4178,
      "theoretical_loss": 3.4187854571987244,
      "tokens_seen": 2083323904
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037505050505050503,
      "loss": 1.5065,
      "theoretical_loss": 3.418776562213699,
      "tokens_seen": 2083389440
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037503030303030305,
      "loss": 1.4285,
      "theoretical_loss": 3.418767667586817,
      "tokens_seen": 2083454976
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000375010101010101,
      "loss": 1.4658,
      "theoretical_loss": 3.418758773318051,
      "tokens_seen": 2083520512
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374989898989899,
      "loss": 1.5077,
      "theoretical_loss": 3.4187498794073763,
      "tokens_seen": 2083586048
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037496969696969697,
      "loss": 1.6018,
      "theoretical_loss": 3.4187409858547673,
      "tokens_seen": 2083651584
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374949494949495,
      "loss": 1.4844,
      "theoretical_loss": 3.418732092660198,
      "tokens_seen": 2083717120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037492929292929294,
      "loss": 1.2961,
      "theoretical_loss": 3.418723199823643,
      "tokens_seen": 2083782656
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003749090909090909,
      "loss": 1.7236,
      "theoretical_loss": 3.418714307345076,
      "tokens_seen": 2083848192
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003748888888888889,
      "loss": 1.4139,
      "theoretical_loss": 3.4187054152244727,
      "tokens_seen": 2083913728
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037486868686868686,
      "loss": 1.3001,
      "theoretical_loss": 3.418696523461806,
      "tokens_seen": 2083979264
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003748484848484849,
      "loss": 1.4182,
      "theoretical_loss": 3.4186876320570505,
      "tokens_seen": 2084044800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037482828282828283,
      "loss": 1.4754,
      "theoretical_loss": 3.4186787410101815,
      "tokens_seen": 2084110336
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037480808080808084,
      "loss": 1.3918,
      "theoretical_loss": 3.418669850321172,
      "tokens_seen": 2084175872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003747878787878788,
      "loss": 1.3437,
      "theoretical_loss": 3.4186609599899977,
      "tokens_seen": 2084241408
    },
    {
      "epoch": 0.26,
      "objective/train/aligned_token_count": 0.3273460268974304,
      "objective/train/avg_token_score": 0.024734986945986748,
      "objective/train/docs_used": 1178778,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1921221017837524,
      "objective/train/original_loss": 4.915249347686768,
      "objective/train/theoretical_loss": 3.418652070016632,
      "objective/train/tokens_used": 443166176,
      "objective/train/unlikelihood_loss": 0.03642764315009117,
      "theoretical_loss": 3.418652070016632,
      "tokens_seen": 2084306944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037476767676767676,
      "loss": 1.6279,
      "theoretical_loss": 3.418652070016632,
      "tokens_seen": 2084306944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037474747474747477,
      "loss": 1.4763,
      "theoretical_loss": 3.418643180401049,
      "tokens_seen": 2084372480
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003747272727272727,
      "loss": 1.4031,
      "theoretical_loss": 3.418634291143224,
      "tokens_seen": 2084438016
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037470707070707073,
      "loss": 1.3153,
      "theoretical_loss": 3.418625402243131,
      "tokens_seen": 2084503552
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003746868686868687,
      "loss": 1.624,
      "theoretical_loss": 3.4186165137007443,
      "tokens_seen": 2084569088
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037466666666666665,
      "loss": 1.161,
      "theoretical_loss": 3.418607625516038,
      "tokens_seen": 2084634624
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037464646464646466,
      "loss": 1.7278,
      "theoretical_loss": 3.4185987376889866,
      "tokens_seen": 2084700160
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037462626262626267,
      "loss": 1.3251,
      "theoretical_loss": 3.418589850219565,
      "tokens_seen": 2084765696
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037460606060606063,
      "loss": 1.4643,
      "theoretical_loss": 3.418580963107746,
      "tokens_seen": 2084831232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745858585858586,
      "loss": 1.6184,
      "theoretical_loss": 3.418572076353506,
      "tokens_seen": 2084896768
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037456565656565654,
      "loss": 1.1971,
      "theoretical_loss": 3.4185631899568185,
      "tokens_seen": 2084962304
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037454545454545455,
      "loss": 1.5665,
      "theoretical_loss": 3.4185543039176576,
      "tokens_seen": 2085027840
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037452525252525256,
      "loss": 1.2791,
      "theoretical_loss": 3.4185454182359982,
      "tokens_seen": 2085093376
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003745050505050505,
      "loss": 1.4199,
      "theoretical_loss": 3.418536532911814,
      "tokens_seen": 2085158912
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744848484848485,
      "loss": 1.3007,
      "theoretical_loss": 3.41852764794508,
      "tokens_seen": 2085224448
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037446464646464643,
      "loss": 1.2452,
      "theoretical_loss": 3.4185187633357703,
      "tokens_seen": 2085289984
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744444444444445,
      "loss": 1.4831,
      "theoretical_loss": 3.4185098790838593,
      "tokens_seen": 2085355520
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037442424242424246,
      "loss": 1.4045,
      "theoretical_loss": 3.418500995189322,
      "tokens_seen": 2085421056
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003744040404040404,
      "loss": 1.6719,
      "theoretical_loss": 3.4184921116521316,
      "tokens_seen": 2085486592
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037438383838383837,
      "loss": 1.3131,
      "theoretical_loss": 3.4184832284722635,
      "tokens_seen": 2085552128
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003743636363636363,
      "loss": 1.4928,
      "theoretical_loss": 3.4184743456496918,
      "tokens_seen": 2085617664
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003743434343434344,
      "loss": 1.5779,
      "theoretical_loss": 3.4184654631843907,
      "tokens_seen": 2085683200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037432323232323235,
      "loss": 1.4453,
      "theoretical_loss": 3.4184565810763345,
      "tokens_seen": 2085748736
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003743030303030303,
      "loss": 1.208,
      "theoretical_loss": 3.4184476993254984,
      "tokens_seen": 2085814272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037428282828282826,
      "loss": 1.4881,
      "theoretical_loss": 3.4184388179318557,
      "tokens_seen": 2085879808
    },
    {
      "epoch": 0.26,
      "objective/train/aligned_token_count": 0.3360215127468109,
      "objective/train/avg_token_score": 0.022043464705348015,
      "objective/train/docs_used": 1179467,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1368883848190308,
      "objective/train/original_loss": 4.677212238311768,
      "objective/train/theoretical_loss": 3.4184299368953814,
      "objective/train/tokens_used": 444804576,
      "objective/train/unlikelihood_loss": 0.06505218893289566,
      "theoretical_loss": 3.4184299368953814,
      "tokens_seen": 2085945344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003742626262626263,
      "loss": 1.338,
      "theoretical_loss": 3.4184299368953814,
      "tokens_seen": 2085945344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003742424242424243,
      "loss": 1.3168,
      "theoretical_loss": 3.4184210562160504,
      "tokens_seen": 2086010880
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037422222222222224,
      "loss": 1.5364,
      "theoretical_loss": 3.4184121758938364,
      "tokens_seen": 2086076416
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003742020202020202,
      "loss": 1.4987,
      "theoretical_loss": 3.418403295928714,
      "tokens_seen": 2086141952
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037418181818181815,
      "loss": 1.2955,
      "theoretical_loss": 3.4183944163206577,
      "tokens_seen": 2086207488
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037416161616161617,
      "loss": 1.5719,
      "theoretical_loss": 3.418385537069642,
      "tokens_seen": 2086273024
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003741414141414142,
      "loss": 1.562,
      "theoretical_loss": 3.4183766581756405,
      "tokens_seen": 2086338560
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037412121212121213,
      "loss": 1.2294,
      "theoretical_loss": 3.418367779638629,
      "tokens_seen": 2086404096
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003741010101010101,
      "loss": 1.5268,
      "theoretical_loss": 3.418358901458581,
      "tokens_seen": 2086469632
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003740808080808081,
      "loss": 1.4696,
      "theoretical_loss": 3.4183500236354716,
      "tokens_seen": 2086535168
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037406060606060606,
      "loss": 1.4381,
      "theoretical_loss": 3.4183411461692748,
      "tokens_seen": 2086600704
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037404040404040407,
      "loss": 1.3861,
      "theoretical_loss": 3.418332269059965,
      "tokens_seen": 2086666240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374020202020202,
      "loss": 1.5148,
      "theoretical_loss": 3.4183233923075163,
      "tokens_seen": 2086731776
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000374,
      "loss": 1.4125,
      "theoretical_loss": 3.418314515911904,
      "tokens_seen": 2086797312
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000373979797979798,
      "loss": 1.3945,
      "theoretical_loss": 3.418305639873102,
      "tokens_seen": 2086862848
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037395959595959595,
      "loss": 1.4482,
      "theoretical_loss": 3.4182967641910853,
      "tokens_seen": 2086928384
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037393939393939396,
      "loss": 1.4271,
      "theoretical_loss": 3.4182878888658275,
      "tokens_seen": 2086993920
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003739191919191919,
      "loss": 1.3665,
      "theoretical_loss": 3.418279013897304,
      "tokens_seen": 2087059456
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037389898989898993,
      "loss": 1.3543,
      "theoretical_loss": 3.418270139285488,
      "tokens_seen": 2087124992
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003738787878787879,
      "loss": 1.5821,
      "theoretical_loss": 3.418261265030355,
      "tokens_seen": 2087190528
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037385858585858584,
      "loss": 1.671,
      "theoretical_loss": 3.4182523911318796,
      "tokens_seen": 2087256064
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037383838383838386,
      "loss": 1.2803,
      "theoretical_loss": 3.418243517590036,
      "tokens_seen": 2087321600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003738181818181818,
      "loss": 1.4464,
      "theoretical_loss": 3.418234644404798,
      "tokens_seen": 2087387136
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003737979797979798,
      "loss": 1.624,
      "theoretical_loss": 3.4182257715761404,
      "tokens_seen": 2087452672
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003737777777777778,
      "loss": 1.2767,
      "theoretical_loss": 3.4182168991040385,
      "tokens_seen": 2087518208
    },
    {
      "epoch": 0.26,
      "objective/train/aligned_token_count": 0.3165322542190552,
      "objective/train/avg_token_score": 0.0110182398930192,
      "objective/train/docs_used": 1180869,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0291162729263306,
      "objective/train/original_loss": 4.169273853302002,
      "objective/train/theoretical_loss": 3.4182080269884656,
      "objective/train/tokens_used": 446442976,
      "objective/train/unlikelihood_loss": 0.08053578436374664,
      "theoretical_loss": 3.4182080269884656,
      "tokens_seen": 2087583744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037375757575757574,
      "loss": 1.2945,
      "theoretical_loss": 3.4182080269884656,
      "tokens_seen": 2087583744
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037373737373737375,
      "loss": 1.4622,
      "theoretical_loss": 3.4181991552293973,
      "tokens_seen": 2087649280
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037371717171717176,
      "loss": 1.4116,
      "theoretical_loss": 3.4181902838268075,
      "tokens_seen": 2087714816
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003736969696969697,
      "loss": 1.4191,
      "theoretical_loss": 3.4181814127806702,
      "tokens_seen": 2087780352
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037367676767676767,
      "loss": 1.4904,
      "theoretical_loss": 3.4181725420909608,
      "tokens_seen": 2087845888
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037365656565656563,
      "loss": 1.3255,
      "theoretical_loss": 3.4181636717576533,
      "tokens_seen": 2087911424
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037363636363636364,
      "loss": 1.4168,
      "theoretical_loss": 3.418154801780722,
      "tokens_seen": 2087976960
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037361616161616165,
      "loss": 1.6034,
      "theoretical_loss": 3.418145932160142,
      "tokens_seen": 2088042496
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003735959595959596,
      "loss": 1.3212,
      "theoretical_loss": 3.4181370628958874,
      "tokens_seen": 2088108032
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037357575757575757,
      "loss": 1.437,
      "theoretical_loss": 3.4181281939879327,
      "tokens_seen": 2088173568
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003735555555555555,
      "loss": 1.3466,
      "theoretical_loss": 3.418119325436253,
      "tokens_seen": 2088239104
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003735353535353536,
      "loss": 1.4281,
      "theoretical_loss": 3.4181104572408216,
      "tokens_seen": 2088304640
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037351515151515154,
      "loss": 1.6064,
      "theoretical_loss": 3.4181015894016142,
      "tokens_seen": 2088370176
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734949494949495,
      "loss": 1.2708,
      "theoretical_loss": 3.4180927219186046,
      "tokens_seen": 2088435712
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037347474747474746,
      "loss": 1.195,
      "theoretical_loss": 3.418083854791768,
      "tokens_seen": 2088501248
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734545454545454,
      "loss": 1.4323,
      "theoretical_loss": 3.4180749880210777,
      "tokens_seen": 2088566784
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003734343434343435,
      "loss": 1.5022,
      "theoretical_loss": 3.4180661216065094,
      "tokens_seen": 2088632320
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037341414141414144,
      "loss": 1.299,
      "theoretical_loss": 3.418057255548037,
      "tokens_seen": 2088697856
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003733939393939394,
      "loss": 1.3051,
      "theoretical_loss": 3.4180483898456355,
      "tokens_seen": 2088763392
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037337373737373735,
      "loss": 1.7103,
      "theoretical_loss": 3.418039524499279,
      "tokens_seen": 2088828928
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003733535353535353,
      "loss": 1.4726,
      "theoretical_loss": 3.418030659508942,
      "tokens_seen": 2088894464
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003733333333333334,
      "loss": 1.8611,
      "theoretical_loss": 3.4180217948745995,
      "tokens_seen": 2088960000
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037331313131313133,
      "loss": 1.5577,
      "theoretical_loss": 3.418012930596226,
      "tokens_seen": 2089025536
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732929292929293,
      "loss": 1.7129,
      "theoretical_loss": 3.418004066673795,
      "tokens_seen": 2089091072
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037327272727272724,
      "loss": 1.6059,
      "theoretical_loss": 3.4179952031072824,
      "tokens_seen": 2089156608
    },
    {
      "epoch": 0.26,
      "objective/train/aligned_token_count": 0.4984726309776306,
      "objective/train/avg_token_score": 0.008942843414843082,
      "objective/train/docs_used": 1181528,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7859277725219727,
      "objective/train/original_loss": 4.169965744018555,
      "objective/train/theoretical_loss": 3.417986339896662,
      "objective/train/tokens_used": 448081376,
      "objective/train/unlikelihood_loss": 0.05391286313533783,
      "theoretical_loss": 3.417986339896662,
      "tokens_seen": 2089222144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732525252525253,
      "loss": 1.5706,
      "theoretical_loss": 3.417986339896662,
      "tokens_seen": 2089222144
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037323232323232327,
      "loss": 1.4367,
      "theoretical_loss": 3.4179774770419087,
      "tokens_seen": 2089287680
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003732121212121212,
      "loss": 1.4038,
      "theoretical_loss": 3.417968614542997,
      "tokens_seen": 2089353216
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731919191919192,
      "loss": 1.4713,
      "theoretical_loss": 3.4179597523999012,
      "tokens_seen": 2089418752
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037317171717171714,
      "loss": 1.5628,
      "theoretical_loss": 3.4179508906125955,
      "tokens_seen": 2089484288
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731515151515152,
      "loss": 1.5289,
      "theoretical_loss": 3.417942029181056,
      "tokens_seen": 2089549824
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037313131313131316,
      "loss": 1.481,
      "theoretical_loss": 3.417933168105255,
      "tokens_seen": 2089615360
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003731111111111111,
      "loss": 1.6574,
      "theoretical_loss": 3.417924307385169,
      "tokens_seen": 2089680896
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037309090909090907,
      "loss": 1.201,
      "theoretical_loss": 3.4179154470207718,
      "tokens_seen": 2089746432
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003730707070707071,
      "loss": 1.2907,
      "theoretical_loss": 3.4179065870120375,
      "tokens_seen": 2089811968
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003730505050505051,
      "loss": 1.3189,
      "theoretical_loss": 3.417897727358942,
      "tokens_seen": 2089877504
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037303030303030305,
      "loss": 1.6222,
      "theoretical_loss": 3.4178888680614583,
      "tokens_seen": 2089943040
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000373010101010101,
      "loss": 1.4959,
      "theoretical_loss": 3.4178800091195622,
      "tokens_seen": 2090008576
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037298989898989896,
      "loss": 1.2521,
      "theoretical_loss": 3.4178711505332275,
      "tokens_seen": 2090074112
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000372969696969697,
      "loss": 1.6979,
      "theoretical_loss": 3.417862292302429,
      "tokens_seen": 2090139648
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000372949494949495,
      "loss": 1.6717,
      "theoretical_loss": 3.4178534344271414,
      "tokens_seen": 2090205184
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037292929292929294,
      "loss": 1.4221,
      "theoretical_loss": 3.4178445769073393,
      "tokens_seen": 2090270720
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003729090909090909,
      "loss": 1.424,
      "theoretical_loss": 3.417835719742998,
      "tokens_seen": 2090336256
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003728888888888889,
      "loss": 1.5732,
      "theoretical_loss": 3.41782686293409,
      "tokens_seen": 2090401792
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037286868686868687,
      "loss": 1.4647,
      "theoretical_loss": 3.417818006480592,
      "tokens_seen": 2090467328
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003728484848484849,
      "loss": 1.6327,
      "theoretical_loss": 3.4178091503824777,
      "tokens_seen": 2090532864
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037282828282828284,
      "loss": 1.4232,
      "theoretical_loss": 3.4178002946397217,
      "tokens_seen": 2090598400
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003728080808080808,
      "loss": 1.383,
      "theoretical_loss": 3.417791439252299,
      "tokens_seen": 2090663936
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003727878787878788,
      "loss": 1.3174,
      "theoretical_loss": 3.4177825842201837,
      "tokens_seen": 2090729472
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037276767676767676,
      "loss": 1.4924,
      "theoretical_loss": 3.4177737295433506,
      "tokens_seen": 2090795008
    },
    {
      "epoch": 0.26,
      "objective/train/aligned_token_count": 0.47623410820961,
      "objective/train/avg_token_score": 0.006108488887548447,
      "objective/train/docs_used": 1182667,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4925786256790161,
      "objective/train/original_loss": 3.958449125289917,
      "objective/train/theoretical_loss": 3.417764875221774,
      "objective/train/tokens_used": 449719776,
      "objective/train/unlikelihood_loss": 0.08952859044075012,
      "theoretical_loss": 3.417764875221774,
      "tokens_seen": 2090860544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037274747474747477,
      "loss": 1.5307,
      "theoretical_loss": 3.417764875221774,
      "tokens_seen": 2090860544
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037272727272727273,
      "loss": 1.456,
      "theoretical_loss": 3.4177560212554297,
      "tokens_seen": 2090926080
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037270707070707074,
      "loss": 1.3673,
      "theoretical_loss": 3.4177471676442908,
      "tokens_seen": 2090991616
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003726868686868687,
      "loss": 1.5375,
      "theoretical_loss": 3.417738314388333,
      "tokens_seen": 2091057152
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037266666666666665,
      "loss": 1.188,
      "theoretical_loss": 3.4177294614875304,
      "tokens_seen": 2091122688
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037264646464646466,
      "loss": 1.4103,
      "theoretical_loss": 3.4177206089418575,
      "tokens_seen": 2091188224
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003726262626262626,
      "loss": 1.5421,
      "theoretical_loss": 3.4177117567512894,
      "tokens_seen": 2091253760
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037260606060606063,
      "loss": 1.17,
      "theoretical_loss": 3.4177029049158003,
      "tokens_seen": 2091319296
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003725858585858586,
      "loss": 1.6131,
      "theoretical_loss": 3.4176940534353655,
      "tokens_seen": 2091384832
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037256565656565655,
      "loss": 1.4615,
      "theoretical_loss": 3.417685202309958,
      "tokens_seen": 2091450368
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037254545454545456,
      "loss": 1.4895,
      "theoretical_loss": 3.417676351539555,
      "tokens_seen": 2091515904
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037252525252525257,
      "loss": 1.3297,
      "theoretical_loss": 3.4176675011241286,
      "tokens_seen": 2091581440
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003725050505050505,
      "loss": 1.2792,
      "theoretical_loss": 3.417658651063655,
      "tokens_seen": 2091646976
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003724848484848485,
      "loss": 1.5627,
      "theoretical_loss": 3.4176498013581087,
      "tokens_seen": 2091712512
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037246464646464644,
      "loss": 1.3112,
      "theoretical_loss": 3.4176409520074635,
      "tokens_seen": 2091778048
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037244444444444445,
      "loss": 1.3046,
      "theoretical_loss": 3.417632103011695,
      "tokens_seen": 2091843584
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037242424242424246,
      "loss": 1.5259,
      "theoretical_loss": 3.417623254370777,
      "tokens_seen": 2091909120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003724040404040404,
      "loss": 1.4383,
      "theoretical_loss": 3.4176144060846854,
      "tokens_seen": 2091974656
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723838383838384,
      "loss": 1.2748,
      "theoretical_loss": 3.4176055581533933,
      "tokens_seen": 2092040192
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037236363636363633,
      "loss": 1.6042,
      "theoretical_loss": 3.417596710576876,
      "tokens_seen": 2092105728
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723434343434344,
      "loss": 1.3628,
      "theoretical_loss": 3.417587863355109,
      "tokens_seen": 2092171264
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037232323232323235,
      "loss": 1.4676,
      "theoretical_loss": 3.417579016488066,
      "tokens_seen": 2092236800
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003723030303030303,
      "loss": 1.334,
      "theoretical_loss": 3.417570169975722,
      "tokens_seen": 2092302336
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037228282828282827,
      "loss": 1.3142,
      "theoretical_loss": 3.417561323818051,
      "tokens_seen": 2092367872
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003722626262626262,
      "loss": 1.3255,
      "theoretical_loss": 3.417552478015029,
      "tokens_seen": 2092433408
    },
    {
      "epoch": 0.26,
      "objective/train/aligned_token_count": 0.4452590346336365,
      "objective/train/avg_token_score": 0.009709584526717663,
      "objective/train/docs_used": 1183291,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5259029865264893,
      "objective/train/original_loss": 4.389416694641113,
      "objective/train/theoretical_loss": 3.4175436325666295,
      "objective/train/tokens_used": 451358176,
      "objective/train/unlikelihood_loss": 0.06972787529230118,
      "theoretical_loss": 3.4175436325666295,
      "tokens_seen": 2092498944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003722424242424243,
      "loss": 1.3672,
      "theoretical_loss": 3.4175436325666295,
      "tokens_seen": 2092498944
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037222222222222225,
      "loss": 1.3865,
      "theoretical_loss": 3.4175347874728277,
      "tokens_seen": 2092564480
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003722020202020202,
      "loss": 1.6353,
      "theoretical_loss": 3.417525942733598,
      "tokens_seen": 2092630016
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037218181818181816,
      "loss": 1.3421,
      "theoretical_loss": 3.417517098348916,
      "tokens_seen": 2092695552
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037216161616161617,
      "loss": 1.3906,
      "theoretical_loss": 3.4175082543187547,
      "tokens_seen": 2092761088
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003721414141414142,
      "loss": 1.4974,
      "theoretical_loss": 3.4174994106430905,
      "tokens_seen": 2092826624
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037212121212121214,
      "loss": 1.567,
      "theoretical_loss": 3.417490567321897,
      "tokens_seen": 2092892160
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003721010101010101,
      "loss": 1.5231,
      "theoretical_loss": 3.4174817243551487,
      "tokens_seen": 2092957696
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037208080808080805,
      "loss": 1.5456,
      "theoretical_loss": 3.4174728817428215,
      "tokens_seen": 2093023232
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037206060606060606,
      "loss": 1.4819,
      "theoretical_loss": 3.417464039484889,
      "tokens_seen": 2093088768
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003720404040404041,
      "loss": 1.5072,
      "theoretical_loss": 3.4174551975813268,
      "tokens_seen": 2093154304
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037202020202020203,
      "loss": 1.5806,
      "theoretical_loss": 3.417446356032109,
      "tokens_seen": 2093219840
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000372,
      "loss": 1.5619,
      "theoretical_loss": 3.4174375148372103,
      "tokens_seen": 2093285376
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.000371979797979798,
      "loss": 1.5147,
      "theoretical_loss": 3.4174286739966053,
      "tokens_seen": 2093350912
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037195959595959596,
      "loss": 1.3901,
      "theoretical_loss": 3.4174198335102695,
      "tokens_seen": 2093416448
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037193939393939397,
      "loss": 1.3022,
      "theoretical_loss": 3.417410993378177,
      "tokens_seen": 2093481984
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003719191919191919,
      "loss": 1.6844,
      "theoretical_loss": 3.4174021536003023,
      "tokens_seen": 2093547520
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003718989898989899,
      "loss": 1.6104,
      "theoretical_loss": 3.4173933141766204,
      "tokens_seen": 2093613056
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003718787878787879,
      "loss": 1.4473,
      "theoretical_loss": 3.417384475107106,
      "tokens_seen": 2093678592
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037185858585858585,
      "loss": 1.3699,
      "theoretical_loss": 3.417375636391734,
      "tokens_seen": 2093744128
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037183838383838386,
      "loss": 1.5994,
      "theoretical_loss": 3.417366798030479,
      "tokens_seen": 2093809664
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003718181818181818,
      "loss": 1.3916,
      "theoretical_loss": 3.4173579600233155,
      "tokens_seen": 2093875200
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037179797979797983,
      "loss": 1.2843,
      "theoretical_loss": 3.4173491223702186,
      "tokens_seen": 2093940736
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003717777777777778,
      "loss": 1.5654,
      "theoretical_loss": 3.417340285071163,
      "tokens_seen": 2094006272
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037175757575757574,
      "loss": 1.6744,
      "theoretical_loss": 3.4173314481261228,
      "tokens_seen": 2094071808
    },
    {
      "epoch": 0.26,
      "objective/train/aligned_token_count": 0.4912634491920471,
      "objective/train/avg_token_score": 0.00644297618418932,
      "objective/train/docs_used": 1183994,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7411361932754517,
      "objective/train/original_loss": 4.177827835083008,
      "objective/train/theoretical_loss": 3.4173226115350737,
      "objective/train/tokens_used": 452996576,
      "objective/train/unlikelihood_loss": 0.0659017562866211,
      "theoretical_loss": 3.4173226115350737,
      "tokens_seen": 2094137344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037173737373737375,
      "loss": 1.5447,
      "theoretical_loss": 3.4173226115350737,
      "tokens_seen": 2094137344
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003717171717171717,
      "loss": 1.4195,
      "theoretical_loss": 3.41731377529799,
      "tokens_seen": 2094202880
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003716969696969697,
      "loss": 1.4795,
      "theoretical_loss": 3.417304939414846,
      "tokens_seen": 2094268416
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003716767676767677,
      "loss": 1.3937,
      "theoretical_loss": 3.4172961038856173,
      "tokens_seen": 2094333952
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037165656565656564,
      "loss": 1.2934,
      "theoretical_loss": 3.417287268710278,
      "tokens_seen": 2094399488
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037163636363636365,
      "loss": 1.4347,
      "theoretical_loss": 3.4172784338888036,
      "tokens_seen": 2094465024
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003716161616161616,
      "loss": 1.357,
      "theoretical_loss": 3.417269599421168,
      "tokens_seen": 2094530560
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003715959595959596,
      "loss": 1.3482,
      "theoretical_loss": 3.417260765307346,
      "tokens_seen": 2094596096
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037157575757575757,
      "loss": 1.4318,
      "theoretical_loss": 3.417251931547313,
      "tokens_seen": 2094661632
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037155555555555553,
      "loss": 1.3884,
      "theoretical_loss": 3.417243098141043,
      "tokens_seen": 2094727168
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037153535353535354,
      "loss": 1.3685,
      "theoretical_loss": 3.4172342650885117,
      "tokens_seen": 2094792704
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037151515151515155,
      "loss": 1.4191,
      "theoretical_loss": 3.417225432389693,
      "tokens_seen": 2094858240
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003714949494949495,
      "loss": 1.2099,
      "theoretical_loss": 3.4172166000445623,
      "tokens_seen": 2094923776
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037147474747474746,
      "loss": 1.491,
      "theoretical_loss": 3.417207768053094,
      "tokens_seen": 2094989312
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003714545454545455,
      "loss": 1.4384,
      "theoretical_loss": 3.4171989364152626,
      "tokens_seen": 2095054848
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037143434343434343,
      "loss": 1.6164,
      "theoretical_loss": 3.417190105131044,
      "tokens_seen": 2095120384
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037141414141414144,
      "loss": 1.4524,
      "theoretical_loss": 3.4171812742004115,
      "tokens_seen": 2095185920
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003713939393939394,
      "loss": 1.3992,
      "theoretical_loss": 3.417172443623341,
      "tokens_seen": 2095251456
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037137373737373736,
      "loss": 1.3847,
      "theoretical_loss": 3.417163613399807,
      "tokens_seen": 2095316992
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037135353535353537,
      "loss": 1.6524,
      "theoretical_loss": 3.417154783529784,
      "tokens_seen": 2095382528
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003713333333333334,
      "loss": 1.468,
      "theoretical_loss": 3.4171459540132467,
      "tokens_seen": 2095448064
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037131313131313134,
      "loss": 1.2954,
      "theoretical_loss": 3.4171371248501705,
      "tokens_seen": 2095513600
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0003712929292929293,
      "loss": 1.559,
      "theoretical_loss": 3.41712829604053,
      "tokens_seen": 2095579136
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037127272727272725,
      "loss": 1.4466,
      "theoretical_loss": 3.4171194675842997,
      "tokens_seen": 2095644672
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00037125252525252526,
      "loss": 1.4198,
      "theoretical_loss": 3.4171106394814545,
      "tokens_seen": 2095710208
    },
    {
      "epoch": 0.26,
      "objective/train/aligned_token_count": 0.28158602118492126,
      "objective/train/avg_token_score": 0.014217756688594818,
      "objective/train/docs_used": 1184873,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8266394734382629,
      "objective/train/original_loss": 5.028609275817871,
      "objective/train/theoretical_loss": 3.41710181173197,
      "objective/train/tokens_used": 454634976,
      "objective/train/unlikelihood_loss": 0.05122791603207588,
      "theoretical_loss": 3.41710181173197,
      "tokens_seen": 2095775744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037123232323232327,
      "loss": 1.256,
      "theoretical_loss": 3.41710181173197,
      "tokens_seen": 2095775744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037121212121212123,
      "loss": 1.3978,
      "theoretical_loss": 3.4170929843358193,
      "tokens_seen": 2095841280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711919191919192,
      "loss": 1.3955,
      "theoretical_loss": 3.4170841572929787,
      "tokens_seen": 2095906816
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037117171717171714,
      "loss": 1.3538,
      "theoretical_loss": 3.4170753306034225,
      "tokens_seen": 2095972352
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711515151515152,
      "loss": 1.2636,
      "theoretical_loss": 3.417066504267125,
      "tokens_seen": 2096037888
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037113131313131316,
      "loss": 1.4575,
      "theoretical_loss": 3.4170576782840625,
      "tokens_seen": 2096103424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003711111111111111,
      "loss": 1.4449,
      "theoretical_loss": 3.417048852654208,
      "tokens_seen": 2096168960
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003710909090909091,
      "loss": 1.2891,
      "theoretical_loss": 3.4170400273775376,
      "tokens_seen": 2096234496
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037107070707070703,
      "loss": 1.4142,
      "theoretical_loss": 3.4170312024540257,
      "tokens_seen": 2096300032
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003710505050505051,
      "loss": 1.4076,
      "theoretical_loss": 3.4170223778836473,
      "tokens_seen": 2096365568
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037103030303030306,
      "loss": 1.2815,
      "theoretical_loss": 3.4170135536663766,
      "tokens_seen": 2096431104
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000371010101010101,
      "loss": 1.3764,
      "theoretical_loss": 3.4170047298021893,
      "tokens_seen": 2096496640
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037098989898989897,
      "loss": 1.6844,
      "theoretical_loss": 3.41699590629106,
      "tokens_seen": 2096562176
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000370969696969697,
      "loss": 1.626,
      "theoretical_loss": 3.4169870831329625,
      "tokens_seen": 2096627712
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000370949494949495,
      "loss": 1.419,
      "theoretical_loss": 3.4169782603278733,
      "tokens_seen": 2096693248
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037092929292929295,
      "loss": 1.4495,
      "theoretical_loss": 3.4169694378757662,
      "tokens_seen": 2096758784
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003709090909090909,
      "loss": 1.3304,
      "theoretical_loss": 3.4169606157766164,
      "tokens_seen": 2096824320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037088888888888886,
      "loss": 1.3579,
      "theoretical_loss": 3.416951794030399,
      "tokens_seen": 2096889856
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003708686868686869,
      "loss": 1.4558,
      "theoretical_loss": 3.416942972637088,
      "tokens_seen": 2096955392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003708484848484849,
      "loss": 1.2601,
      "theoretical_loss": 3.4169341515966583,
      "tokens_seen": 2097020928
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037082828282828284,
      "loss": 1.4484,
      "theoretical_loss": 3.416925330909086,
      "tokens_seen": 2097086464
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003708080808080808,
      "loss": 1.393,
      "theoretical_loss": 3.416916510574345,
      "tokens_seen": 2097152000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707878787878788,
      "loss": 1.3275,
      "theoretical_loss": 3.41690769059241,
      "tokens_seen": 2097217536
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037076767676767677,
      "loss": 1.4141,
      "theoretical_loss": 3.416898870963257,
      "tokens_seen": 2097283072
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707474747474748,
      "loss": 1.5404,
      "theoretical_loss": 3.416890051686859,
      "tokens_seen": 2097348608
    },
    {
      "epoch": 0.27,
      "objective/train/aligned_token_count": 0.5108137726783752,
      "objective/train/avg_token_score": 0.02312566339969635,
      "objective/train/docs_used": 1185597,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7117425203323364,
      "objective/train/original_loss": 4.583292484283447,
      "objective/train/theoretical_loss": 3.4168812327631928,
      "objective/train/tokens_used": 456273376,
      "objective/train/unlikelihood_loss": 0.07427631318569183,
      "theoretical_loss": 3.4168812327631928,
      "tokens_seen": 2097414144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037072727272727273,
      "loss": 1.4591,
      "theoretical_loss": 3.4168812327631928,
      "tokens_seen": 2097414144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003707070707070707,
      "loss": 1.4745,
      "theoretical_loss": 3.416872414192232,
      "tokens_seen": 2097479680
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003706868686868687,
      "loss": 1.4241,
      "theoretical_loss": 3.416863595973952,
      "tokens_seen": 2097545216
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037066666666666666,
      "loss": 1.3937,
      "theoretical_loss": 3.4168547781083274,
      "tokens_seen": 2097610752
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037064646464646467,
      "loss": 1.6599,
      "theoretical_loss": 3.4168459605953334,
      "tokens_seen": 2097676288
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037062626262626263,
      "loss": 1.4604,
      "theoretical_loss": 3.416837143434945,
      "tokens_seen": 2097741824
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037060606060606064,
      "loss": 1.3529,
      "theoretical_loss": 3.416828326627136,
      "tokens_seen": 2097807360
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003705858585858586,
      "loss": 1.3973,
      "theoretical_loss": 3.416819510171883,
      "tokens_seen": 2097872896
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037056565656565655,
      "loss": 1.3854,
      "theoretical_loss": 3.4168106940691594,
      "tokens_seen": 2097938432
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037054545454545456,
      "loss": 1.3052,
      "theoretical_loss": 3.4168018783189407,
      "tokens_seen": 2098003968
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003705252525252525,
      "loss": 1.2121,
      "theoretical_loss": 3.4167930629212018,
      "tokens_seen": 2098069504
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037050505050505053,
      "loss": 1.3941,
      "theoretical_loss": 3.416784247875918,
      "tokens_seen": 2098135040
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003704848484848485,
      "loss": 1.5037,
      "theoretical_loss": 3.416775433183063,
      "tokens_seen": 2098200576
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037046464646464644,
      "loss": 1.4388,
      "theoretical_loss": 3.4167666188426127,
      "tokens_seen": 2098266112
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037044444444444446,
      "loss": 1.4765,
      "theoretical_loss": 3.4167578048545417,
      "tokens_seen": 2098331648
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037042424242424247,
      "loss": 1.5656,
      "theoretical_loss": 3.4167489912188254,
      "tokens_seen": 2098397184
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003704040404040404,
      "loss": 1.3592,
      "theoretical_loss": 3.4167401779354383,
      "tokens_seen": 2098462720
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003703838383838384,
      "loss": 1.6362,
      "theoretical_loss": 3.416731365004355,
      "tokens_seen": 2098528256
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037036363636363634,
      "loss": 1.5223,
      "theoretical_loss": 3.4167225524255507,
      "tokens_seen": 2098593792
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037034343434343435,
      "loss": 1.5873,
      "theoretical_loss": 3.4167137401990004,
      "tokens_seen": 2098659328
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037032323232323236,
      "loss": 1.3729,
      "theoretical_loss": 3.4167049283246786,
      "tokens_seen": 2098724864
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003703030303030303,
      "loss": 1.3783,
      "theoretical_loss": 3.4166961168025614,
      "tokens_seen": 2098790400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003702828282828283,
      "loss": 1.4138,
      "theoretical_loss": 3.4166873056326224,
      "tokens_seen": 2098855936
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037026262626262623,
      "loss": 1.4193,
      "theoretical_loss": 3.416678494814837,
      "tokens_seen": 2098921472
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003702424242424243,
      "loss": 1.5558,
      "theoretical_loss": 3.4166696843491806,
      "tokens_seen": 2098987008
    },
    {
      "epoch": 0.27,
      "objective/train/aligned_token_count": 0.3702346086502075,
      "objective/train/avg_token_score": 0.02101566642522812,
      "objective/train/docs_used": 1186813,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3314353227615356,
      "objective/train/original_loss": 4.864157199859619,
      "objective/train/theoretical_loss": 3.4166608742356273,
      "objective/train/tokens_used": 457911776,
      "objective/train/unlikelihood_loss": 0.05566110089421272,
      "theoretical_loss": 3.4166608742356273,
      "tokens_seen": 2099052544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037022222222222225,
      "loss": 1.2552,
      "theoretical_loss": 3.4166608742356273,
      "tokens_seen": 2099052544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003702020202020202,
      "loss": 1.4825,
      "theoretical_loss": 3.4166520644741523,
      "tokens_seen": 2099118080
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037018181818181817,
      "loss": 1.3166,
      "theoretical_loss": 3.4166432550647308,
      "tokens_seen": 2099183616
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003701616161616161,
      "loss": 1.3974,
      "theoretical_loss": 3.4166344460073375,
      "tokens_seen": 2099249152
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003701414141414142,
      "loss": 1.7226,
      "theoretical_loss": 3.416625637301948,
      "tokens_seen": 2099314688
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037012121212121215,
      "loss": 1.3935,
      "theoretical_loss": 3.416616828948536,
      "tokens_seen": 2099380224
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003701010101010101,
      "loss": 1.3225,
      "theoretical_loss": 3.4166080209470775,
      "tokens_seen": 2099445760
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037008080808080806,
      "loss": 1.3545,
      "theoretical_loss": 3.416599213297547,
      "tokens_seen": 2099511296
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000370060606060606,
      "loss": 1.1832,
      "theoretical_loss": 3.4165904059999193,
      "tokens_seen": 2099576832
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003700404040404041,
      "loss": 1.3616,
      "theoretical_loss": 3.41658159905417,
      "tokens_seen": 2099642368
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037002020202020204,
      "loss": 1.326,
      "theoretical_loss": 3.416572792460274,
      "tokens_seen": 2099707904
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00037,
      "loss": 1.2092,
      "theoretical_loss": 3.4165639862182053,
      "tokens_seen": 2099773440
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036997979797979795,
      "loss": 1.2938,
      "theoretical_loss": 3.4165551803279395,
      "tokens_seen": 2099838976
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036995959595959596,
      "loss": 1.3802,
      "theoretical_loss": 3.416546374789452,
      "tokens_seen": 2099904512
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000369939393939394,
      "loss": 1.4445,
      "theoretical_loss": 3.416537569602717,
      "tokens_seen": 2099970048
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036991919191919193,
      "loss": 1.4575,
      "theoretical_loss": 3.41652876476771,
      "tokens_seen": 2100035584
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003698989898989899,
      "loss": 1.2907,
      "theoretical_loss": 3.4165199602844054,
      "tokens_seen": 2100101120
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036987878787878784,
      "loss": 1.5485,
      "theoretical_loss": 3.416511156152779,
      "tokens_seen": 2100166656
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036985858585858586,
      "loss": 1.3936,
      "theoretical_loss": 3.416502352372805,
      "tokens_seen": 2100232192
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036983838383838387,
      "loss": 1.5031,
      "theoretical_loss": 3.416493548944459,
      "tokens_seen": 2100297728
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003698181818181818,
      "loss": 1.4496,
      "theoretical_loss": 3.4164847458677157,
      "tokens_seen": 2100363264
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697979797979798,
      "loss": 1.4239,
      "theoretical_loss": 3.4164759431425495,
      "tokens_seen": 2100428800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697777777777778,
      "loss": 1.5725,
      "theoretical_loss": 3.4164671407689364,
      "tokens_seen": 2100494336
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697575757575758,
      "loss": 1.5798,
      "theoretical_loss": 3.4164583387468515,
      "tokens_seen": 2100559872
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036973737373737376,
      "loss": 1.3557,
      "theoretical_loss": 3.4164495370762684,
      "tokens_seen": 2100625408
    },
    {
      "epoch": 0.27,
      "objective/train/aligned_token_count": 0.3393206298351288,
      "objective/train/avg_token_score": 0.0204568300396204,
      "objective/train/docs_used": 1187470,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.139434814453125,
      "objective/train/original_loss": 4.641504287719727,
      "objective/train/theoretical_loss": 3.4164407357571633,
      "objective/train/tokens_used": 459550176,
      "objective/train/unlikelihood_loss": 0.058319222182035446,
      "theoretical_loss": 3.4164407357571633,
      "tokens_seen": 2100690944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003697171717171717,
      "loss": 1.3129,
      "theoretical_loss": 3.4164407357571633,
      "tokens_seen": 2100690944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036969696969696967,
      "loss": 1.4859,
      "theoretical_loss": 3.416431934789511,
      "tokens_seen": 2100756480
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003696767676767677,
      "loss": 1.4623,
      "theoretical_loss": 3.416423134173286,
      "tokens_seen": 2100822016
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003696565656565657,
      "loss": 1.4411,
      "theoretical_loss": 3.416414333908464,
      "tokens_seen": 2100887552
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036963636363636365,
      "loss": 1.2275,
      "theoretical_loss": 3.4164055339950195,
      "tokens_seen": 2100953088
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003696161616161616,
      "loss": 1.452,
      "theoretical_loss": 3.4163967344329276,
      "tokens_seen": 2101018624
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003695959595959596,
      "loss": 1.5592,
      "theoretical_loss": 3.416387935222164,
      "tokens_seen": 2101084160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003695757575757576,
      "loss": 1.3659,
      "theoretical_loss": 3.4163791363627025,
      "tokens_seen": 2101149696
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003695555555555556,
      "loss": 1.5916,
      "theoretical_loss": 3.4163703378545187,
      "tokens_seen": 2101215232
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036953535353535354,
      "loss": 1.2753,
      "theoretical_loss": 3.4163615396975877,
      "tokens_seen": 2101280768
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003695151515151515,
      "loss": 1.4703,
      "theoretical_loss": 3.4163527418918846,
      "tokens_seen": 2101346304
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003694949494949495,
      "loss": 1.3113,
      "theoretical_loss": 3.4163439444373846,
      "tokens_seen": 2101411840
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036947474747474747,
      "loss": 1.4597,
      "theoretical_loss": 3.416335147334062,
      "tokens_seen": 2101477376
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003694545454545455,
      "loss": 1.4478,
      "theoretical_loss": 3.416326350581892,
      "tokens_seen": 2101542912
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036943434343434344,
      "loss": 1.4999,
      "theoretical_loss": 3.416317554180851,
      "tokens_seen": 2101608448
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036941414141414145,
      "loss": 1.5301,
      "theoretical_loss": 3.4163087581309117,
      "tokens_seen": 2101673984
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003693939393939394,
      "loss": 1.404,
      "theoretical_loss": 3.416299962432051,
      "tokens_seen": 2101739520
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036937373737373736,
      "loss": 1.3892,
      "theoretical_loss": 3.4162911670842435,
      "tokens_seen": 2101805056
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003693535353535354,
      "loss": 1.569,
      "theoretical_loss": 3.4162823720874633,
      "tokens_seen": 2101870592
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036933333333333333,
      "loss": 1.7534,
      "theoretical_loss": 3.416273577441687,
      "tokens_seen": 2101936128
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036931313131313134,
      "loss": 1.5531,
      "theoretical_loss": 3.416264783146888,
      "tokens_seen": 2102001664
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003692929292929293,
      "loss": 1.6067,
      "theoretical_loss": 3.4162559892030426,
      "tokens_seen": 2102067200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036927272727272725,
      "loss": 1.3843,
      "theoretical_loss": 3.4162471956101257,
      "tokens_seen": 2102132736
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036925252525252527,
      "loss": 1.2177,
      "theoretical_loss": 3.416238402368112,
      "tokens_seen": 2102198272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003692323232323233,
      "loss": 1.6496,
      "theoretical_loss": 3.4162296094769764,
      "tokens_seen": 2102263808
    },
    {
      "epoch": 0.27,
      "objective/train/aligned_token_count": 0.3601539731025696,
      "objective/train/avg_token_score": 0.05041568726301193,
      "objective/train/docs_used": 1188148,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1241708993911743,
      "objective/train/original_loss": 4.18945837020874,
      "objective/train/theoretical_loss": 3.4162208169366943,
      "objective/train/tokens_used": 461188576,
      "objective/train/unlikelihood_loss": 0.07274238765239716,
      "theoretical_loss": 3.4162208169366943,
      "tokens_seen": 2102329344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036921212121212123,
      "loss": 1.5808,
      "theoretical_loss": 3.4162208169366943,
      "tokens_seen": 2102329344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003691919191919192,
      "loss": 1.5366,
      "theoretical_loss": 3.416212024747241,
      "tokens_seen": 2102394880
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036917171717171715,
      "loss": 1.5601,
      "theoretical_loss": 3.416203232908591,
      "tokens_seen": 2102460416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036915151515151516,
      "loss": 1.3181,
      "theoretical_loss": 3.4161944414207195,
      "tokens_seen": 2102525952
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036913131313131317,
      "loss": 1.5095,
      "theoretical_loss": 3.4161856502836017,
      "tokens_seen": 2102591488
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003691111111111111,
      "loss": 1.4926,
      "theoretical_loss": 3.416176859497213,
      "tokens_seen": 2102657024
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003690909090909091,
      "loss": 1.2238,
      "theoretical_loss": 3.416168069061528,
      "tokens_seen": 2102722560
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036907070707070704,
      "loss": 1.233,
      "theoretical_loss": 3.416159278976522,
      "tokens_seen": 2102788096
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003690505050505051,
      "loss": 1.4327,
      "theoretical_loss": 3.4161504892421695,
      "tokens_seen": 2102853632
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036903030303030306,
      "loss": 1.3397,
      "theoretical_loss": 3.4161416998584464,
      "tokens_seen": 2102919168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000369010101010101,
      "loss": 1.5452,
      "theoretical_loss": 3.4161329108253278,
      "tokens_seen": 2102984704
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000368989898989899,
      "loss": 1.1889,
      "theoretical_loss": 3.4161241221427883,
      "tokens_seen": 2103050240
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036896969696969693,
      "loss": 1.3946,
      "theoretical_loss": 3.416115333810803,
      "tokens_seen": 2103115776
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000368949494949495,
      "loss": 1.517,
      "theoretical_loss": 3.416106545829347,
      "tokens_seen": 2103181312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036892929292929295,
      "loss": 1.5197,
      "theoretical_loss": 3.4160977581983962,
      "tokens_seen": 2103246848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003689090909090909,
      "loss": 1.2718,
      "theoretical_loss": 3.4160889709179245,
      "tokens_seen": 2103312384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036888888888888887,
      "loss": 1.408,
      "theoretical_loss": 3.4160801839879076,
      "tokens_seen": 2103377920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688686868686869,
      "loss": 1.4539,
      "theoretical_loss": 3.4160713974083206,
      "tokens_seen": 2103443456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688484848484849,
      "loss": 1.4745,
      "theoretical_loss": 3.4160626111791386,
      "tokens_seen": 2103508992
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036882828282828285,
      "loss": 1.2958,
      "theoretical_loss": 3.4160538253003367,
      "tokens_seen": 2103574528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003688080808080808,
      "loss": 1.1511,
      "theoretical_loss": 3.41604503977189,
      "tokens_seen": 2103640064
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036878787878787876,
      "loss": 1.48,
      "theoretical_loss": 3.4160362545937732,
      "tokens_seen": 2103705600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036876767676767677,
      "loss": 1.3437,
      "theoretical_loss": 3.4160274697659623,
      "tokens_seen": 2103771136
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003687474747474748,
      "loss": 1.6896,
      "theoretical_loss": 3.4160186852884316,
      "tokens_seen": 2103836672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036872727272727274,
      "loss": 1.2146,
      "theoretical_loss": 3.416009901161157,
      "tokens_seen": 2103902208
    },
    {
      "epoch": 0.27,
      "objective/train/aligned_token_count": 0.4290078282356262,
      "objective/train/avg_token_score": 0.015303338877856731,
      "objective/train/docs_used": 1189669,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3741037845611572,
      "objective/train/original_loss": 4.558312892913818,
      "objective/train/theoretical_loss": 3.4160011173841127,
      "objective/train/tokens_used": 462826976,
      "objective/train/unlikelihood_loss": 0.0731637179851532,
      "theoretical_loss": 3.4160011173841127,
      "tokens_seen": 2103967744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003687070707070707,
      "loss": 1.5371,
      "theoretical_loss": 3.4160011173841127,
      "tokens_seen": 2103967744
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686868686868687,
      "loss": 1.2551,
      "theoretical_loss": 3.4159923339572744,
      "tokens_seen": 2104033280
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036866666666666667,
      "loss": 1.5606,
      "theoretical_loss": 3.4159835508806173,
      "tokens_seen": 2104098816
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686464646464647,
      "loss": 1.3252,
      "theoretical_loss": 3.4159747681541166,
      "tokens_seen": 2104164352
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036862626262626263,
      "loss": 1.4486,
      "theoretical_loss": 3.4159659857777465,
      "tokens_seen": 2104229888
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003686060606060606,
      "loss": 1.4635,
      "theoretical_loss": 3.415957203751484,
      "tokens_seen": 2104295424
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003685858585858586,
      "loss": 1.5143,
      "theoretical_loss": 3.415948422075302,
      "tokens_seen": 2104360960
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036856565656565656,
      "loss": 1.3146,
      "theoretical_loss": 3.415939640749177,
      "tokens_seen": 2104426496
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036854545454545457,
      "loss": 1.4438,
      "theoretical_loss": 3.415930859773084,
      "tokens_seen": 2104492032
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003685252525252525,
      "loss": 1.4399,
      "theoretical_loss": 3.415922079146998,
      "tokens_seen": 2104557568
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036850505050505054,
      "loss": 1.4223,
      "theoretical_loss": 3.415913298870894,
      "tokens_seen": 2104623104
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003684848484848485,
      "loss": 1.4879,
      "theoretical_loss": 3.4159045189447474,
      "tokens_seen": 2104688640
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036846464646464645,
      "loss": 1.4589,
      "theoretical_loss": 3.4158957393685334,
      "tokens_seen": 2104754176
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036844444444444446,
      "loss": 1.4606,
      "theoretical_loss": 3.415886960142227,
      "tokens_seen": 2104819712
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003684242424242424,
      "loss": 1.427,
      "theoretical_loss": 3.4158781812658034,
      "tokens_seen": 2104885248
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036840404040404043,
      "loss": 1.4246,
      "theoretical_loss": 3.415869402739238,
      "tokens_seen": 2104950784
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003683838383838384,
      "loss": 1.3288,
      "theoretical_loss": 3.415860624562505,
      "tokens_seen": 2105016320
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036836363636363634,
      "loss": 1.444,
      "theoretical_loss": 3.415851846735581,
      "tokens_seen": 2105081856
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036834343434343435,
      "loss": 1.275,
      "theoretical_loss": 3.4158430692584405,
      "tokens_seen": 2105147392
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003683232323232323,
      "loss": 1.4014,
      "theoretical_loss": 3.4158342921310583,
      "tokens_seen": 2105212928
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003683030303030303,
      "loss": 1.5443,
      "theoretical_loss": 3.4158255153534096,
      "tokens_seen": 2105278464
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003682828282828283,
      "loss": 1.4712,
      "theoretical_loss": 3.4158167389254706,
      "tokens_seen": 2105344000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036826262626262624,
      "loss": 1.5983,
      "theoretical_loss": 3.4158079628472153,
      "tokens_seen": 2105409536
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036824242424242425,
      "loss": 1.37,
      "theoretical_loss": 3.415799187118619,
      "tokens_seen": 2105475072
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036822222222222226,
      "loss": 1.2671,
      "theoretical_loss": 3.415790411739658,
      "tokens_seen": 2105540608
    },
    {
      "epoch": 0.27,
      "objective/train/aligned_token_count": 0.5804007649421692,
      "objective/train/avg_token_score": 0.0030669576954096556,
      "objective/train/docs_used": 1190193,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9711037874221802,
      "objective/train/original_loss": 3.844700574874878,
      "objective/train/theoretical_loss": 3.4157816367103067,
      "objective/train/tokens_used": 464465376,
      "objective/train/unlikelihood_loss": 0.07488252967596054,
      "theoretical_loss": 3.4157816367103067,
      "tokens_seen": 2105606144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003682020202020202,
      "loss": 1.6455,
      "theoretical_loss": 3.4157816367103067,
      "tokens_seen": 2105606144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036818181818181817,
      "loss": 1.5489,
      "theoretical_loss": 3.41577286203054,
      "tokens_seen": 2105671680
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036816161616161613,
      "loss": 1.4038,
      "theoretical_loss": 3.4157640877003335,
      "tokens_seen": 2105737216
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036814141414141414,
      "loss": 1.439,
      "theoretical_loss": 3.415755313719662,
      "tokens_seen": 2105802752
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036812121212121215,
      "loss": 1.5887,
      "theoretical_loss": 3.4157465400885014,
      "tokens_seen": 2105868288
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003681010101010101,
      "loss": 1.3409,
      "theoretical_loss": 3.415737766806826,
      "tokens_seen": 2105933824
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036808080808080806,
      "loss": 1.1507,
      "theoretical_loss": 3.4157289938746125,
      "tokens_seen": 2105999360
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003680606060606061,
      "loss": 1.3687,
      "theoretical_loss": 3.415720221291834,
      "tokens_seen": 2106064896
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003680404040404041,
      "loss": 1.5382,
      "theoretical_loss": 3.4157114490584677,
      "tokens_seen": 2106130432
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036802020202020204,
      "loss": 1.5209,
      "theoretical_loss": 3.4157026771744876,
      "tokens_seen": 2106195968
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000368,
      "loss": 1.3776,
      "theoretical_loss": 3.415693905639869,
      "tokens_seen": 2106261504
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036797979797979796,
      "loss": 1.3829,
      "theoretical_loss": 3.4156851344545878,
      "tokens_seen": 2106327040
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036795959595959597,
      "loss": 1.6695,
      "theoretical_loss": 3.4156763636186183,
      "tokens_seen": 2106392576
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000367939393939394,
      "loss": 1.5646,
      "theoretical_loss": 3.4156675931319365,
      "tokens_seen": 2106458112
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036791919191919194,
      "loss": 1.5348,
      "theoretical_loss": 3.4156588229945175,
      "tokens_seen": 2106523648
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003678989898989899,
      "loss": 1.4938,
      "theoretical_loss": 3.415650053206336,
      "tokens_seen": 2106589184
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036787878787878785,
      "loss": 1.3349,
      "theoretical_loss": 3.4156412837673678,
      "tokens_seen": 2106654720
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003678585858585859,
      "loss": 1.4396,
      "theoretical_loss": 3.415632514677588,
      "tokens_seen": 2106720256
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036783838383838387,
      "loss": 1.4398,
      "theoretical_loss": 3.415623745936972,
      "tokens_seen": 2106785792
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036781818181818183,
      "loss": 1.6979,
      "theoretical_loss": 3.415614977545494,
      "tokens_seen": 2106851328
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003677979797979798,
      "loss": 1.5001,
      "theoretical_loss": 3.4156062095031308,
      "tokens_seen": 2106916864
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036777777777777774,
      "loss": 1.6035,
      "theoretical_loss": 3.4155974418098562,
      "tokens_seen": 2106982400
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003677575757575758,
      "loss": 1.5589,
      "theoretical_loss": 3.4155886744656465,
      "tokens_seen": 2107047936
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036773737373737376,
      "loss": 1.7005,
      "theoretical_loss": 3.4155799074704762,
      "tokens_seen": 2107113472
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003677171717171717,
      "loss": 1.3172,
      "theoretical_loss": 3.415571140824321,
      "tokens_seen": 2107179008
    },
    {
      "epoch": 0.27,
      "objective/train/aligned_token_count": 0.4214320480823517,
      "objective/train/avg_token_score": 0.011986571364104748,
      "objective/train/docs_used": 1191719,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.412983775138855,
      "objective/train/original_loss": 4.426974296569824,
      "objective/train/theoretical_loss": 3.4155623745271564,
      "objective/train/tokens_used": 466103776,
      "objective/train/unlikelihood_loss": 0.0641813576221466,
      "theoretical_loss": 3.4155623745271564,
      "tokens_seen": 2107244544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676969696969697,
      "loss": 1.4396,
      "theoretical_loss": 3.4155623745271564,
      "tokens_seen": 2107244544
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676767676767677,
      "loss": 1.4669,
      "theoretical_loss": 3.415553608578957,
      "tokens_seen": 2107310080
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676565656565657,
      "loss": 1.3826,
      "theoretical_loss": 3.4155448429796986,
      "tokens_seen": 2107375616
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036763636363636366,
      "loss": 1.3475,
      "theoretical_loss": 3.415536077729356,
      "tokens_seen": 2107441152
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003676161616161616,
      "loss": 1.572,
      "theoretical_loss": 3.415527312827905,
      "tokens_seen": 2107506688
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036759595959595957,
      "loss": 1.3952,
      "theoretical_loss": 3.4155185482753203,
      "tokens_seen": 2107572224
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675757575757576,
      "loss": 1.3415,
      "theoretical_loss": 3.4155097840715776,
      "tokens_seen": 2107637760
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675555555555556,
      "loss": 1.5249,
      "theoretical_loss": 3.415501020216652,
      "tokens_seen": 2107703296
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036753535353535355,
      "loss": 1.6025,
      "theoretical_loss": 3.415492256710518,
      "tokens_seen": 2107768832
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003675151515151515,
      "loss": 1.657,
      "theoretical_loss": 3.4154834935531526,
      "tokens_seen": 2107834368
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674949494949495,
      "loss": 1.297,
      "theoretical_loss": 3.41547473074453,
      "tokens_seen": 2107899904
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674747474747475,
      "loss": 1.1955,
      "theoretical_loss": 3.415465968284625,
      "tokens_seen": 2107965440
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674545454545455,
      "loss": 1.6354,
      "theoretical_loss": 3.4154572061734143,
      "tokens_seen": 2108030976
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036743434343434344,
      "loss": 1.145,
      "theoretical_loss": 3.4154484444108717,
      "tokens_seen": 2108096512
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003674141414141414,
      "loss": 1.2601,
      "theoretical_loss": 3.4154396829969733,
      "tokens_seen": 2108162048
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003673939393939394,
      "loss": 1.305,
      "theoretical_loss": 3.4154309219316943,
      "tokens_seen": 2108227584
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036737373737373737,
      "loss": 1.4799,
      "theoretical_loss": 3.4154221612150097,
      "tokens_seen": 2108293120
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003673535353535354,
      "loss": 1.1589,
      "theoretical_loss": 3.4154134008468953,
      "tokens_seen": 2108358656
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036733333333333334,
      "loss": 1.3173,
      "theoretical_loss": 3.415404640827326,
      "tokens_seen": 2108424192
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036731313131313135,
      "loss": 1.5452,
      "theoretical_loss": 3.4153958811562775,
      "tokens_seen": 2108489728
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003672929292929293,
      "loss": 1.4035,
      "theoretical_loss": 3.4153871218337244,
      "tokens_seen": 2108555264
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036727272727272726,
      "loss": 1.4753,
      "theoretical_loss": 3.4153783628596424,
      "tokens_seen": 2108620800
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036725252525252527,
      "loss": 1.4524,
      "theoretical_loss": 3.415369604234007,
      "tokens_seen": 2108686336
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036723232323232323,
      "loss": 1.425,
      "theoretical_loss": 3.4153608459567932,
      "tokens_seen": 2108751872
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036721212121212124,
      "loss": 1.5262,
      "theoretical_loss": 3.415352088027977,
      "tokens_seen": 2108817408
    },
    {
      "epoch": 0.27,
      "objective/train/aligned_token_count": 0.40145406126976013,
      "objective/train/avg_token_score": 0.04017619788646698,
      "objective/train/docs_used": 1192553,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3082767724990845,
      "objective/train/original_loss": 4.763609409332275,
      "objective/train/theoretical_loss": 3.4153433304475325,
      "objective/train/tokens_used": 467742176,
      "objective/train/unlikelihood_loss": 0.044517241418361664,
      "theoretical_loss": 3.4153433304475325,
      "tokens_seen": 2108882944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003671919191919192,
      "loss": 1.3696,
      "theoretical_loss": 3.4153433304475325,
      "tokens_seen": 2108882944
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036717171717171715,
      "loss": 1.2993,
      "theoretical_loss": 3.415334573215436,
      "tokens_seen": 2108948480
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036715151515151516,
      "loss": 1.4364,
      "theoretical_loss": 3.4153258163316624,
      "tokens_seen": 2109014016
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003671313131313132,
      "loss": 1.5369,
      "theoretical_loss": 3.4153170597961875,
      "tokens_seen": 2109079552
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036711111111111113,
      "loss": 1.3395,
      "theoretical_loss": 3.4153083036089855,
      "tokens_seen": 2109145088
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003670909090909091,
      "loss": 1.2902,
      "theoretical_loss": 3.415299547770033,
      "tokens_seen": 2109210624
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036707070707070705,
      "loss": 1.3252,
      "theoretical_loss": 3.415290792279305,
      "tokens_seen": 2109276160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036705050505050506,
      "loss": 1.2963,
      "theoretical_loss": 3.415282037136776,
      "tokens_seen": 2109341696
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036703030303030307,
      "loss": 1.3236,
      "theoretical_loss": 3.4152732823424223,
      "tokens_seen": 2109407232
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000367010101010101,
      "loss": 1.5717,
      "theoretical_loss": 3.415264527896219,
      "tokens_seen": 2109472768
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000366989898989899,
      "loss": 1.4454,
      "theoretical_loss": 3.4152557737981413,
      "tokens_seen": 2109538304
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036696969696969694,
      "loss": 1.4877,
      "theoretical_loss": 3.4152470200481644,
      "tokens_seen": 2109603840
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000366949494949495,
      "loss": 1.5366,
      "theoretical_loss": 3.4152382666462637,
      "tokens_seen": 2109669376
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036692929292929296,
      "loss": 1.2176,
      "theoretical_loss": 3.4152295135924153,
      "tokens_seen": 2109734912
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003669090909090909,
      "loss": 1.1481,
      "theoretical_loss": 3.4152207608865934,
      "tokens_seen": 2109800448
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003668888888888889,
      "loss": 1.5397,
      "theoretical_loss": 3.4152120085287745,
      "tokens_seen": 2109865984
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036686868686868683,
      "loss": 1.5205,
      "theoretical_loss": 3.4152032565189323,
      "tokens_seen": 2109931520
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003668484848484849,
      "loss": 1.7132,
      "theoretical_loss": 3.415194504857044,
      "tokens_seen": 2109997056
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036682828282828285,
      "loss": 1.2513,
      "theoretical_loss": 3.415185753543084,
      "tokens_seen": 2110062592
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003668080808080808,
      "loss": 1.2136,
      "theoretical_loss": 3.415177002577028,
      "tokens_seen": 2110128128
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036678787878787877,
      "loss": 1.4298,
      "theoretical_loss": 3.4151682519588507,
      "tokens_seen": 2110193664
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003667676767676767,
      "loss": 1.468,
      "theoretical_loss": 3.415159501688528,
      "tokens_seen": 2110259200
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003667474747474748,
      "loss": 1.5529,
      "theoretical_loss": 3.4151507517660358,
      "tokens_seen": 2110324736
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036672727272727275,
      "loss": 1.1773,
      "theoretical_loss": 3.4151420021913483,
      "tokens_seen": 2110390272
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003667070707070707,
      "loss": 1.4371,
      "theoretical_loss": 3.415133252964442,
      "tokens_seen": 2110455808
    },
    {
      "epoch": 0.27,
      "objective/train/aligned_token_count": 0.3256353735923767,
      "objective/train/avg_token_score": 0.0284919124096632,
      "objective/train/docs_used": 1194000,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0299172401428223,
      "objective/train/original_loss": 4.219465255737305,
      "objective/train/theoretical_loss": 3.4151245040852913,
      "objective/train/tokens_used": 469380576,
      "objective/train/unlikelihood_loss": 0.08671543002128601,
      "theoretical_loss": 3.4151245040852913,
      "tokens_seen": 2110521344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036668686868686866,
      "loss": 1.2984,
      "theoretical_loss": 3.4151245040852913,
      "tokens_seen": 2110521344
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036666666666666667,
      "loss": 1.3646,
      "theoretical_loss": 3.415115755553872,
      "tokens_seen": 2110586880
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003666464646464647,
      "loss": 1.3163,
      "theoretical_loss": 3.4151070073701595,
      "tokens_seen": 2110652416
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036662626262626264,
      "loss": 1.4469,
      "theoretical_loss": 3.4150982595341297,
      "tokens_seen": 2110717952
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003666060606060606,
      "loss": 1.5699,
      "theoretical_loss": 3.415089512045757,
      "tokens_seen": 2110783488
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036658585858585855,
      "loss": 1.5055,
      "theoretical_loss": 3.415080764905017,
      "tokens_seen": 2110849024
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036656565656565656,
      "loss": 1.4742,
      "theoretical_loss": 3.415072018111886,
      "tokens_seen": 2110914560
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003665454545454546,
      "loss": 1.5452,
      "theoretical_loss": 3.4150632716663387,
      "tokens_seen": 2110980096
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036652525252525253,
      "loss": 1.668,
      "theoretical_loss": 3.41505452556835,
      "tokens_seen": 2111045632
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003665050505050505,
      "loss": 1.3006,
      "theoretical_loss": 3.4150457798178966,
      "tokens_seen": 2111111168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003664848484848485,
      "loss": 1.4793,
      "theoretical_loss": 3.4150370344149525,
      "tokens_seen": 2111176704
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036646464646464646,
      "loss": 1.1601,
      "theoretical_loss": 3.415028289359494,
      "tokens_seen": 2111242240
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036644444444444447,
      "loss": 1.4038,
      "theoretical_loss": 3.4150195446514964,
      "tokens_seen": 2111307776
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003664242424242424,
      "loss": 1.2005,
      "theoretical_loss": 3.415010800290935,
      "tokens_seen": 2111373312
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003664040404040404,
      "loss": 1.4745,
      "theoretical_loss": 3.4150020562777854,
      "tokens_seen": 2111438848
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003663838383838384,
      "loss": 1.3027,
      "theoretical_loss": 3.4149933126120224,
      "tokens_seen": 2111504384
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003663636363636364,
      "loss": 1.3869,
      "theoretical_loss": 3.4149845692936216,
      "tokens_seen": 2111569920
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036634343434343436,
      "loss": 1.3708,
      "theoretical_loss": 3.414975826322559,
      "tokens_seen": 2111635456
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003663232323232323,
      "loss": 1.4073,
      "theoretical_loss": 3.4149670836988095,
      "tokens_seen": 2111700992
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036630303030303033,
      "loss": 1.3876,
      "theoretical_loss": 3.4149583414223486,
      "tokens_seen": 2111766528
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662828282828283,
      "loss": 1.3682,
      "theoretical_loss": 3.414949599493152,
      "tokens_seen": 2111832064
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662626262626263,
      "loss": 1.1863,
      "theoretical_loss": 3.414940857911195,
      "tokens_seen": 2111897600
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00036624242424242425,
      "loss": 1.6353,
      "theoretical_loss": 3.4149321166764532,
      "tokens_seen": 2111963136
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662222222222222,
      "loss": 1.2615,
      "theoretical_loss": 3.4149233757889013,
      "tokens_seen": 2112028672
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0003662020202020202,
      "loss": 1.4495,
      "theoretical_loss": 3.4149146352485156,
      "tokens_seen": 2112094208
    },
    {
      "epoch": 0.27,
      "objective/train/aligned_token_count": 0.43310117721557617,
      "objective/train/avg_token_score": 0.03492316976189613,
      "objective/train/docs_used": 1194755,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5198599100112915,
      "objective/train/original_loss": 4.662668704986572,
      "objective/train/theoretical_loss": 3.414905895055271,
      "objective/train/tokens_used": 471018976,
      "objective/train/unlikelihood_loss": 0.052197135984897614,
      "theoretical_loss": 3.414905895055271,
      "tokens_seen": 2112159744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003661818181818182,
      "loss": 1.4197,
      "theoretical_loss": 3.414905895055271,
      "tokens_seen": 2112159744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003661616161616162,
      "loss": 1.3304,
      "theoretical_loss": 3.414897155209143,
      "tokens_seen": 2112225280
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036614141414141415,
      "loss": 1.4019,
      "theoretical_loss": 3.4148884157101076,
      "tokens_seen": 2112290816
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036612121212121216,
      "loss": 1.6798,
      "theoretical_loss": 3.4148796765581397,
      "tokens_seen": 2112356352
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003661010101010101,
      "loss": 1.1708,
      "theoretical_loss": 3.414870937753215,
      "tokens_seen": 2112421888
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036608080808080807,
      "loss": 1.1303,
      "theoretical_loss": 3.4148621992953085,
      "tokens_seen": 2112487424
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003660606060606061,
      "loss": 1.4747,
      "theoretical_loss": 3.4148534611843964,
      "tokens_seen": 2112552960
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036604040404040404,
      "loss": 1.4416,
      "theoretical_loss": 3.414844723420453,
      "tokens_seen": 2112618496
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036602020202020205,
      "loss": 1.5186,
      "theoretical_loss": 3.414835986003455,
      "tokens_seen": 2112684032
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000366,
      "loss": 1.3175,
      "theoretical_loss": 3.4148272489333773,
      "tokens_seen": 2112749568
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036597979797979796,
      "loss": 1.388,
      "theoretical_loss": 3.4148185122101955,
      "tokens_seen": 2112815104
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000365959595959596,
      "loss": 1.5015,
      "theoretical_loss": 3.414809775833885,
      "tokens_seen": 2112880640
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000365939393939394,
      "loss": 1.3173,
      "theoretical_loss": 3.414801039804421,
      "tokens_seen": 2112946176
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036591919191919194,
      "loss": 1.4154,
      "theoretical_loss": 3.4147923041217796,
      "tokens_seen": 2113011712
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003658989898989899,
      "loss": 1.5216,
      "theoretical_loss": 3.4147835687859356,
      "tokens_seen": 2113077248
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036587878787878786,
      "loss": 1.3413,
      "theoretical_loss": 3.414774833796865,
      "tokens_seen": 2113142784
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036585858585858587,
      "loss": 1.3983,
      "theoretical_loss": 3.414766099154543,
      "tokens_seen": 2113208320
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003658383838383839,
      "loss": 1.3596,
      "theoretical_loss": 3.414757364858945,
      "tokens_seen": 2113273856
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036581818181818183,
      "loss": 1.4247,
      "theoretical_loss": 3.414748630910047,
      "tokens_seen": 2113339392
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657979797979798,
      "loss": 1.4082,
      "theoretical_loss": 3.414739897307824,
      "tokens_seen": 2113404928
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036577777777777775,
      "loss": 1.5664,
      "theoretical_loss": 3.4147311640522515,
      "tokens_seen": 2113470464
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003657575757575758,
      "loss": 1.3231,
      "theoretical_loss": 3.4147224311433053,
      "tokens_seen": 2113536000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036573737373737377,
      "loss": 1.3211,
      "theoretical_loss": 3.41471369858096,
      "tokens_seen": 2113601536
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036571717171717173,
      "loss": 1.4833,
      "theoretical_loss": 3.4147049663651927,
      "tokens_seen": 2113667072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003656969696969697,
      "loss": 1.4256,
      "theoretical_loss": 3.4146962344959775,
      "tokens_seen": 2113732608
    },
    {
      "epoch": 0.28,
      "objective/train/aligned_token_count": 0.4935850501060486,
      "objective/train/avg_token_score": 0.02954690158367157,
      "objective/train/docs_used": 1195575,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5347603559494019,
      "objective/train/original_loss": 4.009374141693115,
      "objective/train/theoretical_loss": 3.4146875029732904,
      "objective/train/tokens_used": 472657376,
      "objective/train/unlikelihood_loss": 0.07271783798933029,
      "theoretical_loss": 3.4146875029732904,
      "tokens_seen": 2113798144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036567676767676764,
      "loss": 1.3484,
      "theoretical_loss": 3.4146875029732904,
      "tokens_seen": 2113798144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003656565656565657,
      "loss": 1.5824,
      "theoretical_loss": 3.4146787717971074,
      "tokens_seen": 2113863680
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036563636363636366,
      "loss": 1.3774,
      "theoretical_loss": 3.414670040967403,
      "tokens_seen": 2113929216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003656161616161616,
      "loss": 1.5458,
      "theoretical_loss": 3.4146613104841532,
      "tokens_seen": 2113994752
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655959595959596,
      "loss": 1.2095,
      "theoretical_loss": 3.414652580347334,
      "tokens_seen": 2114060288
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655757575757576,
      "loss": 1.4644,
      "theoretical_loss": 3.41464385055692,
      "tokens_seen": 2114125824
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655555555555556,
      "loss": 1.4387,
      "theoretical_loss": 3.4146351211128874,
      "tokens_seen": 2114191360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036553535353535356,
      "loss": 1.2225,
      "theoretical_loss": 3.4146263920152116,
      "tokens_seen": 2114256896
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003655151515151515,
      "loss": 1.3472,
      "theoretical_loss": 3.4146176632638676,
      "tokens_seen": 2114322432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036549494949494947,
      "loss": 1.7032,
      "theoretical_loss": 3.4146089348588315,
      "tokens_seen": 2114387968
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003654747474747475,
      "loss": 1.5132,
      "theoretical_loss": 3.4146002068000785,
      "tokens_seen": 2114453504
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003654545454545455,
      "loss": 1.4697,
      "theoretical_loss": 3.4145914790875844,
      "tokens_seen": 2114519040
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036543434343434345,
      "loss": 1.3938,
      "theoretical_loss": 3.414582751721325,
      "tokens_seen": 2114584576
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003654141414141414,
      "loss": 1.4168,
      "theoretical_loss": 3.4145740247012752,
      "tokens_seen": 2114650112
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653939393939394,
      "loss": 1.341,
      "theoretical_loss": 3.414565298027411,
      "tokens_seen": 2114715648
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653737373737374,
      "loss": 1.3186,
      "theoretical_loss": 3.4145565716997073,
      "tokens_seen": 2114781184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653535353535354,
      "loss": 1.0649,
      "theoretical_loss": 3.41454784571814,
      "tokens_seen": 2114846720
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036533333333333334,
      "loss": 1.3758,
      "theoretical_loss": 3.4145391200826856,
      "tokens_seen": 2114912256
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003653131313131313,
      "loss": 1.4244,
      "theoretical_loss": 3.414530394793318,
      "tokens_seen": 2114977792
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003652929292929293,
      "loss": 1.3469,
      "theoretical_loss": 3.4145216698500134,
      "tokens_seen": 2115043328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036527272727272727,
      "loss": 1.2952,
      "theoretical_loss": 3.414512945252748,
      "tokens_seen": 2115108864
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003652525252525253,
      "loss": 1.7349,
      "theoretical_loss": 3.414504221001496,
      "tokens_seen": 2115174400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036523232323232323,
      "loss": 1.4591,
      "theoretical_loss": 3.414495497096235,
      "tokens_seen": 2115239936
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036521212121212124,
      "loss": 1.519,
      "theoretical_loss": 3.4144867735369386,
      "tokens_seen": 2115305472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003651919191919192,
      "loss": 1.4488,
      "theoretical_loss": 3.414478050323583,
      "tokens_seen": 2115371008
    },
    {
      "epoch": 0.28,
      "objective/train/aligned_token_count": 0.386424720287323,
      "objective/train/avg_token_score": 0.013556824065744877,
      "objective/train/docs_used": 1196855,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2169398069381714,
      "objective/train/original_loss": 4.470401763916016,
      "objective/train/theoretical_loss": 3.4144693274561444,
      "objective/train/tokens_used": 474295776,
      "objective/train/unlikelihood_loss": 0.04746641591191292,
      "theoretical_loss": 3.4144693274561444,
      "tokens_seen": 2115436544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036517171717171716,
      "loss": 1.2,
      "theoretical_loss": 3.4144693274561444,
      "tokens_seen": 2115436544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036515151515151517,
      "loss": 1.2002,
      "theoretical_loss": 3.414460604934597,
      "tokens_seen": 2115502080
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003651313131313131,
      "loss": 1.3827,
      "theoretical_loss": 3.414451882758918,
      "tokens_seen": 2115567616
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036511111111111114,
      "loss": 1.5522,
      "theoretical_loss": 3.414443160929082,
      "tokens_seen": 2115633152
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003650909090909091,
      "loss": 1.5936,
      "theoretical_loss": 3.4144344394450643,
      "tokens_seen": 2115698688
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036507070707070705,
      "loss": 1.4197,
      "theoretical_loss": 3.414425718306841,
      "tokens_seen": 2115764224
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036505050505050506,
      "loss": 1.4312,
      "theoretical_loss": 3.4144169975143885,
      "tokens_seen": 2115829760
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000365030303030303,
      "loss": 1.5033,
      "theoretical_loss": 3.4144082770676807,
      "tokens_seen": 2115895296
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036501010101010103,
      "loss": 1.4611,
      "theoretical_loss": 3.4143995569666936,
      "tokens_seen": 2115960832
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000364989898989899,
      "loss": 1.4207,
      "theoretical_loss": 3.4143908372114042,
      "tokens_seen": 2116026368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036496969696969694,
      "loss": 1.4198,
      "theoretical_loss": 3.4143821178017864,
      "tokens_seen": 2116091904
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036494949494949496,
      "loss": 1.0965,
      "theoretical_loss": 3.4143733987378164,
      "tokens_seen": 2116157440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036492929292929297,
      "loss": 1.189,
      "theoretical_loss": 3.41436468001947,
      "tokens_seen": 2116222976
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003649090909090909,
      "loss": 1.2537,
      "theoretical_loss": 3.4143559616467227,
      "tokens_seen": 2116288512
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003648888888888889,
      "loss": 1.4288,
      "theoretical_loss": 3.41434724361955,
      "tokens_seen": 2116354048
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036486868686868684,
      "loss": 1.2104,
      "theoretical_loss": 3.4143385259379277,
      "tokens_seen": 2116419584
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036484848484848485,
      "loss": 1.5573,
      "theoretical_loss": 3.4143298086018308,
      "tokens_seen": 2116485120
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036482828282828286,
      "loss": 1.2042,
      "theoretical_loss": 3.4143210916112356,
      "tokens_seen": 2116550656
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003648080808080808,
      "loss": 1.3208,
      "theoretical_loss": 3.4143123749661175,
      "tokens_seen": 2116616192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036478787878787877,
      "loss": 1.3485,
      "theoretical_loss": 3.414303658666452,
      "tokens_seen": 2116681728
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036476767676767673,
      "loss": 1.2952,
      "theoretical_loss": 3.414294942712215,
      "tokens_seen": 2116747264
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003647474747474748,
      "loss": 1.4681,
      "theoretical_loss": 3.414286227103381,
      "tokens_seen": 2116812800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036472727272727275,
      "loss": 1.3255,
      "theoretical_loss": 3.4142775118399276,
      "tokens_seen": 2116878336
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003647070707070707,
      "loss": 1.228,
      "theoretical_loss": 3.414268796921829,
      "tokens_seen": 2116943872
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036468686868686867,
      "loss": 1.2621,
      "theoretical_loss": 3.414260082349061,
      "tokens_seen": 2117009408
    },
    {
      "epoch": 0.28,
      "objective/train/aligned_token_count": 0.4744623601436615,
      "objective/train/avg_token_score": 0.007280266378074884,
      "objective/train/docs_used": 1197590,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5125036239624023,
      "objective/train/original_loss": 3.8015384674072266,
      "objective/train/theoretical_loss": 3.414251368121599,
      "objective/train/tokens_used": 475934176,
      "objective/train/unlikelihood_loss": 0.0856105238199234,
      "theoretical_loss": 3.414251368121599,
      "tokens_seen": 2117074944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003646666666666666,
      "loss": 1.3927,
      "theoretical_loss": 3.414251368121599,
      "tokens_seen": 2117074944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003646464646464647,
      "loss": 1.2886,
      "theoretical_loss": 3.4142426542394197,
      "tokens_seen": 2117140480
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036462626262626264,
      "loss": 1.3716,
      "theoretical_loss": 3.414233940702498,
      "tokens_seen": 2117206016
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003646060606060606,
      "loss": 1.3096,
      "theoretical_loss": 3.4142252275108094,
      "tokens_seen": 2117271552
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036458585858585856,
      "loss": 1.4137,
      "theoretical_loss": 3.4142165146643295,
      "tokens_seen": 2117337088
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003645656565656566,
      "loss": 1.5802,
      "theoretical_loss": 3.4142078021630344,
      "tokens_seen": 2117402624
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003645454545454546,
      "loss": 1.5419,
      "theoretical_loss": 3.4141990900068997,
      "tokens_seen": 2117468160
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036452525252525254,
      "loss": 1.4169,
      "theoretical_loss": 3.4141903781959004,
      "tokens_seen": 2117533696
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003645050505050505,
      "loss": 1.5674,
      "theoretical_loss": 3.4141816667300127,
      "tokens_seen": 2117599232
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036448484848484845,
      "loss": 1.2841,
      "theoretical_loss": 3.414172955609212,
      "tokens_seen": 2117664768
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003644646464646465,
      "loss": 1.3084,
      "theoretical_loss": 3.414164244833475,
      "tokens_seen": 2117730304
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036444444444444447,
      "loss": 1.271,
      "theoretical_loss": 3.4141555344027754,
      "tokens_seen": 2117795840
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036442424242424243,
      "loss": 1.4627,
      "theoretical_loss": 3.41414682431709,
      "tokens_seen": 2117861376
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003644040404040404,
      "loss": 1.4575,
      "theoretical_loss": 3.414138114576395,
      "tokens_seen": 2117926912
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643838383838384,
      "loss": 1.5042,
      "theoretical_loss": 3.414129405180665,
      "tokens_seen": 2117992448
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643636363636364,
      "loss": 1.5975,
      "theoretical_loss": 3.414120696129876,
      "tokens_seen": 2118057984
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036434343434343437,
      "loss": 1.3464,
      "theoretical_loss": 3.4141119874240036,
      "tokens_seen": 2118123520
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643232323232323,
      "loss": 1.4756,
      "theoretical_loss": 3.414103279063024,
      "tokens_seen": 2118189056
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003643030303030303,
      "loss": 1.6128,
      "theoretical_loss": 3.4140945710469124,
      "tokens_seen": 2118254592
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642828282828283,
      "loss": 1.454,
      "theoretical_loss": 3.414085863375644,
      "tokens_seen": 2118320128
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642626262626263,
      "loss": 1.4898,
      "theoretical_loss": 3.4140771560491956,
      "tokens_seen": 2118385664
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036424242424242426,
      "loss": 1.3478,
      "theoretical_loss": 3.4140684490675426,
      "tokens_seen": 2118451200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642222222222222,
      "loss": 1.5727,
      "theoretical_loss": 3.41405974243066,
      "tokens_seen": 2118516736
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003642020202020202,
      "loss": 1.5289,
      "theoretical_loss": 3.414051036138524,
      "tokens_seen": 2118582272
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641818181818182,
      "loss": 1.5049,
      "theoretical_loss": 3.41404233019111,
      "tokens_seen": 2118647808
    },
    {
      "epoch": 0.28,
      "objective/train/aligned_token_count": 0.4476417303085327,
      "objective/train/avg_token_score": 0.011445474810898304,
      "objective/train/docs_used": 1199039,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3598164319992065,
      "objective/train/original_loss": 3.9978601932525635,
      "objective/train/theoretical_loss": 3.4140336245883933,
      "objective/train/tokens_used": 477572576,
      "objective/train/unlikelihood_loss": 0.08685135096311569,
      "theoretical_loss": 3.4140336245883933,
      "tokens_seen": 2118713344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641616161616162,
      "loss": 1.2644,
      "theoretical_loss": 3.4140336245883933,
      "tokens_seen": 2118713344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036414141414141415,
      "loss": 1.7291,
      "theoretical_loss": 3.414024919330351,
      "tokens_seen": 2118778880
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641212121212121,
      "loss": 1.3314,
      "theoretical_loss": 3.4140162144169577,
      "tokens_seen": 2118844416
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003641010101010101,
      "loss": 1.3058,
      "theoretical_loss": 3.4140075098481892,
      "tokens_seen": 2118909952
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003640808080808081,
      "loss": 1.3949,
      "theoretical_loss": 3.4139988056240216,
      "tokens_seen": 2118975488
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003640606060606061,
      "loss": 1.2031,
      "theoretical_loss": 3.41399010174443,
      "tokens_seen": 2119041024
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036404040404040404,
      "loss": 1.3957,
      "theoretical_loss": 3.4139813982093905,
      "tokens_seen": 2119106560
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036402020202020205,
      "loss": 1.51,
      "theoretical_loss": 3.413972695018879,
      "tokens_seen": 2119172096
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000364,
      "loss": 1.5684,
      "theoretical_loss": 3.413963992172871,
      "tokens_seen": 2119237632
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036397979797979797,
      "loss": 1.2647,
      "theoretical_loss": 3.4139552896713417,
      "tokens_seen": 2119303168
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000363959595959596,
      "loss": 1.4358,
      "theoretical_loss": 3.4139465875142676,
      "tokens_seen": 2119368704
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036393939393939394,
      "loss": 1.6356,
      "theoretical_loss": 3.413937885701624,
      "tokens_seen": 2119434240
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036391919191919195,
      "loss": 1.4228,
      "theoretical_loss": 3.4139291842333863,
      "tokens_seen": 2119499776
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003638989898989899,
      "loss": 1.3778,
      "theoretical_loss": 3.413920483109531,
      "tokens_seen": 2119565312
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036387878787878786,
      "loss": 1.4368,
      "theoretical_loss": 3.413911782330034,
      "tokens_seen": 2119630848
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036385858585858587,
      "loss": 1.4418,
      "theoretical_loss": 3.41390308189487,
      "tokens_seen": 2119696384
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003638383838383839,
      "loss": 1.3635,
      "theoretical_loss": 3.413894381804015,
      "tokens_seen": 2119761920
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036381818181818184,
      "loss": 1.4966,
      "theoretical_loss": 3.413885682057445,
      "tokens_seen": 2119827456
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637979797979798,
      "loss": 1.5134,
      "theoretical_loss": 3.413876982655135,
      "tokens_seen": 2119892992
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036377777777777775,
      "loss": 1.1802,
      "theoretical_loss": 3.4138682835970626,
      "tokens_seen": 2119958528
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036375757575757576,
      "loss": 1.3945,
      "theoretical_loss": 3.413859584883202,
      "tokens_seen": 2120024064
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003637373737373738,
      "loss": 1.3595,
      "theoretical_loss": 3.413850886513529,
      "tokens_seen": 2120089600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036371717171717173,
      "loss": 1.4383,
      "theoretical_loss": 3.4138421884880197,
      "tokens_seen": 2120155136
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003636969696969697,
      "loss": 1.3119,
      "theoretical_loss": 3.41383349080665,
      "tokens_seen": 2120220672
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036367676767676765,
      "loss": 1.356,
      "theoretical_loss": 3.413824793469395,
      "tokens_seen": 2120286208
    },
    {
      "epoch": 0.28,
      "objective/train/aligned_token_count": 0.3866690993309021,
      "objective/train/avg_token_score": 0.043729815632104874,
      "objective/train/docs_used": 1199714,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2710665464401245,
      "objective/train/original_loss": 4.90679931640625,
      "objective/train/theoretical_loss": 3.4138160964762307,
      "objective/train/tokens_used": 479210976,
      "objective/train/unlikelihood_loss": 0.06792816519737244,
      "theoretical_loss": 3.4138160964762307,
      "tokens_seen": 2120351744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003636565656565657,
      "loss": 1.3698,
      "theoretical_loss": 3.4138160964762307,
      "tokens_seen": 2120351744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036363636363636367,
      "loss": 1.9067,
      "theoretical_loss": 3.4138073998271334,
      "tokens_seen": 2120417280
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003636161616161616,
      "loss": 1.2395,
      "theoretical_loss": 3.4137987035220787,
      "tokens_seen": 2120482816
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003635959595959596,
      "loss": 1.3265,
      "theoretical_loss": 3.4137900075610412,
      "tokens_seen": 2120548352
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036357575757575754,
      "loss": 1.2914,
      "theoretical_loss": 3.413781311943999,
      "tokens_seen": 2120613888
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003635555555555556,
      "loss": 1.4336,
      "theoretical_loss": 3.413772616670925,
      "tokens_seen": 2120679424
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036353535353535356,
      "loss": 1.5028,
      "theoretical_loss": 3.413763921741797,
      "tokens_seen": 2120744960
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003635151515151515,
      "loss": 1.5701,
      "theoretical_loss": 3.4137552271565905,
      "tokens_seen": 2120810496
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634949494949495,
      "loss": 1.4403,
      "theoretical_loss": 3.4137465329152805,
      "tokens_seen": 2120876032
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036347474747474743,
      "loss": 1.3683,
      "theoretical_loss": 3.4137378390178434,
      "tokens_seen": 2120941568
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634545454545455,
      "loss": 1.4095,
      "theoretical_loss": 3.413729145464255,
      "tokens_seen": 2121007104
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036343434343434345,
      "loss": 1.5156,
      "theoretical_loss": 3.4137204522544904,
      "tokens_seen": 2121072640
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003634141414141414,
      "loss": 1.4724,
      "theoretical_loss": 3.413711759388526,
      "tokens_seen": 2121138176
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036339393939393937,
      "loss": 1.2512,
      "theoretical_loss": 3.4137030668663373,
      "tokens_seen": 2121203712
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003633737373737374,
      "loss": 1.4421,
      "theoretical_loss": 3.4136943746879007,
      "tokens_seen": 2121269248
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003633535353535354,
      "loss": 1.3767,
      "theoretical_loss": 3.413685682853191,
      "tokens_seen": 2121334784
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036333333333333335,
      "loss": 1.1784,
      "theoretical_loss": 3.413676991362185,
      "tokens_seen": 2121400320
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003633131313131313,
      "loss": 1.2975,
      "theoretical_loss": 3.413668300214858,
      "tokens_seen": 2121465856
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036329292929292926,
      "loss": 1.3971,
      "theoretical_loss": 3.413659609411185,
      "tokens_seen": 2121531392
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036327272727272727,
      "loss": 1.5969,
      "theoretical_loss": 3.4136509189511433,
      "tokens_seen": 2121596928
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003632525252525253,
      "loss": 1.4649,
      "theoretical_loss": 3.4136422288347075,
      "tokens_seen": 2121662464
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036323232323232324,
      "loss": 1.3678,
      "theoretical_loss": 3.4136335390618537,
      "tokens_seen": 2121728000
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003632121212121212,
      "loss": 1.1506,
      "theoretical_loss": 3.4136248496325585,
      "tokens_seen": 2121793536
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003631919191919192,
      "loss": 1.4502,
      "theoretical_loss": 3.413616160546797,
      "tokens_seen": 2121859072
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036317171717171716,
      "loss": 1.3311,
      "theoretical_loss": 3.4136074718045446,
      "tokens_seen": 2121924608
    },
    {
      "epoch": 0.28,
      "objective/train/aligned_token_count": 0.31140029430389404,
      "objective/train/avg_token_score": 0.013784058392047882,
      "objective/train/docs_used": 1200493,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9349797964096069,
      "objective/train/original_loss": 4.567559242248535,
      "objective/train/theoretical_loss": 3.413598783405778,
      "objective/train/tokens_used": 480849376,
      "objective/train/unlikelihood_loss": 0.0624493733048439,
      "theoretical_loss": 3.413598783405778,
      "tokens_seen": 2121990144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003631515151515152,
      "loss": 1.4248,
      "theoretical_loss": 3.413598783405778,
      "tokens_seen": 2121990144
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036313131313131313,
      "loss": 1.2625,
      "theoretical_loss": 3.4135900953504725,
      "tokens_seen": 2122055680
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003631111111111111,
      "loss": 1.2486,
      "theoretical_loss": 3.4135814076386044,
      "tokens_seen": 2122121216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003630909090909091,
      "loss": 1.3548,
      "theoretical_loss": 3.4135727202701487,
      "tokens_seen": 2122186752
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036307070707070706,
      "loss": 1.6199,
      "theoretical_loss": 3.4135640332450814,
      "tokens_seen": 2122252288
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036305050505050507,
      "loss": 1.5875,
      "theoretical_loss": 3.4135553465633794,
      "tokens_seen": 2122317824
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000363030303030303,
      "loss": 1.5659,
      "theoretical_loss": 3.413546660225017,
      "tokens_seen": 2122383360
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036301010101010104,
      "loss": 1.4644,
      "theoretical_loss": 3.413537974229971,
      "tokens_seen": 2122448896
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000362989898989899,
      "loss": 1.3691,
      "theoretical_loss": 3.413529288578217,
      "tokens_seen": 2122514432
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036296969696969695,
      "loss": 1.4405,
      "theoretical_loss": 3.4135206032697307,
      "tokens_seen": 2122579968
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036294949494949496,
      "loss": 1.4858,
      "theoretical_loss": 3.413511918304488,
      "tokens_seen": 2122645504
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003629292929292929,
      "loss": 1.4858,
      "theoretical_loss": 3.413503233682465,
      "tokens_seen": 2122711040
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036290909090909093,
      "loss": 1.364,
      "theoretical_loss": 3.4134945494036373,
      "tokens_seen": 2122776576
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003628888888888889,
      "loss": 1.4413,
      "theoretical_loss": 3.4134858654679805,
      "tokens_seen": 2122842112
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003628686868686869,
      "loss": 1.4418,
      "theoretical_loss": 3.4134771818754714,
      "tokens_seen": 2122907648
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036284848484848485,
      "loss": 1.3524,
      "theoretical_loss": 3.4134684986260844,
      "tokens_seen": 2122973184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036282828282828286,
      "loss": 1.3216,
      "theoretical_loss": 3.413459815719796,
      "tokens_seen": 2123038720
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003628080808080808,
      "loss": 1.2786,
      "theoretical_loss": 3.413451133156583,
      "tokens_seen": 2123104256
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003627878787878788,
      "loss": 1.3786,
      "theoretical_loss": 3.41344245093642,
      "tokens_seen": 2123169792
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003627676767676768,
      "loss": 1.1093,
      "theoretical_loss": 3.4134337690592833,
      "tokens_seen": 2123235328
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036274747474747475,
      "loss": 1.4715,
      "theoretical_loss": 3.413425087525149,
      "tokens_seen": 2123300864
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036272727272727276,
      "loss": 1.5278,
      "theoretical_loss": 3.413416406333992,
      "tokens_seen": 2123366400
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003627070707070707,
      "loss": 1.3447,
      "theoretical_loss": 3.413407725485789,
      "tokens_seen": 2123431936
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036268686868686867,
      "loss": 1.3902,
      "theoretical_loss": 3.4133990449805163,
      "tokens_seen": 2123497472
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003626666666666667,
      "loss": 1.4581,
      "theoretical_loss": 3.4133903648181487,
      "tokens_seen": 2123563008
    },
    {
      "epoch": 0.28,
      "objective/train/aligned_token_count": 0.46126589179039,
      "objective/train/avg_token_score": 0.017873598262667656,
      "objective/train/docs_used": 1201963,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.436392903327942,
      "objective/train/original_loss": 4.299962043762207,
      "objective/train/theoretical_loss": 3.413381684998663,
      "objective/train/tokens_used": 482487776,
      "objective/train/unlikelihood_loss": 0.07796698063611984,
      "theoretical_loss": 3.413381684998663,
      "tokens_seen": 2123628544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003626464646464647,
      "loss": 1.3903,
      "theoretical_loss": 3.413381684998663,
      "tokens_seen": 2123628544
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036262626262626265,
      "loss": 1.5286,
      "theoretical_loss": 3.4133730055220344,
      "tokens_seen": 2123694080
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003626060606060606,
      "loss": 1.5028,
      "theoretical_loss": 3.4133643263882387,
      "tokens_seen": 2123759616
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036258585858585856,
      "loss": 1.4162,
      "theoretical_loss": 3.413355647597253,
      "tokens_seen": 2123825152
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003625656565656566,
      "loss": 1.4244,
      "theoretical_loss": 3.4133469691490514,
      "tokens_seen": 2123890688
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003625454545454546,
      "loss": 1.8553,
      "theoretical_loss": 3.413338291043611,
      "tokens_seen": 2123956224
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036252525252525254,
      "loss": 1.3402,
      "theoretical_loss": 3.4133296132809074,
      "tokens_seen": 2124021760
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003625050505050505,
      "loss": 1.3649,
      "theoretical_loss": 3.4133209358609164,
      "tokens_seen": 2124087296
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036248484848484846,
      "loss": 1.3512,
      "theoretical_loss": 3.413312258783614,
      "tokens_seen": 2124152832
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003624646464646465,
      "loss": 1.2718,
      "theoretical_loss": 3.413303582048976,
      "tokens_seen": 2124218368
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003624444444444445,
      "loss": 1.498,
      "theoretical_loss": 3.4132949056569784,
      "tokens_seen": 2124283904
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036242424242424244,
      "loss": 1.4112,
      "theoretical_loss": 3.4132862296075968,
      "tokens_seen": 2124349440
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003624040404040404,
      "loss": 1.38,
      "theoretical_loss": 3.413277553900808,
      "tokens_seen": 2124414976
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036238383838383835,
      "loss": 1.3447,
      "theoretical_loss": 3.413268878536586,
      "tokens_seen": 2124480512
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003623636363636364,
      "loss": 1.5402,
      "theoretical_loss": 3.413260203514909,
      "tokens_seen": 2124546048
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036234343434343437,
      "loss": 1.4928,
      "theoretical_loss": 3.413251528835751,
      "tokens_seen": 2124611584
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036232323232323233,
      "loss": 1.3285,
      "theoretical_loss": 3.4132428544990896,
      "tokens_seen": 2124677120
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003623030303030303,
      "loss": 1.2969,
      "theoretical_loss": 3.4132341805048996,
      "tokens_seen": 2124742656
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003622828282828283,
      "loss": 1.5159,
      "theoretical_loss": 3.4132255068531574,
      "tokens_seen": 2124808192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003622626262626263,
      "loss": 1.5201,
      "theoretical_loss": 3.413216833543838,
      "tokens_seen": 2124873728
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036224242424242426,
      "loss": 1.2455,
      "theoretical_loss": 3.413208160576919,
      "tokens_seen": 2124939264
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003622222222222222,
      "loss": 1.436,
      "theoretical_loss": 3.413199487952374,
      "tokens_seen": 2125004800
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003622020202020202,
      "loss": 1.4685,
      "theoretical_loss": 3.4131908156701813,
      "tokens_seen": 2125070336
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003621818181818182,
      "loss": 1.4574,
      "theoretical_loss": 3.4131821437303156,
      "tokens_seen": 2125135872
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003621616161616162,
      "loss": 1.4075,
      "theoretical_loss": 3.4131734721327533,
      "tokens_seen": 2125201408
    },
    {
      "epoch": 0.28,
      "objective/train/aligned_token_count": 0.5243768095970154,
      "objective/train/avg_token_score": 0.016918325796723366,
      "objective/train/docs_used": 1202675,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5344035625457764,
      "objective/train/original_loss": 4.4746246337890625,
      "objective/train/theoretical_loss": 3.41316480087747,
      "objective/train/tokens_used": 484126176,
      "objective/train/unlikelihood_loss": 0.03440330550074577,
      "theoretical_loss": 3.41316480087747,
      "tokens_seen": 2125266944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036214141414141416,
      "loss": 1.4035,
      "theoretical_loss": 3.41316480087747,
      "tokens_seen": 2125266944
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003621212121212121,
      "loss": 1.4559,
      "theoretical_loss": 3.4131561299644417,
      "tokens_seen": 2125332480
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003621010101010101,
      "loss": 1.4123,
      "theoretical_loss": 3.413147459393644,
      "tokens_seen": 2125398016
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003620808080808081,
      "loss": 1.5865,
      "theoretical_loss": 3.4131387891650533,
      "tokens_seen": 2125463552
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003620606060606061,
      "loss": 1.4184,
      "theoretical_loss": 3.4131301192786454,
      "tokens_seen": 2125529088
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036204040404040405,
      "loss": 1.3117,
      "theoretical_loss": 3.4131214497343967,
      "tokens_seen": 2125594624
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000362020202020202,
      "loss": 1.1812,
      "theoretical_loss": 3.4131127805322823,
      "tokens_seen": 2125660160
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000362,
      "loss": 1.3233,
      "theoretical_loss": 3.4131041116722787,
      "tokens_seen": 2125725696
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000361979797979798,
      "loss": 1.5804,
      "theoretical_loss": 3.413095443154362,
      "tokens_seen": 2125791232
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.000361959595959596,
      "loss": 1.3498,
      "theoretical_loss": 3.4130867749785074,
      "tokens_seen": 2125856768
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036193939393939394,
      "loss": 1.4536,
      "theoretical_loss": 3.4130781071446914,
      "tokens_seen": 2125922304
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003619191919191919,
      "loss": 1.5518,
      "theoretical_loss": 3.4130694396528902,
      "tokens_seen": 2125987840
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003618989898989899,
      "loss": 1.494,
      "theoretical_loss": 3.413060772503079,
      "tokens_seen": 2126053376
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036187878787878787,
      "loss": 1.244,
      "theoretical_loss": 3.4130521056952348,
      "tokens_seen": 2126118912
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003618585858585859,
      "loss": 1.5502,
      "theoretical_loss": 3.413043439229333,
      "tokens_seen": 2126184448
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036183838383838383,
      "loss": 1.2781,
      "theoretical_loss": 3.413034773105349,
      "tokens_seen": 2126249984
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036181818181818185,
      "loss": 1.5193,
      "theoretical_loss": 3.41302610732326,
      "tokens_seen": 2126315520
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003617979797979798,
      "loss": 1.2846,
      "theoretical_loss": 3.4130174418830403,
      "tokens_seen": 2126381056
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036177777777777776,
      "loss": 1.5768,
      "theoretical_loss": 3.4130087767846677,
      "tokens_seen": 2126446592
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036175757575757577,
      "loss": 1.714,
      "theoretical_loss": 3.4130001120281173,
      "tokens_seen": 2126512128
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036173737373737373,
      "loss": 1.4666,
      "theoretical_loss": 3.412991447613365,
      "tokens_seen": 2126577664
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036171717171717174,
      "loss": 1.5119,
      "theoretical_loss": 3.412982783540387,
      "tokens_seen": 2126643200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003616969696969697,
      "loss": 1.3709,
      "theoretical_loss": 3.412974119809159,
      "tokens_seen": 2126708736
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036167676767676765,
      "loss": 1.6,
      "theoretical_loss": 3.4129654564196574,
      "tokens_seen": 2126774272
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036165656565656566,
      "loss": 1.327,
      "theoretical_loss": 3.4129567933718583,
      "tokens_seen": 2126839808
    },
    {
      "epoch": 0.28,
      "objective/train/aligned_token_count": 0.4833821952342987,
      "objective/train/avg_token_score": 0.004767133388668299,
      "objective/train/docs_used": 1203810,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5523133277893066,
      "objective/train/original_loss": 3.932647705078125,
      "objective/train/theoretical_loss": 3.412948130665737,
      "objective/train/tokens_used": 485764576,
      "objective/train/unlikelihood_loss": 0.051562901586294174,
      "theoretical_loss": 3.412948130665737,
      "tokens_seen": 2126905344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003616363636363637,
      "loss": 1.5143,
      "theoretical_loss": 3.412948130665737,
      "tokens_seen": 2126905344
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036161616161616163,
      "loss": 1.3193,
      "theoretical_loss": 3.4129394683012704,
      "tokens_seen": 2126970880
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003615959595959596,
      "loss": 1.4429,
      "theoretical_loss": 3.4129308062784336,
      "tokens_seen": 2127036416
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036157575757575754,
      "loss": 1.263,
      "theoretical_loss": 3.4129221445972027,
      "tokens_seen": 2127101952
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036155555555555556,
      "loss": 1.4309,
      "theoretical_loss": 3.4129134832575545,
      "tokens_seen": 2127167488
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036153535353535357,
      "loss": 1.5379,
      "theoretical_loss": 3.4129048222594642,
      "tokens_seen": 2127233024
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003615151515151515,
      "loss": 1.5309,
      "theoretical_loss": 3.4128961616029083,
      "tokens_seen": 2127298560
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003614949494949495,
      "loss": 1.3326,
      "theoretical_loss": 3.4128875012878623,
      "tokens_seen": 2127364096
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036147474747474744,
      "loss": 1.2591,
      "theoretical_loss": 3.4128788413143027,
      "tokens_seen": 2127429632
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003614545454545455,
      "loss": 1.1578,
      "theoretical_loss": 3.412870181682206,
      "tokens_seen": 2127495168
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036143434343434346,
      "loss": 1.4151,
      "theoretical_loss": 3.412861522391547,
      "tokens_seen": 2127560704
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003614141414141414,
      "loss": 1.2281,
      "theoretical_loss": 3.412852863442302,
      "tokens_seen": 2127626240
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003613939393939394,
      "loss": 1.3528,
      "theoretical_loss": 3.412844204834448,
      "tokens_seen": 2127691776
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036137373737373733,
      "loss": 1.1927,
      "theoretical_loss": 3.4128355465679596,
      "tokens_seen": 2127757312
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003613535353535354,
      "loss": 1.3806,
      "theoretical_loss": 3.412826888642814,
      "tokens_seen": 2127822848
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036133333333333335,
      "loss": 1.3415,
      "theoretical_loss": 3.4128182310589867,
      "tokens_seen": 2127888384
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003613131313131313,
      "loss": 1.6989,
      "theoretical_loss": 3.4128095738164537,
      "tokens_seen": 2127953920
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036129292929292927,
      "loss": 1.6076,
      "theoretical_loss": 3.4128009169151916,
      "tokens_seen": 2128019456
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003612727272727273,
      "loss": 1.6315,
      "theoretical_loss": 3.4127922603551752,
      "tokens_seen": 2128084992
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003612525252525253,
      "loss": 1.2592,
      "theoretical_loss": 3.412783604136382,
      "tokens_seen": 2128150528
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036123232323232324,
      "loss": 1.4463,
      "theoretical_loss": 3.4127749482587872,
      "tokens_seen": 2128216064
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003612121212121212,
      "loss": 1.1009,
      "theoretical_loss": 3.412766292722367,
      "tokens_seen": 2128281600
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036119191919191916,
      "loss": 1.3194,
      "theoretical_loss": 3.412757637527098,
      "tokens_seen": 2128347136
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003611717171717172,
      "loss": 1.4327,
      "theoretical_loss": 3.412748982672955,
      "tokens_seen": 2128412672
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0003611515151515152,
      "loss": 1.4485,
      "theoretical_loss": 3.412740328159915,
      "tokens_seen": 2128478208
    },
    {
      "epoch": 0.28,
      "objective/train/aligned_token_count": 0.39565005898475647,
      "objective/train/avg_token_score": 0.030088288709521294,
      "objective/train/docs_used": 1204521,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.358384609222412,
      "objective/train/original_loss": 4.55379056930542,
      "objective/train/theoretical_loss": 3.4127316739879543,
      "objective/train/tokens_used": 487402976,
      "objective/train/unlikelihood_loss": 0.047692324966192245,
      "theoretical_loss": 3.4127316739879543,
      "tokens_seen": 2128543744
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00036113131313131314,
      "loss": 1.3064,
      "theoretical_loss": 3.4127316739879543,
      "tokens_seen": 2128543744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003611111111111111,
      "loss": 1.4247,
      "theoretical_loss": 3.4127230201570478,
      "tokens_seen": 2128609280
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003610909090909091,
      "loss": 1.5585,
      "theoretical_loss": 3.4127143666671724,
      "tokens_seen": 2128674816
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003610707070707071,
      "loss": 1.6389,
      "theoretical_loss": 3.4127057135183043,
      "tokens_seen": 2128740352
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003610505050505051,
      "loss": 1.4532,
      "theoretical_loss": 3.412697060710419,
      "tokens_seen": 2128805888
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036103030303030303,
      "loss": 1.6741,
      "theoretical_loss": 3.4126884082434934,
      "tokens_seen": 2128871424
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000361010101010101,
      "loss": 1.4898,
      "theoretical_loss": 3.412679756117502,
      "tokens_seen": 2128936960
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360989898989899,
      "loss": 1.4162,
      "theoretical_loss": 3.4126711043324227,
      "tokens_seen": 2129002496
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360969696969697,
      "loss": 1.5396,
      "theoretical_loss": 3.4126624528882306,
      "tokens_seen": 2129068032
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036094949494949497,
      "loss": 1.4113,
      "theoretical_loss": 3.4126538017849017,
      "tokens_seen": 2129133568
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003609292929292929,
      "loss": 1.4209,
      "theoretical_loss": 3.4126451510224127,
      "tokens_seen": 2129199104
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036090909090909093,
      "loss": 1.4796,
      "theoretical_loss": 3.4126365006007386,
      "tokens_seen": 2129264640
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003608888888888889,
      "loss": 1.5643,
      "theoretical_loss": 3.412627850519857,
      "tokens_seen": 2129330176
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003608686868686869,
      "loss": 1.5414,
      "theoretical_loss": 3.412619200779743,
      "tokens_seen": 2129395712
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036084848484848486,
      "loss": 1.4031,
      "theoretical_loss": 3.4126105513803724,
      "tokens_seen": 2129461248
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003608282828282828,
      "loss": 1.5691,
      "theoretical_loss": 3.4126019023217222,
      "tokens_seen": 2129526784
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003608080808080808,
      "loss": 1.3082,
      "theoretical_loss": 3.4125932536037675,
      "tokens_seen": 2129592320
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003607878787878788,
      "loss": 1.3075,
      "theoretical_loss": 3.4125846052264857,
      "tokens_seen": 2129657856
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003607676767676768,
      "loss": 1.3246,
      "theoretical_loss": 3.4125759571898513,
      "tokens_seen": 2129723392
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036074747474747475,
      "loss": 1.6082,
      "theoretical_loss": 3.412567309493842,
      "tokens_seen": 2129788928
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036072727272727276,
      "loss": 1.364,
      "theoretical_loss": 3.4125586621384327,
      "tokens_seen": 2129854464
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003607070707070707,
      "loss": 1.4541,
      "theoretical_loss": 3.4125500151236,
      "tokens_seen": 2129920000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003606868686868687,
      "loss": 1.3068,
      "theoretical_loss": 3.41254136844932,
      "tokens_seen": 2129985536
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003606666666666667,
      "loss": 1.4795,
      "theoretical_loss": 3.412532722115569,
      "tokens_seen": 2130051072
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036064646464646464,
      "loss": 1.253,
      "theoretical_loss": 3.412524076122322,
      "tokens_seen": 2130116608
    },
    {
      "epoch": 0.29,
      "objective/train/aligned_token_count": 0.5087365508079529,
      "objective/train/avg_token_score": 0.009626802057027817,
      "objective/train/docs_used": 1206012,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.630181074142456,
      "objective/train/original_loss": 3.9676575660705566,
      "objective/train/theoretical_loss": 3.412515430469557,
      "objective/train/tokens_used": 489041376,
      "objective/train/unlikelihood_loss": 0.08346080034971237,
      "theoretical_loss": 3.412515430469557,
      "tokens_seen": 2130182144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036062626262626266,
      "loss": 1.5235,
      "theoretical_loss": 3.412515430469557,
      "tokens_seen": 2130182144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003606060606060606,
      "loss": 1.5016,
      "theoretical_loss": 3.412506785157248,
      "tokens_seen": 2130247680
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036058585858585857,
      "loss": 1.5602,
      "theoretical_loss": 3.4124981401853733,
      "tokens_seen": 2130313216
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003605656565656566,
      "loss": 1.456,
      "theoretical_loss": 3.412489495553908,
      "tokens_seen": 2130378752
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003605454545454546,
      "loss": 1.5841,
      "theoretical_loss": 3.4124808512628277,
      "tokens_seen": 2130444288
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036052525252525255,
      "loss": 1.5069,
      "theoretical_loss": 3.412472207312109,
      "tokens_seen": 2130509824
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003605050505050505,
      "loss": 1.3602,
      "theoretical_loss": 3.4124635637017278,
      "tokens_seen": 2130575360
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036048484848484846,
      "loss": 1.5518,
      "theoretical_loss": 3.4124549204316605,
      "tokens_seen": 2130640896
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036046464646464647,
      "loss": 1.5326,
      "theoretical_loss": 3.412446277501884,
      "tokens_seen": 2130706432
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003604444444444445,
      "loss": 1.4563,
      "theoretical_loss": 3.4124376349123726,
      "tokens_seen": 2130771968
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036042424242424244,
      "loss": 1.5847,
      "theoretical_loss": 3.412428992663104,
      "tokens_seen": 2130837504
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003604040404040404,
      "loss": 1.7093,
      "theoretical_loss": 3.4124203507540534,
      "tokens_seen": 2130903040
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036038383838383835,
      "loss": 1.6105,
      "theoretical_loss": 3.4124117091851978,
      "tokens_seen": 2130968576
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003603636363636364,
      "loss": 1.5254,
      "theoretical_loss": 3.4124030679565127,
      "tokens_seen": 2131034112
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003603434343434344,
      "loss": 1.482,
      "theoretical_loss": 3.4123944270679742,
      "tokens_seen": 2131099648
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036032323232323233,
      "loss": 1.5934,
      "theoretical_loss": 3.4123857865195593,
      "tokens_seen": 2131165184
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003603030303030303,
      "loss": 1.3605,
      "theoretical_loss": 3.4123771463112433,
      "tokens_seen": 2131230720
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036028282828282825,
      "loss": 1.4594,
      "theoretical_loss": 3.412368506443002,
      "tokens_seen": 2131296256
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003602626262626263,
      "loss": 1.4644,
      "theoretical_loss": 3.412359866914813,
      "tokens_seen": 2131361792
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036024242424242427,
      "loss": 1.5003,
      "theoretical_loss": 3.4123512277266514,
      "tokens_seen": 2131427328
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003602222222222222,
      "loss": 1.3692,
      "theoretical_loss": 3.4123425888784933,
      "tokens_seen": 2131492864
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003602020202020202,
      "loss": 1.3248,
      "theoretical_loss": 3.4123339503703156,
      "tokens_seen": 2131558400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036018181818181814,
      "loss": 1.5451,
      "theoretical_loss": 3.4123253122020936,
      "tokens_seen": 2131623936
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601616161616162,
      "loss": 1.5646,
      "theoretical_loss": 3.4123166743738045,
      "tokens_seen": 2131689472
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036014141414141416,
      "loss": 1.2047,
      "theoretical_loss": 3.412308036885423,
      "tokens_seen": 2131755008
    },
    {
      "epoch": 0.29,
      "objective/train/aligned_token_count": 0.5309139490127563,
      "objective/train/avg_token_score": 0.013998433016240597,
      "objective/train/docs_used": 1206751,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8272676467895508,
      "objective/train/original_loss": 4.273675441741943,
      "objective/train/theoretical_loss": 3.412299399736927,
      "objective/train/tokens_used": 490679776,
      "objective/train/unlikelihood_loss": 0.04324044659733772,
      "theoretical_loss": 3.412299399736927,
      "tokens_seen": 2131820544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601212121212121,
      "loss": 1.5616,
      "theoretical_loss": 3.412299399736927,
      "tokens_seen": 2131820544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003601010101010101,
      "loss": 1.165,
      "theoretical_loss": 3.4122907629282913,
      "tokens_seen": 2131886080
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003600808080808081,
      "loss": 1.4589,
      "theoretical_loss": 3.412282126459493,
      "tokens_seen": 2131951616
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003600606060606061,
      "loss": 1.4226,
      "theoretical_loss": 3.412273490330507,
      "tokens_seen": 2132017152
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00036004040404040405,
      "loss": 1.5225,
      "theoretical_loss": 3.412264854541311,
      "tokens_seen": 2132082688
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000360020202020202,
      "loss": 1.4629,
      "theoretical_loss": 3.412256219091881,
      "tokens_seen": 2132148224
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035999999999999997,
      "loss": 1.3966,
      "theoretical_loss": 3.412247583982192,
      "tokens_seen": 2132213760
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359979797979798,
      "loss": 1.4012,
      "theoretical_loss": 3.412238949212221,
      "tokens_seen": 2132279296
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359959595959596,
      "loss": 1.2143,
      "theoretical_loss": 3.4122303147819446,
      "tokens_seen": 2132344832
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035993939393939395,
      "loss": 1.4501,
      "theoretical_loss": 3.412221680691338,
      "tokens_seen": 2132410368
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003599191919191919,
      "loss": 1.4592,
      "theoretical_loss": 3.412213046940378,
      "tokens_seen": 2132475904
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003598989898989899,
      "loss": 1.4762,
      "theoretical_loss": 3.4122044135290412,
      "tokens_seen": 2132541440
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035987878787878787,
      "loss": 1.3354,
      "theoretical_loss": 3.412195780457303,
      "tokens_seen": 2132606976
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003598585858585859,
      "loss": 1.6177,
      "theoretical_loss": 3.41218714772514,
      "tokens_seen": 2132672512
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035983838383838384,
      "loss": 1.5753,
      "theoretical_loss": 3.412178515332528,
      "tokens_seen": 2132738048
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003598181818181818,
      "loss": 1.4519,
      "theoretical_loss": 3.412169883279444,
      "tokens_seen": 2132803584
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003597979797979798,
      "loss": 1.4036,
      "theoretical_loss": 3.412161251565864,
      "tokens_seen": 2132869120
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035977777777777776,
      "loss": 1.4998,
      "theoretical_loss": 3.4121526201917636,
      "tokens_seen": 2132934656
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003597575757575758,
      "loss": 1.5271,
      "theoretical_loss": 3.4121439891571193,
      "tokens_seen": 2133000192
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035973737373737373,
      "loss": 1.4258,
      "theoretical_loss": 3.4121353584619074,
      "tokens_seen": 2133065728
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035971717171717174,
      "loss": 1.2605,
      "theoretical_loss": 3.412126728106104,
      "tokens_seen": 2133131264
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003596969696969697,
      "loss": 1.1889,
      "theoretical_loss": 3.412118098089686,
      "tokens_seen": 2133196800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035967676767676766,
      "loss": 1.4107,
      "theoretical_loss": 3.4121094684126287,
      "tokens_seen": 2133262336
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035965656565656567,
      "loss": 1.5936,
      "theoretical_loss": 3.412100839074909,
      "tokens_seen": 2133327872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003596363636363636,
      "loss": 1.5137,
      "theoretical_loss": 3.412092210076503,
      "tokens_seen": 2133393408
    },
    {
      "epoch": 0.29,
      "objective/train/aligned_token_count": 0.2601417303085327,
      "objective/train/avg_token_score": 0.018833760172128677,
      "objective/train/docs_used": 1207876,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8731229305267334,
      "objective/train/original_loss": 4.842316150665283,
      "objective/train/theoretical_loss": 3.412083581417386,
      "objective/train/tokens_used": 492318176,
      "objective/train/unlikelihood_loss": 0.054733455181121826,
      "theoretical_loss": 3.412083581417386,
      "tokens_seen": 2133458944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035961616161616164,
      "loss": 1.3005,
      "theoretical_loss": 3.412083581417386,
      "tokens_seen": 2133458944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003595959595959596,
      "loss": 1.4467,
      "theoretical_loss": 3.412074953097536,
      "tokens_seen": 2133524480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035957575757575755,
      "loss": 1.376,
      "theoretical_loss": 3.412066325116928,
      "tokens_seen": 2133590016
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035955555555555556,
      "loss": 1.6853,
      "theoretical_loss": 3.4120576974755386,
      "tokens_seen": 2133655552
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035953535353535357,
      "loss": 1.3437,
      "theoretical_loss": 3.4120490701733432,
      "tokens_seen": 2133721088
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035951515151515153,
      "loss": 1.3724,
      "theoretical_loss": 3.4120404432103193,
      "tokens_seen": 2133786624
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594949494949495,
      "loss": 1.5865,
      "theoretical_loss": 3.4120318165864427,
      "tokens_seen": 2133852160
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594747474747475,
      "loss": 1.4658,
      "theoretical_loss": 3.41202319030169,
      "tokens_seen": 2133917696
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035945454545454545,
      "loss": 1.4402,
      "theoretical_loss": 3.4120145643560367,
      "tokens_seen": 2133983232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035943434343434347,
      "loss": 1.2873,
      "theoretical_loss": 3.4120059387494592,
      "tokens_seen": 2134048768
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003594141414141414,
      "loss": 1.3218,
      "theoretical_loss": 3.411997313481934,
      "tokens_seen": 2134114304
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003593939393939394,
      "loss": 1.4666,
      "theoretical_loss": 3.411988688553438,
      "tokens_seen": 2134179840
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003593737373737374,
      "loss": 1.4111,
      "theoretical_loss": 3.411980063963946,
      "tokens_seen": 2134245376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003593535353535354,
      "loss": 1.4017,
      "theoretical_loss": 3.4119714397134357,
      "tokens_seen": 2134310912
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035933333333333336,
      "loss": 1.3271,
      "theoretical_loss": 3.4119628158018824,
      "tokens_seen": 2134376448
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003593131313131313,
      "loss": 1.4653,
      "theoretical_loss": 3.411954192229263,
      "tokens_seen": 2134441984
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035929292929292927,
      "loss": 1.4801,
      "theoretical_loss": 3.411945568995553,
      "tokens_seen": 2134507520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003592727272727273,
      "loss": 1.4413,
      "theoretical_loss": 3.411936946100729,
      "tokens_seen": 2134573056
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003592525252525253,
      "loss": 1.3266,
      "theoretical_loss": 3.4119283235447684,
      "tokens_seen": 2134638592
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035923232323232325,
      "loss": 1.2132,
      "theoretical_loss": 3.411919701327646,
      "tokens_seen": 2134704128
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003592121212121212,
      "loss": 1.1523,
      "theoretical_loss": 3.4119110794493386,
      "tokens_seen": 2134769664
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035919191919191916,
      "loss": 1.4889,
      "theoretical_loss": 3.4119024579098225,
      "tokens_seen": 2134835200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035917171717171723,
      "loss": 1.3673,
      "theoretical_loss": 3.4118938367090736,
      "tokens_seen": 2134900736
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003591515151515152,
      "loss": 1.4432,
      "theoretical_loss": 3.411885215847069,
      "tokens_seen": 2134966272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035913131313131314,
      "loss": 1.4495,
      "theoretical_loss": 3.4118765953237844,
      "tokens_seen": 2135031808
    },
    {
      "epoch": 0.29,
      "objective/train/aligned_token_count": 0.43487292528152466,
      "objective/train/avg_token_score": 0.015729127451777458,
      "objective/train/docs_used": 1208575,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4166325330734253,
      "objective/train/original_loss": 3.7783589363098145,
      "objective/train/theoretical_loss": 3.4118679751391965,
      "objective/train/tokens_used": 493956576,
      "objective/train/unlikelihood_loss": 0.09019943326711655,
      "theoretical_loss": 3.4118679751391965,
      "tokens_seen": 2135097344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003591111111111111,
      "loss": 1.3516,
      "theoretical_loss": 3.4118679751391965,
      "tokens_seen": 2135097344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035909090909090906,
      "loss": 1.6024,
      "theoretical_loss": 3.411859355293281,
      "tokens_seen": 2135162880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003590707070707071,
      "loss": 1.5744,
      "theoretical_loss": 3.4118507357860146,
      "tokens_seen": 2135228416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003590505050505051,
      "loss": 1.2641,
      "theoretical_loss": 3.4118421166173736,
      "tokens_seen": 2135293952
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035903030303030304,
      "loss": 1.4731,
      "theoretical_loss": 3.4118334977873346,
      "tokens_seen": 2135359488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000359010101010101,
      "loss": 1.7547,
      "theoretical_loss": 3.4118248792958736,
      "tokens_seen": 2135425024
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358989898989899,
      "loss": 1.5753,
      "theoretical_loss": 3.4118162611429663,
      "tokens_seen": 2135490560
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358969696969697,
      "loss": 1.5385,
      "theoretical_loss": 3.41180764332859,
      "tokens_seen": 2135556096
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035894949494949497,
      "loss": 1.6049,
      "theoretical_loss": 3.41179902585272,
      "tokens_seen": 2135621632
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035892929292929293,
      "loss": 1.5225,
      "theoretical_loss": 3.411790408715334,
      "tokens_seen": 2135687168
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003589090909090909,
      "loss": 1.5541,
      "theoretical_loss": 3.411781791916407,
      "tokens_seen": 2135752704
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003588888888888889,
      "loss": 1.3487,
      "theoretical_loss": 3.411773175455916,
      "tokens_seen": 2135818240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003588686868686869,
      "loss": 1.56,
      "theoretical_loss": 3.4117645593338377,
      "tokens_seen": 2135883776
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035884848484848486,
      "loss": 1.4217,
      "theoretical_loss": 3.411755943550147,
      "tokens_seen": 2135949312
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003588282828282828,
      "loss": 1.5536,
      "theoretical_loss": 3.4117473281048216,
      "tokens_seen": 2136014848
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035880808080808083,
      "loss": 1.3469,
      "theoretical_loss": 3.4117387129978374,
      "tokens_seen": 2136080384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587878787878788,
      "loss": 1.5176,
      "theoretical_loss": 3.4117300982291705,
      "tokens_seen": 2136145920
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587676767676768,
      "loss": 1.3404,
      "theoretical_loss": 3.4117214837987975,
      "tokens_seen": 2136211456
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035874747474747476,
      "loss": 1.4812,
      "theoretical_loss": 3.4117128697066947,
      "tokens_seen": 2136276992
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587272727272727,
      "loss": 1.4882,
      "theoretical_loss": 3.4117042559528388,
      "tokens_seen": 2136342528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003587070707070707,
      "loss": 1.6542,
      "theoretical_loss": 3.411695642537205,
      "tokens_seen": 2136408064
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586868686868687,
      "loss": 1.5705,
      "theoretical_loss": 3.411687029459771,
      "tokens_seen": 2136473600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586666666666667,
      "loss": 1.5797,
      "theoretical_loss": 3.4116784167205125,
      "tokens_seen": 2136539136
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035864646464646465,
      "loss": 1.3391,
      "theoretical_loss": 3.4116698043194056,
      "tokens_seen": 2136604672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586262626262626,
      "loss": 1.4513,
      "theoretical_loss": 3.411661192256427,
      "tokens_seen": 2136670208
    },
    {
      "epoch": 0.29,
      "objective/train/aligned_token_count": 0.5114858150482178,
      "objective/train/avg_token_score": 0.026600955054163933,
      "objective/train/docs_used": 1209997,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.747342824935913,
      "objective/train/original_loss": 4.3426594734191895,
      "objective/train/theoretical_loss": 3.4116525805315527,
      "objective/train/tokens_used": 495594976,
      "objective/train/unlikelihood_loss": 0.055273815989494324,
      "theoretical_loss": 3.4116525805315527,
      "tokens_seen": 2136735744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003586060606060606,
      "loss": 1.6058,
      "theoretical_loss": 3.4116525805315527,
      "tokens_seen": 2136735744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003585858585858586,
      "loss": 1.3531,
      "theoretical_loss": 3.4116439691447598,
      "tokens_seen": 2136801280
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003585656565656566,
      "loss": 1.5037,
      "theoretical_loss": 3.4116353580960244,
      "tokens_seen": 2136866816
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035854545454545454,
      "loss": 1.4249,
      "theoretical_loss": 3.411626747385322,
      "tokens_seen": 2136932352
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035852525252525255,
      "loss": 1.2405,
      "theoretical_loss": 3.4116181370126304,
      "tokens_seen": 2136997888
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003585050505050505,
      "loss": 1.2637,
      "theoretical_loss": 3.4116095269779247,
      "tokens_seen": 2137063424
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035848484848484847,
      "loss": 1.3767,
      "theoretical_loss": 3.4116009172811816,
      "tokens_seen": 2137128960
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003584646464646465,
      "loss": 1.465,
      "theoretical_loss": 3.411592307922378,
      "tokens_seen": 2137194496
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035844444444444444,
      "loss": 1.4217,
      "theoretical_loss": 3.41158369890149,
      "tokens_seen": 2137260032
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035842424242424245,
      "loss": 1.7356,
      "theoretical_loss": 3.411575090218494,
      "tokens_seen": 2137325568
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003584040404040404,
      "loss": 1.3519,
      "theoretical_loss": 3.4115664818733658,
      "tokens_seen": 2137391104
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035838383838383836,
      "loss": 1.399,
      "theoretical_loss": 3.411557873866083,
      "tokens_seen": 2137456640
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035836363636363637,
      "loss": 1.4519,
      "theoretical_loss": 3.4115492661966202,
      "tokens_seen": 2137522176
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003583434343434344,
      "loss": 1.5747,
      "theoretical_loss": 3.411540658864955,
      "tokens_seen": 2137587712
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035832323232323234,
      "loss": 1.3938,
      "theoretical_loss": 3.411532051871064,
      "tokens_seen": 2137653248
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003583030303030303,
      "loss": 1.4424,
      "theoretical_loss": 3.4115234452149235,
      "tokens_seen": 2137718784
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035828282828282825,
      "loss": 1.0837,
      "theoretical_loss": 3.411514838896509,
      "tokens_seen": 2137784320
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035826262626262626,
      "loss": 1.4307,
      "theoretical_loss": 3.411506232915798,
      "tokens_seen": 2137849856
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003582424242424243,
      "loss": 1.3948,
      "theoretical_loss": 3.411497627272766,
      "tokens_seen": 2137915392
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035822222222222223,
      "loss": 1.3995,
      "theoretical_loss": 3.41148902196739,
      "tokens_seen": 2137980928
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003582020202020202,
      "loss": 1.5618,
      "theoretical_loss": 3.4114804169996464,
      "tokens_seen": 2138046464
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035818181818181815,
      "loss": 1.4169,
      "theoretical_loss": 3.411471812369511,
      "tokens_seen": 2138112000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003581616161616162,
      "loss": 1.6591,
      "theoretical_loss": 3.4114632080769605,
      "tokens_seen": 2138177536
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035814141414141417,
      "loss": 1.5279,
      "theoretical_loss": 3.4114546041219715,
      "tokens_seen": 2138243072
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003581212121212121,
      "loss": 1.2843,
      "theoretical_loss": 3.4114460005045206,
      "tokens_seen": 2138308608
    },
    {
      "epoch": 0.29,
      "objective/train/aligned_token_count": 0.43316227197647095,
      "objective/train/avg_token_score": 0.01263100653886795,
      "objective/train/docs_used": 1210838,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.406528353691101,
      "objective/train/original_loss": 5.164313793182373,
      "objective/train/theoretical_loss": 3.4114373972245837,
      "objective/train/tokens_used": 497233376,
      "objective/train/unlikelihood_loss": 0.04017988219857216,
      "theoretical_loss": 3.4114373972245837,
      "tokens_seen": 2138374144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003581010101010101,
      "loss": 1.2272,
      "theoretical_loss": 3.4114373972245837,
      "tokens_seen": 2138374144
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035808080808080804,
      "loss": 1.497,
      "theoretical_loss": 3.4114287942821377,
      "tokens_seen": 2138439680
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003580606060606061,
      "loss": 1.4253,
      "theoretical_loss": 3.4114201916771583,
      "tokens_seen": 2138505216
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035804040404040406,
      "loss": 1.4705,
      "theoretical_loss": 3.4114115894096226,
      "tokens_seen": 2138570752
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358020202020202,
      "loss": 1.4205,
      "theoretical_loss": 3.411402987479507,
      "tokens_seen": 2138636288
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000358,
      "loss": 1.4689,
      "theoretical_loss": 3.4113943858867875,
      "tokens_seen": 2138701824
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000357979797979798,
      "loss": 1.4757,
      "theoretical_loss": 3.411385784631441,
      "tokens_seen": 2138767360
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000357959595959596,
      "loss": 1.5008,
      "theoretical_loss": 3.4113771837134435,
      "tokens_seen": 2138832896
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035793939393939395,
      "loss": 1.3656,
      "theoretical_loss": 3.411368583132772,
      "tokens_seen": 2138898432
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003579191919191919,
      "loss": 1.459,
      "theoretical_loss": 3.411359982889402,
      "tokens_seen": 2138963968
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035789898989898987,
      "loss": 1.1593,
      "theoretical_loss": 3.4113513829833106,
      "tokens_seen": 2139029504
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578787878787879,
      "loss": 1.395,
      "theoretical_loss": 3.411342783414474,
      "tokens_seen": 2139095040
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578585858585859,
      "loss": 1.3069,
      "theoretical_loss": 3.411334184182869,
      "tokens_seen": 2139160576
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035783838383838385,
      "loss": 1.5065,
      "theoretical_loss": 3.411325585288472,
      "tokens_seen": 2139226112
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003578181818181818,
      "loss": 1.4873,
      "theoretical_loss": 3.411316986731259,
      "tokens_seen": 2139291648
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003577979797979798,
      "loss": 1.5681,
      "theoretical_loss": 3.411308388511207,
      "tokens_seen": 2139357184
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035777777777777777,
      "loss": 1.456,
      "theoretical_loss": 3.4112997906282923,
      "tokens_seen": 2139422720
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003577575757575758,
      "loss": 1.5545,
      "theoretical_loss": 3.4112911930824907,
      "tokens_seen": 2139488256
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035773737373737374,
      "loss": 1.448,
      "theoretical_loss": 3.4112825958737796,
      "tokens_seen": 2139553792
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003577171717171717,
      "loss": 1.5469,
      "theoretical_loss": 3.411273999002135,
      "tokens_seen": 2139619328
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003576969696969697,
      "loss": 1.6172,
      "theoretical_loss": 3.4112654024675333,
      "tokens_seen": 2139684864
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003576767676767677,
      "loss": 1.4533,
      "theoretical_loss": 3.411256806269951,
      "tokens_seen": 2139750400
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003576565656565657,
      "loss": 1.5875,
      "theoretical_loss": 3.411248210409364,
      "tokens_seen": 2139815936
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035763636363636363,
      "loss": 1.3138,
      "theoretical_loss": 3.4112396148857504,
      "tokens_seen": 2139881472
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035761616161616164,
      "loss": 1.3831,
      "theoretical_loss": 3.411231019699086,
      "tokens_seen": 2139947008
    },
    {
      "epoch": 0.29,
      "objective/train/aligned_token_count": 0.38550829887390137,
      "objective/train/avg_token_score": 0.010996246710419655,
      "objective/train/docs_used": 1212386,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3142551183700562,
      "objective/train/original_loss": 4.66601037979126,
      "objective/train/theoretical_loss": 3.411222424849346,
      "objective/train/tokens_used": 498871776,
      "objective/train/unlikelihood_loss": 0.042978379875421524,
      "theoretical_loss": 3.411222424849346,
      "tokens_seen": 2140012544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575959595959596,
      "loss": 1.4413,
      "theoretical_loss": 3.411222424849346,
      "tokens_seen": 2140012544
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575757575757576,
      "loss": 1.4538,
      "theoretical_loss": 3.411213830336508,
      "tokens_seen": 2140078080
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035755555555555557,
      "loss": 1.3922,
      "theoretical_loss": 3.4112052361605487,
      "tokens_seen": 2140143616
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003575353535353535,
      "loss": 1.3603,
      "theoretical_loss": 3.4111966423214435,
      "tokens_seen": 2140209152
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035751515151515153,
      "loss": 1.5785,
      "theoretical_loss": 3.41118804881917,
      "tokens_seen": 2140274688
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003574949494949495,
      "loss": 1.6913,
      "theoretical_loss": 3.4111794556537047,
      "tokens_seen": 2140340224
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003574747474747475,
      "loss": 1.4119,
      "theoretical_loss": 3.411170862825023,
      "tokens_seen": 2140405760
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035745454545454546,
      "loss": 1.2982,
      "theoretical_loss": 3.4111622703331017,
      "tokens_seen": 2140471296
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035743434343434347,
      "loss": 1.2424,
      "theoretical_loss": 3.4111536781779184,
      "tokens_seen": 2140536832
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035741414141414143,
      "loss": 1.5727,
      "theoretical_loss": 3.4111450863594484,
      "tokens_seen": 2140602368
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573939393939394,
      "loss": 1.649,
      "theoretical_loss": 3.4111364948776686,
      "tokens_seen": 2140667904
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573737373737374,
      "loss": 1.4136,
      "theoretical_loss": 3.411127903732556,
      "tokens_seen": 2140733440
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035735353535353535,
      "loss": 1.4533,
      "theoretical_loss": 3.411119312924086,
      "tokens_seen": 2140798976
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035733333333333336,
      "loss": 1.5858,
      "theoretical_loss": 3.4111107224522357,
      "tokens_seen": 2140864512
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003573131313131313,
      "loss": 1.4188,
      "theoretical_loss": 3.411102132316982,
      "tokens_seen": 2140930048
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003572929292929293,
      "loss": 1.2086,
      "theoretical_loss": 3.411093542518301,
      "tokens_seen": 2140995584
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003572727272727273,
      "loss": 1.5396,
      "theoretical_loss": 3.411084953056169,
      "tokens_seen": 2141061120
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003572525252525253,
      "loss": 1.3101,
      "theoretical_loss": 3.411076363930563,
      "tokens_seen": 2141126656
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035723232323232326,
      "loss": 1.448,
      "theoretical_loss": 3.411067775141459,
      "tokens_seen": 2141192192
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003572121212121212,
      "loss": 1.5575,
      "theoretical_loss": 3.411059186688834,
      "tokens_seen": 2141257728
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035719191919191917,
      "loss": 1.4125,
      "theoretical_loss": 3.411050598572664,
      "tokens_seen": 2141323264
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003571717171717172,
      "loss": 1.8047,
      "theoretical_loss": 3.4110420107929262,
      "tokens_seen": 2141388800
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003571515151515152,
      "loss": 1.4474,
      "theoretical_loss": 3.4110334233495965,
      "tokens_seen": 2141454336
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035713131313131315,
      "loss": 1.6298,
      "theoretical_loss": 3.4110248362426514,
      "tokens_seen": 2141519872
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003571111111111111,
      "loss": 1.6019,
      "theoretical_loss": 3.4110162494720684,
      "tokens_seen": 2141585408
    },
    {
      "epoch": 0.29,
      "objective/train/aligned_token_count": 0.3673631548881531,
      "objective/train/avg_token_score": 0.01528299879282713,
      "objective/train/docs_used": 1213144,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2732948064804077,
      "objective/train/original_loss": 4.2406415939331055,
      "objective/train/theoretical_loss": 3.4110076630378225,
      "objective/train/tokens_used": 500510176,
      "objective/train/unlikelihood_loss": 0.07076766341924667,
      "theoretical_loss": 3.4110076630378225,
      "tokens_seen": 2141650944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035709090909090906,
      "loss": 1.3349,
      "theoretical_loss": 3.4110076630378225,
      "tokens_seen": 2141650944
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035707070707070713,
      "loss": 1.3803,
      "theoretical_loss": 3.410999076939892,
      "tokens_seen": 2141716480
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003570505050505051,
      "loss": 1.4504,
      "theoretical_loss": 3.410990491178252,
      "tokens_seen": 2141782016
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035703030303030304,
      "loss": 1.5414,
      "theoretical_loss": 3.4109819057528794,
      "tokens_seen": 2141847552
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000357010101010101,
      "loss": 1.4387,
      "theoretical_loss": 3.4109733206637514,
      "tokens_seen": 2141913088
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035698989898989896,
      "loss": 1.3521,
      "theoretical_loss": 3.4109647359108437,
      "tokens_seen": 2141978624
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000356969696969697,
      "loss": 1.3918,
      "theoretical_loss": 3.4109561514941333,
      "tokens_seen": 2142044160
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.000356949494949495,
      "loss": 1.4157,
      "theoretical_loss": 3.4109475674135963,
      "tokens_seen": 2142109696
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035692929292929293,
      "loss": 1.3731,
      "theoretical_loss": 3.41093898366921,
      "tokens_seen": 2142175232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003569090909090909,
      "loss": 1.732,
      "theoretical_loss": 3.41093040026095,
      "tokens_seen": 2142240768
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035688888888888885,
      "loss": 1.3342,
      "theoretical_loss": 3.410921817188794,
      "tokens_seen": 2142306304
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568686868686869,
      "loss": 1.462,
      "theoretical_loss": 3.4109132344527175,
      "tokens_seen": 2142371840
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035684848484848487,
      "loss": 1.5058,
      "theoretical_loss": 3.4109046520526975,
      "tokens_seen": 2142437376
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568282828282828,
      "loss": 1.4065,
      "theoretical_loss": 3.410896069988711,
      "tokens_seen": 2142502912
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003568080808080808,
      "loss": 1.5842,
      "theoretical_loss": 3.410887488260734,
      "tokens_seen": 2142568448
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567878787878788,
      "loss": 1.4824,
      "theoretical_loss": 3.410878906868743,
      "tokens_seen": 2142633984
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567676767676768,
      "loss": 1.3311,
      "theoretical_loss": 3.4108703258127147,
      "tokens_seen": 2142699520
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035674747474747476,
      "loss": 1.405,
      "theoretical_loss": 3.4108617450926255,
      "tokens_seen": 2142765056
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567272727272727,
      "loss": 1.6685,
      "theoretical_loss": 3.4108531647084526,
      "tokens_seen": 2142830592
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003567070707070707,
      "loss": 1.2828,
      "theoretical_loss": 3.4108445846601727,
      "tokens_seen": 2142896128
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003566868686868687,
      "loss": 1.3849,
      "theoretical_loss": 3.410836004947761,
      "tokens_seen": 2142961664
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003566666666666667,
      "loss": 1.2875,
      "theoretical_loss": 3.4108274255711954,
      "tokens_seen": 2143027200
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035664646464646466,
      "loss": 1.4101,
      "theoretical_loss": 3.410818846530452,
      "tokens_seen": 2143092736
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003566262626262626,
      "loss": 1.4154,
      "theoretical_loss": 3.4108102678255072,
      "tokens_seen": 2143158272
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003566060606060606,
      "loss": 1.1729,
      "theoretical_loss": 3.410801689456338,
      "tokens_seen": 2143223808
    },
    {
      "epoch": 0.29,
      "objective/train/aligned_token_count": 0.34463587403297424,
      "objective/train/avg_token_score": 0.016467271372675896,
      "objective/train/docs_used": 1213818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0075488090515137,
      "objective/train/original_loss": 4.430059909820557,
      "objective/train/theoretical_loss": 3.4107931114229206,
      "objective/train/tokens_used": 502148576,
      "objective/train/unlikelihood_loss": 0.06894190609455109,
      "theoretical_loss": 3.4107931114229206,
      "tokens_seen": 2143289344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565858585858586,
      "loss": 1.397,
      "theoretical_loss": 3.4107931114229206,
      "tokens_seen": 2143289344
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565656565656566,
      "loss": 1.3651,
      "theoretical_loss": 3.410784533725232,
      "tokens_seen": 2143354880
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035654545454545455,
      "loss": 1.5144,
      "theoretical_loss": 3.4107759563632487,
      "tokens_seen": 2143420416
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565252525252525,
      "loss": 1.5995,
      "theoretical_loss": 3.4107673793369466,
      "tokens_seen": 2143485952
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003565050505050505,
      "loss": 1.5493,
      "theoretical_loss": 3.4107588026463036,
      "tokens_seen": 2143551488
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564848484848485,
      "loss": 1.3524,
      "theoretical_loss": 3.410750226291295,
      "tokens_seen": 2143617024
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564646464646465,
      "loss": 1.4136,
      "theoretical_loss": 3.4107416502718984,
      "tokens_seen": 2143682560
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035644444444444444,
      "loss": 1.5132,
      "theoretical_loss": 3.41073307458809,
      "tokens_seen": 2143748096
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035642424242424245,
      "loss": 1.421,
      "theoretical_loss": 3.4107244992398464,
      "tokens_seen": 2143813632
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003564040404040404,
      "loss": 1.6114,
      "theoretical_loss": 3.4107159242271443,
      "tokens_seen": 2143879168
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035638383838383837,
      "loss": 1.5623,
      "theoretical_loss": 3.41070734954996,
      "tokens_seen": 2143944704
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003563636363636364,
      "loss": 1.5388,
      "theoretical_loss": 3.41069877520827,
      "tokens_seen": 2144010240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035634343434343433,
      "loss": 1.2676,
      "theoretical_loss": 3.410690201202052,
      "tokens_seen": 2144075776
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035632323232323234,
      "loss": 1.4555,
      "theoretical_loss": 3.4106816275312815,
      "tokens_seen": 2144141312
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003563030303030303,
      "loss": 1.322,
      "theoretical_loss": 3.410673054195936,
      "tokens_seen": 2144206848
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035628282828282826,
      "loss": 1.5607,
      "theoretical_loss": 3.410664481195991,
      "tokens_seen": 2144272384
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035626262626262627,
      "loss": 1.4592,
      "theoretical_loss": 3.4106559085314236,
      "tokens_seen": 2144337920
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003562424242424243,
      "loss": 1.5127,
      "theoretical_loss": 3.4106473362022114,
      "tokens_seen": 2144403456
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035622222222222224,
      "loss": 1.4354,
      "theoretical_loss": 3.41063876420833,
      "tokens_seen": 2144468992
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003562020202020202,
      "loss": 1.6413,
      "theoretical_loss": 3.410630192549756,
      "tokens_seen": 2144534528
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035618181818181815,
      "loss": 1.6882,
      "theoretical_loss": 3.4106216212264666,
      "tokens_seen": 2144600064
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035616161616161616,
      "loss": 1.4265,
      "theoretical_loss": 3.4106130502384375,
      "tokens_seen": 2144665600
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003561414141414142,
      "loss": 1.3411,
      "theoretical_loss": 3.410604479585647,
      "tokens_seen": 2144731136
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035612121212121213,
      "loss": 1.2964,
      "theoretical_loss": 3.4105959092680695,
      "tokens_seen": 2144796672
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0003561010101010101,
      "loss": 1.3232,
      "theoretical_loss": 3.410587339285684,
      "tokens_seen": 2144862208
    },
    {
      "epoch": 0.29,
      "objective/train/aligned_token_count": 0.3159213066101074,
      "objective/train/avg_token_score": 0.012928572483360767,
      "objective/train/docs_used": 1215022,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9320014715194702,
      "objective/train/original_loss": 4.363121032714844,
      "objective/train/theoretical_loss": 3.4105787696384655,
      "objective/train/tokens_used": 503786976,
      "objective/train/unlikelihood_loss": 0.05617682263255119,
      "theoretical_loss": 3.4105787696384655,
      "tokens_seen": 2144927744
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00035608080808080804,
      "loss": 1.3469,
      "theoretical_loss": 3.4105787696384655,
      "tokens_seen": 2144927744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003560606060606061,
      "loss": 1.4266,
      "theoretical_loss": 3.4105702003263914,
      "tokens_seen": 2144993280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035604040404040407,
      "loss": 1.4661,
      "theoretical_loss": 3.410561631349438,
      "tokens_seen": 2145058816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000356020202020202,
      "loss": 1.4141,
      "theoretical_loss": 3.410553062707582,
      "tokens_seen": 2145124352
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000356,
      "loss": 1.3925,
      "theoretical_loss": 3.4105444944008,
      "tokens_seen": 2145189888
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000355979797979798,
      "loss": 1.3971,
      "theoretical_loss": 3.4105359264290693,
      "tokens_seen": 2145255424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000355959595959596,
      "loss": 1.5648,
      "theoretical_loss": 3.4105273587923657,
      "tokens_seen": 2145320960
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035593939393939396,
      "loss": 1.5217,
      "theoretical_loss": 3.4105187914906665,
      "tokens_seen": 2145386496
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003559191919191919,
      "loss": 1.3943,
      "theoretical_loss": 3.4105102245239483,
      "tokens_seen": 2145452032
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035589898989898987,
      "loss": 1.2379,
      "theoretical_loss": 3.410501657892187,
      "tokens_seen": 2145517568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035587878787878794,
      "loss": 1.6016,
      "theoretical_loss": 3.4104930915953604,
      "tokens_seen": 2145583104
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003558585858585859,
      "loss": 1.1341,
      "theoretical_loss": 3.410484525633444,
      "tokens_seen": 2145648640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035583838383838385,
      "loss": 1.6203,
      "theoretical_loss": 3.4104759600064156,
      "tokens_seen": 2145714176
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003558181818181818,
      "loss": 1.5268,
      "theoretical_loss": 3.410467394714251,
      "tokens_seen": 2145779712
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035579797979797977,
      "loss": 1.193,
      "theoretical_loss": 3.4104588297569283,
      "tokens_seen": 2145845248
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035577777777777783,
      "loss": 1.3764,
      "theoretical_loss": 3.410450265134422,
      "tokens_seen": 2145910784
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003557575757575758,
      "loss": 1.3695,
      "theoretical_loss": 3.41044170084671,
      "tokens_seen": 2145976320
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035573737373737374,
      "loss": 1.3551,
      "theoretical_loss": 3.4104331368937695,
      "tokens_seen": 2146041856
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003557171717171717,
      "loss": 1.6121,
      "theoretical_loss": 3.4104245732755762,
      "tokens_seen": 2146107392
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003556969696969697,
      "loss": 1.4464,
      "theoretical_loss": 3.4104160099921077,
      "tokens_seen": 2146172928
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003556767676767677,
      "loss": 1.4414,
      "theoretical_loss": 3.41040744704334,
      "tokens_seen": 2146238464
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003556565656565657,
      "loss": 1.4752,
      "theoretical_loss": 3.4103988844292497,
      "tokens_seen": 2146304000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035563636363636364,
      "loss": 1.4039,
      "theoretical_loss": 3.410390322149814,
      "tokens_seen": 2146369536
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003556161616161616,
      "loss": 1.3964,
      "theoretical_loss": 3.4103817602050093,
      "tokens_seen": 2146435072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003555959595959596,
      "loss": 1.6103,
      "theoretical_loss": 3.4103731985948125,
      "tokens_seen": 2146500608
    },
    {
      "epoch": 0.3,
      "objective/train/aligned_token_count": 0.31359970569610596,
      "objective/train/avg_token_score": 0.02061370201408863,
      "objective/train/docs_used": 1215651,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.065711259841919,
      "objective/train/original_loss": 4.7896342277526855,
      "objective/train/theoretical_loss": 3.4103646373192005,
      "objective/train/tokens_used": 505425376,
      "objective/train/unlikelihood_loss": 0.05333314463496208,
      "theoretical_loss": 3.4103646373192005,
      "tokens_seen": 2146566144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003555757575757576,
      "loss": 1.3544,
      "theoretical_loss": 3.4103646373192005,
      "tokens_seen": 2146566144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035555555555555557,
      "loss": 1.3426,
      "theoretical_loss": 3.4103560763781493,
      "tokens_seen": 2146631680
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035553535353535353,
      "loss": 1.3554,
      "theoretical_loss": 3.4103475157716363,
      "tokens_seen": 2146697216
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035551515151515154,
      "loss": 1.7167,
      "theoretical_loss": 3.410338955499638,
      "tokens_seen": 2146762752
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003554949494949495,
      "loss": 1.5226,
      "theoretical_loss": 3.4103303955621307,
      "tokens_seen": 2146828288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003554747474747475,
      "loss": 1.6862,
      "theoretical_loss": 3.410321835959092,
      "tokens_seen": 2146893824
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035545454545454547,
      "loss": 1.5275,
      "theoretical_loss": 3.4103132766904976,
      "tokens_seen": 2146959360
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003554343434343434,
      "loss": 1.1603,
      "theoretical_loss": 3.4103047177563255,
      "tokens_seen": 2147024896
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035541414141414143,
      "loss": 1.5835,
      "theoretical_loss": 3.410296159156551,
      "tokens_seen": 2147090432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003553939393939394,
      "loss": 1.4474,
      "theoretical_loss": 3.410287600891152,
      "tokens_seen": 2147155968
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003553737373737374,
      "loss": 1.6101,
      "theoretical_loss": 3.410279042960104,
      "tokens_seen": 2147221504
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035535353535353536,
      "loss": 1.4196,
      "theoretical_loss": 3.4102704853633847,
      "tokens_seen": 2147287040
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003553333333333333,
      "loss": 1.2991,
      "theoretical_loss": 3.4102619281009714,
      "tokens_seen": 2147352576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003553131313131313,
      "loss": 1.4851,
      "theoretical_loss": 3.410253371172839,
      "tokens_seen": 2147418112
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003552929292929293,
      "loss": 1.5103,
      "theoretical_loss": 3.4102448145789657,
      "tokens_seen": 2147483648
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003552727272727273,
      "loss": 1.4754,
      "theoretical_loss": 3.410236258319328,
      "tokens_seen": 2147549184
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035525252525252525,
      "loss": 1.3945,
      "theoretical_loss": 3.4102277023939025,
      "tokens_seen": 2147614720
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035523232323232326,
      "loss": 1.3597,
      "theoretical_loss": 3.4102191468026657,
      "tokens_seen": 2147680256
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003552121212121212,
      "loss": 1.6051,
      "theoretical_loss": 3.4102105915455945,
      "tokens_seen": 2147745792
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003551919191919192,
      "loss": 1.4595,
      "theoretical_loss": 3.4102020366226657,
      "tokens_seen": 2147811328
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003551717171717172,
      "loss": 1.4641,
      "theoretical_loss": 3.410193482033856,
      "tokens_seen": 2147876864
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035515151515151514,
      "loss": 1.4089,
      "theoretical_loss": 3.4101849277791425,
      "tokens_seen": 2147942400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035513131313131315,
      "loss": 1.2692,
      "theoretical_loss": 3.4101763738585014,
      "tokens_seen": 2148007936
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003551111111111111,
      "loss": 1.3462,
      "theoretical_loss": 3.4101678202719103,
      "tokens_seen": 2148073472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035509090909090907,
      "loss": 1.634,
      "theoretical_loss": 3.4101592670193446,
      "tokens_seen": 2148139008
    },
    {
      "epoch": 0.3,
      "objective/train/aligned_token_count": 0.45839443802833557,
      "objective/train/avg_token_score": 0.009575318545103073,
      "objective/train/docs_used": 1216251,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5462843179702759,
      "objective/train/original_loss": 4.091464996337891,
      "objective/train/theoretical_loss": 3.4101507141007827,
      "objective/train/tokens_used": 507063776,
      "objective/train/unlikelihood_loss": 0.06715980917215347,
      "theoretical_loss": 3.4101507141007827,
      "tokens_seen": 2148204544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003550707070707071,
      "loss": 1.4466,
      "theoretical_loss": 3.4101507141007827,
      "tokens_seen": 2148204544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003550505050505051,
      "loss": 1.5371,
      "theoretical_loss": 3.4101421615162,
      "tokens_seen": 2148270080
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035503030303030305,
      "loss": 1.6062,
      "theoretical_loss": 3.410133609265574,
      "tokens_seen": 2148335616
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000355010101010101,
      "loss": 1.2978,
      "theoretical_loss": 3.410125057348881,
      "tokens_seen": 2148401152
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035498989898989896,
      "loss": 1.3665,
      "theoretical_loss": 3.4101165057660987,
      "tokens_seen": 2148466688
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035496969696969697,
      "loss": 1.4329,
      "theoretical_loss": 3.4101079545172026,
      "tokens_seen": 2148532224
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000354949494949495,
      "loss": 1.4802,
      "theoretical_loss": 3.41009940360217,
      "tokens_seen": 2148597760
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035492929292929294,
      "loss": 1.3216,
      "theoretical_loss": 3.4100908530209786,
      "tokens_seen": 2148663296
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003549090909090909,
      "loss": 1.5478,
      "theoretical_loss": 3.4100823027736036,
      "tokens_seen": 2148728832
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035488888888888885,
      "loss": 1.4956,
      "theoretical_loss": 3.410073752860023,
      "tokens_seen": 2148794368
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548686868686869,
      "loss": 1.5093,
      "theoretical_loss": 3.4100652032802135,
      "tokens_seen": 2148859904
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548484848484849,
      "loss": 1.1582,
      "theoretical_loss": 3.410056654034151,
      "tokens_seen": 2148925440
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035482828282828283,
      "loss": 1.5054,
      "theoretical_loss": 3.4100481051218132,
      "tokens_seen": 2148990976
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003548080808080808,
      "loss": 1.2743,
      "theoretical_loss": 3.4100395565431763,
      "tokens_seen": 2149056512
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035478787878787875,
      "loss": 1.4463,
      "theoretical_loss": 3.4100310082982173,
      "tokens_seen": 2149122048
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003547676767676768,
      "loss": 1.2762,
      "theoretical_loss": 3.410022460386913,
      "tokens_seen": 2149187584
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035474747474747477,
      "loss": 1.3302,
      "theoretical_loss": 3.4100139128092404,
      "tokens_seen": 2149253120
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003547272727272727,
      "loss": 1.3001,
      "theoretical_loss": 3.4100053655651763,
      "tokens_seen": 2149318656
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003547070707070707,
      "loss": 1.5577,
      "theoretical_loss": 3.409996818654697,
      "tokens_seen": 2149384192
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003546868686868687,
      "loss": 1.2636,
      "theoretical_loss": 3.4099882720777797,
      "tokens_seen": 2149449728
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003546666666666667,
      "loss": 1.2444,
      "theoretical_loss": 3.4099797258344013,
      "tokens_seen": 2149515264
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035464646464646466,
      "loss": 1.1734,
      "theoretical_loss": 3.4099711799245385,
      "tokens_seen": 2149580800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003546262626262626,
      "loss": 1.2099,
      "theoretical_loss": 3.4099626343481684,
      "tokens_seen": 2149646336
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003546060606060606,
      "loss": 1.3326,
      "theoretical_loss": 3.409954089105267,
      "tokens_seen": 2149711872
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003545858585858586,
      "loss": 1.5443,
      "theoretical_loss": 3.4099455441958124,
      "tokens_seen": 2149777408
    },
    {
      "epoch": 0.3,
      "objective/train/aligned_token_count": 0.3271627426147461,
      "objective/train/avg_token_score": 0.021921349689364433,
      "objective/train/docs_used": 1217415,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.14596688747406,
      "objective/train/original_loss": 4.5187530517578125,
      "objective/train/theoretical_loss": 3.40993699961978,
      "objective/train/tokens_used": 508702176,
      "objective/train/unlikelihood_loss": 0.05059804767370224,
      "theoretical_loss": 3.40993699961978,
      "tokens_seen": 2149842944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003545656565656566,
      "loss": 1.4055,
      "theoretical_loss": 3.40993699961978,
      "tokens_seen": 2149842944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035454545454545455,
      "loss": 1.3152,
      "theoretical_loss": 3.4099284553771474,
      "tokens_seen": 2149908480
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003545252525252525,
      "loss": 1.3024,
      "theoretical_loss": 3.4099199114678918,
      "tokens_seen": 2149974016
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003545050505050505,
      "loss": 1.5076,
      "theoretical_loss": 3.409911367891989,
      "tokens_seen": 2150039552
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003544848484848485,
      "loss": 1.4309,
      "theoretical_loss": 3.4099028246494165,
      "tokens_seen": 2150105088
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003544646464646465,
      "loss": 1.2262,
      "theoretical_loss": 3.409894281740151,
      "tokens_seen": 2150170624
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035444444444444445,
      "loss": 1.4737,
      "theoretical_loss": 3.40988573916417,
      "tokens_seen": 2150236160
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003544242424242424,
      "loss": 1.424,
      "theoretical_loss": 3.4098771969214487,
      "tokens_seen": 2150301696
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003544040404040404,
      "loss": 1.4056,
      "theoretical_loss": 3.409868655011966,
      "tokens_seen": 2150367232
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035438383838383837,
      "loss": 1.4112,
      "theoretical_loss": 3.4098601134356965,
      "tokens_seen": 2150432768
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003543636363636364,
      "loss": 1.5308,
      "theoretical_loss": 3.4098515721926193,
      "tokens_seen": 2150498304
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035434343434343434,
      "loss": 1.3105,
      "theoretical_loss": 3.4098430312827097,
      "tokens_seen": 2150563840
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035432323232323235,
      "loss": 1.4338,
      "theoretical_loss": 3.4098344907059452,
      "tokens_seen": 2150629376
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003543030303030303,
      "loss": 1.3089,
      "theoretical_loss": 3.4098259504623023,
      "tokens_seen": 2150694912
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003542828282828283,
      "loss": 1.2264,
      "theoretical_loss": 3.4098174105517582,
      "tokens_seen": 2150760448
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003542626262626263,
      "loss": 1.1657,
      "theoretical_loss": 3.40980887097429,
      "tokens_seen": 2150825984
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035424242424242423,
      "loss": 1.5638,
      "theoretical_loss": 3.4098003317298735,
      "tokens_seen": 2150891520
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035422222222222224,
      "loss": 1.4518,
      "theoretical_loss": 3.4097917928184867,
      "tokens_seen": 2150957056
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003542020202020202,
      "loss": 1.2664,
      "theoretical_loss": 3.4097832542401063,
      "tokens_seen": 2151022592
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003541818181818182,
      "loss": 1.4255,
      "theoretical_loss": 3.4097747159947085,
      "tokens_seen": 2151088128
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035416161616161617,
      "loss": 1.3902,
      "theoretical_loss": 3.40976617808227,
      "tokens_seen": 2151153664
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003541414141414142,
      "loss": 1.3423,
      "theoretical_loss": 3.409757640502769,
      "tokens_seen": 2151219200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035412121212121214,
      "loss": 1.5539,
      "theoretical_loss": 3.409749103256181,
      "tokens_seen": 2151284736
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003541010101010101,
      "loss": 1.4584,
      "theoretical_loss": 3.4097405663424842,
      "tokens_seen": 2151350272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003540808080808081,
      "loss": 1.3504,
      "theoretical_loss": 3.4097320297616545,
      "tokens_seen": 2151415808
    },
    {
      "epoch": 0.3,
      "objective/train/aligned_token_count": 0.5538856387138367,
      "objective/train/avg_token_score": 0.012418524362146854,
      "objective/train/docs_used": 1217907,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7489923238754272,
      "objective/train/original_loss": 3.712601661682129,
      "objective/train/theoretical_loss": 3.4097234935136687,
      "objective/train/tokens_used": 510340576,
      "objective/train/unlikelihood_loss": 0.07619179040193558,
      "theoretical_loss": 3.4097234935136687,
      "tokens_seen": 2151481344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035406060606060606,
      "loss": 1.4716,
      "theoretical_loss": 3.4097234935136687,
      "tokens_seen": 2151481344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035404040404040407,
      "loss": 1.369,
      "theoretical_loss": 3.4097149575985037,
      "tokens_seen": 2151546880
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035402020202020203,
      "loss": 1.4984,
      "theoretical_loss": 3.4097064220161375,
      "tokens_seen": 2151612416
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000354,
      "loss": 1.4327,
      "theoretical_loss": 3.409697886766546,
      "tokens_seen": 2151677952
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000353979797979798,
      "loss": 1.3699,
      "theoretical_loss": 3.409689351849706,
      "tokens_seen": 2151743488
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000353959595959596,
      "loss": 1.164,
      "theoretical_loss": 3.409680817265595,
      "tokens_seen": 2151809024
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035393939393939396,
      "loss": 1.3928,
      "theoretical_loss": 3.4096722830141895,
      "tokens_seen": 2151874560
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003539191919191919,
      "loss": 1.5867,
      "theoretical_loss": 3.409663749095466,
      "tokens_seen": 2151940096
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003538989898989899,
      "loss": 1.1484,
      "theoretical_loss": 3.4096552155094026,
      "tokens_seen": 2152005632
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003538787878787879,
      "loss": 1.285,
      "theoretical_loss": 3.409646682255975,
      "tokens_seen": 2152071168
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003538585858585859,
      "loss": 1.5314,
      "theoretical_loss": 3.409638149335161,
      "tokens_seen": 2152136704
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035383838383838386,
      "loss": 1.3147,
      "theoretical_loss": 3.4096296167469364,
      "tokens_seen": 2152202240
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003538181818181818,
      "loss": 1.4144,
      "theoretical_loss": 3.4096210844912793,
      "tokens_seen": 2152267776
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035379797979797977,
      "loss": 1.151,
      "theoretical_loss": 3.409612552568166,
      "tokens_seen": 2152333312
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035377777777777784,
      "loss": 1.1738,
      "theoretical_loss": 3.4096040209775733,
      "tokens_seen": 2152398848
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003537575757575758,
      "loss": 1.3645,
      "theoretical_loss": 3.4095954897194787,
      "tokens_seen": 2152464384
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035373737373737375,
      "loss": 1.4558,
      "theoretical_loss": 3.4095869587938585,
      "tokens_seen": 2152529920
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003537171717171717,
      "loss": 1.2696,
      "theoretical_loss": 3.40957842820069,
      "tokens_seen": 2152595456
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035369696969696966,
      "loss": 1.3487,
      "theoretical_loss": 3.4095698979399502,
      "tokens_seen": 2152660992
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035367676767676773,
      "loss": 1.449,
      "theoretical_loss": 3.4095613680116155,
      "tokens_seen": 2152726528
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003536565656565657,
      "loss": 1.3081,
      "theoretical_loss": 3.4095528384156633,
      "tokens_seen": 2152792064
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035363636363636364,
      "loss": 1.3938,
      "theoretical_loss": 3.4095443091520705,
      "tokens_seen": 2152857600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003536161616161616,
      "loss": 1.3076,
      "theoretical_loss": 3.4095357802208133,
      "tokens_seen": 2152923136
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035359595959595956,
      "loss": 1.2976,
      "theoretical_loss": 3.4095272516218698,
      "tokens_seen": 2152988672
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003535757575757576,
      "loss": 1.324,
      "theoretical_loss": 3.4095187233552164,
      "tokens_seen": 2153054208
    },
    {
      "debugging/Self-BLEU-5": 0.4342780019633275,
      "debugging/distinct-1-grams": 0.8078002938067941,
      "debugging/distinct-2-grams": 0.9689768537714897,
      "debugging/entropy-1-grams": 5.9515082084159445,
      "debugging/entropy-2-grams": 6.74050317919558,
      "debugging/length": 469.15384615384613,
      "debugging/num_segments": 13,
      "epoch": 0.3,
      "objective/train/aligned_token_count": 0.574474573135376,
      "objective/train/avg_token_score": 0.01960972137749195,
      "objective/train/docs_used": 1219413,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7889963388442993,
      "objective/train/original_loss": 3.6546130180358887,
      "objective/train/theoretical_loss": 3.40951019542083,
      "objective/train/tokens_used": 511978976,
      "objective/train/unlikelihood_loss": 0.05853184312582016,
      "theoretical_loss": 3.40951019542083,
      "tokens_seen": 2153119744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003535555555555556,
      "loss": 1.5971,
      "theoretical_loss": 3.40951019542083,
      "tokens_seen": 2153119744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035353535353535354,
      "loss": 1.6635,
      "theoretical_loss": 3.409501667818687,
      "tokens_seen": 2153185280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003535151515151515,
      "loss": 1.7406,
      "theoretical_loss": 3.409493140548766,
      "tokens_seen": 2153250816
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003534949494949495,
      "loss": 1.2769,
      "theoretical_loss": 3.409484613611042,
      "tokens_seen": 2153316352
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003534747474747475,
      "loss": 1.1519,
      "theoretical_loss": 3.409476087005493,
      "tokens_seen": 2153381888
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035345454545454547,
      "loss": 1.4715,
      "theoretical_loss": 3.4094675607320957,
      "tokens_seen": 2153447424
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035343434343434343,
      "loss": 1.6237,
      "theoretical_loss": 3.4094590347908276,
      "tokens_seen": 2153512960
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003534141414141414,
      "loss": 1.4639,
      "theoretical_loss": 3.4094505091816645,
      "tokens_seen": 2153578496
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003533939393939394,
      "loss": 1.6356,
      "theoretical_loss": 3.4094419839045846,
      "tokens_seen": 2153644032
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003533737373737374,
      "loss": 1.5859,
      "theoretical_loss": 3.409433458959564,
      "tokens_seen": 2153709568
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035335353535353536,
      "loss": 1.1783,
      "theoretical_loss": 3.40942493434658,
      "tokens_seen": 2153775104
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003533333333333333,
      "loss": 1.5161,
      "theoretical_loss": 3.4094164100656092,
      "tokens_seen": 2153840640
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035331313131313133,
      "loss": 1.3464,
      "theoretical_loss": 3.4094078861166297,
      "tokens_seen": 2153906176
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532929292929293,
      "loss": 1.1118,
      "theoretical_loss": 3.4093993624996166,
      "tokens_seen": 2153971712
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532727272727273,
      "loss": 1.3196,
      "theoretical_loss": 3.4093908392145487,
      "tokens_seen": 2154037248
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035325252525252526,
      "loss": 1.4936,
      "theoretical_loss": 3.409382316261402,
      "tokens_seen": 2154102784
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532323232323232,
      "loss": 1.6132,
      "theoretical_loss": 3.4093737936401536,
      "tokens_seen": 2154168320
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003532121212121212,
      "loss": 1.5541,
      "theoretical_loss": 3.4093652713507803,
      "tokens_seen": 2154233856
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531919191919192,
      "loss": 1.3592,
      "theoretical_loss": 3.4093567493932597,
      "tokens_seen": 2154299392
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531717171717172,
      "loss": 1.5112,
      "theoretical_loss": 3.4093482277675684,
      "tokens_seen": 2154364928
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035315151515151515,
      "loss": 1.5003,
      "theoretical_loss": 3.4093397064736832,
      "tokens_seen": 2154430464
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035313131313131316,
      "loss": 1.6506,
      "theoretical_loss": 3.4093311855115815,
      "tokens_seen": 2154496000
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003531111111111111,
      "loss": 1.5629,
      "theoretical_loss": 3.4093226648812403,
      "tokens_seen": 2154561536
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003530909090909091,
      "loss": 1.4242,
      "theoretical_loss": 3.409314144582636,
      "tokens_seen": 2154627072
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003530707070707071,
      "loss": 1.3268,
      "theoretical_loss": 3.4093056246157456,
      "tokens_seen": 2154692608
    },
    {
      "epoch": 0.3,
      "objective/train/aligned_token_count": 0.2871456444263458,
      "objective/train/avg_token_score": 0.014688773080706596,
      "objective/train/docs_used": 1220074,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.890113115310669,
      "objective/train/original_loss": 5.298357963562012,
      "objective/train/theoretical_loss": 3.4092971049805474,
      "objective/train/tokens_used": 513617376,
      "objective/train/unlikelihood_loss": 0.041902169585227966,
      "theoretical_loss": 3.4092971049805474,
      "tokens_seen": 2154758144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035305050505050504,
      "loss": 1.1781,
      "theoretical_loss": 3.4092971049805474,
      "tokens_seen": 2154758144
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035303030303030305,
      "loss": 1.5121,
      "theoretical_loss": 3.4092885856770168,
      "tokens_seen": 2154823680
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000353010101010101,
      "loss": 1.4291,
      "theoretical_loss": 3.4092800667051315,
      "tokens_seen": 2154889216
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035298989898989897,
      "loss": 1.4243,
      "theoretical_loss": 3.4092715480648685,
      "tokens_seen": 2154954752
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352969696969697,
      "loss": 1.6243,
      "theoretical_loss": 3.4092630297562048,
      "tokens_seen": 2155020288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352949494949495,
      "loss": 1.7129,
      "theoretical_loss": 3.4092545117791175,
      "tokens_seen": 2155085824
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035292929292929295,
      "loss": 1.5722,
      "theoretical_loss": 3.4092459941335838,
      "tokens_seen": 2155151360
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003529090909090909,
      "loss": 1.3569,
      "theoretical_loss": 3.4092374768195794,
      "tokens_seen": 2155216896
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035288888888888886,
      "loss": 1.4478,
      "theoretical_loss": 3.4092289598370833,
      "tokens_seen": 2155282432
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035286868686868687,
      "loss": 1.3799,
      "theoretical_loss": 3.4092204431860713,
      "tokens_seen": 2155347968
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003528484848484849,
      "loss": 1.3461,
      "theoretical_loss": 3.4092119268665204,
      "tokens_seen": 2155413504
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035282828282828284,
      "loss": 1.4725,
      "theoretical_loss": 3.409203410878408,
      "tokens_seen": 2155479040
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003528080808080808,
      "loss": 1.4399,
      "theoretical_loss": 3.4091948952217113,
      "tokens_seen": 2155544576
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035278787878787875,
      "loss": 1.1607,
      "theoretical_loss": 3.4091863798964064,
      "tokens_seen": 2155610112
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527676767676768,
      "loss": 1.6022,
      "theoretical_loss": 3.409177864902471,
      "tokens_seen": 2155675648
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527474747474748,
      "loss": 1.5076,
      "theoretical_loss": 3.4091693502398828,
      "tokens_seen": 2155741184
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035272727272727273,
      "loss": 1.4086,
      "theoretical_loss": 3.409160835908618,
      "tokens_seen": 2155806720
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003527070707070707,
      "loss": 1.3605,
      "theoretical_loss": 3.4091523219086532,
      "tokens_seen": 2155872256
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035268686868686864,
      "loss": 1.5611,
      "theoretical_loss": 3.4091438082399668,
      "tokens_seen": 2155937792
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526666666666667,
      "loss": 1.4342,
      "theoretical_loss": 3.4091352949025344,
      "tokens_seen": 2156003328
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035264646464646467,
      "loss": 1.239,
      "theoretical_loss": 3.4091267818963336,
      "tokens_seen": 2156068864
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526262626262626,
      "loss": 1.3897,
      "theoretical_loss": 3.4091182692213415,
      "tokens_seen": 2156134400
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003526060606060606,
      "loss": 1.5341,
      "theoretical_loss": 3.4091097568775357,
      "tokens_seen": 2156199936
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035258585858585865,
      "loss": 1.328,
      "theoretical_loss": 3.4091012448648925,
      "tokens_seen": 2156265472
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003525656565656566,
      "loss": 1.2071,
      "theoretical_loss": 3.409092733183389,
      "tokens_seen": 2156331008
    },
    {
      "epoch": 0.3,
      "objective/train/aligned_token_count": 0.5043988227844238,
      "objective/train/avg_token_score": 0.01132004614919424,
      "objective/train/docs_used": 1221515,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7465596199035645,
      "objective/train/original_loss": 4.144891738891602,
      "objective/train/theoretical_loss": 3.4090842218330026,
      "objective/train/tokens_used": 515255776,
      "objective/train/unlikelihood_loss": 0.058709658682346344,
      "theoretical_loss": 3.4090842218330026,
      "tokens_seen": 2156396544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035254545454545456,
      "loss": 1.577,
      "theoretical_loss": 3.4090842218330026,
      "tokens_seen": 2156396544
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003525252525252525,
      "loss": 1.4144,
      "theoretical_loss": 3.40907571081371,
      "tokens_seen": 2156462080
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003525050505050505,
      "loss": 1.622,
      "theoretical_loss": 3.409067200125489,
      "tokens_seen": 2156527616
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035248484848484854,
      "loss": 1.4798,
      "theoretical_loss": 3.4090586897683153,
      "tokens_seen": 2156593152
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003524646464646465,
      "loss": 1.3924,
      "theoretical_loss": 3.4090501797421675,
      "tokens_seen": 2156658688
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035244444444444445,
      "loss": 1.4228,
      "theoretical_loss": 3.4090416700470216,
      "tokens_seen": 2156724224
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003524242424242424,
      "loss": 1.3212,
      "theoretical_loss": 3.409033160682855,
      "tokens_seen": 2156789760
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003524040404040404,
      "loss": 1.2613,
      "theoretical_loss": 3.409024651649645,
      "tokens_seen": 2156855296
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035238383838383843,
      "loss": 1.3956,
      "theoretical_loss": 3.409016142947368,
      "tokens_seen": 2156920832
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003523636363636364,
      "loss": 1.3073,
      "theoretical_loss": 3.409007634576002,
      "tokens_seen": 2156986368
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035234343434343434,
      "loss": 1.3321,
      "theoretical_loss": 3.408999126535523,
      "tokens_seen": 2157051904
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003523232323232323,
      "loss": 1.5444,
      "theoretical_loss": 3.4089906188259094,
      "tokens_seen": 2157117440
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003523030303030303,
      "loss": 1.4442,
      "theoretical_loss": 3.408982111447137,
      "tokens_seen": 2157182976
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003522828282828283,
      "loss": 1.347,
      "theoretical_loss": 3.4089736043991836,
      "tokens_seen": 2157248512
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003522626262626263,
      "loss": 1.5098,
      "theoretical_loss": 3.4089650976820263,
      "tokens_seen": 2157314048
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035224242424242424,
      "loss": 1.5813,
      "theoretical_loss": 3.4089565912956417,
      "tokens_seen": 2157379584
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035222222222222225,
      "loss": 1.3246,
      "theoretical_loss": 3.4089480852400076,
      "tokens_seen": 2157445120
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003522020202020202,
      "loss": 1.2625,
      "theoretical_loss": 3.408939579515101,
      "tokens_seen": 2157510656
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521818181818182,
      "loss": 1.407,
      "theoretical_loss": 3.408931074120898,
      "tokens_seen": 2157576192
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521616161616162,
      "loss": 1.4867,
      "theoretical_loss": 3.4089225690573763,
      "tokens_seen": 2157641728
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035214141414141413,
      "loss": 1.3851,
      "theoretical_loss": 3.4089140643245135,
      "tokens_seen": 2157707264
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035212121212121214,
      "loss": 1.5038,
      "theoretical_loss": 3.408905559922286,
      "tokens_seen": 2157772800
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003521010101010101,
      "loss": 1.292,
      "theoretical_loss": 3.4088970558506717,
      "tokens_seen": 2157838336
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003520808080808081,
      "loss": 1.359,
      "theoretical_loss": 3.408888552109647,
      "tokens_seen": 2157903872
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035206060606060607,
      "loss": 1.4124,
      "theoretical_loss": 3.408880048699189,
      "tokens_seen": 2157969408
    },
    {
      "epoch": 0.3,
      "objective/train/aligned_token_count": 0.4323069453239441,
      "objective/train/avg_token_score": 0.02164054848253727,
      "objective/train/docs_used": 1222265,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3648401498794556,
      "objective/train/original_loss": 4.0131659507751465,
      "objective/train/theoretical_loss": 3.4088715456192755,
      "objective/train/tokens_used": 516894176,
      "objective/train/unlikelihood_loss": 0.05729663372039795,
      "theoretical_loss": 3.4088715456192755,
      "tokens_seen": 2158034944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352040404040404,
      "loss": 1.4769,
      "theoretical_loss": 3.4088715456192755,
      "tokens_seen": 2158034944
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035202020202020203,
      "loss": 1.4269,
      "theoretical_loss": 3.4088630428698825,
      "tokens_seen": 2158100480
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000352,
      "loss": 1.4243,
      "theoretical_loss": 3.4088545404509882,
      "tokens_seen": 2158166016
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000351979797979798,
      "loss": 1.3082,
      "theoretical_loss": 3.408846038362569,
      "tokens_seen": 2158231552
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035195959595959596,
      "loss": 1.6036,
      "theoretical_loss": 3.4088375366046026,
      "tokens_seen": 2158297088
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035193939393939397,
      "loss": 1.5378,
      "theoretical_loss": 3.4088290351770656,
      "tokens_seen": 2158362624
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003519191919191919,
      "loss": 1.5653,
      "theoretical_loss": 3.4088205340799353,
      "tokens_seen": 2158428160
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003518989898989899,
      "loss": 1.6522,
      "theoretical_loss": 3.4088120333131893,
      "tokens_seen": 2158493696
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003518787878787879,
      "loss": 1.4333,
      "theoretical_loss": 3.408803532876804,
      "tokens_seen": 2158559232
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035185858585858585,
      "loss": 1.5149,
      "theoretical_loss": 3.408795032770757,
      "tokens_seen": 2158624768
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035183838383838386,
      "loss": 1.4609,
      "theoretical_loss": 3.408786532995025,
      "tokens_seen": 2158690304
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003518181818181818,
      "loss": 1.4859,
      "theoretical_loss": 3.408778033549585,
      "tokens_seen": 2158755840
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517979797979798,
      "loss": 1.2978,
      "theoretical_loss": 3.4087695344344153,
      "tokens_seen": 2158821376
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517777777777778,
      "loss": 1.3379,
      "theoretical_loss": 3.4087610356494924,
      "tokens_seen": 2158886912
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517575757575758,
      "loss": 1.4553,
      "theoretical_loss": 3.408752537194793,
      "tokens_seen": 2158952448
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035173737373737376,
      "loss": 1.164,
      "theoretical_loss": 3.408744039070294,
      "tokens_seen": 2159017984
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003517171717171717,
      "loss": 1.5104,
      "theoretical_loss": 3.408735541275974,
      "tokens_seen": 2159083520
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035169696969696967,
      "loss": 1.7157,
      "theoretical_loss": 3.408727043811809,
      "tokens_seen": 2159149056
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516767676767677,
      "loss": 1.6672,
      "theoretical_loss": 3.4087185466777763,
      "tokens_seen": 2159214592
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516565656565657,
      "loss": 1.3043,
      "theoretical_loss": 3.4087100498738536,
      "tokens_seen": 2159280128
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035163636363636365,
      "loss": 1.2514,
      "theoretical_loss": 3.408701553400017,
      "tokens_seen": 2159345664
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003516161616161616,
      "loss": 1.5195,
      "theoretical_loss": 3.4086930572562446,
      "tokens_seen": 2159411200
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035159595959595956,
      "loss": 1.5695,
      "theoretical_loss": 3.4086845614425134,
      "tokens_seen": 2159476736
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003515757575757576,
      "loss": 1.4327,
      "theoretical_loss": 3.4086760659588,
      "tokens_seen": 2159542272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003515555555555556,
      "loss": 1.633,
      "theoretical_loss": 3.408667570805082,
      "tokens_seen": 2159607808
    },
    {
      "epoch": 0.3,
      "objective/train/aligned_token_count": 0.4694525897502899,
      "objective/train/avg_token_score": 0.0066945492289960384,
      "objective/train/docs_used": 1223379,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5661883354187012,
      "objective/train/original_loss": 4.237104415893555,
      "objective/train/theoretical_loss": 3.408659075981337,
      "objective/train/tokens_used": 518532576,
      "objective/train/unlikelihood_loss": 0.050316303968429565,
      "theoretical_loss": 3.408659075981337,
      "tokens_seen": 2159673344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035153535353535354,
      "loss": 1.3222,
      "theoretical_loss": 3.408659075981337,
      "tokens_seen": 2159673344
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003515151515151515,
      "loss": 1.4225,
      "theoretical_loss": 3.4086505814875414,
      "tokens_seen": 2159738880
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035149494949494945,
      "loss": 1.571,
      "theoretical_loss": 3.408642087323673,
      "tokens_seen": 2159804416
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514747474747475,
      "loss": 1.6036,
      "theoretical_loss": 3.4086335934897085,
      "tokens_seen": 2159869952
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514545454545455,
      "loss": 1.3161,
      "theoretical_loss": 3.4086250999856254,
      "tokens_seen": 2159935488
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035143434343434343,
      "loss": 1.4086,
      "theoretical_loss": 3.4086166068114006,
      "tokens_seen": 2160001024
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003514141414141414,
      "loss": 1.5212,
      "theoretical_loss": 3.408608113967011,
      "tokens_seen": 2160066560
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513939393939394,
      "loss": 1.5512,
      "theoretical_loss": 3.408599621452435,
      "tokens_seen": 2160132096
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513737373737374,
      "loss": 1.6539,
      "theoretical_loss": 3.4085911292676485,
      "tokens_seen": 2160197632
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035135353535353537,
      "loss": 1.5601,
      "theoretical_loss": 3.4085826374126293,
      "tokens_seen": 2160263168
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513333333333333,
      "loss": 1.4409,
      "theoretical_loss": 3.4085741458873544,
      "tokens_seen": 2160328704
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003513131313131313,
      "loss": 1.2464,
      "theoretical_loss": 3.408565654691801,
      "tokens_seen": 2160394240
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512929292929293,
      "loss": 1.1461,
      "theoretical_loss": 3.408557163825946,
      "tokens_seen": 2160459776
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512727272727273,
      "loss": 1.2523,
      "theoretical_loss": 3.4085486732897676,
      "tokens_seen": 2160525312
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035125252525252526,
      "loss": 1.363,
      "theoretical_loss": 3.4085401830832422,
      "tokens_seen": 2160590848
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003512323232323232,
      "loss": 1.4663,
      "theoretical_loss": 3.408531693206347,
      "tokens_seen": 2160656384
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035121212121212123,
      "loss": 1.4173,
      "theoretical_loss": 3.408523203659059,
      "tokens_seen": 2160721920
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511919191919192,
      "loss": 1.5237,
      "theoretical_loss": 3.4085147144413566,
      "tokens_seen": 2160787456
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511717171717172,
      "loss": 1.2192,
      "theoretical_loss": 3.4085062255532157,
      "tokens_seen": 2160852992
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035115151515151515,
      "loss": 1.4062,
      "theoretical_loss": 3.408497736994614,
      "tokens_seen": 2160918528
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511313131313131,
      "loss": 1.4445,
      "theoretical_loss": 3.4084892487655285,
      "tokens_seen": 2160984064
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003511111111111111,
      "loss": 1.4997,
      "theoretical_loss": 3.408480760865937,
      "tokens_seen": 2161049600
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003510909090909091,
      "loss": 1.5424,
      "theoretical_loss": 3.408472273295816,
      "tokens_seen": 2161115136
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0003510707070707071,
      "loss": 1.3217,
      "theoretical_loss": 3.408463786055143,
      "tokens_seen": 2161180672
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035105050505050505,
      "loss": 1.2627,
      "theoretical_loss": 3.408455299143896,
      "tokens_seen": 2161246208
    },
    {
      "epoch": 0.3,
      "objective/train/aligned_token_count": 0.34714075922966003,
      "objective/train/avg_token_score": 0.013145077042281628,
      "objective/train/docs_used": 1223894,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0526677370071411,
      "objective/train/original_loss": 4.279311180114746,
      "objective/train/theoretical_loss": 3.408446812562051,
      "objective/train/tokens_used": 520170976,
      "objective/train/unlikelihood_loss": 0.0752415880560875,
      "theoretical_loss": 3.408446812562051,
      "tokens_seen": 2161311744
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00035103030303030306,
      "loss": 1.3306,
      "theoretical_loss": 3.408446812562051,
      "tokens_seen": 2161311744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000351010101010101,
      "loss": 1.3538,
      "theoretical_loss": 3.4084383263095854,
      "tokens_seen": 2161377280
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035098989898989897,
      "loss": 1.5626,
      "theoretical_loss": 3.4084298403864772,
      "tokens_seen": 2161442816
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000350969696969697,
      "loss": 1.1488,
      "theoretical_loss": 3.408421354792703,
      "tokens_seen": 2161508352
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035094949494949494,
      "loss": 1.5153,
      "theoretical_loss": 3.4084128695282403,
      "tokens_seen": 2161573888
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035092929292929295,
      "loss": 1.4964,
      "theoretical_loss": 3.408404384593066,
      "tokens_seen": 2161639424
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003509090909090909,
      "loss": 1.4712,
      "theoretical_loss": 3.4083958999871578,
      "tokens_seen": 2161704960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035088888888888886,
      "loss": 1.3121,
      "theoretical_loss": 3.408387415710493,
      "tokens_seen": 2161770496
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508686868686869,
      "loss": 1.3962,
      "theoretical_loss": 3.408378931763048,
      "tokens_seen": 2161836032
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508484848484849,
      "loss": 1.4988,
      "theoretical_loss": 3.4083704481448014,
      "tokens_seen": 2161901568
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035082828282828284,
      "loss": 1.22,
      "theoretical_loss": 3.408361964855729,
      "tokens_seen": 2161967104
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003508080808080808,
      "loss": 1.5712,
      "theoretical_loss": 3.408353481895809,
      "tokens_seen": 2162032640
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003507878787878788,
      "loss": 1.311,
      "theoretical_loss": 3.4083449992650183,
      "tokens_seen": 2162098176
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035076767676767677,
      "loss": 1.4028,
      "theoretical_loss": 3.408336516963334,
      "tokens_seen": 2162163712
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003507474747474748,
      "loss": 1.351,
      "theoretical_loss": 3.408328034990734,
      "tokens_seen": 2162229248
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035072727272727274,
      "loss": 1.3889,
      "theoretical_loss": 3.4083195533471953,
      "tokens_seen": 2162294784
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003507070707070707,
      "loss": 1.4702,
      "theoretical_loss": 3.4083110720326943,
      "tokens_seen": 2162360320
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003506868686868687,
      "loss": 1.4916,
      "theoretical_loss": 3.4083025910472093,
      "tokens_seen": 2162425856
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003506666666666667,
      "loss": 1.5523,
      "theoretical_loss": 3.4082941103907176,
      "tokens_seen": 2162491392
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035064646464646467,
      "loss": 1.5491,
      "theoretical_loss": 3.4082856300631956,
      "tokens_seen": 2162556928
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035062626262626263,
      "loss": 1.5712,
      "theoretical_loss": 3.408277150064621,
      "tokens_seen": 2162622464
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003506060606060606,
      "loss": 1.5236,
      "theoretical_loss": 3.4082686703949716,
      "tokens_seen": 2162688000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505858585858586,
      "loss": 1.3416,
      "theoretical_loss": 3.4082601910542243,
      "tokens_seen": 2162753536
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505656565656566,
      "loss": 1.5501,
      "theoretical_loss": 3.4082517120423557,
      "tokens_seen": 2162819072
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035054545454545457,
      "loss": 1.3997,
      "theoretical_loss": 3.408243233359344,
      "tokens_seen": 2162884608
    },
    {
      "epoch": 0.31,
      "objective/train/aligned_token_count": 0.5213831663131714,
      "objective/train/avg_token_score": 0.011658943258225918,
      "objective/train/docs_used": 1225280,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.716713309288025,
      "objective/train/original_loss": 3.8835885524749756,
      "objective/train/theoretical_loss": 3.408234755005166,
      "objective/train/tokens_used": 521809376,
      "objective/train/unlikelihood_loss": 0.06082100793719292,
      "theoretical_loss": 3.408234755005166,
      "tokens_seen": 2162950144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505252525252525,
      "loss": 1.6555,
      "theoretical_loss": 3.408234755005166,
      "tokens_seen": 2162950144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003505050505050505,
      "loss": 1.5661,
      "theoretical_loss": 3.4082262769797995,
      "tokens_seen": 2163015680
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035048484848484854,
      "loss": 1.3639,
      "theoretical_loss": 3.4082177992832214,
      "tokens_seen": 2163081216
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003504646464646465,
      "loss": 1.3825,
      "theoretical_loss": 3.408209321915409,
      "tokens_seen": 2163146752
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035044444444444446,
      "loss": 1.1873,
      "theoretical_loss": 3.4082008448763395,
      "tokens_seen": 2163212288
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003504242424242424,
      "loss": 1.7825,
      "theoretical_loss": 3.40819236816599,
      "tokens_seen": 2163277824
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035040404040404037,
      "loss": 1.491,
      "theoretical_loss": 3.408183891784338,
      "tokens_seen": 2163343360
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035038383838383844,
      "loss": 1.6984,
      "theoretical_loss": 3.4081754157313617,
      "tokens_seen": 2163408896
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003503636363636364,
      "loss": 1.4559,
      "theoretical_loss": 3.408166940007037,
      "tokens_seen": 2163474432
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035034343434343435,
      "loss": 1.4566,
      "theoretical_loss": 3.408158464611342,
      "tokens_seen": 2163539968
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003503232323232323,
      "loss": 1.3875,
      "theoretical_loss": 3.4081499895442535,
      "tokens_seen": 2163605504
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035030303030303026,
      "loss": 1.3506,
      "theoretical_loss": 3.4081415148057497,
      "tokens_seen": 2163671040
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035028282828282833,
      "loss": 1.4746,
      "theoretical_loss": 3.4081330403958066,
      "tokens_seen": 2163736576
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003502626262626263,
      "loss": 1.3826,
      "theoretical_loss": 3.4081245663144024,
      "tokens_seen": 2163802112
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035024242424242424,
      "loss": 1.4831,
      "theoretical_loss": 3.4081160925615146,
      "tokens_seen": 2163867648
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003502222222222222,
      "loss": 1.5708,
      "theoretical_loss": 3.40810761913712,
      "tokens_seen": 2163933184
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003502020202020202,
      "loss": 1.4254,
      "theoretical_loss": 3.408099146041196,
      "tokens_seen": 2163998720
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501818181818182,
      "loss": 1.3181,
      "theoretical_loss": 3.40809067327372,
      "tokens_seen": 2164064256
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501616161616162,
      "loss": 1.3429,
      "theoretical_loss": 3.408082200834669,
      "tokens_seen": 2164129792
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035014141414141414,
      "loss": 1.3053,
      "theoretical_loss": 3.408073728724021,
      "tokens_seen": 2164195328
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501212121212121,
      "loss": 1.5331,
      "theoretical_loss": 3.408065256941753,
      "tokens_seen": 2164260864
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003501010101010101,
      "loss": 1.3846,
      "theoretical_loss": 3.408056785487842,
      "tokens_seen": 2164326400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003500808080808081,
      "loss": 1.35,
      "theoretical_loss": 3.408048314362266,
      "tokens_seen": 2164391936
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035006060606060607,
      "loss": 1.4042,
      "theoretical_loss": 3.408039843565002,
      "tokens_seen": 2164457472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035004040404040403,
      "loss": 1.3902,
      "theoretical_loss": 3.4080313730960268,
      "tokens_seen": 2164523008
    },
    {
      "epoch": 0.31,
      "objective/train/aligned_token_count": 0.5018328428268433,
      "objective/train/avg_token_score": 0.010568901896476746,
      "objective/train/docs_used": 1225734,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.518566608428955,
      "objective/train/original_loss": 3.887605667114258,
      "objective/train/theoretical_loss": 3.4080229029553184,
      "objective/train/tokens_used": 523447776,
      "objective/train/unlikelihood_loss": 0.06365974247455597,
      "theoretical_loss": 3.4080229029553184,
      "tokens_seen": 2164588544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035002020202020204,
      "loss": 1.4157,
      "theoretical_loss": 3.4080229029553184,
      "tokens_seen": 2164588544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00035,
      "loss": 1.5878,
      "theoretical_loss": 3.408014433142854,
      "tokens_seen": 2164654080
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000349979797979798,
      "loss": 1.4493,
      "theoretical_loss": 3.408005963658611,
      "tokens_seen": 2164719616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034995959595959596,
      "loss": 1.4054,
      "theoretical_loss": 3.407997494502567,
      "tokens_seen": 2164785152
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003499393939393939,
      "loss": 1.4806,
      "theoretical_loss": 3.4079890256746985,
      "tokens_seen": 2164850688
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034991919191919193,
      "loss": 1.7433,
      "theoretical_loss": 3.4079805571749833,
      "tokens_seen": 2164916224
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498989898989899,
      "loss": 1.5107,
      "theoretical_loss": 3.4079720890033993,
      "tokens_seen": 2164981760
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498787878787879,
      "loss": 1.4595,
      "theoretical_loss": 3.4079636211599227,
      "tokens_seen": 2165047296
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034985858585858586,
      "loss": 1.5988,
      "theoretical_loss": 3.407955153644532,
      "tokens_seen": 2165112832
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034983838383838387,
      "loss": 1.3644,
      "theoretical_loss": 3.407946686457204,
      "tokens_seen": 2165178368
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003498181818181818,
      "loss": 1.316,
      "theoretical_loss": 3.407938219597916,
      "tokens_seen": 2165243904
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003497979797979798,
      "loss": 1.5045,
      "theoretical_loss": 3.4079297530666457,
      "tokens_seen": 2165309440
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003497777777777778,
      "loss": 1.2512,
      "theoretical_loss": 3.40792128686337,
      "tokens_seen": 2165374976
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034975757575757575,
      "loss": 1.6495,
      "theoretical_loss": 3.4079128209880665,
      "tokens_seen": 2165440512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034973737373737376,
      "loss": 1.3468,
      "theoretical_loss": 3.4079043554407127,
      "tokens_seen": 2165506048
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003497171717171717,
      "loss": 1.5198,
      "theoretical_loss": 3.407895890221286,
      "tokens_seen": 2165571584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003496969696969697,
      "loss": 1.3962,
      "theoretical_loss": 3.4078874253297635,
      "tokens_seen": 2165637120
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003496767676767677,
      "loss": 1.2735,
      "theoretical_loss": 3.407878960766123,
      "tokens_seen": 2165702656
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003496565656565657,
      "loss": 1.6461,
      "theoretical_loss": 3.407870496530341,
      "tokens_seen": 2165768192
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034963636363636365,
      "loss": 1.5251,
      "theoretical_loss": 3.4078620326223956,
      "tokens_seen": 2165833728
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003496161616161616,
      "loss": 1.402,
      "theoretical_loss": 3.407853569042264,
      "tokens_seen": 2165899264
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034959595959595957,
      "loss": 1.1895,
      "theoretical_loss": 3.4078451057899244,
      "tokens_seen": 2165964800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003495757575757576,
      "loss": 1.4859,
      "theoretical_loss": 3.4078366428653526,
      "tokens_seen": 2166030336
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003495555555555556,
      "loss": 1.4303,
      "theoretical_loss": 3.4078281802685275,
      "tokens_seen": 2166095872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034953535353535355,
      "loss": 1.6138,
      "theoretical_loss": 3.407819717999425,
      "tokens_seen": 2166161408
    },
    {
      "epoch": 0.31,
      "objective/train/aligned_token_count": 0.3477517068386078,
      "objective/train/avg_token_score": 0.01290730107575655,
      "objective/train/docs_used": 1227178,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1927399635314941,
      "objective/train/original_loss": 4.659249782562256,
      "objective/train/theoretical_loss": 3.4078112560580243,
      "objective/train/tokens_used": 525086176,
      "objective/train/unlikelihood_loss": 0.06020481511950493,
      "theoretical_loss": 3.4078112560580243,
      "tokens_seen": 2166226944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003495151515151515,
      "loss": 1.3426,
      "theoretical_loss": 3.4078112560580243,
      "tokens_seen": 2166226944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034949494949494946,
      "loss": 1.4095,
      "theoretical_loss": 3.4078027944443012,
      "tokens_seen": 2166292480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494747474747475,
      "loss": 1.4766,
      "theoretical_loss": 3.407794333158234,
      "tokens_seen": 2166358016
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494545454545455,
      "loss": 1.3463,
      "theoretical_loss": 3.407785872199799,
      "tokens_seen": 2166423552
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034943434343434344,
      "loss": 1.2822,
      "theoretical_loss": 3.407777411568975,
      "tokens_seen": 2166489088
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003494141414141414,
      "loss": 1.196,
      "theoretical_loss": 3.407768951265739,
      "tokens_seen": 2166554624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034939393939393935,
      "loss": 1.2755,
      "theoretical_loss": 3.407760491290068,
      "tokens_seen": 2166620160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003493737373737374,
      "loss": 1.5409,
      "theoretical_loss": 3.4077520316419396,
      "tokens_seen": 2166685696
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003493535353535354,
      "loss": 1.4348,
      "theoretical_loss": 3.4077435723213316,
      "tokens_seen": 2166751232
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034933333333333333,
      "loss": 1.3459,
      "theoretical_loss": 3.4077351133282203,
      "tokens_seen": 2166816768
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003493131313131313,
      "loss": 1.5812,
      "theoretical_loss": 3.407726654662585,
      "tokens_seen": 2166882304
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003492929292929293,
      "loss": 1.5209,
      "theoretical_loss": 3.4077181963244008,
      "tokens_seen": 2166947840
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003492727272727273,
      "loss": 1.5643,
      "theoretical_loss": 3.4077097383136468,
      "tokens_seen": 2167013376
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034925252525252527,
      "loss": 1.315,
      "theoretical_loss": 3.4077012806302998,
      "tokens_seen": 2167078912
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003492323232323232,
      "loss": 1.5792,
      "theoretical_loss": 3.4076928232743375,
      "tokens_seen": 2167144448
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003492121212121212,
      "loss": 1.3122,
      "theoretical_loss": 3.4076843662457375,
      "tokens_seen": 2167209984
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003491919191919192,
      "loss": 1.2635,
      "theoretical_loss": 3.4076759095444764,
      "tokens_seen": 2167275520
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003491717171717172,
      "loss": 1.408,
      "theoretical_loss": 3.407667453170532,
      "tokens_seen": 2167341056
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034915151515151516,
      "loss": 1.5328,
      "theoretical_loss": 3.4076589971238826,
      "tokens_seen": 2167406592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003491313131313131,
      "loss": 1.7338,
      "theoretical_loss": 3.4076505414045046,
      "tokens_seen": 2167472128
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034911111111111113,
      "loss": 1.5082,
      "theoretical_loss": 3.4076420860123755,
      "tokens_seen": 2167537664
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034909090909090914,
      "loss": 1.3222,
      "theoretical_loss": 3.4076336309474735,
      "tokens_seen": 2167603200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003490707070707071,
      "loss": 1.3801,
      "theoretical_loss": 3.407625176209775,
      "tokens_seen": 2167668736
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034905050505050505,
      "loss": 1.5104,
      "theoretical_loss": 3.407616721799258,
      "tokens_seen": 2167734272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000349030303030303,
      "loss": 1.3728,
      "theoretical_loss": 3.4076082677159003,
      "tokens_seen": 2167799808
    },
    {
      "epoch": 0.31,
      "objective/train/aligned_token_count": 0.41104593873023987,
      "objective/train/avg_token_score": 0.0095455227419734,
      "objective/train/docs_used": 1227923,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2222384214401245,
      "objective/train/original_loss": 4.229890823364258,
      "objective/train/theoretical_loss": 3.4075998139596786,
      "objective/train/tokens_used": 526724576,
      "objective/train/unlikelihood_loss": 0.07498421519994736,
      "theoretical_loss": 3.4075998139596786,
      "tokens_seen": 2167865344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000349010101010101,
      "loss": 1.4891,
      "theoretical_loss": 3.4075998139596786,
      "tokens_seen": 2167865344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034898989898989903,
      "loss": 1.4768,
      "theoretical_loss": 3.407591360530571,
      "tokens_seen": 2167930880
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000348969696969697,
      "loss": 1.1792,
      "theoretical_loss": 3.4075829074285546,
      "tokens_seen": 2167996416
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034894949494949495,
      "loss": 1.3942,
      "theoretical_loss": 3.4075744546536075,
      "tokens_seen": 2168061952
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034892929292929296,
      "loss": 1.4931,
      "theoretical_loss": 3.407566002205706,
      "tokens_seen": 2168127488
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003489090909090909,
      "loss": 1.4215,
      "theoretical_loss": 3.407557550084828,
      "tokens_seen": 2168193024
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003488888888888889,
      "loss": 1.157,
      "theoretical_loss": 3.4075490982909518,
      "tokens_seen": 2168258560
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003488686868686869,
      "loss": 1.3789,
      "theoretical_loss": 3.4075406468240534,
      "tokens_seen": 2168324096
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034884848484848484,
      "loss": 1.4203,
      "theoretical_loss": 3.4075321956841114,
      "tokens_seen": 2168389632
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034882828282828285,
      "loss": 1.4127,
      "theoretical_loss": 3.4075237448711033,
      "tokens_seen": 2168455168
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003488080808080808,
      "loss": 1.484,
      "theoretical_loss": 3.4075152943850058,
      "tokens_seen": 2168520704
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003487878787878788,
      "loss": 1.6545,
      "theoretical_loss": 3.407506844225797,
      "tokens_seen": 2168586240
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003487676767676768,
      "loss": 1.5071,
      "theoretical_loss": 3.4074983943934543,
      "tokens_seen": 2168651776
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034874747474747473,
      "loss": 1.5304,
      "theoretical_loss": 3.4074899448879545,
      "tokens_seen": 2168717312
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034872727272727274,
      "loss": 1.4292,
      "theoretical_loss": 3.407481495709276,
      "tokens_seen": 2168782848
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003487070707070707,
      "loss": 1.387,
      "theoretical_loss": 3.4074730468573957,
      "tokens_seen": 2168848384
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003486868686868687,
      "loss": 1.7382,
      "theoretical_loss": 3.4074645983322913,
      "tokens_seen": 2168913920
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034866666666666667,
      "loss": 1.4861,
      "theoretical_loss": 3.4074561501339407,
      "tokens_seen": 2168979456
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003486464646464647,
      "loss": 1.4377,
      "theoretical_loss": 3.4074477022623206,
      "tokens_seen": 2169044992
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034862626262626263,
      "loss": 1.3233,
      "theoretical_loss": 3.407439254717409,
      "tokens_seen": 2169110528
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003486060606060606,
      "loss": 1.5383,
      "theoretical_loss": 3.407430807499183,
      "tokens_seen": 2169176064
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003485858585858586,
      "loss": 1.2294,
      "theoretical_loss": 3.4074223606076206,
      "tokens_seen": 2169241600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034856565656565656,
      "loss": 1.4041,
      "theoretical_loss": 3.407413914042699,
      "tokens_seen": 2169307136
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034854545454545457,
      "loss": 1.5116,
      "theoretical_loss": 3.407405467804396,
      "tokens_seen": 2169372672
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034852525252525253,
      "loss": 1.4211,
      "theoretical_loss": 3.407397021892688,
      "tokens_seen": 2169438208
    },
    {
      "epoch": 0.31,
      "objective/train/aligned_token_count": 0.4577834904193878,
      "objective/train/avg_token_score": 0.005549462512135506,
      "objective/train/docs_used": 1229408,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4827572107315063,
      "objective/train/original_loss": 3.7893826961517334,
      "objective/train/theoretical_loss": 3.407388576307554,
      "objective/train/tokens_used": 528362976,
      "objective/train/unlikelihood_loss": 0.0773424282670021,
      "theoretical_loss": 3.407388576307554,
      "tokens_seen": 2169503744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003485050505050505,
      "loss": 1.4088,
      "theoretical_loss": 3.407388576307554,
      "tokens_seen": 2169503744
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484848484848485,
      "loss": 1.5524,
      "theoretical_loss": 3.407380131048971,
      "tokens_seen": 2169569280
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484646464646465,
      "loss": 1.3356,
      "theoretical_loss": 3.407371686116916,
      "tokens_seen": 2169634816
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034844444444444446,
      "loss": 1.4496,
      "theoretical_loss": 3.4073632415113675,
      "tokens_seen": 2169700352
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484242424242424,
      "loss": 1.3436,
      "theoretical_loss": 3.4073547972323017,
      "tokens_seen": 2169765888
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003484040404040404,
      "loss": 1.4492,
      "theoretical_loss": 3.407346353279697,
      "tokens_seen": 2169831424
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003483838383838384,
      "loss": 1.4561,
      "theoretical_loss": 3.4073379096535312,
      "tokens_seen": 2169896960
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003483636363636364,
      "loss": 1.5412,
      "theoretical_loss": 3.4073294663537808,
      "tokens_seen": 2169962496
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034834343434343436,
      "loss": 1.3392,
      "theoretical_loss": 3.4073210233804243,
      "tokens_seen": 2170028032
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003483232323232323,
      "loss": 1.3465,
      "theoretical_loss": 3.4073125807334383,
      "tokens_seen": 2170093568
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034830303030303027,
      "loss": 1.4699,
      "theoretical_loss": 3.407304138412801,
      "tokens_seen": 2170159104
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034828282828282834,
      "loss": 1.1744,
      "theoretical_loss": 3.4072956964184904,
      "tokens_seen": 2170224640
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003482626262626263,
      "loss": 1.5251,
      "theoretical_loss": 3.4072872547504827,
      "tokens_seen": 2170290176
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034824242424242425,
      "loss": 1.4806,
      "theoretical_loss": 3.4072788134087566,
      "tokens_seen": 2170355712
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003482222222222222,
      "loss": 1.5599,
      "theoretical_loss": 3.407270372393289,
      "tokens_seen": 2170421248
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034820202020202016,
      "loss": 1.3233,
      "theoretical_loss": 3.4072619317040576,
      "tokens_seen": 2170486784
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034818181818181823,
      "loss": 1.2296,
      "theoretical_loss": 3.4072534913410397,
      "tokens_seen": 2170552320
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003481616161616162,
      "loss": 1.4637,
      "theoretical_loss": 3.4072450513042134,
      "tokens_seen": 2170617856
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034814141414141414,
      "loss": 1.4726,
      "theoretical_loss": 3.407236611593556,
      "tokens_seen": 2170683392
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003481212121212121,
      "loss": 1.3315,
      "theoretical_loss": 3.4072281722090447,
      "tokens_seen": 2170748928
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003481010101010101,
      "loss": 1.4103,
      "theoretical_loss": 3.4072197331506575,
      "tokens_seen": 2170814464
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003480808080808081,
      "loss": 1.3799,
      "theoretical_loss": 3.407211294418372,
      "tokens_seen": 2170880000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003480606060606061,
      "loss": 1.6105,
      "theoretical_loss": 3.4072028560121654,
      "tokens_seen": 2170945536
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034804040404040403,
      "loss": 1.4788,
      "theoretical_loss": 3.4071944179320153,
      "tokens_seen": 2171011072
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000348020202020202,
      "loss": 1.2984,
      "theoretical_loss": 3.4071859801778994,
      "tokens_seen": 2171076608
    },
    {
      "epoch": 0.31,
      "objective/train/aligned_token_count": 0.3588709533214569,
      "objective/train/avg_token_score": 0.02379283681511879,
      "objective/train/docs_used": 1230033,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1464951038360596,
      "objective/train/original_loss": 4.28645658493042,
      "objective/train/theoretical_loss": 3.4071775427497952,
      "objective/train/tokens_used": 530001376,
      "objective/train/unlikelihood_loss": 0.05953194946050644,
      "theoretical_loss": 3.4071775427497952,
      "tokens_seen": 2171142144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000348,
      "loss": 1.4746,
      "theoretical_loss": 3.4071775427497952,
      "tokens_seen": 2171142144
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000347979797979798,
      "loss": 1.3192,
      "theoretical_loss": 3.4071691056476805,
      "tokens_seen": 2171207680
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034795959595959597,
      "loss": 1.4024,
      "theoretical_loss": 3.4071606688715326,
      "tokens_seen": 2171273216
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003479393939393939,
      "loss": 1.1892,
      "theoretical_loss": 3.4071522324213293,
      "tokens_seen": 2171338752
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034791919191919194,
      "loss": 1.1662,
      "theoretical_loss": 3.4071437962970474,
      "tokens_seen": 2171404288
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003478989898989899,
      "loss": 1.4422,
      "theoretical_loss": 3.407135360498666,
      "tokens_seen": 2171469824
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003478787878787879,
      "loss": 1.2939,
      "theoretical_loss": 3.4071269250261604,
      "tokens_seen": 2171535360
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034785858585858586,
      "loss": 1.4748,
      "theoretical_loss": 3.407118489879511,
      "tokens_seen": 2171600896
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003478383838383838,
      "loss": 1.4642,
      "theoretical_loss": 3.407110055058693,
      "tokens_seen": 2171666432
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034781818181818183,
      "loss": 1.2982,
      "theoretical_loss": 3.407101620563685,
      "tokens_seen": 2171731968
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003477979797979798,
      "loss": 1.3944,
      "theoretical_loss": 3.4070931863944645,
      "tokens_seen": 2171797504
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003477777777777778,
      "loss": 1.2743,
      "theoretical_loss": 3.4070847525510093,
      "tokens_seen": 2171863040
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034775757575757576,
      "loss": 1.6257,
      "theoretical_loss": 3.4070763190332967,
      "tokens_seen": 2171928576
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034773737373737377,
      "loss": 1.3762,
      "theoretical_loss": 3.407067885841304,
      "tokens_seen": 2171994112
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003477171717171717,
      "loss": 1.5096,
      "theoretical_loss": 3.4070594529750093,
      "tokens_seen": 2172059648
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003476969696969697,
      "loss": 1.3406,
      "theoretical_loss": 3.4070510204343902,
      "tokens_seen": 2172125184
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003476767676767677,
      "loss": 1.3564,
      "theoretical_loss": 3.407042588219424,
      "tokens_seen": 2172190720
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034765656565656565,
      "loss": 1.2878,
      "theoretical_loss": 3.407034156330088,
      "tokens_seen": 2172256256
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034763636363636366,
      "loss": 1.365,
      "theoretical_loss": 3.4070257247663607,
      "tokens_seen": 2172321792
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003476161616161616,
      "loss": 1.389,
      "theoretical_loss": 3.4070172935282192,
      "tokens_seen": 2172387328
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034759595959595957,
      "loss": 1.4986,
      "theoretical_loss": 3.407008862615641,
      "tokens_seen": 2172452864
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003475757575757576,
      "loss": 1.5085,
      "theoretical_loss": 3.4070004320286036,
      "tokens_seen": 2172518400
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003475555555555556,
      "loss": 1.4882,
      "theoretical_loss": 3.4069920017670854,
      "tokens_seen": 2172583936
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034753535353535355,
      "loss": 1.5162,
      "theoretical_loss": 3.4069835718310633,
      "tokens_seen": 2172649472
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003475151515151515,
      "loss": 1.3297,
      "theoretical_loss": 3.4069751422205146,
      "tokens_seen": 2172715008
    },
    {
      "epoch": 0.31,
      "objective/train/aligned_token_count": 0.4213709533214569,
      "objective/train/avg_token_score": 0.008849565871059895,
      "objective/train/docs_used": 1231439,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.314804196357727,
      "objective/train/original_loss": 3.911224603652954,
      "objective/train/theoretical_loss": 3.406966712935418,
      "objective/train/tokens_used": 531639776,
      "objective/train/unlikelihood_loss": 0.05618034675717354,
      "theoretical_loss": 3.406966712935418,
      "tokens_seen": 2172780544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034749494949494947,
      "loss": 1.5124,
      "theoretical_loss": 3.406966712935418,
      "tokens_seen": 2172780544
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474747474747475,
      "loss": 1.4383,
      "theoretical_loss": 3.40695828397575,
      "tokens_seen": 2172846080
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474545454545455,
      "loss": 1.4544,
      "theoretical_loss": 3.406949855341489,
      "tokens_seen": 2172911616
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034743434343434344,
      "loss": 1.217,
      "theoretical_loss": 3.4069414270326126,
      "tokens_seen": 2172977152
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003474141414141414,
      "loss": 1.6647,
      "theoretical_loss": 3.4069329990490975,
      "tokens_seen": 2173042688
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003473939393939394,
      "loss": 1.4098,
      "theoretical_loss": 3.4069245713909226,
      "tokens_seen": 2173108224
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003473737373737374,
      "loss": 1.1227,
      "theoretical_loss": 3.406916144058065,
      "tokens_seen": 2173173760
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003473535353535354,
      "loss": 1.6007,
      "theoretical_loss": 3.4069077170505015,
      "tokens_seen": 2173239296
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034733333333333334,
      "loss": 1.4299,
      "theoretical_loss": 3.4068992903682114,
      "tokens_seen": 2173304832
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003473131313131313,
      "loss": 1.255,
      "theoretical_loss": 3.406890864011171,
      "tokens_seen": 2173370368
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003472929292929293,
      "loss": 1.4924,
      "theoretical_loss": 3.406882437979358,
      "tokens_seen": 2173435904
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003472727272727273,
      "loss": 1.5084,
      "theoretical_loss": 3.406874012272751,
      "tokens_seen": 2173501440
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003472525252525253,
      "loss": 1.5913,
      "theoretical_loss": 3.4068655868913273,
      "tokens_seen": 2173566976
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034723232323232323,
      "loss": 1.4332,
      "theoretical_loss": 3.406857161835064,
      "tokens_seen": 2173632512
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003472121212121212,
      "loss": 1.383,
      "theoretical_loss": 3.406848737103939,
      "tokens_seen": 2173698048
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034719191919191925,
      "loss": 1.2416,
      "theoretical_loss": 3.4068403126979296,
      "tokens_seen": 2173763584
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003471717171717172,
      "loss": 1.3397,
      "theoretical_loss": 3.406831888617014,
      "tokens_seen": 2173829120
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034715151515151517,
      "loss": 1.4482,
      "theoretical_loss": 3.40682346486117,
      "tokens_seen": 2173894656
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003471313131313131,
      "loss": 1.5174,
      "theoretical_loss": 3.4068150414303755,
      "tokens_seen": 2173960192
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003471111111111111,
      "loss": 1.6404,
      "theoretical_loss": 3.406806618324607,
      "tokens_seen": 2174025728
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034709090909090914,
      "loss": 1.2623,
      "theoretical_loss": 3.4067981955438427,
      "tokens_seen": 2174091264
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003470707070707071,
      "loss": 1.4303,
      "theoretical_loss": 3.4067897730880605,
      "tokens_seen": 2174156800
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034705050505050506,
      "loss": 1.506,
      "theoretical_loss": 3.406781350957238,
      "tokens_seen": 2174222336
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000347030303030303,
      "loss": 1.448,
      "theoretical_loss": 3.406772929151353,
      "tokens_seen": 2174287872
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034701010101010097,
      "loss": 1.2877,
      "theoretical_loss": 3.4067645076703825,
      "tokens_seen": 2174353408
    },
    {
      "epoch": 0.31,
      "objective/train/aligned_token_count": 0.3757942318916321,
      "objective/train/avg_token_score": 0.03156242519617081,
      "objective/train/docs_used": 1232037,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2827579975128174,
      "objective/train/original_loss": 4.592061519622803,
      "objective/train/theoretical_loss": 3.4067560865143047,
      "objective/train/tokens_used": 533278176,
      "objective/train/unlikelihood_loss": 0.05171503126621246,
      "theoretical_loss": 3.4067560865143047,
      "tokens_seen": 2174418944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034698989898989904,
      "loss": 1.3391,
      "theoretical_loss": 3.4067560865143047,
      "tokens_seen": 2174418944
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000346969696969697,
      "loss": 1.3103,
      "theoretical_loss": 3.4067476656830973,
      "tokens_seen": 2174484480
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034694949494949495,
      "loss": 1.333,
      "theoretical_loss": 3.406739245176738,
      "tokens_seen": 2174550016
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003469292929292929,
      "loss": 1.4198,
      "theoretical_loss": 3.406730824995204,
      "tokens_seen": 2174615552
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003469090909090909,
      "loss": 1.2975,
      "theoretical_loss": 3.406722405138474,
      "tokens_seen": 2174681088
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034688888888888893,
      "loss": 1.5594,
      "theoretical_loss": 3.4067139856065247,
      "tokens_seen": 2174746624
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003468686868686869,
      "loss": 1.6118,
      "theoretical_loss": 3.406705566399334,
      "tokens_seen": 2174812160
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034684848484848484,
      "loss": 1.5357,
      "theoretical_loss": 3.40669714751688,
      "tokens_seen": 2174877696
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003468282828282828,
      "loss": 1.4127,
      "theoretical_loss": 3.4066887289591397,
      "tokens_seen": 2174943232
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003468080808080808,
      "loss": 1.2804,
      "theoretical_loss": 3.4066803107260912,
      "tokens_seen": 2175008768
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003467878787878788,
      "loss": 1.3672,
      "theoretical_loss": 3.4066718928177124,
      "tokens_seen": 2175074304
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003467676767676768,
      "loss": 1.3516,
      "theoretical_loss": 3.4066634752339806,
      "tokens_seen": 2175139840
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034674747474747474,
      "loss": 1.4755,
      "theoretical_loss": 3.4066550579748736,
      "tokens_seen": 2175205376
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034672727272727275,
      "loss": 1.5279,
      "theoretical_loss": 3.4066466410403695,
      "tokens_seen": 2175270912
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003467070707070707,
      "loss": 1.4523,
      "theoretical_loss": 3.4066382244304454,
      "tokens_seen": 2175336448
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003466868686868687,
      "loss": 1.5743,
      "theoretical_loss": 3.4066298081450794,
      "tokens_seen": 2175401984
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034666666666666667,
      "loss": 1.4397,
      "theoretical_loss": 3.406621392184249,
      "tokens_seen": 2175467520
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034664646464646463,
      "loss": 1.4411,
      "theoretical_loss": 3.4066129765479323,
      "tokens_seen": 2175533056
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034662626262626264,
      "loss": 1.4024,
      "theoretical_loss": 3.406604561236106,
      "tokens_seen": 2175598592
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003466060606060606,
      "loss": 1.4038,
      "theoretical_loss": 3.406596146248749,
      "tokens_seen": 2175664128
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003465858585858586,
      "loss": 1.5053,
      "theoretical_loss": 3.4065877315858386,
      "tokens_seen": 2175729664
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034656565656565657,
      "loss": 1.3374,
      "theoretical_loss": 3.4065793172473526,
      "tokens_seen": 2175795200
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003465454545454546,
      "loss": 1.398,
      "theoretical_loss": 3.406570903233268,
      "tokens_seen": 2175860736
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034652525252525253,
      "loss": 1.4793,
      "theoretical_loss": 3.4065624895435636,
      "tokens_seen": 2175926272
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003465050505050505,
      "loss": 1.4454,
      "theoretical_loss": 3.4065540761782165,
      "tokens_seen": 2175991808
    },
    {
      "epoch": 0.31,
      "objective/train/aligned_token_count": 0.4927908182144165,
      "objective/train/avg_token_score": 0.012171359732747078,
      "objective/train/docs_used": 1233356,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5692194700241089,
      "objective/train/original_loss": 3.8782966136932373,
      "objective/train/theoretical_loss": 3.406545663137204,
      "objective/train/tokens_used": 534916576,
      "objective/train/unlikelihood_loss": 0.07065261155366898,
      "theoretical_loss": 3.406545663137204,
      "tokens_seen": 2176057344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003464848484848485,
      "loss": 1.4187,
      "theoretical_loss": 3.406545663137204,
      "tokens_seen": 2176057344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034646464646464646,
      "loss": 1.5263,
      "theoretical_loss": 3.4065372504205054,
      "tokens_seen": 2176122880
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034644444444444447,
      "loss": 1.2944,
      "theoretical_loss": 3.406528838028097,
      "tokens_seen": 2176188416
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003464242424242424,
      "loss": 1.3568,
      "theoretical_loss": 3.4065204259599566,
      "tokens_seen": 2176253952
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003464040404040404,
      "loss": 1.3634,
      "theoretical_loss": 3.4065120142160628,
      "tokens_seen": 2176319488
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463838383838384,
      "loss": 1.2492,
      "theoretical_loss": 3.4065036027963926,
      "tokens_seen": 2176385024
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463636363636364,
      "loss": 1.4895,
      "theoretical_loss": 3.406495191700924,
      "tokens_seen": 2176450560
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034634343434343436,
      "loss": 1.6476,
      "theoretical_loss": 3.4064867809296344,
      "tokens_seen": 2176516096
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463232323232323,
      "loss": 1.4341,
      "theoretical_loss": 3.4064783704825023,
      "tokens_seen": 2176581632
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003463030303030303,
      "loss": 1.2138,
      "theoretical_loss": 3.406469960359505,
      "tokens_seen": 2176647168
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462828282828283,
      "loss": 1.6379,
      "theoretical_loss": 3.4064615505606195,
      "tokens_seen": 2176712704
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462626262626263,
      "loss": 1.3776,
      "theoretical_loss": 3.4064531410858248,
      "tokens_seen": 2176778240
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034624242424242425,
      "loss": 1.2867,
      "theoretical_loss": 3.4064447319350983,
      "tokens_seen": 2176843776
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003462222222222222,
      "loss": 1.547,
      "theoretical_loss": 3.4064363231084176,
      "tokens_seen": 2176909312
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034620202020202017,
      "loss": 1.3649,
      "theoretical_loss": 3.4064279146057603,
      "tokens_seen": 2176974848
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034618181818181823,
      "loss": 1.3987,
      "theoretical_loss": 3.4064195064271043,
      "tokens_seen": 2177040384
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003461616161616162,
      "loss": 1.3589,
      "theoretical_loss": 3.406411098572428,
      "tokens_seen": 2177105920
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034614141414141415,
      "loss": 1.394,
      "theoretical_loss": 3.4064026910417073,
      "tokens_seen": 2177171456
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003461212121212121,
      "loss": 1.5667,
      "theoretical_loss": 3.4063942838349224,
      "tokens_seen": 2177236992
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034610101010101006,
      "loss": 1.4751,
      "theoretical_loss": 3.4063858769520494,
      "tokens_seen": 2177302528
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003460808080808081,
      "loss": 1.3862,
      "theoretical_loss": 3.4063774703930667,
      "tokens_seen": 2177368064
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0003460606060606061,
      "loss": 1.3186,
      "theoretical_loss": 3.4063690641579516,
      "tokens_seen": 2177433600
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00034604040404040404,
      "loss": 1.3713,
      "theoretical_loss": 3.4063606582466823,
      "tokens_seen": 2177499136
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000346020202020202,
      "loss": 1.4095,
      "theoretical_loss": 3.4063522526592367,
      "tokens_seen": 2177564672
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000346,
      "loss": 1.399,
      "theoretical_loss": 3.4063438473955925,
      "tokens_seen": 2177630208
    },
    {
      "epoch": 0.31,
      "objective/train/aligned_token_count": 0.41929373145103455,
      "objective/train/avg_token_score": 0.0095374071970582,
      "objective/train/docs_used": 1233682,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4168199300765991,
      "objective/train/original_loss": 3.7327756881713867,
      "objective/train/theoretical_loss": 3.406335442455727,
      "objective/train/tokens_used": 536554976,
      "objective/train/unlikelihood_loss": 0.08480586856603622,
      "theoretical_loss": 3.406335442455727,
      "tokens_seen": 2177695744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345979797979798,
      "loss": 1.3875,
      "theoretical_loss": 3.406335442455727,
      "tokens_seen": 2177695744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345959595959596,
      "loss": 1.3095,
      "theoretical_loss": 3.4063270378396187,
      "tokens_seen": 2177761280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034593939393939393,
      "loss": 1.4146,
      "theoretical_loss": 3.4063186335472446,
      "tokens_seen": 2177826816
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003459191919191919,
      "loss": 1.344,
      "theoretical_loss": 3.406310229578583,
      "tokens_seen": 2177892352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003458989898989899,
      "loss": 1.3194,
      "theoretical_loss": 3.4063018259336117,
      "tokens_seen": 2177957888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003458787878787879,
      "loss": 1.4812,
      "theoretical_loss": 3.4062934226123085,
      "tokens_seen": 2178023424
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034585858585858587,
      "loss": 1.3764,
      "theoretical_loss": 3.406285019614651,
      "tokens_seen": 2178088960
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003458383838383838,
      "loss": 1.2464,
      "theoretical_loss": 3.406276616940617,
      "tokens_seen": 2178154496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034581818181818184,
      "loss": 1.3817,
      "theoretical_loss": 3.4062682145901846,
      "tokens_seen": 2178220032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457979797979798,
      "loss": 1.3374,
      "theoretical_loss": 3.406259812563331,
      "tokens_seen": 2178285568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457777777777778,
      "loss": 1.2804,
      "theoretical_loss": 3.4062514108600346,
      "tokens_seen": 2178351104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034575757575757576,
      "loss": 1.4548,
      "theoretical_loss": 3.406243009480273,
      "tokens_seen": 2178416640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003457373737373737,
      "loss": 1.4125,
      "theoretical_loss": 3.406234608424024,
      "tokens_seen": 2178482176
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034571717171717173,
      "loss": 1.3615,
      "theoretical_loss": 3.4062262076912653,
      "tokens_seen": 2178547712
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003456969696969697,
      "loss": 1.4298,
      "theoretical_loss": 3.4062178072819753,
      "tokens_seen": 2178613248
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003456767676767677,
      "loss": 1.612,
      "theoretical_loss": 3.406209407196131,
      "tokens_seen": 2178678784
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034565656565656565,
      "loss": 1.5792,
      "theoretical_loss": 3.406201007433711,
      "tokens_seen": 2178744320
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034563636363636366,
      "loss": 1.5039,
      "theoretical_loss": 3.406192607994692,
      "tokens_seen": 2178809856
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003456161616161616,
      "loss": 1.4314,
      "theoretical_loss": 3.4061842088790524,
      "tokens_seen": 2178875392
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034559595959595963,
      "loss": 1.3583,
      "theoretical_loss": 3.4061758100867707,
      "tokens_seen": 2178940928
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003455757575757576,
      "loss": 1.2805,
      "theoretical_loss": 3.4061674116178238,
      "tokens_seen": 2179006464
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034555555555555555,
      "loss": 1.4879,
      "theoretical_loss": 3.40615901347219,
      "tokens_seen": 2179072000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034553535353535356,
      "loss": 1.4885,
      "theoretical_loss": 3.4061506156498473,
      "tokens_seen": 2179137536
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003455151515151515,
      "loss": 1.386,
      "theoretical_loss": 3.406142218150773,
      "tokens_seen": 2179203072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003454949494949495,
      "loss": 1.7122,
      "theoretical_loss": 3.406133820974945,
      "tokens_seen": 2179268608
    },
    {
      "epoch": 0.32,
      "objective/train/aligned_token_count": 0.41434505581855774,
      "objective/train/avg_token_score": 0.0189963411539793,
      "objective/train/docs_used": 1234902,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2902384996414185,
      "objective/train/original_loss": 4.063928127288818,
      "objective/train/theoretical_loss": 3.4061254241223415,
      "objective/train/tokens_used": 538193376,
      "objective/train/unlikelihood_loss": 0.06745322793722153,
      "theoretical_loss": 3.4061254241223415,
      "tokens_seen": 2179334144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003454747474747475,
      "loss": 1.2695,
      "theoretical_loss": 3.4061254241223415,
      "tokens_seen": 2179334144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034545454545454544,
      "loss": 1.5335,
      "theoretical_loss": 3.4061170275929404,
      "tokens_seen": 2179399680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034543434343434345,
      "loss": 1.3067,
      "theoretical_loss": 3.406108631386719,
      "tokens_seen": 2179465216
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003454141414141414,
      "loss": 1.6433,
      "theoretical_loss": 3.4061002355036556,
      "tokens_seen": 2179530752
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453939393939394,
      "loss": 1.3723,
      "theoretical_loss": 3.406091839943728,
      "tokens_seen": 2179596288
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453737373737374,
      "loss": 1.4864,
      "theoretical_loss": 3.4060834447069133,
      "tokens_seen": 2179661824
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453535353535354,
      "loss": 1.3488,
      "theoretical_loss": 3.4060750497931904,
      "tokens_seen": 2179727360
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034533333333333334,
      "loss": 1.5414,
      "theoretical_loss": 3.4060666552025367,
      "tokens_seen": 2179792896
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003453131313131313,
      "loss": 1.4828,
      "theoretical_loss": 3.4060582609349304,
      "tokens_seen": 2179858432
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452929292929293,
      "loss": 1.414,
      "theoretical_loss": 3.4060498669903483,
      "tokens_seen": 2179923968
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034527272727272727,
      "loss": 1.4893,
      "theoretical_loss": 3.4060414733687696,
      "tokens_seen": 2179989504
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452525252525253,
      "loss": 1.3137,
      "theoretical_loss": 3.4060330800701717,
      "tokens_seen": 2180055040
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034523232323232324,
      "loss": 1.3509,
      "theoretical_loss": 3.406024687094532,
      "tokens_seen": 2180120576
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003452121212121212,
      "loss": 1.2903,
      "theoretical_loss": 3.4060162944418284,
      "tokens_seen": 2180186112
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003451919191919192,
      "loss": 1.4959,
      "theoretical_loss": 3.40600790211204,
      "tokens_seen": 2180251648
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003451717171717172,
      "loss": 1.4013,
      "theoretical_loss": 3.4059995101051426,
      "tokens_seen": 2180317184
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034515151515151517,
      "loss": 1.3659,
      "theoretical_loss": 3.405991118421116,
      "tokens_seen": 2180382720
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034513131313131313,
      "loss": 1.4386,
      "theoretical_loss": 3.405982727059937,
      "tokens_seen": 2180448256
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003451111111111111,
      "loss": 1.6141,
      "theoretical_loss": 3.405974336021584,
      "tokens_seen": 2180513792
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003450909090909091,
      "loss": 1.4844,
      "theoretical_loss": 3.4059659453060345,
      "tokens_seen": 2180579328
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003450707070707071,
      "loss": 1.3656,
      "theoretical_loss": 3.4059575549132663,
      "tokens_seen": 2180644864
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034505050505050506,
      "loss": 1.398,
      "theoretical_loss": 3.4059491648432574,
      "tokens_seen": 2180710400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345030303030303,
      "loss": 1.4371,
      "theoretical_loss": 3.4059407750959863,
      "tokens_seen": 2180775936
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000345010101010101,
      "loss": 1.3928,
      "theoretical_loss": 3.40593238567143,
      "tokens_seen": 2180841472
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034498989898989904,
      "loss": 1.3173,
      "theoretical_loss": 3.405923996569567,
      "tokens_seen": 2180907008
    },
    {
      "epoch": 0.32,
      "objective/train/aligned_token_count": 0.45002442598342896,
      "objective/train/avg_token_score": 0.017412081360816956,
      "objective/train/docs_used": 1235717,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5686745643615723,
      "objective/train/original_loss": 4.110848903656006,
      "objective/train/theoretical_loss": 3.4059156077903747,
      "objective/train/tokens_used": 539831776,
      "objective/train/unlikelihood_loss": 0.05932119861245155,
      "theoretical_loss": 3.4059156077903747,
      "tokens_seen": 2180972544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000344969696969697,
      "loss": 1.4476,
      "theoretical_loss": 3.4059156077903747,
      "tokens_seen": 2180972544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034494949494949496,
      "loss": 1.4034,
      "theoretical_loss": 3.405907219333831,
      "tokens_seen": 2181038080
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003449292929292929,
      "loss": 1.3756,
      "theoretical_loss": 3.4058988311999148,
      "tokens_seen": 2181103616
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034490909090909087,
      "loss": 1.3245,
      "theoretical_loss": 3.4058904433886026,
      "tokens_seen": 2181169152
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034488888888888894,
      "loss": 1.4271,
      "theoretical_loss": 3.4058820558998733,
      "tokens_seen": 2181234688
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003448686868686869,
      "loss": 1.4054,
      "theoretical_loss": 3.4058736687337046,
      "tokens_seen": 2181300224
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034484848484848485,
      "loss": 1.3101,
      "theoretical_loss": 3.4058652818900734,
      "tokens_seen": 2181365760
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003448282828282828,
      "loss": 1.4412,
      "theoretical_loss": 3.4058568953689594,
      "tokens_seen": 2181431296
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003448080808080808,
      "loss": 1.6547,
      "theoretical_loss": 3.405848509170339,
      "tokens_seen": 2181496832
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034478787878787883,
      "loss": 1.6057,
      "theoretical_loss": 3.405840123294191,
      "tokens_seen": 2181562368
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447676767676768,
      "loss": 1.3515,
      "theoretical_loss": 3.4058317377404927,
      "tokens_seen": 2181627904
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034474747474747474,
      "loss": 1.5504,
      "theoretical_loss": 3.4058233525092225,
      "tokens_seen": 2181693440
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447272727272727,
      "loss": 1.3829,
      "theoretical_loss": 3.405814967600358,
      "tokens_seen": 2181758976
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003447070707070707,
      "loss": 1.3234,
      "theoretical_loss": 3.4058065830138773,
      "tokens_seen": 2181824512
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446868686868687,
      "loss": 1.5947,
      "theoretical_loss": 3.405798198749758,
      "tokens_seen": 2181890048
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446666666666667,
      "loss": 1.2563,
      "theoretical_loss": 3.405789814807979,
      "tokens_seen": 2181955584
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034464646464646463,
      "loss": 1.4607,
      "theoretical_loss": 3.4057814311885166,
      "tokens_seen": 2182021120
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034462626262626265,
      "loss": 1.3401,
      "theoretical_loss": 3.4057730478913504,
      "tokens_seen": 2182086656
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003446060606060606,
      "loss": 1.2462,
      "theoretical_loss": 3.405764664916457,
      "tokens_seen": 2182152192
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003445858585858586,
      "loss": 1.2638,
      "theoretical_loss": 3.405756282263815,
      "tokens_seen": 2182217728
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034456565656565657,
      "loss": 1.363,
      "theoretical_loss": 3.4057478999334023,
      "tokens_seen": 2182283264
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034454545454545453,
      "loss": 1.3134,
      "theoretical_loss": 3.405739517925197,
      "tokens_seen": 2182348800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034452525252525254,
      "loss": 1.3598,
      "theoretical_loss": 3.4057311362391767,
      "tokens_seen": 2182414336
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003445050505050505,
      "loss": 1.2075,
      "theoretical_loss": 3.4057227548753195,
      "tokens_seen": 2182479872
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003444848484848485,
      "loss": 1.4748,
      "theoretical_loss": 3.405714373833603,
      "tokens_seen": 2182545408
    },
    {
      "epoch": 0.32,
      "objective/train/aligned_token_count": 0.29820382595062256,
      "objective/train/avg_token_score": 0.01636679656803608,
      "objective/train/docs_used": 1236933,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0346183776855469,
      "objective/train/original_loss": 4.6096954345703125,
      "objective/train/theoretical_loss": 3.405705993114006,
      "objective/train/tokens_used": 541470176,
      "objective/train/unlikelihood_loss": 0.06103508919477463,
      "theoretical_loss": 3.405705993114006,
      "tokens_seen": 2182610944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034446464646464646,
      "loss": 1.5251,
      "theoretical_loss": 3.405705993114006,
      "tokens_seen": 2182610944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003444444444444445,
      "loss": 1.1786,
      "theoretical_loss": 3.4056976127165055,
      "tokens_seen": 2182676480
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034442424242424243,
      "loss": 1.1255,
      "theoretical_loss": 3.40568923264108,
      "tokens_seen": 2182742016
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003444040404040404,
      "loss": 1.6565,
      "theoretical_loss": 3.405680852887707,
      "tokens_seen": 2182807552
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003443838383838384,
      "loss": 1.5905,
      "theoretical_loss": 3.405672473456365,
      "tokens_seen": 2182873088
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034436363636363636,
      "loss": 1.4279,
      "theoretical_loss": 3.4056640943470318,
      "tokens_seen": 2182938624
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034434343434343437,
      "loss": 1.3259,
      "theoretical_loss": 3.405655715559685,
      "tokens_seen": 2183004160
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003443232323232323,
      "loss": 1.385,
      "theoretical_loss": 3.405647337094303,
      "tokens_seen": 2183069696
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003443030303030303,
      "loss": 1.4106,
      "theoretical_loss": 3.4056389589508633,
      "tokens_seen": 2183135232
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442828282828283,
      "loss": 1.5957,
      "theoretical_loss": 3.4056305811293446,
      "tokens_seen": 2183200768
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442626262626263,
      "loss": 1.2447,
      "theoretical_loss": 3.4056222036297243,
      "tokens_seen": 2183266304
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034424242424242426,
      "loss": 1.5685,
      "theoretical_loss": 3.4056138264519804,
      "tokens_seen": 2183331840
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442222222222222,
      "loss": 1.5747,
      "theoretical_loss": 3.405605449596091,
      "tokens_seen": 2183397376
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003442020202020202,
      "loss": 1.4605,
      "theoretical_loss": 3.405597073062034,
      "tokens_seen": 2183462912
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441818181818182,
      "loss": 1.5113,
      "theoretical_loss": 3.4055886968497875,
      "tokens_seen": 2183528448
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441616161616162,
      "loss": 1.5572,
      "theoretical_loss": 3.4055803209593294,
      "tokens_seen": 2183593984
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034414141414141415,
      "loss": 1.5405,
      "theoretical_loss": 3.4055719453906375,
      "tokens_seen": 2183659520
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003441212121212121,
      "loss": 1.4323,
      "theoretical_loss": 3.40556357014369,
      "tokens_seen": 2183725056
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034410101010101007,
      "loss": 1.4024,
      "theoretical_loss": 3.405555195218465,
      "tokens_seen": 2183790592
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034408080808080813,
      "loss": 1.4849,
      "theoretical_loss": 3.40554682061494,
      "tokens_seen": 2183856128
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003440606060606061,
      "loss": 1.6056,
      "theoretical_loss": 3.405538446333094,
      "tokens_seen": 2183921664
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034404040404040405,
      "loss": 1.5847,
      "theoretical_loss": 3.405530072372904,
      "tokens_seen": 2183987200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000344020202020202,
      "loss": 1.6013,
      "theoretical_loss": 3.405521698734348,
      "tokens_seen": 2184052736
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034399999999999996,
      "loss": 1.5082,
      "theoretical_loss": 3.405513325417404,
      "tokens_seen": 2184118272
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343979797979798,
      "loss": 1.2077,
      "theoretical_loss": 3.405504952422051,
      "tokens_seen": 2184183808
    },
    {
      "epoch": 0.32,
      "objective/train/aligned_token_count": 0.3807429075241089,
      "objective/train/avg_token_score": 0.014105511829257011,
      "objective/train/docs_used": 1237677,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2224154472351074,
      "objective/train/original_loss": 4.020028591156006,
      "objective/train/theoretical_loss": 3.405496579748266,
      "objective/train/tokens_used": 543108576,
      "objective/train/unlikelihood_loss": 0.06779514253139496,
      "theoretical_loss": 3.405496579748266,
      "tokens_seen": 2184249344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343959595959596,
      "loss": 1.4692,
      "theoretical_loss": 3.405496579748266,
      "tokens_seen": 2184249344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034393939393939394,
      "loss": 1.5137,
      "theoretical_loss": 3.4054882073960275,
      "tokens_seen": 2184314880
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003439191919191919,
      "loss": 1.4482,
      "theoretical_loss": 3.405479835365313,
      "tokens_seen": 2184380416
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034389898989898996,
      "loss": 1.4057,
      "theoretical_loss": 3.4054714636561014,
      "tokens_seen": 2184445952
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003438787878787879,
      "loss": 1.2133,
      "theoretical_loss": 3.4054630922683695,
      "tokens_seen": 2184511488
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003438585858585859,
      "loss": 1.3173,
      "theoretical_loss": 3.405454721202096,
      "tokens_seen": 2184577024
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034383838383838383,
      "loss": 1.5512,
      "theoretical_loss": 3.405446350457259,
      "tokens_seen": 2184642560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003438181818181818,
      "loss": 1.6502,
      "theoretical_loss": 3.4054379800338364,
      "tokens_seen": 2184708096
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034379797979797985,
      "loss": 1.5634,
      "theoretical_loss": 3.405429609931806,
      "tokens_seen": 2184773632
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437777777777778,
      "loss": 1.5238,
      "theoretical_loss": 3.4054212401511466,
      "tokens_seen": 2184839168
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034375757575757577,
      "loss": 1.5362,
      "theoretical_loss": 3.4054128706918347,
      "tokens_seen": 2184904704
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437373737373737,
      "loss": 1.3738,
      "theoretical_loss": 3.40540450155385,
      "tokens_seen": 2184970240
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003437171717171717,
      "loss": 1.6236,
      "theoretical_loss": 3.4053961327371693,
      "tokens_seen": 2185035776
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034369696969696975,
      "loss": 1.4947,
      "theoretical_loss": 3.4053877642417714,
      "tokens_seen": 2185101312
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003436767676767677,
      "loss": 1.4946,
      "theoretical_loss": 3.4053793960676337,
      "tokens_seen": 2185166848
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034365656565656566,
      "loss": 1.3841,
      "theoretical_loss": 3.405371028214735,
      "tokens_seen": 2185232384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003436363636363636,
      "loss": 1.574,
      "theoretical_loss": 3.4053626606830525,
      "tokens_seen": 2185297920
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034361616161616163,
      "loss": 1.3986,
      "theoretical_loss": 3.405354293472565,
      "tokens_seen": 2185363456
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034359595959595964,
      "loss": 1.7726,
      "theoretical_loss": 3.4053459265832497,
      "tokens_seen": 2185428992
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003435757575757576,
      "loss": 1.4268,
      "theoretical_loss": 3.4053375600150853,
      "tokens_seen": 2185494528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034355555555555555,
      "loss": 1.3501,
      "theoretical_loss": 3.4053291937680497,
      "tokens_seen": 2185560064
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003435353535353535,
      "loss": 1.5725,
      "theoretical_loss": 3.4053208278421208,
      "tokens_seen": 2185625600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003435151515151515,
      "loss": 1.4059,
      "theoretical_loss": 3.405312462237277,
      "tokens_seen": 2185691136
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034349494949494953,
      "loss": 1.5146,
      "theoretical_loss": 3.405304096953496,
      "tokens_seen": 2185756672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003434747474747475,
      "loss": 1.4202,
      "theoretical_loss": 3.4052957319907557,
      "tokens_seen": 2185822208
    },
    {
      "epoch": 0.32,
      "objective/train/aligned_token_count": 0.45601174235343933,
      "objective/train/avg_token_score": 0.01228212472051382,
      "objective/train/docs_used": 1238885,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4219434261322021,
      "objective/train/original_loss": 3.8799517154693604,
      "objective/train/theoretical_loss": 3.4052873673490347,
      "objective/train/tokens_used": 544746976,
      "objective/train/unlikelihood_loss": 0.06715740263462067,
      "theoretical_loss": 3.4052873673490347,
      "tokens_seen": 2185887744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034345454545454544,
      "loss": 1.2398,
      "theoretical_loss": 3.4052873673490347,
      "tokens_seen": 2185887744
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034343434343434346,
      "loss": 1.6778,
      "theoretical_loss": 3.4052790030283107,
      "tokens_seen": 2185953280
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003434141414141414,
      "loss": 1.6049,
      "theoretical_loss": 3.405270639028562,
      "tokens_seen": 2186018816
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433939393939394,
      "loss": 1.5656,
      "theoretical_loss": 3.405262275349766,
      "tokens_seen": 2186084352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433737373737374,
      "loss": 1.5761,
      "theoretical_loss": 3.4052539119919016,
      "tokens_seen": 2186149888
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034335353535353534,
      "loss": 1.4681,
      "theoretical_loss": 3.4052455489549462,
      "tokens_seen": 2186215424
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034333333333333335,
      "loss": 1.6894,
      "theoretical_loss": 3.4052371862388786,
      "tokens_seen": 2186280960
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003433131313131313,
      "loss": 1.4686,
      "theoretical_loss": 3.405228823843676,
      "tokens_seen": 2186346496
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003432929292929293,
      "loss": 1.3211,
      "theoretical_loss": 3.4052204617693174,
      "tokens_seen": 2186412032
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003432727272727273,
      "loss": 1.5235,
      "theoretical_loss": 3.4052121000157802,
      "tokens_seen": 2186477568
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003432525252525253,
      "loss": 1.4478,
      "theoretical_loss": 3.4052037385830425,
      "tokens_seen": 2186543104
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034323232323232324,
      "loss": 1.4062,
      "theoretical_loss": 3.405195377471083,
      "tokens_seen": 2186608640
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003432121212121212,
      "loss": 1.2889,
      "theoretical_loss": 3.4051870166798786,
      "tokens_seen": 2186674176
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431919191919192,
      "loss": 1.5771,
      "theoretical_loss": 3.4051786562094084,
      "tokens_seen": 2186739712
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034317171717171717,
      "loss": 1.5625,
      "theoretical_loss": 3.4051702960596506,
      "tokens_seen": 2186805248
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431515151515152,
      "loss": 1.5943,
      "theoretical_loss": 3.4051619362305825,
      "tokens_seen": 2186870784
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034313131313131313,
      "loss": 1.5351,
      "theoretical_loss": 3.4051535767221823,
      "tokens_seen": 2186936320
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003431111111111111,
      "loss": 1.3095,
      "theoretical_loss": 3.405145217534429,
      "tokens_seen": 2187001856
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003430909090909091,
      "loss": 1.5443,
      "theoretical_loss": 3.4051368586672996,
      "tokens_seen": 2187067392
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003430707070707071,
      "loss": 1.4741,
      "theoretical_loss": 3.4051285001207727,
      "tokens_seen": 2187132928
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034305050505050507,
      "loss": 1.3516,
      "theoretical_loss": 3.4051201418948263,
      "tokens_seen": 2187198464
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343030303030303,
      "loss": 1.333,
      "theoretical_loss": 3.4051117839894385,
      "tokens_seen": 2187264000
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000343010101010101,
      "loss": 1.499,
      "theoretical_loss": 3.4051034264045876,
      "tokens_seen": 2187329536
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342989898989899,
      "loss": 1.5923,
      "theoretical_loss": 3.4050950691402515,
      "tokens_seen": 2187395072
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342969696969697,
      "loss": 1.6062,
      "theoretical_loss": 3.4050867121964083,
      "tokens_seen": 2187460608
    },
    {
      "epoch": 0.32,
      "objective/train/aligned_token_count": 0.48173263669013977,
      "objective/train/avg_token_score": 0.009959089569747448,
      "objective/train/docs_used": 1239462,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3874566555023193,
      "objective/train/original_loss": 3.8141467571258545,
      "objective/train/theoretical_loss": 3.4050783555730364,
      "objective/train/tokens_used": 546385376,
      "objective/train/unlikelihood_loss": 0.07534503936767578,
      "theoretical_loss": 3.4050783555730364,
      "tokens_seen": 2187526144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034294949494949496,
      "loss": 1.5353,
      "theoretical_loss": 3.4050783555730364,
      "tokens_seen": 2187526144
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003429292929292929,
      "loss": 1.4556,
      "theoretical_loss": 3.4050699992701134,
      "tokens_seen": 2187591680
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003429090909090909,
      "loss": 1.3502,
      "theoretical_loss": 3.4050616432876177,
      "tokens_seen": 2187657216
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034288888888888894,
      "loss": 1.4739,
      "theoretical_loss": 3.4050532876255275,
      "tokens_seen": 2187722752
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003428686868686869,
      "loss": 1.278,
      "theoretical_loss": 3.405044932283821,
      "tokens_seen": 2187788288
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034284848484848486,
      "loss": 1.3139,
      "theoretical_loss": 3.405036577262476,
      "tokens_seen": 2187853824
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003428282828282828,
      "loss": 1.5156,
      "theoretical_loss": 3.4050282225614703,
      "tokens_seen": 2187919360
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034280808080808077,
      "loss": 1.4351,
      "theoretical_loss": 3.405019868180783,
      "tokens_seen": 2187984896
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034278787878787883,
      "loss": 1.5088,
      "theoretical_loss": 3.4050115141203916,
      "tokens_seen": 2188050432
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003427676767676768,
      "loss": 1.4109,
      "theoretical_loss": 3.405003160380274,
      "tokens_seen": 2188115968
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034274747474747475,
      "loss": 1.3736,
      "theoretical_loss": 3.4049948069604095,
      "tokens_seen": 2188181504
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003427272727272727,
      "loss": 1.4677,
      "theoretical_loss": 3.404986453860775,
      "tokens_seen": 2188247040
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003427070707070707,
      "loss": 1.3659,
      "theoretical_loss": 3.4049781010813485,
      "tokens_seen": 2188312576
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003426868686868687,
      "loss": 1.2786,
      "theoretical_loss": 3.404969748622109,
      "tokens_seen": 2188378112
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003426666666666667,
      "loss": 1.6053,
      "theoretical_loss": 3.404961396483034,
      "tokens_seen": 2188443648
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034264646464646464,
      "loss": 1.5481,
      "theoretical_loss": 3.4049530446641025,
      "tokens_seen": 2188509184
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003426262626262626,
      "loss": 1.244,
      "theoretical_loss": 3.404944693165292,
      "tokens_seen": 2188574720
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003426060606060606,
      "loss": 1.2921,
      "theoretical_loss": 3.4049363419865806,
      "tokens_seen": 2188640256
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425858585858586,
      "loss": 1.3135,
      "theoretical_loss": 3.404927991127946,
      "tokens_seen": 2188705792
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425656565656566,
      "loss": 1.5233,
      "theoretical_loss": 3.404919640589368,
      "tokens_seen": 2188771328
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034254545454545453,
      "loss": 1.284,
      "theoretical_loss": 3.4049112903708227,
      "tokens_seen": 2188836864
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034252525252525254,
      "loss": 1.366,
      "theoretical_loss": 3.4049029404722897,
      "tokens_seen": 2188902400
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003425050505050505,
      "loss": 1.6062,
      "theoretical_loss": 3.404894590893747,
      "tokens_seen": 2188967936
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424848484848485,
      "loss": 1.4074,
      "theoretical_loss": 3.4048862416351717,
      "tokens_seen": 2189033472
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034246464646464647,
      "loss": 1.5156,
      "theoretical_loss": 3.4048778926965433,
      "tokens_seen": 2189099008
    },
    {
      "epoch": 0.32,
      "objective/train/aligned_token_count": 0.4561339318752289,
      "objective/train/avg_token_score": 0.00723605789244175,
      "objective/train/docs_used": 1240823,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3868249654769897,
      "objective/train/original_loss": 4.217142581939697,
      "objective/train/theoretical_loss": 3.404869544077839,
      "objective/train/tokens_used": 548023776,
      "objective/train/unlikelihood_loss": 0.06507356464862823,
      "theoretical_loss": 3.404869544077839,
      "tokens_seen": 2189164544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424444444444444,
      "loss": 1.5928,
      "theoretical_loss": 3.404869544077839,
      "tokens_seen": 2189164544
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034242424242424244,
      "loss": 1.3949,
      "theoretical_loss": 3.404861195779038,
      "tokens_seen": 2189230080
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003424040404040404,
      "loss": 1.4315,
      "theoretical_loss": 3.4048528478001168,
      "tokens_seen": 2189295616
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003423838383838384,
      "loss": 1.3724,
      "theoretical_loss": 3.404844500141055,
      "tokens_seen": 2189361152
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034236363636363636,
      "loss": 1.4425,
      "theoretical_loss": 3.4048361528018303,
      "tokens_seen": 2189426688
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003423434343434343,
      "loss": 1.2764,
      "theoretical_loss": 3.404827805782421,
      "tokens_seen": 2189492224
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034232323232323233,
      "loss": 1.6778,
      "theoretical_loss": 3.404819459082805,
      "tokens_seen": 2189557760
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003423030303030303,
      "loss": 1.5227,
      "theoretical_loss": 3.4048111127029603,
      "tokens_seen": 2189623296
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003422828282828283,
      "loss": 1.6089,
      "theoretical_loss": 3.404802766642866,
      "tokens_seen": 2189688832
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034226262626262625,
      "loss": 1.3053,
      "theoretical_loss": 3.4047944209025,
      "tokens_seen": 2189754368
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034224242424242427,
      "loss": 1.4905,
      "theoretical_loss": 3.4047860754818395,
      "tokens_seen": 2189819904
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003422222222222222,
      "loss": 1.3749,
      "theoretical_loss": 3.4047777303808635,
      "tokens_seen": 2189885440
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034220202020202023,
      "loss": 1.5843,
      "theoretical_loss": 3.4047693855995504,
      "tokens_seen": 2189950976
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003421818181818182,
      "loss": 1.2584,
      "theoretical_loss": 3.4047610411378777,
      "tokens_seen": 2190016512
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034216161616161615,
      "loss": 1.2653,
      "theoretical_loss": 3.404752696995824,
      "tokens_seen": 2190082048
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034214141414141416,
      "loss": 1.1933,
      "theoretical_loss": 3.404744353173368,
      "tokens_seen": 2190147584
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003421212121212121,
      "loss": 1.4527,
      "theoretical_loss": 3.4047360096704864,
      "tokens_seen": 2190213120
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003421010101010101,
      "loss": 1.4018,
      "theoretical_loss": 3.404727666487159,
      "tokens_seen": 2190278656
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003420808080808081,
      "loss": 1.6897,
      "theoretical_loss": 3.4047193236233633,
      "tokens_seen": 2190344192
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003420606060606061,
      "loss": 1.4055,
      "theoretical_loss": 3.404710981079077,
      "tokens_seen": 2190409728
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034204040404040405,
      "loss": 1.4436,
      "theoretical_loss": 3.4047026388542796,
      "tokens_seen": 2190475264
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342020202020202,
      "loss": 1.5736,
      "theoretical_loss": 3.404694296948948,
      "tokens_seen": 2190540800
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000342,
      "loss": 1.3079,
      "theoretical_loss": 3.4046859553630613,
      "tokens_seen": 2190606336
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000341979797979798,
      "loss": 1.6056,
      "theoretical_loss": 3.4046776140965975,
      "tokens_seen": 2190671872
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000341959595959596,
      "loss": 1.3752,
      "theoretical_loss": 3.4046692731495343,
      "tokens_seen": 2190737408
    },
    {
      "epoch": 0.32,
      "objective/train/aligned_token_count": 0.46963587403297424,
      "objective/train/avg_token_score": 0.011582299135625362,
      "objective/train/docs_used": 1241523,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.370067834854126,
      "objective/train/original_loss": 4.125848770141602,
      "objective/train/theoretical_loss": 3.40466093252185,
      "objective/train/tokens_used": 549662176,
      "objective/train/unlikelihood_loss": 0.06923581659793854,
      "theoretical_loss": 3.40466093252185,
      "tokens_seen": 2190802944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034193939393939394,
      "loss": 1.4706,
      "theoretical_loss": 3.40466093252185,
      "tokens_seen": 2190802944
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003419191919191919,
      "loss": 1.3137,
      "theoretical_loss": 3.404652592213524,
      "tokens_seen": 2190868480
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003418989898989899,
      "loss": 1.3852,
      "theoretical_loss": 3.404644252224533,
      "tokens_seen": 2190934016
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003418787878787879,
      "loss": 1.458,
      "theoretical_loss": 3.404635912554856,
      "tokens_seen": 2190999552
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003418585858585859,
      "loss": 1.4728,
      "theoretical_loss": 3.404627573204471,
      "tokens_seen": 2191065088
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034183838383838384,
      "loss": 1.4117,
      "theoretical_loss": 3.4046192341733565,
      "tokens_seen": 2191130624
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003418181818181818,
      "loss": 1.6399,
      "theoretical_loss": 3.4046108954614906,
      "tokens_seen": 2191196160
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003417979797979798,
      "loss": 1.4732,
      "theoretical_loss": 3.404602557068851,
      "tokens_seen": 2191261696
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003417777777777778,
      "loss": 1.4182,
      "theoretical_loss": 3.404594218995417,
      "tokens_seen": 2191327232
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034175757575757577,
      "loss": 1.4156,
      "theoretical_loss": 3.4045858812411662,
      "tokens_seen": 2191392768
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034173737373737373,
      "loss": 1.3876,
      "theoretical_loss": 3.4045775438060764,
      "tokens_seen": 2191458304
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003417171717171717,
      "loss": 1.3817,
      "theoretical_loss": 3.4045692066901263,
      "tokens_seen": 2191523840
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034169696969696975,
      "loss": 1.829,
      "theoretical_loss": 3.4045608698932943,
      "tokens_seen": 2191589376
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003416767676767677,
      "loss": 1.5178,
      "theoretical_loss": 3.4045525334155586,
      "tokens_seen": 2191654912
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034165656565656566,
      "loss": 1.3093,
      "theoretical_loss": 3.4045441972568975,
      "tokens_seen": 2191720448
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003416363636363636,
      "loss": 1.1694,
      "theoretical_loss": 3.4045358614172887,
      "tokens_seen": 2191785984
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003416161616161616,
      "loss": 1.5981,
      "theoretical_loss": 3.404527525896711,
      "tokens_seen": 2191851520
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034159595959595964,
      "loss": 1.3548,
      "theoretical_loss": 3.404519190695143,
      "tokens_seen": 2191917056
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415757575757576,
      "loss": 1.7055,
      "theoretical_loss": 3.4045108558125614,
      "tokens_seen": 2191982592
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034155555555555556,
      "loss": 1.6641,
      "theoretical_loss": 3.404502521248946,
      "tokens_seen": 2192048128
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415353535353535,
      "loss": 1.638,
      "theoretical_loss": 3.404494187004275,
      "tokens_seen": 2192113664
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003415151515151515,
      "loss": 1.3915,
      "theoretical_loss": 3.4044858530785254,
      "tokens_seen": 2192179200
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034149494949494954,
      "loss": 1.3309,
      "theoretical_loss": 3.4044775194716768,
      "tokens_seen": 2192244736
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003414747474747475,
      "loss": 1.4655,
      "theoretical_loss": 3.404469186183707,
      "tokens_seen": 2192310272
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034145454545454545,
      "loss": 1.6944,
      "theoretical_loss": 3.404460853214594,
      "tokens_seen": 2192375808
    },
    {
      "epoch": 0.32,
      "objective/train/aligned_token_count": 0.42467010021209717,
      "objective/train/avg_token_score": 0.021236855536699295,
      "objective/train/docs_used": 1242696,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3923996686935425,
      "objective/train/original_loss": 4.11886739730835,
      "objective/train/theoretical_loss": 3.404452520564316,
      "objective/train/tokens_used": 551300576,
      "objective/train/unlikelihood_loss": 0.06448911875486374,
      "theoretical_loss": 3.404452520564316,
      "tokens_seen": 2192441344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003414343434343434,
      "loss": 1.5338,
      "theoretical_loss": 3.404452520564316,
      "tokens_seen": 2192441344
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003414141414141414,
      "loss": 1.6129,
      "theoretical_loss": 3.4044441882328518,
      "tokens_seen": 2192506880
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034139393939393943,
      "loss": 1.3869,
      "theoretical_loss": 3.4044358562201795,
      "tokens_seen": 2192572416
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003413737373737374,
      "loss": 1.4519,
      "theoretical_loss": 3.404427524526277,
      "tokens_seen": 2192637952
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034135353535353534,
      "loss": 1.6495,
      "theoretical_loss": 3.4044191931511234,
      "tokens_seen": 2192703488
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034133333333333335,
      "loss": 1.5969,
      "theoretical_loss": 3.404410862094696,
      "tokens_seen": 2192769024
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003413131313131313,
      "loss": 1.3766,
      "theoretical_loss": 3.4044025313569737,
      "tokens_seen": 2192834560
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003412929292929293,
      "loss": 1.4505,
      "theoretical_loss": 3.4043942009379347,
      "tokens_seen": 2192900096
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003412727272727273,
      "loss": 1.2071,
      "theoretical_loss": 3.404385870837557,
      "tokens_seen": 2192965632
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034125252525252524,
      "loss": 1.5274,
      "theoretical_loss": 3.404377541055819,
      "tokens_seen": 2193031168
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034123232323232325,
      "loss": 1.4255,
      "theoretical_loss": 3.4043692115926993,
      "tokens_seen": 2193096704
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003412121212121212,
      "loss": 1.2616,
      "theoretical_loss": 3.4043608824481764,
      "tokens_seen": 2193162240
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003411919191919192,
      "loss": 1.4435,
      "theoretical_loss": 3.4043525536222274,
      "tokens_seen": 2193227776
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034117171717171717,
      "loss": 1.6325,
      "theoretical_loss": 3.4043442251148317,
      "tokens_seen": 2193293312
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003411515151515152,
      "loss": 1.5136,
      "theoretical_loss": 3.404335896925967,
      "tokens_seen": 2193358848
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034113131313131314,
      "loss": 1.4908,
      "theoretical_loss": 3.4043275690556123,
      "tokens_seen": 2193424384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003411111111111111,
      "loss": 1.4977,
      "theoretical_loss": 3.404319241503745,
      "tokens_seen": 2193489920
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003410909090909091,
      "loss": 1.5878,
      "theoretical_loss": 3.404310914270344,
      "tokens_seen": 2193555456
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034107070707070706,
      "loss": 1.5565,
      "theoretical_loss": 3.404302587355388,
      "tokens_seen": 2193620992
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0003410505050505051,
      "loss": 1.436,
      "theoretical_loss": 3.404294260758854,
      "tokens_seen": 2193686528
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034103030303030303,
      "loss": 1.361,
      "theoretical_loss": 3.4042859344807215,
      "tokens_seen": 2193752064
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000341010101010101,
      "loss": 1.3996,
      "theoretical_loss": 3.4042776085209683,
      "tokens_seen": 2193817600
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000340989898989899,
      "loss": 1.3523,
      "theoretical_loss": 3.404269282879573,
      "tokens_seen": 2193883136
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.000340969696969697,
      "loss": 1.4947,
      "theoretical_loss": 3.4042609575565135,
      "tokens_seen": 2193948672
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.00034094949494949497,
      "loss": 1.3769,
      "theoretical_loss": 3.404252632551769,
      "tokens_seen": 2194014208
    },
    {
      "epoch": 0.32,
      "objective/train/aligned_token_count": 0.5155791640281677,
      "objective/train/avg_token_score": 0.0181901715695858,
      "objective/train/docs_used": 1243185,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.684660792350769,
      "objective/train/original_loss": 4.145715713500977,
      "objective/train/theoretical_loss": 3.404244307865316,
      "objective/train/tokens_used": 552938976,
      "objective/train/unlikelihood_loss": 0.03865066543221474,
      "theoretical_loss": 3.404244307865316,
      "tokens_seen": 2194079744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003409292929292929,
      "loss": 1.5198,
      "theoretical_loss": 3.404244307865316,
      "tokens_seen": 2194079744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003409090909090909,
      "loss": 1.4695,
      "theoretical_loss": 3.4042359834971347,
      "tokens_seen": 2194145280
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003408888888888889,
      "loss": 1.588,
      "theoretical_loss": 3.4042276594472027,
      "tokens_seen": 2194210816
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003408686868686869,
      "loss": 1.5145,
      "theoretical_loss": 3.4042193357154984,
      "tokens_seen": 2194276352
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034084848484848486,
      "loss": 1.589,
      "theoretical_loss": 3.404211012302,
      "tokens_seen": 2194341888
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003408282828282828,
      "loss": 1.3098,
      "theoretical_loss": 3.404202689206686,
      "tokens_seen": 2194407424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003408080808080808,
      "loss": 1.3523,
      "theoretical_loss": 3.4041943664295347,
      "tokens_seen": 2194472960
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034078787878787884,
      "loss": 1.4896,
      "theoretical_loss": 3.4041860439705243,
      "tokens_seen": 2194538496
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003407676767676768,
      "loss": 1.4257,
      "theoretical_loss": 3.404177721829633,
      "tokens_seen": 2194604032
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034074747474747475,
      "loss": 1.4387,
      "theoretical_loss": 3.4041694000068397,
      "tokens_seen": 2194669568
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003407272727272727,
      "loss": 1.4608,
      "theoretical_loss": 3.4041610785021224,
      "tokens_seen": 2194735104
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034070707070707067,
      "loss": 1.4205,
      "theoretical_loss": 3.4041527573154595,
      "tokens_seen": 2194800640
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034068686868686873,
      "loss": 1.4233,
      "theoretical_loss": 3.404144436446829,
      "tokens_seen": 2194866176
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003406666666666667,
      "loss": 1.4875,
      "theoretical_loss": 3.40413611589621,
      "tokens_seen": 2194931712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034064646464646465,
      "loss": 1.4283,
      "theoretical_loss": 3.40412779566358,
      "tokens_seen": 2194997248
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003406262626262626,
      "loss": 1.6025,
      "theoretical_loss": 3.4041194757489177,
      "tokens_seen": 2195062784
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034060606060606056,
      "loss": 1.5291,
      "theoretical_loss": 3.4041111561522017,
      "tokens_seen": 2195128320
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003405858585858586,
      "loss": 1.6309,
      "theoretical_loss": 3.4041028368734105,
      "tokens_seen": 2195193856
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003405656565656566,
      "loss": 1.5667,
      "theoretical_loss": 3.4040945179125215,
      "tokens_seen": 2195259392
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034054545454545454,
      "loss": 1.4927,
      "theoretical_loss": 3.404086199269514,
      "tokens_seen": 2195324928
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003405252525252525,
      "loss": 1.5115,
      "theoretical_loss": 3.4040778809443664,
      "tokens_seen": 2195390464
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034050505050505056,
      "loss": 1.3444,
      "theoretical_loss": 3.4040695629370563,
      "tokens_seen": 2195456000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003404848484848485,
      "loss": 1.6481,
      "theoretical_loss": 3.4040612452475627,
      "tokens_seen": 2195521536
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003404646464646465,
      "loss": 1.5376,
      "theoretical_loss": 3.4040529278758638,
      "tokens_seen": 2195587072
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034044444444444443,
      "loss": 1.5275,
      "theoretical_loss": 3.4040446108219378,
      "tokens_seen": 2195652608
    },
    {
      "epoch": 0.33,
      "objective/train/aligned_token_count": 0.3767717480659485,
      "objective/train/avg_token_score": 0.024272341281175613,
      "objective/train/docs_used": 1244493,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2653477191925049,
      "objective/train/original_loss": 4.593385219573975,
      "objective/train/theoretical_loss": 3.4040362940857634,
      "objective/train/tokens_used": 554577376,
      "objective/train/unlikelihood_loss": 0.10035522282123566,
      "theoretical_loss": 3.4040362940857634,
      "tokens_seen": 2195718144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003404242424242424,
      "loss": 1.3536,
      "theoretical_loss": 3.4040362940857634,
      "tokens_seen": 2195718144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034040404040404045,
      "loss": 1.5952,
      "theoretical_loss": 3.4040279776673183,
      "tokens_seen": 2195783680
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003403838383838384,
      "loss": 1.4917,
      "theoretical_loss": 3.404019661566582,
      "tokens_seen": 2195849216
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034036363636363637,
      "loss": 1.5986,
      "theoretical_loss": 3.4040113457835317,
      "tokens_seen": 2195914752
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003403434343434343,
      "loss": 1.4393,
      "theoretical_loss": 3.4040030303181465,
      "tokens_seen": 2195980288
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034032323232323234,
      "loss": 1.4309,
      "theoretical_loss": 3.403994715170405,
      "tokens_seen": 2196045824
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034030303030303035,
      "loss": 1.372,
      "theoretical_loss": 3.403986400340285,
      "tokens_seen": 2196111360
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003402828282828283,
      "loss": 1.426,
      "theoretical_loss": 3.403978085827765,
      "tokens_seen": 2196176896
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034026262626262626,
      "loss": 1.46,
      "theoretical_loss": 3.4039697716328234,
      "tokens_seen": 2196242432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003402424242424242,
      "loss": 1.2766,
      "theoretical_loss": 3.4039614577554387,
      "tokens_seen": 2196307968
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034022222222222223,
      "loss": 1.5789,
      "theoretical_loss": 3.4039531441955897,
      "tokens_seen": 2196373504
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034020202020202024,
      "loss": 1.4842,
      "theoretical_loss": 3.403944830953254,
      "tokens_seen": 2196439040
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003401818181818182,
      "loss": 1.3811,
      "theoretical_loss": 3.4039365180284102,
      "tokens_seen": 2196504576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034016161616161615,
      "loss": 1.5219,
      "theoretical_loss": 3.403928205421037,
      "tokens_seen": 2196570112
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034014141414141416,
      "loss": 1.4855,
      "theoretical_loss": 3.403919893131113,
      "tokens_seen": 2196635648
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003401212121212121,
      "loss": 1.5378,
      "theoretical_loss": 3.4039115811586163,
      "tokens_seen": 2196701184
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034010101010101013,
      "loss": 1.5319,
      "theoretical_loss": 3.403903269503525,
      "tokens_seen": 2196766720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003400808080808081,
      "loss": 1.3808,
      "theoretical_loss": 3.403894958165818,
      "tokens_seen": 2196832256
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034006060606060605,
      "loss": 1.5228,
      "theoretical_loss": 3.4038866471454736,
      "tokens_seen": 2196897792
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034004040404040406,
      "loss": 1.2662,
      "theoretical_loss": 3.4038783364424696,
      "tokens_seen": 2196963328
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000340020202020202,
      "loss": 1.5777,
      "theoretical_loss": 3.4038700260567856,
      "tokens_seen": 2197028864
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00034,
      "loss": 1.563,
      "theoretical_loss": 3.4038617159883993,
      "tokens_seen": 2197094400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000339979797979798,
      "loss": 1.4058,
      "theoretical_loss": 3.403853406237289,
      "tokens_seen": 2197159936
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000339959595959596,
      "loss": 1.2944,
      "theoretical_loss": 3.4038450968034333,
      "tokens_seen": 2197225472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033993939393939395,
      "loss": 1.395,
      "theoretical_loss": 3.4038367876868105,
      "tokens_seen": 2197291008
    },
    {
      "epoch": 0.33,
      "objective/train/aligned_token_count": 0.49163001775741577,
      "objective/train/avg_token_score": 0.007524600252509117,
      "objective/train/docs_used": 1244898,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5049396753311157,
      "objective/train/original_loss": 3.5132977962493896,
      "objective/train/theoretical_loss": 3.4038284788873994,
      "objective/train/tokens_used": 556215776,
      "objective/train/unlikelihood_loss": 0.09745852649211884,
      "theoretical_loss": 3.4038284788873994,
      "tokens_seen": 2197356544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003399191919191919,
      "loss": 1.4767,
      "theoretical_loss": 3.4038284788873994,
      "tokens_seen": 2197356544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398989898989899,
      "loss": 1.491,
      "theoretical_loss": 3.403820170405178,
      "tokens_seen": 2197422080
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398787878787879,
      "loss": 1.4289,
      "theoretical_loss": 3.403811862240125,
      "tokens_seen": 2197487616
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398585858585859,
      "loss": 1.6081,
      "theoretical_loss": 3.403803554392219,
      "tokens_seen": 2197553152
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033983838383838384,
      "loss": 1.4993,
      "theoretical_loss": 3.403795246861438,
      "tokens_seen": 2197618688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003398181818181818,
      "loss": 1.3579,
      "theoretical_loss": 3.4037869396477607,
      "tokens_seen": 2197684224
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003397979797979798,
      "loss": 1.3509,
      "theoretical_loss": 3.4037786327511657,
      "tokens_seen": 2197749760
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003397777777777778,
      "loss": 1.4229,
      "theoretical_loss": 3.4037703261716308,
      "tokens_seen": 2197815296
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003397575757575758,
      "loss": 1.1829,
      "theoretical_loss": 3.403762019909135,
      "tokens_seen": 2197880832
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033973737373737373,
      "loss": 1.6943,
      "theoretical_loss": 3.4037537139636567,
      "tokens_seen": 2197946368
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003397171717171717,
      "loss": 1.3778,
      "theoretical_loss": 3.4037454083351744,
      "tokens_seen": 2198011904
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396969696969697,
      "loss": 1.3798,
      "theoretical_loss": 3.403737103023666,
      "tokens_seen": 2198077440
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396767676767677,
      "loss": 1.3101,
      "theoretical_loss": 3.403728798029111,
      "tokens_seen": 2198142976
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033965656565656567,
      "loss": 1.3486,
      "theoretical_loss": 3.4037204933514866,
      "tokens_seen": 2198208512
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033963636363636363,
      "loss": 1.1413,
      "theoretical_loss": 3.403712188990772,
      "tokens_seen": 2198274048
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003396161616161616,
      "loss": 1.2034,
      "theoretical_loss": 3.403703884946946,
      "tokens_seen": 2198339584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033959595959595965,
      "loss": 1.5705,
      "theoretical_loss": 3.4036955812199863,
      "tokens_seen": 2198405120
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003395757575757576,
      "loss": 1.5245,
      "theoretical_loss": 3.4036872778098717,
      "tokens_seen": 2198470656
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033955555555555556,
      "loss": 1.3485,
      "theoretical_loss": 3.4036789747165805,
      "tokens_seen": 2198536192
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003395353535353535,
      "loss": 1.3644,
      "theoretical_loss": 3.403670671940091,
      "tokens_seen": 2198601728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003395151515151515,
      "loss": 1.5472,
      "theoretical_loss": 3.403662369480383,
      "tokens_seen": 2198667264
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033949494949494954,
      "loss": 1.2719,
      "theoretical_loss": 3.403654067337433,
      "tokens_seen": 2198732800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003394747474747475,
      "loss": 1.3775,
      "theoretical_loss": 3.4036457655112207,
      "tokens_seen": 2198798336
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033945454545454546,
      "loss": 1.4753,
      "theoretical_loss": 3.403637464001724,
      "tokens_seen": 2198863872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003394343434343434,
      "loss": 1.3135,
      "theoretical_loss": 3.4036291628089224,
      "tokens_seen": 2198929408
    },
    {
      "epoch": 0.33,
      "objective/train/aligned_token_count": 0.47904446721076965,
      "objective/train/avg_token_score": 0.011134693399071693,
      "objective/train/docs_used": 1246280,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.650755763053894,
      "objective/train/original_loss": 4.499822616577148,
      "objective/train/theoretical_loss": 3.403620861932793,
      "objective/train/tokens_used": 557854176,
      "objective/train/unlikelihood_loss": 0.06235527992248535,
      "theoretical_loss": 3.403620861932793,
      "tokens_seen": 2198994944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003394141414141414,
      "loss": 1.4781,
      "theoretical_loss": 3.403620861932793,
      "tokens_seen": 2198994944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033939393939393943,
      "loss": 1.2679,
      "theoretical_loss": 3.4036125613733152,
      "tokens_seen": 2199060480
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003393737373737374,
      "loss": 1.5773,
      "theoretical_loss": 3.403604261130467,
      "tokens_seen": 2199126016
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033935353535353535,
      "loss": 1.2901,
      "theoretical_loss": 3.4035959612042275,
      "tokens_seen": 2199191552
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003393333333333333,
      "loss": 1.4478,
      "theoretical_loss": 3.4035876615945746,
      "tokens_seen": 2199257088
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003393131313131313,
      "loss": 1.3237,
      "theoretical_loss": 3.4035793623014867,
      "tokens_seen": 2199322624
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033929292929292933,
      "loss": 1.4784,
      "theoretical_loss": 3.403571063324943,
      "tokens_seen": 2199388160
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003392727272727273,
      "loss": 1.4187,
      "theoretical_loss": 3.403562764664921,
      "tokens_seen": 2199453696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033925252525252524,
      "loss": 1.5308,
      "theoretical_loss": 3.4035544663214004,
      "tokens_seen": 2199519232
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033923232323232325,
      "loss": 1.3365,
      "theoretical_loss": 3.403546168294359,
      "tokens_seen": 2199584768
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003392121212121212,
      "loss": 1.3832,
      "theoretical_loss": 3.403537870583775,
      "tokens_seen": 2199650304
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003391919191919192,
      "loss": 1.6406,
      "theoretical_loss": 3.4035295731896276,
      "tokens_seen": 2199715840
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003391717171717172,
      "loss": 1.3674,
      "theoretical_loss": 3.4035212761118947,
      "tokens_seen": 2199781376
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033915151515151513,
      "loss": 1.3445,
      "theoretical_loss": 3.4035129793505554,
      "tokens_seen": 2199846912
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033913131313131315,
      "loss": 1.4655,
      "theoretical_loss": 3.403504682905588,
      "tokens_seen": 2199912448
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003391111111111111,
      "loss": 1.4708,
      "theoretical_loss": 3.40349638677697,
      "tokens_seen": 2199977984
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003390909090909091,
      "loss": 1.3476,
      "theoretical_loss": 3.403488090964682,
      "tokens_seen": 2200043520
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033907070707070707,
      "loss": 1.4378,
      "theoretical_loss": 3.4034797954687006,
      "tokens_seen": 2200109056
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000339050505050505,
      "loss": 1.522,
      "theoretical_loss": 3.403471500289005,
      "tokens_seen": 2200174592
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033903030303030304,
      "loss": 1.4803,
      "theoretical_loss": 3.4034632054255742,
      "tokens_seen": 2200240128
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000339010101010101,
      "loss": 1.5468,
      "theoretical_loss": 3.403454910878386,
      "tokens_seen": 2200305664
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000338989898989899,
      "loss": 1.5615,
      "theoretical_loss": 3.4034466166474195,
      "tokens_seen": 2200371200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033896969696969696,
      "loss": 1.5067,
      "theoretical_loss": 3.4034383227326526,
      "tokens_seen": 2200436736
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000338949494949495,
      "loss": 1.4688,
      "theoretical_loss": 3.4034300291340647,
      "tokens_seen": 2200502272
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033892929292929293,
      "loss": 1.4446,
      "theoretical_loss": 3.403421735851633,
      "tokens_seen": 2200567808
    },
    {
      "epoch": 0.33,
      "objective/train/aligned_token_count": 0.43071848154067993,
      "objective/train/avg_token_score": 0.009768429212272167,
      "objective/train/docs_used": 1246987,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2987014055252075,
      "objective/train/original_loss": 4.015090465545654,
      "objective/train/theoretical_loss": 3.4034134428853378,
      "objective/train/tokens_used": 559492576,
      "objective/train/unlikelihood_loss": 0.08210985362529755,
      "theoretical_loss": 3.4034134428853378,
      "tokens_seen": 2200633344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003389090909090909,
      "loss": 1.3704,
      "theoretical_loss": 3.4034134428853378,
      "tokens_seen": 2200633344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003388888888888889,
      "loss": 1.3965,
      "theoretical_loss": 3.403405150235156,
      "tokens_seen": 2200698880
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033886868686868686,
      "loss": 1.3968,
      "theoretical_loss": 3.403396857901067,
      "tokens_seen": 2200764416
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033884848484848487,
      "loss": 1.4845,
      "theoretical_loss": 3.4033885658830494,
      "tokens_seen": 2200829952
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003388282828282828,
      "loss": 1.339,
      "theoretical_loss": 3.403380274181081,
      "tokens_seen": 2200895488
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003388080808080808,
      "loss": 1.4995,
      "theoretical_loss": 3.403371982795141,
      "tokens_seen": 2200961024
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387878787878788,
      "loss": 1.52,
      "theoretical_loss": 3.403363691725208,
      "tokens_seen": 2201026560
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387676767676768,
      "loss": 1.7699,
      "theoretical_loss": 3.40335540097126,
      "tokens_seen": 2201092096
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033874747474747476,
      "loss": 1.4613,
      "theoretical_loss": 3.403347110533276,
      "tokens_seen": 2201157632
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387272727272727,
      "loss": 1.4194,
      "theoretical_loss": 3.4033388204112347,
      "tokens_seen": 2201223168
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003387070707070707,
      "loss": 1.3421,
      "theoretical_loss": 3.403330530605114,
      "tokens_seen": 2201288704
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386868686868687,
      "loss": 1.4333,
      "theoretical_loss": 3.403322241114893,
      "tokens_seen": 2201354240
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386666666666667,
      "loss": 1.3605,
      "theoretical_loss": 3.40331395194055,
      "tokens_seen": 2201419776
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033864646464646465,
      "loss": 1.5161,
      "theoretical_loss": 3.403305663082064,
      "tokens_seen": 2201485312
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386262626262626,
      "loss": 1.5184,
      "theoretical_loss": 3.403297374539413,
      "tokens_seen": 2201550848
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003386060606060606,
      "loss": 1.6142,
      "theoretical_loss": 3.4032890863125758,
      "tokens_seen": 2201616384
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033858585858585863,
      "loss": 1.5271,
      "theoretical_loss": 3.403280798401531,
      "tokens_seen": 2201681920
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003385656565656566,
      "loss": 1.518,
      "theoretical_loss": 3.403272510806257,
      "tokens_seen": 2201747456
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033854545454545454,
      "loss": 1.4669,
      "theoretical_loss": 3.4032642235267327,
      "tokens_seen": 2201812992
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003385252525252525,
      "loss": 1.3459,
      "theoretical_loss": 3.4032559365629362,
      "tokens_seen": 2201878528
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003385050505050505,
      "loss": 1.5996,
      "theoretical_loss": 3.4032476499148467,
      "tokens_seen": 2201944064
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003384848484848485,
      "loss": 1.4353,
      "theoretical_loss": 3.4032393635824416,
      "tokens_seen": 2202009600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003384646464646465,
      "loss": 1.3004,
      "theoretical_loss": 3.4032310775657013,
      "tokens_seen": 2202075136
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033844444444444444,
      "loss": 1.5033,
      "theoretical_loss": 3.4032227918646027,
      "tokens_seen": 2202140672
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003384242424242424,
      "loss": 1.4705,
      "theoretical_loss": 3.4032145064791255,
      "tokens_seen": 2202206208
    },
    {
      "epoch": 0.33,
      "objective/train/aligned_token_count": 0.48478737473487854,
      "objective/train/avg_token_score": 0.010392933152616024,
      "objective/train/docs_used": 1247624,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6772665977478027,
      "objective/train/original_loss": 4.189998626708984,
      "objective/train/theoretical_loss": 3.4032062214092473,
      "objective/train/tokens_used": 561130976,
      "objective/train/unlikelihood_loss": 0.05547265708446503,
      "theoretical_loss": 3.4032062214092473,
      "tokens_seen": 2202271744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033840404040404046,
      "loss": 1.4897,
      "theoretical_loss": 3.4032062214092473,
      "tokens_seen": 2202271744
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383838383838384,
      "loss": 1.6519,
      "theoretical_loss": 3.403197936654948,
      "tokens_seen": 2202337280
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383636363636364,
      "loss": 1.5678,
      "theoretical_loss": 3.403189652216205,
      "tokens_seen": 2202402816
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033834343434343433,
      "loss": 1.5457,
      "theoretical_loss": 3.403181368092997,
      "tokens_seen": 2202468352
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003383232323232323,
      "loss": 1.6627,
      "theoretical_loss": 3.4031730842853034,
      "tokens_seen": 2202533888
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033830303030303035,
      "loss": 1.376,
      "theoretical_loss": 3.403164800793102,
      "tokens_seen": 2202599424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003382828282828283,
      "loss": 1.4554,
      "theoretical_loss": 3.4031565176163716,
      "tokens_seen": 2202664960
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033826262626262627,
      "loss": 1.4945,
      "theoretical_loss": 3.403148234755091,
      "tokens_seen": 2202730496
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003382424242424242,
      "loss": 1.524,
      "theoretical_loss": 3.403139952209239,
      "tokens_seen": 2202796032
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033822222222222223,
      "loss": 1.3153,
      "theoretical_loss": 3.4031316699787935,
      "tokens_seen": 2202861568
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033820202020202024,
      "loss": 1.3698,
      "theoretical_loss": 3.403123388063734,
      "tokens_seen": 2202927104
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003381818181818182,
      "loss": 1.4519,
      "theoretical_loss": 3.403115106464038,
      "tokens_seen": 2202992640
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033816161616161616,
      "loss": 1.4998,
      "theoretical_loss": 3.403106825179685,
      "tokens_seen": 2203058176
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003381414141414141,
      "loss": 1.4254,
      "theoretical_loss": 3.403098544210654,
      "tokens_seen": 2203123712
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003381212121212121,
      "loss": 1.1569,
      "theoretical_loss": 3.403090263556922,
      "tokens_seen": 2203189248
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033810101010101014,
      "loss": 1.8033,
      "theoretical_loss": 3.403081983218469,
      "tokens_seen": 2203254784
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003380808080808081,
      "loss": 1.7503,
      "theoretical_loss": 3.403073703195273,
      "tokens_seen": 2203320320
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033806060606060605,
      "loss": 1.3272,
      "theoretical_loss": 3.403065423487313,
      "tokens_seen": 2203385856
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033804040404040406,
      "loss": 1.4063,
      "theoretical_loss": 3.4030571440945674,
      "tokens_seen": 2203451392
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000338020202020202,
      "loss": 1.2819,
      "theoretical_loss": 3.4030488650170145,
      "tokens_seen": 2203516928
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033800000000000003,
      "loss": 1.547,
      "theoretical_loss": 3.4030405862546336,
      "tokens_seen": 2203582464
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000337979797979798,
      "loss": 1.55,
      "theoretical_loss": 3.403032307807403,
      "tokens_seen": 2203648000
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033795959595959594,
      "loss": 1.3724,
      "theoretical_loss": 3.4030240296753016,
      "tokens_seen": 2203713536
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033793939393939395,
      "loss": 1.346,
      "theoretical_loss": 3.4030157518583075,
      "tokens_seen": 2203779072
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003379191919191919,
      "loss": 1.5798,
      "theoretical_loss": 3.4030074743563996,
      "tokens_seen": 2203844608
    },
    {
      "epoch": 0.33,
      "objective/train/aligned_token_count": 0.3628421425819397,
      "objective/train/avg_token_score": 0.01902676746249199,
      "objective/train/docs_used": 1248832,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1231106519699097,
      "objective/train/original_loss": 4.619980812072754,
      "objective/train/theoretical_loss": 3.4029991971695566,
      "objective/train/tokens_used": 562769376,
      "objective/train/unlikelihood_loss": 0.06308392435312271,
      "theoretical_loss": 3.4029991971695566,
      "tokens_seen": 2203910144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378989898989899,
      "loss": 1.1364,
      "theoretical_loss": 3.4029991971695566,
      "tokens_seen": 2203910144
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378787878787879,
      "loss": 1.3944,
      "theoretical_loss": 3.402990920297757,
      "tokens_seen": 2203975680
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378585858585859,
      "loss": 1.441,
      "theoretical_loss": 3.40298264374098,
      "tokens_seen": 2204041216
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033783838383838385,
      "loss": 1.3021,
      "theoretical_loss": 3.402974367499203,
      "tokens_seen": 2204106752
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003378181818181818,
      "loss": 1.4349,
      "theoretical_loss": 3.402966091572406,
      "tokens_seen": 2204172288
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003377979797979798,
      "loss": 1.3899,
      "theoretical_loss": 3.4029578159605665,
      "tokens_seen": 2204237824
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033777777777777777,
      "loss": 1.3816,
      "theoretical_loss": 3.4029495406636645,
      "tokens_seen": 2204303360
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003377575757575758,
      "loss": 1.331,
      "theoretical_loss": 3.4029412656816773,
      "tokens_seen": 2204368896
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033773737373737374,
      "loss": 1.1773,
      "theoretical_loss": 3.4029329910145845,
      "tokens_seen": 2204434432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003377171717171717,
      "loss": 1.6429,
      "theoretical_loss": 3.4029247166623637,
      "tokens_seen": 2204499968
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003376969696969697,
      "loss": 1.4772,
      "theoretical_loss": 3.402916442624995,
      "tokens_seen": 2204565504
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003376767676767677,
      "loss": 1.2524,
      "theoretical_loss": 3.402908168902456,
      "tokens_seen": 2204631040
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003376565656565657,
      "loss": 1.3954,
      "theoretical_loss": 3.4028998954947256,
      "tokens_seen": 2204696576
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033763636363636363,
      "loss": 1.4426,
      "theoretical_loss": 3.4028916224017824,
      "tokens_seen": 2204762112
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003376161616161616,
      "loss": 1.5947,
      "theoretical_loss": 3.4028833496236057,
      "tokens_seen": 2204827648
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003375959595959596,
      "loss": 1.4298,
      "theoretical_loss": 3.4028750771601732,
      "tokens_seen": 2204893184
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003375757575757576,
      "loss": 1.5938,
      "theoretical_loss": 3.402866805011464,
      "tokens_seen": 2204958720
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033755555555555557,
      "loss": 1.3275,
      "theoretical_loss": 3.402858533177457,
      "tokens_seen": 2205024256
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003375353535353535,
      "loss": 1.4658,
      "theoretical_loss": 3.4028502616581306,
      "tokens_seen": 2205089792
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003375151515151515,
      "loss": 1.6698,
      "theoretical_loss": 3.4028419904534637,
      "tokens_seen": 2205155328
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033749494949494955,
      "loss": 1.2632,
      "theoretical_loss": 3.402833719563435,
      "tokens_seen": 2205220864
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003374747474747475,
      "loss": 1.4245,
      "theoretical_loss": 3.4028254489880227,
      "tokens_seen": 2205286400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033745454545454546,
      "loss": 1.6392,
      "theoretical_loss": 3.4028171787272057,
      "tokens_seen": 2205351936
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003374343434343434,
      "loss": 1.311,
      "theoretical_loss": 3.402808908780963,
      "tokens_seen": 2205417472
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003374141414141414,
      "loss": 1.441,
      "theoretical_loss": 3.402800639149273,
      "tokens_seen": 2205483008
    },
    {
      "epoch": 0.33,
      "objective/train/aligned_token_count": 0.30547410249710083,
      "objective/train/avg_token_score": 0.031316936016082764,
      "objective/train/docs_used": 1249474,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0070712566375732,
      "objective/train/original_loss": 5.012324333190918,
      "objective/train/theoretical_loss": 3.4027923698321145,
      "objective/train/tokens_used": 564407776,
      "objective/train/unlikelihood_loss": 0.06484723091125488,
      "theoretical_loss": 3.4027923698321145,
      "tokens_seen": 2205548544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033739393939393944,
      "loss": 1.4265,
      "theoretical_loss": 3.4027923698321145,
      "tokens_seen": 2205548544
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003373737373737374,
      "loss": 1.4553,
      "theoretical_loss": 3.402784100829466,
      "tokens_seen": 2205614080
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033735353535353535,
      "loss": 1.4714,
      "theoretical_loss": 3.402775832141307,
      "tokens_seen": 2205679616
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003373333333333333,
      "loss": 1.5625,
      "theoretical_loss": 3.4027675637676147,
      "tokens_seen": 2205745152
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033731313131313127,
      "loss": 1.5293,
      "theoretical_loss": 3.402759295708369,
      "tokens_seen": 2205810688
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033729292929292933,
      "loss": 1.7782,
      "theoretical_loss": 3.4027510279635482,
      "tokens_seen": 2205876224
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003372727272727273,
      "loss": 1.4207,
      "theoretical_loss": 3.402742760533131,
      "tokens_seen": 2205941760
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033725252525252525,
      "loss": 1.5677,
      "theoretical_loss": 3.402734493417096,
      "tokens_seen": 2206007296
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003372323232323232,
      "loss": 1.5556,
      "theoretical_loss": 3.402726226615423,
      "tokens_seen": 2206072832
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003372121212121212,
      "loss": 1.2242,
      "theoretical_loss": 3.4027179601280886,
      "tokens_seen": 2206138368
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371919191919192,
      "loss": 1.6313,
      "theoretical_loss": 3.402709693955073,
      "tokens_seen": 2206203904
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371717171717172,
      "loss": 1.3596,
      "theoretical_loss": 3.402701428096355,
      "tokens_seen": 2206269440
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033715151515151514,
      "loss": 1.5203,
      "theoretical_loss": 3.402693162551912,
      "tokens_seen": 2206334976
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371313131313131,
      "loss": 1.6972,
      "theoretical_loss": 3.4026848973217243,
      "tokens_seen": 2206400512
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003371111111111111,
      "loss": 1.2215,
      "theoretical_loss": 3.40267663240577,
      "tokens_seen": 2206466048
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003370909090909091,
      "loss": 1.4475,
      "theoretical_loss": 3.4026683678040275,
      "tokens_seen": 2206531584
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003370707070707071,
      "loss": 1.6499,
      "theoretical_loss": 3.402660103516476,
      "tokens_seen": 2206597120
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033705050505050503,
      "loss": 1.2895,
      "theoretical_loss": 3.402651839543094,
      "tokens_seen": 2206662656
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033703030303030304,
      "loss": 1.3408,
      "theoretical_loss": 3.4026435758838596,
      "tokens_seen": 2206728192
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033701010101010105,
      "loss": 1.4557,
      "theoretical_loss": 3.4026353125387527,
      "tokens_seen": 2206793728
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000336989898989899,
      "loss": 1.3329,
      "theoretical_loss": 3.4026270495077515,
      "tokens_seen": 2206859264
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033696969696969697,
      "loss": 1.4916,
      "theoretical_loss": 3.4026187867908346,
      "tokens_seen": 2206924800
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003369494949494949,
      "loss": 1.4133,
      "theoretical_loss": 3.4026105243879807,
      "tokens_seen": 2206990336
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033692929292929294,
      "loss": 1.6488,
      "theoretical_loss": 3.402602262299169,
      "tokens_seen": 2207055872
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033690909090909095,
      "loss": 1.5597,
      "theoretical_loss": 3.4025940005243775,
      "tokens_seen": 2207121408
    },
    {
      "epoch": 0.33,
      "objective/train/aligned_token_count": 0.45894429087638855,
      "objective/train/avg_token_score": 0.053512170910835266,
      "objective/train/docs_used": 1250567,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.429175615310669,
      "objective/train/original_loss": 4.429356575012207,
      "objective/train/theoretical_loss": 3.4025857390635856,
      "objective/train/tokens_used": 566046176,
      "objective/train/unlikelihood_loss": 0.061125341802835464,
      "theoretical_loss": 3.4025857390635856,
      "tokens_seen": 2207186944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003368888888888889,
      "loss": 1.4504,
      "theoretical_loss": 3.4025857390635856,
      "tokens_seen": 2207186944
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033686868686868686,
      "loss": 1.4028,
      "theoretical_loss": 3.402577477916772,
      "tokens_seen": 2207252480
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033684848484848487,
      "loss": 1.2642,
      "theoretical_loss": 3.402569217083915,
      "tokens_seen": 2207318016
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033682828282828283,
      "loss": 1.4275,
      "theoretical_loss": 3.4025609565649937,
      "tokens_seen": 2207383552
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033680808080808084,
      "loss": 1.2492,
      "theoretical_loss": 3.4025526963599866,
      "tokens_seen": 2207449088
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003367878787878788,
      "loss": 1.5749,
      "theoretical_loss": 3.402544436468873,
      "tokens_seen": 2207514624
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033676767676767675,
      "loss": 1.647,
      "theoretical_loss": 3.402536176891631,
      "tokens_seen": 2207580160
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033674747474747476,
      "loss": 1.633,
      "theoretical_loss": 3.4025279176282397,
      "tokens_seen": 2207645696
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003367272727272727,
      "loss": 1.26,
      "theoretical_loss": 3.4025196586786777,
      "tokens_seen": 2207711232
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033670707070707073,
      "loss": 1.3001,
      "theoretical_loss": 3.4025114000429237,
      "tokens_seen": 2207776768
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366868686868687,
      "loss": 1.6467,
      "theoretical_loss": 3.402503141720957,
      "tokens_seen": 2207842304
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366666666666667,
      "loss": 1.6317,
      "theoretical_loss": 3.4024948837127558,
      "tokens_seen": 2207907840
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033664646464646466,
      "loss": 1.6675,
      "theoretical_loss": 3.4024866260182987,
      "tokens_seen": 2207973376
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366262626262626,
      "loss": 1.524,
      "theoretical_loss": 3.4024783686375653,
      "tokens_seen": 2208038912
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003366060606060606,
      "loss": 1.3489,
      "theoretical_loss": 3.4024701115705334,
      "tokens_seen": 2208104448
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365858585858586,
      "loss": 1.3666,
      "theoretical_loss": 3.4024618548171826,
      "tokens_seen": 2208169984
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365656565656566,
      "loss": 1.2017,
      "theoretical_loss": 3.402453598377491,
      "tokens_seen": 2208235520
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033654545454545455,
      "loss": 1.5482,
      "theoretical_loss": 3.402445342251438,
      "tokens_seen": 2208301056
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365252525252525,
      "loss": 1.5986,
      "theoretical_loss": 3.4024370864390017,
      "tokens_seen": 2208366592
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003365050505050505,
      "loss": 1.5721,
      "theoretical_loss": 3.4024288309401616,
      "tokens_seen": 2208432128
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033648484848484853,
      "loss": 1.5885,
      "theoretical_loss": 3.402420575754896,
      "tokens_seen": 2208497664
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003364646464646465,
      "loss": 1.4714,
      "theoretical_loss": 3.4024123208831836,
      "tokens_seen": 2208563200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033644444444444444,
      "loss": 1.2546,
      "theoretical_loss": 3.402404066325004,
      "tokens_seen": 2208628736
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003364242424242424,
      "loss": 1.1764,
      "theoretical_loss": 3.402395812080335,
      "tokens_seen": 2208694272
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003364040404040404,
      "loss": 1.3605,
      "theoretical_loss": 3.4023875581491554,
      "tokens_seen": 2208759808
    },
    {
      "epoch": 0.33,
      "objective/train/aligned_token_count": 0.47605082392692566,
      "objective/train/avg_token_score": 0.010573615320026875,
      "objective/train/docs_used": 1251703,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.720569372177124,
      "objective/train/original_loss": 4.1492743492126465,
      "objective/train/theoretical_loss": 3.4023793045314448,
      "objective/train/tokens_used": 567684576,
      "objective/train/unlikelihood_loss": 0.05719097703695297,
      "theoretical_loss": 3.4023793045314448,
      "tokens_seen": 2208825344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363838383838384,
      "loss": 1.5406,
      "theoretical_loss": 3.4023793045314448,
      "tokens_seen": 2208825344
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363636363636364,
      "loss": 1.4489,
      "theoretical_loss": 3.402371051227181,
      "tokens_seen": 2208890880
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033634343434343434,
      "loss": 1.5043,
      "theoretical_loss": 3.4023627982363442,
      "tokens_seen": 2208956416
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003363232323232323,
      "loss": 1.6471,
      "theoretical_loss": 3.402354545558912,
      "tokens_seen": 2209021952
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033630303030303036,
      "loss": 1.6153,
      "theoretical_loss": 3.4023462931948636,
      "tokens_seen": 2209087488
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003362828282828283,
      "loss": 1.4899,
      "theoretical_loss": 3.4023380411441777,
      "tokens_seen": 2209153024
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033626262626262627,
      "loss": 1.2756,
      "theoretical_loss": 3.402329789406833,
      "tokens_seen": 2209218560
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033624242424242423,
      "loss": 1.4633,
      "theoretical_loss": 3.402321537982809,
      "tokens_seen": 2209284096
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003362222222222222,
      "loss": 1.3959,
      "theoretical_loss": 3.402313286872084,
      "tokens_seen": 2209349632
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033620202020202025,
      "loss": 1.4587,
      "theoretical_loss": 3.4023050360746363,
      "tokens_seen": 2209415168
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003361818181818182,
      "loss": 1.4366,
      "theoretical_loss": 3.4022967855904453,
      "tokens_seen": 2209480704
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033616161616161616,
      "loss": 1.4386,
      "theoretical_loss": 3.4022885354194896,
      "tokens_seen": 2209546240
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003361414141414141,
      "loss": 1.2041,
      "theoretical_loss": 3.4022802855617487,
      "tokens_seen": 2209611776
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033612121212121213,
      "loss": 1.3323,
      "theoretical_loss": 3.4022720360172007,
      "tokens_seen": 2209677312
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033610101010101014,
      "loss": 1.3245,
      "theoretical_loss": 3.402263786785824,
      "tokens_seen": 2209742848
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003360808080808081,
      "loss": 1.527,
      "theoretical_loss": 3.4022555378675987,
      "tokens_seen": 2209808384
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033606060606060606,
      "loss": 1.2025,
      "theoretical_loss": 3.402247289262503,
      "tokens_seen": 2209873920
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000336040404040404,
      "loss": 1.2066,
      "theoretical_loss": 3.402239040970515,
      "tokens_seen": 2209939456
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000336020202020202,
      "loss": 1.4186,
      "theoretical_loss": 3.4022307929916145,
      "tokens_seen": 2210004992
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033600000000000004,
      "loss": 1.3214,
      "theoretical_loss": 3.40222254532578,
      "tokens_seen": 2210070528
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000335979797979798,
      "loss": 1.3915,
      "theoretical_loss": 3.4022142979729906,
      "tokens_seen": 2210136064
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033595959595959595,
      "loss": 1.3737,
      "theoretical_loss": 3.402206050933225,
      "tokens_seen": 2210201600
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033593939393939396,
      "loss": 1.4327,
      "theoretical_loss": 3.4021978042064616,
      "tokens_seen": 2210267136
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0003359191919191919,
      "loss": 1.4071,
      "theoretical_loss": 3.4021895577926795,
      "tokens_seen": 2210332672
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.00033589898989898993,
      "loss": 1.5206,
      "theoretical_loss": 3.402181311691858,
      "tokens_seen": 2210398208
    },
    {
      "epoch": 0.33,
      "objective/train/aligned_token_count": 0.41006842255592346,
      "objective/train/avg_token_score": 0.009129072539508343,
      "objective/train/docs_used": 1252332,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.273158311843872,
      "objective/train/original_loss": 4.655575752258301,
      "objective/train/theoretical_loss": 3.402173065903975,
      "objective/train/tokens_used": 569322976,
      "objective/train/unlikelihood_loss": 0.07809732854366302,
      "theoretical_loss": 3.402173065903975,
      "tokens_seen": 2210463744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003358787878787879,
      "loss": 1.4114,
      "theoretical_loss": 3.402173065903975,
      "tokens_seen": 2210463744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033585858585858584,
      "loss": 1.4299,
      "theoretical_loss": 3.4021648204290105,
      "tokens_seen": 2210529280
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033583838383838385,
      "loss": 1.5837,
      "theoretical_loss": 3.4021565752669423,
      "tokens_seen": 2210594816
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003358181818181818,
      "loss": 1.1277,
      "theoretical_loss": 3.40214833041775,
      "tokens_seen": 2210660352
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003357979797979798,
      "loss": 1.3065,
      "theoretical_loss": 3.4021400858814124,
      "tokens_seen": 2210725888
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003357777777777778,
      "loss": 1.1862,
      "theoretical_loss": 3.4021318416579076,
      "tokens_seen": 2210791424
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033575757575757573,
      "loss": 1.2753,
      "theoretical_loss": 3.4021235977472153,
      "tokens_seen": 2210856960
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033573737373737375,
      "loss": 1.3462,
      "theoretical_loss": 3.4021153541493137,
      "tokens_seen": 2210922496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003357171717171717,
      "loss": 1.5668,
      "theoretical_loss": 3.4021071108641823,
      "tokens_seen": 2210988032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356969696969697,
      "loss": 1.2746,
      "theoretical_loss": 3.4020988678918,
      "tokens_seen": 2211053568
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033567676767676767,
      "loss": 1.6774,
      "theoretical_loss": 3.4020906252321446,
      "tokens_seen": 2211119104
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356565656565657,
      "loss": 1.3521,
      "theoretical_loss": 3.4020823828851956,
      "tokens_seen": 2211184640
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033563636363636364,
      "loss": 1.5306,
      "theoretical_loss": 3.4020741408509325,
      "tokens_seen": 2211250176
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003356161616161616,
      "loss": 1.4494,
      "theoretical_loss": 3.4020658991293335,
      "tokens_seen": 2211315712
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355959595959596,
      "loss": 1.616,
      "theoretical_loss": 3.402057657720378,
      "tokens_seen": 2211381248
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033557575757575756,
      "loss": 1.5117,
      "theoretical_loss": 3.4020494166240436,
      "tokens_seen": 2211446784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355555555555556,
      "loss": 1.6153,
      "theoretical_loss": 3.402041175840311,
      "tokens_seen": 2211512320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033553535353535353,
      "loss": 1.3602,
      "theoretical_loss": 3.4020329353691574,
      "tokens_seen": 2211577856
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003355151515151515,
      "loss": 1.328,
      "theoretical_loss": 3.402024695210563,
      "tokens_seen": 2211643392
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354949494949495,
      "loss": 1.3722,
      "theoretical_loss": 3.4020164553645054,
      "tokens_seen": 2211708928
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354747474747475,
      "loss": 1.3442,
      "theoretical_loss": 3.4020082158309646,
      "tokens_seen": 2211774464
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033545454545454547,
      "loss": 1.2491,
      "theoretical_loss": 3.4019999766099187,
      "tokens_seen": 2211840000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354343434343434,
      "loss": 1.3416,
      "theoretical_loss": 3.4019917377013473,
      "tokens_seen": 2211905536
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003354141414141414,
      "loss": 1.296,
      "theoretical_loss": 3.401983499105229,
      "tokens_seen": 2211971072
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353939393939394,
      "loss": 1.3672,
      "theoretical_loss": 3.4019752608215423,
      "tokens_seen": 2212036608
    },
    {
      "epoch": 0.34,
      "objective/train/aligned_token_count": 0.563599705696106,
      "objective/train/avg_token_score": 0.01548867579549551,
      "objective/train/docs_used": 1253043,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9471914768218994,
      "objective/train/original_loss": 3.8710668087005615,
      "objective/train/theoretical_loss": 3.401967022850267,
      "objective/train/tokens_used": 570961376,
      "objective/train/unlikelihood_loss": 0.06240874528884888,
      "theoretical_loss": 3.401967022850267,
      "tokens_seen": 2212102144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353737373737374,
      "loss": 1.6499,
      "theoretical_loss": 3.401967022850267,
      "tokens_seen": 2212102144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033535353535353536,
      "loss": 1.5985,
      "theoretical_loss": 3.401958785191381,
      "tokens_seen": 2212167680
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003353333333333333,
      "loss": 1.3165,
      "theoretical_loss": 3.401950547844864,
      "tokens_seen": 2212233216
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033531313131313133,
      "loss": 1.3792,
      "theoretical_loss": 3.4019423108106945,
      "tokens_seen": 2212298752
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033529292929292934,
      "loss": 1.4826,
      "theoretical_loss": 3.4019340740888513,
      "tokens_seen": 2212364288
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003352727272727273,
      "loss": 1.0717,
      "theoretical_loss": 3.401925837679313,
      "tokens_seen": 2212429824
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033525252525252525,
      "loss": 1.4916,
      "theoretical_loss": 3.40191760158206,
      "tokens_seen": 2212495360
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003352323232323232,
      "loss": 1.4251,
      "theoretical_loss": 3.401909365797069,
      "tokens_seen": 2212560896
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003352121212121212,
      "loss": 1.3757,
      "theoretical_loss": 3.4019011303243207,
      "tokens_seen": 2212626432
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033519191919191923,
      "loss": 1.2861,
      "theoretical_loss": 3.4018928951637935,
      "tokens_seen": 2212691968
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003351717171717172,
      "loss": 1.4191,
      "theoretical_loss": 3.401884660315466,
      "tokens_seen": 2212757504
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033515151515151515,
      "loss": 1.3832,
      "theoretical_loss": 3.4018764257793173,
      "tokens_seen": 2212823040
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003351313131313131,
      "loss": 1.4821,
      "theoretical_loss": 3.4018681915553266,
      "tokens_seen": 2212888576
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033511111111111117,
      "loss": 1.3594,
      "theoretical_loss": 3.4018599576434725,
      "tokens_seen": 2212954112
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350909090909091,
      "loss": 1.279,
      "theoretical_loss": 3.4018517240437336,
      "tokens_seen": 2213019648
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003350707070707071,
      "loss": 1.3921,
      "theoretical_loss": 3.4018434907560895,
      "tokens_seen": 2213085184
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033505050505050504,
      "loss": 1.4204,
      "theoretical_loss": 3.401835257780519,
      "tokens_seen": 2213150720
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000335030303030303,
      "loss": 1.3904,
      "theoretical_loss": 3.401827025117001,
      "tokens_seen": 2213216256
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033501010101010106,
      "loss": 1.5145,
      "theoretical_loss": 3.401818792765514,
      "tokens_seen": 2213281792
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334989898989899,
      "loss": 1.3289,
      "theoretical_loss": 3.4018105607260374,
      "tokens_seen": 2213347328
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334969696969697,
      "loss": 1.3862,
      "theoretical_loss": 3.40180232899855,
      "tokens_seen": 2213412864
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033494949494949493,
      "loss": 1.3783,
      "theoretical_loss": 3.4017940975830308,
      "tokens_seen": 2213478400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033492929292929294,
      "loss": 1.3888,
      "theoretical_loss": 3.4017858664794582,
      "tokens_seen": 2213543936
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033490909090909095,
      "loss": 1.6184,
      "theoretical_loss": 3.401777635687812,
      "tokens_seen": 2213609472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003348888888888889,
      "loss": 1.6573,
      "theoretical_loss": 3.4017694052080705,
      "tokens_seen": 2213675008
    },
    {
      "epoch": 0.34,
      "objective/train/aligned_token_count": 0.3229472041130066,
      "objective/train/avg_token_score": 0.03513987362384796,
      "objective/train/docs_used": 1254214,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9982045292854309,
      "objective/train/original_loss": 5.482150077819824,
      "objective/train/theoretical_loss": 3.4017611750402135,
      "objective/train/tokens_used": 572599776,
      "objective/train/unlikelihood_loss": 0.03328944370150566,
      "theoretical_loss": 3.4017611750402135,
      "tokens_seen": 2213740544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033486868686868687,
      "loss": 1.4133,
      "theoretical_loss": 3.4017611750402135,
      "tokens_seen": 2213740544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003348484848484848,
      "loss": 1.577,
      "theoretical_loss": 3.401752945184219,
      "tokens_seen": 2213806080
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033482828282828283,
      "loss": 1.388,
      "theoretical_loss": 3.401744715640066,
      "tokens_seen": 2213871616
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033480808080808085,
      "loss": 1.4758,
      "theoretical_loss": 3.401736486407734,
      "tokens_seen": 2213937152
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003347878787878788,
      "loss": 1.3819,
      "theoretical_loss": 3.4017282574872016,
      "tokens_seen": 2214002688
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033476767676767676,
      "loss": 1.4519,
      "theoretical_loss": 3.401720028878448,
      "tokens_seen": 2214068224
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033474747474747477,
      "loss": 1.5277,
      "theoretical_loss": 3.4017118005814524,
      "tokens_seen": 2214133760
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003347272727272727,
      "loss": 1.5364,
      "theoretical_loss": 3.4017035725961926,
      "tokens_seen": 2214199296
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033470707070707074,
      "loss": 1.5913,
      "theoretical_loss": 3.401695344922649,
      "tokens_seen": 2214264832
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003346868686868687,
      "loss": 1.052,
      "theoretical_loss": 3.4016871175607997,
      "tokens_seen": 2214330368
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033466666666666665,
      "loss": 1.5706,
      "theoretical_loss": 3.401678890510624,
      "tokens_seen": 2214395904
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033464646464646466,
      "loss": 1.2212,
      "theoretical_loss": 3.4016706637721006,
      "tokens_seen": 2214461440
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003346262626262626,
      "loss": 1.4005,
      "theoretical_loss": 3.4016624373452085,
      "tokens_seen": 2214526976
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033460606060606063,
      "loss": 1.3723,
      "theoretical_loss": 3.401654211229927,
      "tokens_seen": 2214592512
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345858585858586,
      "loss": 1.3648,
      "theoretical_loss": 3.4016459854262346,
      "tokens_seen": 2214658048
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345656565656566,
      "loss": 1.6713,
      "theoretical_loss": 3.4016377599341108,
      "tokens_seen": 2214723584
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033454545454545456,
      "loss": 1.6673,
      "theoretical_loss": 3.4016295347535346,
      "tokens_seen": 2214789120
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345252525252525,
      "loss": 1.411,
      "theoretical_loss": 3.4016213098844843,
      "tokens_seen": 2214854656
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003345050505050505,
      "loss": 1.2615,
      "theoretical_loss": 3.4016130853269395,
      "tokens_seen": 2214920192
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003344848484848485,
      "loss": 1.2258,
      "theoretical_loss": 3.4016048610808785,
      "tokens_seen": 2214985728
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003344646464646465,
      "loss": 1.4159,
      "theoretical_loss": 3.401596637146281,
      "tokens_seen": 2215051264
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033444444444444445,
      "loss": 1.4202,
      "theoretical_loss": 3.4015884135231262,
      "tokens_seen": 2215116800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003344242424242424,
      "loss": 1.5943,
      "theoretical_loss": 3.4015801902113925,
      "tokens_seen": 2215182336
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003344040404040404,
      "loss": 1.1998,
      "theoretical_loss": 3.401571967211059,
      "tokens_seen": 2215247872
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033438383838383843,
      "loss": 1.2044,
      "theoretical_loss": 3.4015637445221043,
      "tokens_seen": 2215313408
    },
    {
      "epoch": 0.34,
      "objective/train/aligned_token_count": 0.4114125072956085,
      "objective/train/avg_token_score": 0.012347203679382801,
      "objective/train/docs_used": 1254739,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3653172254562378,
      "objective/train/original_loss": 4.256288051605225,
      "objective/train/theoretical_loss": 3.4015555221445086,
      "objective/train/tokens_used": 574238176,
      "objective/train/unlikelihood_loss": 0.06681492924690247,
      "theoretical_loss": 3.4015555221445086,
      "tokens_seen": 2215378944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003343636363636364,
      "loss": 1.2438,
      "theoretical_loss": 3.4015555221445086,
      "tokens_seen": 2215378944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033434343434343434,
      "loss": 1.6333,
      "theoretical_loss": 3.4015473000782497,
      "tokens_seen": 2215444480
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003343232323232323,
      "loss": 1.1555,
      "theoretical_loss": 3.401539078323307,
      "tokens_seen": 2215510016
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003343030303030303,
      "loss": 1.2994,
      "theoretical_loss": 3.40153085687966,
      "tokens_seen": 2215575552
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342828282828283,
      "loss": 1.3789,
      "theoretical_loss": 3.4015226357472867,
      "tokens_seen": 2215641088
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342626262626263,
      "loss": 1.2177,
      "theoretical_loss": 3.401514414926167,
      "tokens_seen": 2215706624
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033424242424242423,
      "loss": 1.3103,
      "theoretical_loss": 3.401506194416279,
      "tokens_seen": 2215772160
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003342222222222222,
      "loss": 1.6004,
      "theoretical_loss": 3.4014979742176026,
      "tokens_seen": 2215837696
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033420202020202026,
      "loss": 1.4937,
      "theoretical_loss": 3.401489754330117,
      "tokens_seen": 2215903232
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003341818181818182,
      "loss": 1.305,
      "theoretical_loss": 3.4014815347538,
      "tokens_seen": 2215968768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033416161616161617,
      "loss": 1.5056,
      "theoretical_loss": 3.401473315488632,
      "tokens_seen": 2216034304
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003341414141414141,
      "loss": 1.5286,
      "theoretical_loss": 3.4014650965345905,
      "tokens_seen": 2216099840
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003341212121212121,
      "loss": 1.443,
      "theoretical_loss": 3.401456877891656,
      "tokens_seen": 2216165376
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033410101010101015,
      "loss": 1.4169,
      "theoretical_loss": 3.401448659559807,
      "tokens_seen": 2216230912
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003340808080808081,
      "loss": 1.3985,
      "theoretical_loss": 3.401440441539022,
      "tokens_seen": 2216296448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033406060606060606,
      "loss": 1.3602,
      "theoretical_loss": 3.4014322238292807,
      "tokens_seen": 2216361984
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334040404040404,
      "loss": 1.4641,
      "theoretical_loss": 3.401424006430562,
      "tokens_seen": 2216427520
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000334020202020202,
      "loss": 1.354,
      "theoretical_loss": 3.4014157893428445,
      "tokens_seen": 2216493056
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033400000000000004,
      "loss": 1.3428,
      "theoretical_loss": 3.4014075725661077,
      "tokens_seen": 2216558592
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000333979797979798,
      "loss": 1.3749,
      "theoretical_loss": 3.40139935610033,
      "tokens_seen": 2216624128
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033395959595959595,
      "loss": 1.4769,
      "theoretical_loss": 3.401391139945492,
      "tokens_seen": 2216689664
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003339393939393939,
      "loss": 1.4624,
      "theoretical_loss": 3.401382924101571,
      "tokens_seen": 2216755200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003339191919191919,
      "loss": 1.6038,
      "theoretical_loss": 3.4013747085685466,
      "tokens_seen": 2216820736
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033389898989898993,
      "loss": 1.4452,
      "theoretical_loss": 3.4013664933463983,
      "tokens_seen": 2216886272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003338787878787879,
      "loss": 1.321,
      "theoretical_loss": 3.4013582784351044,
      "tokens_seen": 2216951808
    },
    {
      "epoch": 0.34,
      "objective/train/aligned_token_count": 0.34060361981391907,
      "objective/train/avg_token_score": 0.0206461101770401,
      "objective/train/docs_used": 1256282,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.191922903060913,
      "objective/train/original_loss": 4.967258453369141,
      "objective/train/theoretical_loss": 3.401350063834645,
      "objective/train/tokens_used": 575876576,
      "objective/train/unlikelihood_loss": 0.0664689689874649,
      "theoretical_loss": 3.401350063834645,
      "tokens_seen": 2217017344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033385858585858585,
      "loss": 1.3583,
      "theoretical_loss": 3.401350063834645,
      "tokens_seen": 2217017344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003338383838383838,
      "loss": 1.2539,
      "theoretical_loss": 3.401341849544998,
      "tokens_seen": 2217082880
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003338181818181818,
      "loss": 1.3948,
      "theoretical_loss": 3.401333635566143,
      "tokens_seen": 2217148416
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003337979797979798,
      "loss": 1.5806,
      "theoretical_loss": 3.401325421898059,
      "tokens_seen": 2217213952
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003337777777777778,
      "loss": 1.4388,
      "theoretical_loss": 3.4013172085407253,
      "tokens_seen": 2217279488
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033375757575757574,
      "loss": 1.3891,
      "theoretical_loss": 3.4013089954941202,
      "tokens_seen": 2217345024
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033373737373737375,
      "loss": 1.4754,
      "theoretical_loss": 3.401300782758224,
      "tokens_seen": 2217410560
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003337171717171717,
      "loss": 1.5519,
      "theoretical_loss": 3.4012925703330144,
      "tokens_seen": 2217476096
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003336969696969697,
      "loss": 1.5402,
      "theoretical_loss": 3.4012843582184717,
      "tokens_seen": 2217541632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003336767676767677,
      "loss": 1.4875,
      "theoretical_loss": 3.401276146414574,
      "tokens_seen": 2217607168
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033365656565656563,
      "loss": 1.5459,
      "theoretical_loss": 3.401267934921301,
      "tokens_seen": 2217672704
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033363636363636364,
      "loss": 1.5937,
      "theoretical_loss": 3.4012597237386313,
      "tokens_seen": 2217738240
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033361616161616166,
      "loss": 1.2698,
      "theoretical_loss": 3.401251512866544,
      "tokens_seen": 2217803776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003335959595959596,
      "loss": 1.4168,
      "theoretical_loss": 3.4012433023050193,
      "tokens_seen": 2217869312
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033357575757575757,
      "loss": 1.3082,
      "theoretical_loss": 3.4012350920540344,
      "tokens_seen": 2217934848
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003335555555555556,
      "loss": 1.4065,
      "theoretical_loss": 3.4012268821135696,
      "tokens_seen": 2218000384
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033353535353535354,
      "loss": 1.3413,
      "theoretical_loss": 3.4012186724836035,
      "tokens_seen": 2218065920
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033351515151515155,
      "loss": 1.2611,
      "theoretical_loss": 3.401210463164116,
      "tokens_seen": 2218131456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003334949494949495,
      "loss": 1.1196,
      "theoretical_loss": 3.401202254155085,
      "tokens_seen": 2218196992
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033347474747474746,
      "loss": 1.526,
      "theoretical_loss": 3.4011940454564904,
      "tokens_seen": 2218262528
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033345454545454547,
      "loss": 1.3485,
      "theoretical_loss": 3.401185837068311,
      "tokens_seen": 2218328064
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033343434343434343,
      "loss": 1.453,
      "theoretical_loss": 3.401177628990526,
      "tokens_seen": 2218393600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033341414141414144,
      "loss": 1.5991,
      "theoretical_loss": 3.4011694212231145,
      "tokens_seen": 2218459136
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003333939393939394,
      "loss": 1.4885,
      "theoretical_loss": 3.401161213766055,
      "tokens_seen": 2218524672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003333737373737374,
      "loss": 1.4899,
      "theoretical_loss": 3.401153006619328,
      "tokens_seen": 2218590208
    },
    {
      "epoch": 0.34,
      "objective/train/aligned_token_count": 0.5152736902236938,
      "objective/train/avg_token_score": 0.008905273862183094,
      "objective/train/docs_used": 1256744,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6531444787979126,
      "objective/train/original_loss": 4.157623767852783,
      "objective/train/theoretical_loss": 3.4011447997829114,
      "objective/train/tokens_used": 577514976,
      "objective/train/unlikelihood_loss": 0.05239066481590271,
      "theoretical_loss": 3.4011447997829114,
      "tokens_seen": 2218655744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033335353535353537,
      "loss": 1.501,
      "theoretical_loss": 3.4011447997829114,
      "tokens_seen": 2218655744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003333333333333333,
      "loss": 1.37,
      "theoretical_loss": 3.4011365932567843,
      "tokens_seen": 2218721280
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033331313131313133,
      "loss": 1.2405,
      "theoretical_loss": 3.4011283870409263,
      "tokens_seen": 2218786816
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003332929292929293,
      "loss": 1.6276,
      "theoretical_loss": 3.4011201811353167,
      "tokens_seen": 2218852352
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003332727272727273,
      "loss": 1.2401,
      "theoretical_loss": 3.4011119755399335,
      "tokens_seen": 2218917888
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033325252525252526,
      "loss": 1.6331,
      "theoretical_loss": 3.4011037702547573,
      "tokens_seen": 2218983424
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003332323232323232,
      "loss": 1.5946,
      "theoretical_loss": 3.4010955652797663,
      "tokens_seen": 2219048960
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003332121212121212,
      "loss": 1.5945,
      "theoretical_loss": 3.4010873606149397,
      "tokens_seen": 2219114496
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033319191919191924,
      "loss": 1.3913,
      "theoretical_loss": 3.4010791562602565,
      "tokens_seen": 2219180032
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003331717171717172,
      "loss": 1.6308,
      "theoretical_loss": 3.4010709522156963,
      "tokens_seen": 2219245568
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033315151515151515,
      "loss": 1.4065,
      "theoretical_loss": 3.401062748481238,
      "tokens_seen": 2219311104
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003331313131313131,
      "loss": 1.2411,
      "theoretical_loss": 3.4010545450568603,
      "tokens_seen": 2219376640
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003331111111111111,
      "loss": 1.5126,
      "theoretical_loss": 3.4010463419425427,
      "tokens_seen": 2219442176
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033309090909090913,
      "loss": 1.4424,
      "theoretical_loss": 3.4010381391382647,
      "tokens_seen": 2219507712
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003330707070707071,
      "loss": 1.4699,
      "theoretical_loss": 3.401029936644005,
      "tokens_seen": 2219573248
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033305050505050504,
      "loss": 1.1831,
      "theoretical_loss": 3.4010217344597424,
      "tokens_seen": 2219638784
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000333030303030303,
      "loss": 1.389,
      "theoretical_loss": 3.4010135325854565,
      "tokens_seen": 2219704320
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033301010101010107,
      "loss": 1.4318,
      "theoretical_loss": 3.4010053310211266,
      "tokens_seen": 2219769856
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000332989898989899,
      "loss": 1.6328,
      "theoretical_loss": 3.4009971297667314,
      "tokens_seen": 2219835392
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000332969696969697,
      "loss": 1.5143,
      "theoretical_loss": 3.40098892882225,
      "tokens_seen": 2219900928
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033294949494949494,
      "loss": 1.442,
      "theoretical_loss": 3.4009807281876623,
      "tokens_seen": 2219966464
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003329292929292929,
      "loss": 1.464,
      "theoretical_loss": 3.4009725278629466,
      "tokens_seen": 2220032000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033290909090909096,
      "loss": 1.2395,
      "theoretical_loss": 3.400964327848082,
      "tokens_seen": 2220097536
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003328888888888889,
      "loss": 1.5497,
      "theoretical_loss": 3.4009561281430485,
      "tokens_seen": 2220163072
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033286868686868687,
      "loss": 1.2372,
      "theoretical_loss": 3.4009479287478244,
      "tokens_seen": 2220228608
    },
    {
      "epoch": 0.34,
      "objective/train/aligned_token_count": 0.41874387860298157,
      "objective/train/avg_token_score": 0.012948398478329182,
      "objective/train/docs_used": 1258010,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2894787788391113,
      "objective/train/original_loss": 4.343560218811035,
      "objective/train/theoretical_loss": 3.4009397296623893,
      "objective/train/tokens_used": 579153376,
      "objective/train/unlikelihood_loss": 0.0524880550801754,
      "theoretical_loss": 3.4009397296623893,
      "tokens_seen": 2220294144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033284848484848483,
      "loss": 1.2411,
      "theoretical_loss": 3.4009397296623893,
      "tokens_seen": 2220294144
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033282828282828284,
      "loss": 1.333,
      "theoretical_loss": 3.400931530886722,
      "tokens_seen": 2220359680
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033280808080808085,
      "loss": 1.5778,
      "theoretical_loss": 3.4009233324208026,
      "tokens_seen": 2220425216
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003327878787878788,
      "loss": 1.3843,
      "theoretical_loss": 3.4009151342646087,
      "tokens_seen": 2220490752
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033276767676767676,
      "loss": 1.3702,
      "theoretical_loss": 3.400906936418121,
      "tokens_seen": 2220556288
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003327474747474747,
      "loss": 1.2843,
      "theoretical_loss": 3.400898738881317,
      "tokens_seen": 2220621824
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033272727272727273,
      "loss": 1.3721,
      "theoretical_loss": 3.400890541654178,
      "tokens_seen": 2220687360
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033270707070707074,
      "loss": 1.5032,
      "theoretical_loss": 3.400882344736681,
      "tokens_seen": 2220752896
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003326868686868687,
      "loss": 1.3053,
      "theoretical_loss": 3.400874148128807,
      "tokens_seen": 2220818432
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033266666666666666,
      "loss": 1.6282,
      "theoretical_loss": 3.4008659518305335,
      "tokens_seen": 2220883968
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033264646464646467,
      "loss": 1.3721,
      "theoretical_loss": 3.400857755841841,
      "tokens_seen": 2220949504
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003326262626262626,
      "loss": 1.5673,
      "theoretical_loss": 3.4008495601627082,
      "tokens_seen": 2221015040
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033260606060606064,
      "loss": 1.3497,
      "theoretical_loss": 3.4008413647931137,
      "tokens_seen": 2221080576
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003325858585858586,
      "loss": 1.5404,
      "theoretical_loss": 3.400833169733038,
      "tokens_seen": 2221146112
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033256565656565655,
      "loss": 1.5884,
      "theoretical_loss": 3.400824974982459,
      "tokens_seen": 2221211648
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033254545454545456,
      "loss": 1.5904,
      "theoretical_loss": 3.4008167805413563,
      "tokens_seen": 2221277184
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003325252525252525,
      "loss": 1.6638,
      "theoretical_loss": 3.4008085864097093,
      "tokens_seen": 2221342720
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033250505050505053,
      "loss": 1.4543,
      "theoretical_loss": 3.400800392587497,
      "tokens_seen": 2221408256
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003324848484848485,
      "loss": 1.3463,
      "theoretical_loss": 3.4007921990746985,
      "tokens_seen": 2221473792
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033246464646464644,
      "loss": 1.1855,
      "theoretical_loss": 3.4007840058712935,
      "tokens_seen": 2221539328
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033244444444444445,
      "loss": 1.4074,
      "theoretical_loss": 3.4007758129772605,
      "tokens_seen": 2221604864
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003324242424242424,
      "loss": 1.417,
      "theoretical_loss": 3.400767620392579,
      "tokens_seen": 2221670400
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003324040404040404,
      "loss": 1.2211,
      "theoretical_loss": 3.400759428117228,
      "tokens_seen": 2221735936
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323838383838384,
      "loss": 1.2802,
      "theoretical_loss": 3.4007512361511876,
      "tokens_seen": 2221801472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323636363636364,
      "loss": 1.4802,
      "theoretical_loss": 3.400743044494436,
      "tokens_seen": 2221867008
    },
    {
      "epoch": 0.34,
      "objective/train/aligned_token_count": 0.3404814302921295,
      "objective/train/avg_token_score": 0.012891263701021671,
      "objective/train/docs_used": 1258716,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0845504999160767,
      "objective/train/original_loss": 4.005323886871338,
      "objective/train/theoretical_loss": 3.400734853146952,
      "objective/train/tokens_used": 580791776,
      "objective/train/unlikelihood_loss": 0.06375692039728165,
      "theoretical_loss": 3.400734853146952,
      "tokens_seen": 2221932544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033234343434343435,
      "loss": 1.3922,
      "theoretical_loss": 3.400734853146952,
      "tokens_seen": 2221932544
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323232323232323,
      "loss": 1.5284,
      "theoretical_loss": 3.400726662108716,
      "tokens_seen": 2221998080
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003323030303030303,
      "loss": 1.6639,
      "theoretical_loss": 3.4007184713797067,
      "tokens_seen": 2222063616
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033228282828282827,
      "loss": 1.3728,
      "theoretical_loss": 3.4007102809599035,
      "tokens_seen": 2222129152
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003322626262626263,
      "loss": 1.4825,
      "theoretical_loss": 3.4007020908492853,
      "tokens_seen": 2222194688
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033224242424242424,
      "loss": 1.4907,
      "theoretical_loss": 3.4006939010478314,
      "tokens_seen": 2222260224
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003322222222222222,
      "loss": 1.4475,
      "theoretical_loss": 3.400685711555521,
      "tokens_seen": 2222325760
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003322020202020202,
      "loss": 1.4844,
      "theoretical_loss": 3.4006775223723333,
      "tokens_seen": 2222391296
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321818181818182,
      "loss": 1.3429,
      "theoretical_loss": 3.4006693334982474,
      "tokens_seen": 2222456832
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321616161616162,
      "loss": 1.3801,
      "theoretical_loss": 3.4006611449332427,
      "tokens_seen": 2222522368
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033214141414141413,
      "loss": 1.3121,
      "theoretical_loss": 3.4006529566772987,
      "tokens_seen": 2222587904
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321212121212121,
      "loss": 1.3211,
      "theoretical_loss": 3.4006447687303942,
      "tokens_seen": 2222653440
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003321010101010101,
      "loss": 1.454,
      "theoretical_loss": 3.4006365810925088,
      "tokens_seen": 2222718976
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003320808080808081,
      "loss": 1.4087,
      "theoretical_loss": 3.400628393763621,
      "tokens_seen": 2222784512
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033206060606060607,
      "loss": 1.4161,
      "theoretical_loss": 3.4006202067437106,
      "tokens_seen": 2222850048
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000332040404040404,
      "loss": 1.2653,
      "theoretical_loss": 3.400612020032757,
      "tokens_seen": 2222915584
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000332020202020202,
      "loss": 1.3887,
      "theoretical_loss": 3.400603833630739,
      "tokens_seen": 2222981120
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033200000000000005,
      "loss": 1.5272,
      "theoretical_loss": 3.400595647537636,
      "tokens_seen": 2223046656
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000331979797979798,
      "loss": 1.589,
      "theoretical_loss": 3.4005874617534273,
      "tokens_seen": 2223112192
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033195959595959596,
      "loss": 1.7044,
      "theoretical_loss": 3.400579276278092,
      "tokens_seen": 2223177728
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003319393939393939,
      "loss": 1.4815,
      "theoretical_loss": 3.400571091111609,
      "tokens_seen": 2223243264
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003319191919191919,
      "loss": 1.5482,
      "theoretical_loss": 3.400562906253959,
      "tokens_seen": 2223308800
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033189898989898994,
      "loss": 1.4904,
      "theoretical_loss": 3.4005547217051193,
      "tokens_seen": 2223374336
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318787878787879,
      "loss": 1.4421,
      "theoretical_loss": 3.4005465374650705,
      "tokens_seen": 2223439872
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033185858585858585,
      "loss": 1.6641,
      "theoretical_loss": 3.400538353533791,
      "tokens_seen": 2223505408
    },
    {
      "epoch": 0.34,
      "objective/train/aligned_token_count": 0.4250366687774658,
      "objective/train/avg_token_score": 0.009826776571571827,
      "objective/train/docs_used": 1259970,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5176795721054077,
      "objective/train/original_loss": 4.1156463623046875,
      "objective/train/theoretical_loss": 3.400530169911261,
      "objective/train/tokens_used": 582430176,
      "objective/train/unlikelihood_loss": 0.06612487137317657,
      "theoretical_loss": 3.400530169911261,
      "tokens_seen": 2223570944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318383838383838,
      "loss": 1.6582,
      "theoretical_loss": 3.400530169911261,
      "tokens_seen": 2223570944
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003318181818181819,
      "loss": 1.5203,
      "theoretical_loss": 3.4005219865974587,
      "tokens_seen": 2223636480
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033179797979797983,
      "loss": 1.682,
      "theoretical_loss": 3.4005138035923643,
      "tokens_seen": 2223702016
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003317777777777778,
      "loss": 1.4011,
      "theoretical_loss": 3.4005056208959563,
      "tokens_seen": 2223767552
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033175757575757575,
      "loss": 1.3513,
      "theoretical_loss": 3.4004974385082143,
      "tokens_seen": 2223833088
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003317373737373737,
      "loss": 1.581,
      "theoretical_loss": 3.400489256429118,
      "tokens_seen": 2223898624
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033171717171717177,
      "loss": 1.52,
      "theoretical_loss": 3.4004810746586456,
      "tokens_seen": 2223964160
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003316969696969697,
      "loss": 1.5779,
      "theoretical_loss": 3.400472893196777,
      "tokens_seen": 2224029696
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003316767676767677,
      "loss": 1.2395,
      "theoretical_loss": 3.4004647120434917,
      "tokens_seen": 2224095232
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033165656565656564,
      "loss": 1.3382,
      "theoretical_loss": 3.4004565311987687,
      "tokens_seen": 2224160768
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033163636363636365,
      "loss": 1.3184,
      "theoretical_loss": 3.400448350662587,
      "tokens_seen": 2224226304
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033161616161616166,
      "loss": 1.3241,
      "theoretical_loss": 3.4004401704349263,
      "tokens_seen": 2224291840
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003315959595959596,
      "loss": 1.0203,
      "theoretical_loss": 3.400431990515766,
      "tokens_seen": 2224357376
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003315757575757576,
      "loss": 1.5063,
      "theoretical_loss": 3.400423810905085,
      "tokens_seen": 2224422912
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033155555555555553,
      "loss": 1.4688,
      "theoretical_loss": 3.400415631602862,
      "tokens_seen": 2224488448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033153535353535354,
      "loss": 1.2984,
      "theoretical_loss": 3.4004074526090777,
      "tokens_seen": 2224553984
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033151515151515155,
      "loss": 1.3075,
      "theoretical_loss": 3.4003992739237106,
      "tokens_seen": 2224619520
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003314949494949495,
      "loss": 1.5274,
      "theoretical_loss": 3.4003910955467394,
      "tokens_seen": 2224685056
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033147474747474747,
      "loss": 1.3229,
      "theoretical_loss": 3.4003829174781446,
      "tokens_seen": 2224750592
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003314545454545455,
      "loss": 1.3988,
      "theoretical_loss": 3.400374739717905,
      "tokens_seen": 2224816128
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033143434343434344,
      "loss": 1.5734,
      "theoretical_loss": 3.400366562265999,
      "tokens_seen": 2224881664
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033141414141414145,
      "loss": 1.3292,
      "theoretical_loss": 3.4003583851224075,
      "tokens_seen": 2224947200
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003313939393939394,
      "loss": 1.2925,
      "theoretical_loss": 3.4003502082871084,
      "tokens_seen": 2225012736
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033137373737373736,
      "loss": 1.2722,
      "theoretical_loss": 3.400342031760082,
      "tokens_seen": 2225078272
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033135353535353537,
      "loss": 1.4306,
      "theoretical_loss": 3.400333855541307,
      "tokens_seen": 2225143808
    },
    {
      "epoch": 0.34,
      "objective/train/aligned_token_count": 0.46169355511665344,
      "objective/train/avg_token_score": 0.004667219705879688,
      "objective/train/docs_used": 1260594,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4731347560882568,
      "objective/train/original_loss": 4.219857215881348,
      "objective/train/theoretical_loss": 3.400325679630763,
      "objective/train/tokens_used": 584068576,
      "objective/train/unlikelihood_loss": 0.04526223614811897,
      "theoretical_loss": 3.400325679630763,
      "tokens_seen": 2225209344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033133333333333333,
      "loss": 1.446,
      "theoretical_loss": 3.400325679630763,
      "tokens_seen": 2225209344
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033131313131313134,
      "loss": 1.4518,
      "theoretical_loss": 3.400317504028429,
      "tokens_seen": 2225274880
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003312929292929293,
      "loss": 1.5685,
      "theoretical_loss": 3.4003093287342843,
      "tokens_seen": 2225340416
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003312727272727273,
      "loss": 1.0434,
      "theoretical_loss": 3.4003011537483085,
      "tokens_seen": 2225405952
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033125252525252526,
      "loss": 1.1862,
      "theoretical_loss": 3.400292979070481,
      "tokens_seen": 2225471488
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003312323232323232,
      "loss": 1.2856,
      "theoretical_loss": 3.4002848047007808,
      "tokens_seen": 2225537024
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033121212121212123,
      "loss": 1.3302,
      "theoretical_loss": 3.4002766306391874,
      "tokens_seen": 2225602560
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003311919191919192,
      "loss": 1.1986,
      "theoretical_loss": 3.40026845688568,
      "tokens_seen": 2225668096
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003311717171717172,
      "loss": 1.5208,
      "theoretical_loss": 3.400260283440238,
      "tokens_seen": 2225733632
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033115151515151516,
      "loss": 1.3456,
      "theoretical_loss": 3.4002521103028407,
      "tokens_seen": 2225799168
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003311313131313131,
      "loss": 1.4009,
      "theoretical_loss": 3.400243937473467,
      "tokens_seen": 2225864704
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003311111111111111,
      "loss": 1.4486,
      "theoretical_loss": 3.4002357649520967,
      "tokens_seen": 2225930240
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033109090909090914,
      "loss": 1.2706,
      "theoretical_loss": 3.4002275927387093,
      "tokens_seen": 2225995776
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003310707070707071,
      "loss": 1.4595,
      "theoretical_loss": 3.4002194208332837,
      "tokens_seen": 2226061312
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033105050505050505,
      "loss": 1.5573,
      "theoretical_loss": 3.4002112492357996,
      "tokens_seen": 2226126848
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000331030303030303,
      "loss": 1.4708,
      "theoretical_loss": 3.400203077946236,
      "tokens_seen": 2226192384
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000331010101010101,
      "loss": 1.2841,
      "theoretical_loss": 3.4001949069645723,
      "tokens_seen": 2226257920
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033098989898989903,
      "loss": 1.5849,
      "theoretical_loss": 3.400186736290788,
      "tokens_seen": 2226323456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.000330969696969697,
      "loss": 1.5127,
      "theoretical_loss": 3.400178565924862,
      "tokens_seen": 2226388992
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033094949494949494,
      "loss": 1.3832,
      "theoretical_loss": 3.400170395866774,
      "tokens_seen": 2226454528
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003309292929292929,
      "loss": 1.4887,
      "theoretical_loss": 3.400162226116504,
      "tokens_seen": 2226520064
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003309292929292929,
      "loss": 1.218,
      "theoretical_loss": 3.40015405667403,
      "tokens_seen": 2226585600
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033090909090909096,
      "loss": 1.3501,
      "theoretical_loss": 3.4001458875393324,
      "tokens_seen": 2226651136
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003308888888888889,
      "loss": 1.4561,
      "theoretical_loss": 3.40013771871239,
      "tokens_seen": 2226716672
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0003308686868686869,
      "loss": 1.3576,
      "theoretical_loss": 3.4001295501931823,
      "tokens_seen": 2226782208
    },
    {
      "epoch": 0.34,
      "objective/train/aligned_token_count": 0.34139785170555115,
      "objective/train/avg_token_score": 0.02149634063243866,
      "objective/train/docs_used": 1261463,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1288074254989624,
      "objective/train/original_loss": 4.781523704528809,
      "objective/train/theoretical_loss": 3.4001213819816885,
      "objective/train/tokens_used": 585706976,
      "objective/train/unlikelihood_loss": 0.06013362854719162,
      "theoretical_loss": 3.4001213819816885,
      "tokens_seen": 2226847744
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00033084848484848483,
      "loss": 1.2436,
      "theoretical_loss": 3.4001213819816885,
      "tokens_seen": 2226847744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003308282828282828,
      "loss": 1.2607,
      "theoretical_loss": 3.400113214077888,
      "tokens_seen": 2226913280
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033080808080808086,
      "loss": 1.412,
      "theoretical_loss": 3.40010504648176,
      "tokens_seen": 2226978816
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003307878787878788,
      "loss": 1.3093,
      "theoretical_loss": 3.400096879193285,
      "tokens_seen": 2227044352
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033076767676767677,
      "loss": 1.4431,
      "theoretical_loss": 3.400088712212441,
      "tokens_seen": 2227109888
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033074747474747473,
      "loss": 1.2175,
      "theoretical_loss": 3.4000805455392076,
      "tokens_seen": 2227175424
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003307272727272727,
      "loss": 1.2903,
      "theoretical_loss": 3.4000723791735643,
      "tokens_seen": 2227240960
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033070707070707075,
      "loss": 1.6026,
      "theoretical_loss": 3.400064213115491,
      "tokens_seen": 2227306496
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003306868686868687,
      "loss": 1.3501,
      "theoretical_loss": 3.4000560473649664,
      "tokens_seen": 2227372032
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033066666666666666,
      "loss": 1.2625,
      "theoretical_loss": 3.40004788192197,
      "tokens_seen": 2227437568
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003306464646464646,
      "loss": 1.4832,
      "theoretical_loss": 3.4000397167864813,
      "tokens_seen": 2227503104
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033062626262626263,
      "loss": 1.3911,
      "theoretical_loss": 3.40003155195848,
      "tokens_seen": 2227568640
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033060606060606064,
      "loss": 1.4554,
      "theoretical_loss": 3.400023387437945,
      "tokens_seen": 2227634176
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003305858585858586,
      "loss": 1.3723,
      "theoretical_loss": 3.4000152232248553,
      "tokens_seen": 2227699712
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033056565656565656,
      "loss": 1.5842,
      "theoretical_loss": 3.400007059319191,
      "tokens_seen": 2227765248
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003305454545454545,
      "loss": 1.4194,
      "theoretical_loss": 3.3999988957209313,
      "tokens_seen": 2227830784
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003305252525252525,
      "loss": 1.4042,
      "theoretical_loss": 3.3999907324300556,
      "tokens_seen": 2227896320
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033050505050505053,
      "loss": 1.4382,
      "theoretical_loss": 3.399982569446543,
      "tokens_seen": 2227961856
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304848484848485,
      "loss": 1.4225,
      "theoretical_loss": 3.399974406770373,
      "tokens_seen": 2228027392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033046464646464645,
      "loss": 1.2719,
      "theoretical_loss": 3.3999662444015257,
      "tokens_seen": 2228092928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033044444444444446,
      "loss": 1.6624,
      "theoretical_loss": 3.399958082339979,
      "tokens_seen": 2228158464
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003304242424242424,
      "loss": 1.5653,
      "theoretical_loss": 3.3999499205857138,
      "tokens_seen": 2228224000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033040404040404043,
      "loss": 1.4108,
      "theoretical_loss": 3.399941759138709,
      "tokens_seen": 2228289536
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303838383838384,
      "loss": 1.4695,
      "theoretical_loss": 3.399933597998943,
      "tokens_seen": 2228355072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033036363636363634,
      "loss": 1.4881,
      "theoretical_loss": 3.399925437166397,
      "tokens_seen": 2228420608
    },
    {
      "epoch": 0.35,
      "objective/train/aligned_token_count": 0.42467010021209717,
      "objective/train/avg_token_score": 0.007075726520270109,
      "objective/train/docs_used": 1262743,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3123854398727417,
      "objective/train/original_loss": 3.991392135620117,
      "objective/train/theoretical_loss": 3.399917276641049,
      "objective/train/tokens_used": 587345376,
      "objective/train/unlikelihood_loss": 0.06691636145114899,
      "theoretical_loss": 3.399917276641049,
      "tokens_seen": 2228486144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033034343434343435,
      "loss": 1.5216,
      "theoretical_loss": 3.399917276641049,
      "tokens_seen": 2228486144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303232323232323,
      "loss": 1.5061,
      "theoretical_loss": 3.399909116422879,
      "tokens_seen": 2228551680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003303030303030303,
      "loss": 1.2332,
      "theoretical_loss": 3.399900956511866,
      "tokens_seen": 2228617216
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302828282828283,
      "loss": 1.4849,
      "theoretical_loss": 3.3998927969079897,
      "tokens_seen": 2228682752
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302626262626263,
      "loss": 1.5272,
      "theoretical_loss": 3.3998846376112297,
      "tokens_seen": 2228748288
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033024242424242424,
      "loss": 1.2654,
      "theoretical_loss": 3.3998764786215654,
      "tokens_seen": 2228813824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302222222222222,
      "loss": 1.4573,
      "theoretical_loss": 3.3998683199389754,
      "tokens_seen": 2228879360
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003302020202020202,
      "loss": 1.2836,
      "theoretical_loss": 3.3998601615634403,
      "tokens_seen": 2228944896
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033018181818181817,
      "loss": 1.3309,
      "theoretical_loss": 3.3998520034949387,
      "tokens_seen": 2229010432
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003301616161616162,
      "loss": 1.5113,
      "theoretical_loss": 3.39984384573345,
      "tokens_seen": 2229075968
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033014141414141414,
      "loss": 1.5353,
      "theoretical_loss": 3.3998356882789538,
      "tokens_seen": 2229141504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033012121212121215,
      "loss": 1.4845,
      "theoretical_loss": 3.39982753113143,
      "tokens_seen": 2229207040
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003301010101010101,
      "loss": 1.2403,
      "theoretical_loss": 3.3998193742908573,
      "tokens_seen": 2229272576
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003300808080808081,
      "loss": 1.5131,
      "theoretical_loss": 3.3998112177572155,
      "tokens_seen": 2229338112
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003300606060606061,
      "loss": 1.1798,
      "theoretical_loss": 3.399803061530484,
      "tokens_seen": 2229403648
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033004040404040403,
      "loss": 1.2516,
      "theoretical_loss": 3.3997949056106425,
      "tokens_seen": 2229469184
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033002020202020204,
      "loss": 1.2812,
      "theoretical_loss": 3.39978674999767,
      "tokens_seen": 2229534720
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00033,
      "loss": 1.4255,
      "theoretical_loss": 3.399778594691546,
      "tokens_seen": 2229600256
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329979797979798,
      "loss": 1.3537,
      "theoretical_loss": 3.39977043969225,
      "tokens_seen": 2229665792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032995959595959597,
      "loss": 1.3319,
      "theoretical_loss": 3.3997622849997615,
      "tokens_seen": 2229731328
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003299393939393939,
      "loss": 1.5274,
      "theoretical_loss": 3.3997541306140597,
      "tokens_seen": 2229796864
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032991919191919193,
      "loss": 1.2534,
      "theoretical_loss": 3.399745976535124,
      "tokens_seen": 2229862400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032989898989898995,
      "loss": 1.3359,
      "theoretical_loss": 3.3997378227629347,
      "tokens_seen": 2229927936
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003298787878787879,
      "loss": 1.2958,
      "theoretical_loss": 3.39972966929747,
      "tokens_seen": 2229993472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032985858585858586,
      "loss": 1.5058,
      "theoretical_loss": 3.3997215161387104,
      "tokens_seen": 2230059008
    },
    {
      "epoch": 0.35,
      "objective/train/aligned_token_count": 0.47983869910240173,
      "objective/train/avg_token_score": 0.014847558923065662,
      "objective/train/docs_used": 1263995,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5770068168640137,
      "objective/train/original_loss": 4.253453731536865,
      "objective/train/theoretical_loss": 3.3997133632866348,
      "objective/train/tokens_used": 588983776,
      "objective/train/unlikelihood_loss": 0.047522395849227905,
      "theoretical_loss": 3.3997133632866348,
      "tokens_seen": 2230124544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003298383838383838,
      "loss": 1.4607,
      "theoretical_loss": 3.3997133632866348,
      "tokens_seen": 2230124544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003298181818181818,
      "loss": 1.4597,
      "theoretical_loss": 3.3997052107412227,
      "tokens_seen": 2230190080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032979797979797984,
      "loss": 1.391,
      "theoretical_loss": 3.3996970585024537,
      "tokens_seen": 2230255616
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297777777777778,
      "loss": 1.4077,
      "theoretical_loss": 3.399688906570307,
      "tokens_seen": 2230321152
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032975757575757575,
      "loss": 1.2367,
      "theoretical_loss": 3.3996807549447627,
      "tokens_seen": 2230386688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297373737373737,
      "loss": 1.2571,
      "theoretical_loss": 3.3996726036257994,
      "tokens_seen": 2230452224
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003297171717171718,
      "loss": 1.4495,
      "theoretical_loss": 3.399664452613397,
      "tokens_seen": 2230517760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032969696969696973,
      "loss": 1.3635,
      "theoretical_loss": 3.3996563019075348,
      "tokens_seen": 2230583296
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003296767676767677,
      "loss": 1.6259,
      "theoretical_loss": 3.3996481515081927,
      "tokens_seen": 2230648832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032965656565656564,
      "loss": 1.4773,
      "theoretical_loss": 3.3996400014153494,
      "tokens_seen": 2230714368
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003296363636363636,
      "loss": 1.6106,
      "theoretical_loss": 3.399631851628985,
      "tokens_seen": 2230779904
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032961616161616167,
      "loss": 1.4901,
      "theoretical_loss": 3.399623702149079,
      "tokens_seen": 2230845440
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003295959595959596,
      "loss": 1.4583,
      "theoretical_loss": 3.3996155529756105,
      "tokens_seen": 2230910976
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003295757575757576,
      "loss": 1.5332,
      "theoretical_loss": 3.399607404108559,
      "tokens_seen": 2230976512
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032955555555555554,
      "loss": 1.3963,
      "theoretical_loss": 3.3995992555479044,
      "tokens_seen": 2231042048
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032953535353535355,
      "loss": 1.546,
      "theoretical_loss": 3.3995911072936256,
      "tokens_seen": 2231107584
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032951515151515156,
      "loss": 1.3306,
      "theoretical_loss": 3.399582959345703,
      "tokens_seen": 2231173120
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003294949494949495,
      "loss": 1.3427,
      "theoretical_loss": 3.399574811704115,
      "tokens_seen": 2231238656
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032947474747474747,
      "loss": 1.6377,
      "theoretical_loss": 3.3995666643688414,
      "tokens_seen": 2231304192
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032945454545454543,
      "loss": 1.5174,
      "theoretical_loss": 3.399558517339862,
      "tokens_seen": 2231369728
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032943434343434344,
      "loss": 1.5219,
      "theoretical_loss": 3.399550370617156,
      "tokens_seen": 2231435264
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032941414141414145,
      "loss": 1.2786,
      "theoretical_loss": 3.3995422242007036,
      "tokens_seen": 2231500800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003293939393939394,
      "loss": 1.309,
      "theoretical_loss": 3.399534078090483,
      "tokens_seen": 2231566336
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032937373737373737,
      "loss": 1.5592,
      "theoretical_loss": 3.399525932286475,
      "tokens_seen": 2231631872
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003293535353535354,
      "loss": 1.552,
      "theoretical_loss": 3.3995177867886577,
      "tokens_seen": 2231697408
    },
    {
      "epoch": 0.35,
      "objective/train/aligned_token_count": 0.48106059432029724,
      "objective/train/avg_token_score": 0.021318404003977776,
      "objective/train/docs_used": 1264337,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.607075572013855,
      "objective/train/original_loss": 4.34976863861084,
      "objective/train/theoretical_loss": 3.399509641597012,
      "objective/train/tokens_used": 590622176,
      "objective/train/unlikelihood_loss": 0.05099575221538544,
      "theoretical_loss": 3.399509641597012,
      "tokens_seen": 2231762944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032933333333333333,
      "loss": 1.5886,
      "theoretical_loss": 3.399509641597012,
      "tokens_seen": 2231762944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032931313131313134,
      "loss": 1.3315,
      "theoretical_loss": 3.399501496711517,
      "tokens_seen": 2231828480
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003292929292929293,
      "loss": 1.4497,
      "theoretical_loss": 3.3994933521321515,
      "tokens_seen": 2231894016
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032927272727272726,
      "loss": 1.507,
      "theoretical_loss": 3.3994852078588957,
      "tokens_seen": 2231959552
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032925252525252527,
      "loss": 1.5324,
      "theoretical_loss": 3.3994770638917293,
      "tokens_seen": 2232025088
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003292323232323232,
      "loss": 1.5057,
      "theoretical_loss": 3.399468920230631,
      "tokens_seen": 2232090624
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032921212121212124,
      "loss": 1.5817,
      "theoretical_loss": 3.3994607768755807,
      "tokens_seen": 2232156160
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003291919191919192,
      "loss": 1.2276,
      "theoretical_loss": 3.399452633826558,
      "tokens_seen": 2232221696
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032917171717171715,
      "loss": 1.4426,
      "theoretical_loss": 3.3994444910835426,
      "tokens_seen": 2232287232
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032915151515151516,
      "loss": 1.6156,
      "theoretical_loss": 3.3994363486465136,
      "tokens_seen": 2232352768
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003291313131313131,
      "loss": 1.4836,
      "theoretical_loss": 3.399428206515451,
      "tokens_seen": 2232418304
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032911111111111113,
      "loss": 1.48,
      "theoretical_loss": 3.3994200646903336,
      "tokens_seen": 2232483840
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003290909090909091,
      "loss": 1.0801,
      "theoretical_loss": 3.3994119231711415,
      "tokens_seen": 2232549376
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003290707070707071,
      "loss": 1.4356,
      "theoretical_loss": 3.3994037819578544,
      "tokens_seen": 2232614912
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032905050505050505,
      "loss": 1.3566,
      "theoretical_loss": 3.399395641050451,
      "tokens_seen": 2232680448
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329030303030303,
      "loss": 1.3146,
      "theoretical_loss": 3.3993875004489116,
      "tokens_seen": 2232745984
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000329010101010101,
      "loss": 1.5466,
      "theoretical_loss": 3.3993793601532154,
      "tokens_seen": 2232811520
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000328989898989899,
      "loss": 1.4732,
      "theoretical_loss": 3.399371220163342,
      "tokens_seen": 2232877056
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000328969696969697,
      "loss": 1.383,
      "theoretical_loss": 3.399363080479271,
      "tokens_seen": 2232942592
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032894949494949495,
      "loss": 1.3416,
      "theoretical_loss": 3.399354941100982,
      "tokens_seen": 2233008128
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003289292929292929,
      "loss": 1.5784,
      "theoretical_loss": 3.399346802028454,
      "tokens_seen": 2233073664
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003289090909090909,
      "loss": 1.4897,
      "theoretical_loss": 3.3993386632616676,
      "tokens_seen": 2233139200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288888888888889,
      "loss": 1.4759,
      "theoretical_loss": 3.399330524800601,
      "tokens_seen": 2233204736
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288686868686869,
      "loss": 1.5653,
      "theoretical_loss": 3.3993223866452347,
      "tokens_seen": 2233270272
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032884848484848484,
      "loss": 1.4404,
      "theoretical_loss": 3.399314248795548,
      "tokens_seen": 2233335808
    },
    {
      "epoch": 0.35,
      "objective/train/aligned_token_count": 0.4806940257549286,
      "objective/train/avg_token_score": 0.02901451662182808,
      "objective/train/docs_used": 1265500,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.530691146850586,
      "objective/train/original_loss": 4.268867015838623,
      "objective/train/theoretical_loss": 3.3993061112515206,
      "objective/train/tokens_used": 592260576,
      "objective/train/unlikelihood_loss": 0.06376490741968155,
      "theoretical_loss": 3.3993061112515206,
      "tokens_seen": 2233401344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288282828282828,
      "loss": 1.2595,
      "theoretical_loss": 3.3993061112515206,
      "tokens_seen": 2233401344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003288080808080808,
      "loss": 1.2418,
      "theoretical_loss": 3.3992979740131317,
      "tokens_seen": 2233466880
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287878787878788,
      "loss": 1.4604,
      "theoretical_loss": 3.399289837080361,
      "tokens_seen": 2233532416
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287676767676768,
      "loss": 1.4523,
      "theoretical_loss": 3.399281700453188,
      "tokens_seen": 2233597952
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032874747474747473,
      "loss": 1.3612,
      "theoretical_loss": 3.399273564131593,
      "tokens_seen": 2233663488
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003287272727272727,
      "loss": 1.3548,
      "theoretical_loss": 3.3992654281155543,
      "tokens_seen": 2233729024
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032870707070707075,
      "loss": 1.3796,
      "theoretical_loss": 3.399257292405052,
      "tokens_seen": 2233794560
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286868686868687,
      "loss": 1.4942,
      "theoretical_loss": 3.399249157000066,
      "tokens_seen": 2233860096
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032866666666666667,
      "loss": 1.5047,
      "theoretical_loss": 3.3992410219005755,
      "tokens_seen": 2233925632
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286464646464646,
      "loss": 1.4878,
      "theoretical_loss": 3.3992328871065602,
      "tokens_seen": 2233991168
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003286262626262626,
      "loss": 1.4635,
      "theoretical_loss": 3.3992247526179997,
      "tokens_seen": 2234056704
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032860606060606065,
      "loss": 1.3078,
      "theoretical_loss": 3.3992166184348735,
      "tokens_seen": 2234122240
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003285858585858586,
      "loss": 1.5763,
      "theoretical_loss": 3.3992084845571613,
      "tokens_seen": 2234187776
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032856565656565656,
      "loss": 1.5913,
      "theoretical_loss": 3.399200350984842,
      "tokens_seen": 2234253312
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003285454545454545,
      "loss": 1.2942,
      "theoretical_loss": 3.3991922177178964,
      "tokens_seen": 2234318848
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032852525252525253,
      "loss": 1.2939,
      "theoretical_loss": 3.399184084756303,
      "tokens_seen": 2234384384
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032850505050505054,
      "loss": 1.5869,
      "theoretical_loss": 3.3991759521000415,
      "tokens_seen": 2234449920
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284848484848485,
      "loss": 1.4194,
      "theoretical_loss": 3.399167819749092,
      "tokens_seen": 2234515456
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032846464646464645,
      "loss": 1.4075,
      "theoretical_loss": 3.3991596877034342,
      "tokens_seen": 2234580992
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284444444444444,
      "loss": 1.2223,
      "theoretical_loss": 3.399151555963047,
      "tokens_seen": 2234646528
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003284242424242425,
      "loss": 1.5779,
      "theoretical_loss": 3.3991434245279106,
      "tokens_seen": 2234712064
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032840404040404043,
      "loss": 1.1823,
      "theoretical_loss": 3.399135293398004,
      "tokens_seen": 2234777600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003283838383838384,
      "loss": 1.5818,
      "theoretical_loss": 3.3991271625733073,
      "tokens_seen": 2234843136
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032836363636363635,
      "loss": 1.5505,
      "theoretical_loss": 3.3991190320538,
      "tokens_seen": 2234908672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032834343434343436,
      "loss": 1.5451,
      "theoretical_loss": 3.399110901839461,
      "tokens_seen": 2234974208
    },
    {
      "epoch": 0.35,
      "objective/train/aligned_token_count": 0.45081865787506104,
      "objective/train/avg_token_score": 0.015752630308270454,
      "objective/train/docs_used": 1266280,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5672367811203003,
      "objective/train/original_loss": 4.046272277832031,
      "objective/train/theoretical_loss": 3.3991027719302713,
      "objective/train/tokens_used": 593898976,
      "objective/train/unlikelihood_loss": 0.09052581340074539,
      "theoretical_loss": 3.3991027719302713,
      "tokens_seen": 2235039744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032832323232323237,
      "loss": 1.4206,
      "theoretical_loss": 3.3991027719302713,
      "tokens_seen": 2235039744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003283030303030303,
      "loss": 1.4861,
      "theoretical_loss": 3.399094642326209,
      "tokens_seen": 2235105280
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003282828282828283,
      "loss": 1.4738,
      "theoretical_loss": 3.399086513027255,
      "tokens_seen": 2235170816
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032826262626262624,
      "loss": 1.5669,
      "theoretical_loss": 3.399078384033388,
      "tokens_seen": 2235236352
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032824242424242425,
      "loss": 1.4535,
      "theoretical_loss": 3.3990702553445877,
      "tokens_seen": 2235301888
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032822222222222226,
      "loss": 1.3679,
      "theoretical_loss": 3.399062126960834,
      "tokens_seen": 2235367424
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003282020202020202,
      "loss": 1.4059,
      "theoretical_loss": 3.3990539988821067,
      "tokens_seen": 2235432960
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281818181818182,
      "loss": 1.379,
      "theoretical_loss": 3.399045871108385,
      "tokens_seen": 2235498496
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281616161616162,
      "loss": 1.4641,
      "theoretical_loss": 3.399037743639649,
      "tokens_seen": 2235564032
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032814141414141414,
      "loss": 1.6168,
      "theoretical_loss": 3.399029616475877,
      "tokens_seen": 2235629568
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032812121212121215,
      "loss": 1.2751,
      "theoretical_loss": 3.3990214896170503,
      "tokens_seen": 2235695104
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003281010101010101,
      "loss": 1.1475,
      "theoretical_loss": 3.3990133630631476,
      "tokens_seen": 2235760640
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032808080808080807,
      "loss": 1.4256,
      "theoretical_loss": 3.399005236814149,
      "tokens_seen": 2235826176
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003280606060606061,
      "loss": 1.4543,
      "theoretical_loss": 3.3989971108700336,
      "tokens_seen": 2235891712
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032804040404040404,
      "loss": 1.4303,
      "theoretical_loss": 3.3989889852307815,
      "tokens_seen": 2235957248
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032802020202020205,
      "loss": 1.6602,
      "theoretical_loss": 3.3989808598963718,
      "tokens_seen": 2236022784
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000328,
      "loss": 1.3516,
      "theoretical_loss": 3.3989727348667844,
      "tokens_seen": 2236088320
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000327979797979798,
      "loss": 1.5232,
      "theoretical_loss": 3.398964610141999,
      "tokens_seen": 2236153856
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032795959595959597,
      "loss": 1.5443,
      "theoretical_loss": 3.3989564857219956,
      "tokens_seen": 2236219392
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032793939393939393,
      "loss": 1.5781,
      "theoretical_loss": 3.398948361606753,
      "tokens_seen": 2236284928
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032791919191919194,
      "loss": 1.6518,
      "theoretical_loss": 3.398940237796251,
      "tokens_seen": 2236350464
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003278989898989899,
      "loss": 1.586,
      "theoretical_loss": 3.3989321142904703,
      "tokens_seen": 2236416000
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003278787878787879,
      "loss": 1.2997,
      "theoretical_loss": 3.398923991089389,
      "tokens_seen": 2236481536
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032785858585858586,
      "loss": 1.3937,
      "theoretical_loss": 3.3989158681929883,
      "tokens_seen": 2236547072
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003278383838383838,
      "loss": 1.646,
      "theoretical_loss": 3.3989077456012464,
      "tokens_seen": 2236612608
    },
    {
      "epoch": 0.35,
      "objective/train/aligned_token_count": 0.30241936445236206,
      "objective/train/avg_token_score": 0.0227317176759243,
      "objective/train/docs_used": 1267649,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0899677276611328,
      "objective/train/original_loss": 5.205410480499268,
      "objective/train/theoretical_loss": 3.3988996233141435,
      "objective/train/tokens_used": 595537376,
      "objective/train/unlikelihood_loss": 0.05280158668756485,
      "theoretical_loss": 3.3988996233141435,
      "tokens_seen": 2236678144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032781818181818183,
      "loss": 1.4046,
      "theoretical_loss": 3.3988996233141435,
      "tokens_seen": 2236678144
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032779797979797984,
      "loss": 1.6461,
      "theoretical_loss": 3.3988915013316596,
      "tokens_seen": 2236743680
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003277777777777778,
      "loss": 1.4129,
      "theoretical_loss": 3.398883379653774,
      "tokens_seen": 2236809216
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032775757575757576,
      "loss": 1.432,
      "theoretical_loss": 3.3988752582804667,
      "tokens_seen": 2236874752
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003277373737373737,
      "loss": 1.502,
      "theoretical_loss": 3.3988671372117167,
      "tokens_seen": 2236940288
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003277171717171717,
      "loss": 1.4962,
      "theoretical_loss": 3.398859016447504,
      "tokens_seen": 2237005824
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032769696969696974,
      "loss": 1.3661,
      "theoretical_loss": 3.3988508959878088,
      "tokens_seen": 2237071360
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003276767676767677,
      "loss": 1.6559,
      "theoretical_loss": 3.39884277583261,
      "tokens_seen": 2237136896
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032765656565656565,
      "loss": 1.62,
      "theoretical_loss": 3.3988346559818874,
      "tokens_seen": 2237202432
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003276363636363636,
      "loss": 1.5972,
      "theoretical_loss": 3.3988265364356205,
      "tokens_seen": 2237267968
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032761616161616167,
      "loss": 1.4235,
      "theoretical_loss": 3.3988184171937896,
      "tokens_seen": 2237333504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032759595959595963,
      "loss": 1.4608,
      "theoretical_loss": 3.3988102982563744,
      "tokens_seen": 2237399040
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003275757575757576,
      "loss": 1.5237,
      "theoretical_loss": 3.3988021796233534,
      "tokens_seen": 2237464576
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032755555555555554,
      "loss": 1.5319,
      "theoretical_loss": 3.3987940612947076,
      "tokens_seen": 2237530112
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003275353535353535,
      "loss": 1.4315,
      "theoretical_loss": 3.398785943270416,
      "tokens_seen": 2237595648
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032751515151515156,
      "loss": 1.4836,
      "theoretical_loss": 3.3987778255504586,
      "tokens_seen": 2237661184
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003274949494949495,
      "loss": 1.467,
      "theoretical_loss": 3.3987697081348145,
      "tokens_seen": 2237726720
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003274747474747475,
      "loss": 1.5088,
      "theoretical_loss": 3.3987615910234643,
      "tokens_seen": 2237792256
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032745454545454544,
      "loss": 1.3256,
      "theoretical_loss": 3.3987534742163867,
      "tokens_seen": 2237857792
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003274343434343434,
      "loss": 1.5592,
      "theoretical_loss": 3.398745357713562,
      "tokens_seen": 2237923328
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032741414141414146,
      "loss": 1.2363,
      "theoretical_loss": 3.39873724151497,
      "tokens_seen": 2237988864
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003273939393939394,
      "loss": 1.4674,
      "theoretical_loss": 3.39872912562059,
      "tokens_seen": 2238054400
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032737373737373737,
      "loss": 1.4063,
      "theoretical_loss": 3.3987210100304015,
      "tokens_seen": 2238119936
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032735353535353533,
      "loss": 1.4315,
      "theoretical_loss": 3.3987128947443845,
      "tokens_seen": 2238185472
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032733333333333334,
      "loss": 1.3067,
      "theoretical_loss": 3.398704779762519,
      "tokens_seen": 2238251008
    },
    {
      "epoch": 0.35,
      "objective/train/aligned_token_count": 0.4437316656112671,
      "objective/train/avg_token_score": 0.029380444437265396,
      "objective/train/docs_used": 1268187,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5166982412338257,
      "objective/train/original_loss": 4.38730001449585,
      "objective/train/theoretical_loss": 3.398696665084784,
      "objective/train/tokens_used": 597175776,
      "objective/train/unlikelihood_loss": 0.0650297999382019,
      "theoretical_loss": 3.398696665084784,
      "tokens_seen": 2238316544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032731313131313135,
      "loss": 1.2785,
      "theoretical_loss": 3.398696665084784,
      "tokens_seen": 2238316544
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003272929292929293,
      "loss": 1.3596,
      "theoretical_loss": 3.39868855071116,
      "tokens_seen": 2238382080
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032727272727272726,
      "loss": 1.6296,
      "theoretical_loss": 3.398680436641626,
      "tokens_seen": 2238447616
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003272525252525252,
      "loss": 1.3991,
      "theoretical_loss": 3.398672322876162,
      "tokens_seen": 2238513152
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032723232323232323,
      "loss": 1.533,
      "theoretical_loss": 3.398664209414748,
      "tokens_seen": 2238578688
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032721212121212124,
      "loss": 1.3945,
      "theoretical_loss": 3.398656096257363,
      "tokens_seen": 2238644224
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271919191919192,
      "loss": 1.5214,
      "theoretical_loss": 3.398647983403987,
      "tokens_seen": 2238709760
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032717171717171716,
      "loss": 1.3918,
      "theoretical_loss": 3.3986398708546,
      "tokens_seen": 2238775296
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032715151515151517,
      "loss": 1.1824,
      "theoretical_loss": 3.3986317586091817,
      "tokens_seen": 2238840832
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003271313131313131,
      "loss": 1.113,
      "theoretical_loss": 3.398623646667712,
      "tokens_seen": 2238906368
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032711111111111114,
      "loss": 1.4291,
      "theoretical_loss": 3.3986155350301694,
      "tokens_seen": 2238971904
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003270909090909091,
      "loss": 1.4128,
      "theoretical_loss": 3.398607423696535,
      "tokens_seen": 2239037440
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032707070707070705,
      "loss": 1.4798,
      "theoretical_loss": 3.3985993126667875,
      "tokens_seen": 2239102976
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032705050505050506,
      "loss": 1.3742,
      "theoretical_loss": 3.3985912019409077,
      "tokens_seen": 2239168512
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000327030303030303,
      "loss": 1.3399,
      "theoretical_loss": 3.3985830915188746,
      "tokens_seen": 2239234048
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032701010101010103,
      "loss": 1.3975,
      "theoretical_loss": 3.398574981400668,
      "tokens_seen": 2239299584
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000326989898989899,
      "loss": 1.3945,
      "theoretical_loss": 3.3985668715862674,
      "tokens_seen": 2239365120
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000326969696969697,
      "loss": 1.3231,
      "theoretical_loss": 3.398558762075653,
      "tokens_seen": 2239430656
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032694949494949495,
      "loss": 1.2998,
      "theoretical_loss": 3.3985506528688045,
      "tokens_seen": 2239496192
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003269292929292929,
      "loss": 1.3864,
      "theoretical_loss": 3.3985425439657013,
      "tokens_seen": 2239561728
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003269090909090909,
      "loss": 1.5142,
      "theoretical_loss": 3.398534435366323,
      "tokens_seen": 2239627264
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003268888888888889,
      "loss": 1.2941,
      "theoretical_loss": 3.3985263270706505,
      "tokens_seen": 2239692800
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003268686868686869,
      "loss": 1.1525,
      "theoretical_loss": 3.398518219078662,
      "tokens_seen": 2239758336
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032684848484848485,
      "loss": 1.3459,
      "theoretical_loss": 3.398510111390338,
      "tokens_seen": 2239823872
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003268282828282828,
      "loss": 1.147,
      "theoretical_loss": 3.3985020040056586,
      "tokens_seen": 2239889408
    },
    {
      "epoch": 0.35,
      "objective/train/aligned_token_count": 0.43835532665252686,
      "objective/train/avg_token_score": 0.02898171916604042,
      "objective/train/docs_used": 1268791,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4798280000686646,
      "objective/train/original_loss": 4.313960552215576,
      "objective/train/theoretical_loss": 3.3984938969246024,
      "objective/train/tokens_used": 598814176,
      "objective/train/unlikelihood_loss": 0.06326164305210114,
      "theoretical_loss": 3.3984938969246024,
      "tokens_seen": 2239954944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003268080808080808,
      "loss": 1.4562,
      "theoretical_loss": 3.3984938969246024,
      "tokens_seen": 2239954944
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267878787878788,
      "loss": 1.6054,
      "theoretical_loss": 3.3984857901471504,
      "tokens_seen": 2240020480
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267676767676768,
      "loss": 1.454,
      "theoretical_loss": 3.3984776836732817,
      "tokens_seen": 2240086016
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032674747474747474,
      "loss": 1.4615,
      "theoretical_loss": 3.398469577502976,
      "tokens_seen": 2240151552
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267272727272727,
      "loss": 1.2361,
      "theoretical_loss": 3.3984614716362134,
      "tokens_seen": 2240217088
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003267070707070707,
      "loss": 1.3255,
      "theoretical_loss": 3.398453366072973,
      "tokens_seen": 2240282624
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003266868686868687,
      "loss": 1.6919,
      "theoretical_loss": 3.3984452608132356,
      "tokens_seen": 2240348160
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003266666666666667,
      "loss": 1.3944,
      "theoretical_loss": 3.39843715585698,
      "tokens_seen": 2240413696
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032664646464646463,
      "loss": 1.3787,
      "theoretical_loss": 3.398429051204187,
      "tokens_seen": 2240479232
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032662626262626264,
      "loss": 1.4012,
      "theoretical_loss": 3.398420946854835,
      "tokens_seen": 2240544768
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032660606060606065,
      "loss": 1.1689,
      "theoretical_loss": 3.398412842808905,
      "tokens_seen": 2240610304
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003265858585858586,
      "loss": 1.4378,
      "theoretical_loss": 3.3984047390663754,
      "tokens_seen": 2240675840
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032656565656565657,
      "loss": 1.4191,
      "theoretical_loss": 3.3983966356272277,
      "tokens_seen": 2240741376
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003265454545454545,
      "loss": 1.3231,
      "theoretical_loss": 3.39838853249144,
      "tokens_seen": 2240806912
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032652525252525253,
      "loss": 1.4912,
      "theoretical_loss": 3.3983804296589932,
      "tokens_seen": 2240872448
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032650505050505055,
      "loss": 1.542,
      "theoretical_loss": 3.398372327129867,
      "tokens_seen": 2240937984
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003264848484848485,
      "loss": 1.4301,
      "theoretical_loss": 3.3983642249040407,
      "tokens_seen": 2241003520
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032646464646464646,
      "loss": 1.409,
      "theoretical_loss": 3.398356122981494,
      "tokens_seen": 2241069056
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003264444444444444,
      "loss": 1.3458,
      "theoretical_loss": 3.398348021362207,
      "tokens_seen": 2241134592
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003264242424242425,
      "loss": 1.5005,
      "theoretical_loss": 3.3983399200461597,
      "tokens_seen": 2241200128
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032640404040404044,
      "loss": 1.4791,
      "theoretical_loss": 3.3983318190333316,
      "tokens_seen": 2241265664
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003263838383838384,
      "loss": 1.1348,
      "theoretical_loss": 3.3983237183237023,
      "tokens_seen": 2241331200
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032636363636363635,
      "loss": 1.5084,
      "theoretical_loss": 3.398315617917252,
      "tokens_seen": 2241396736
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003263434343434343,
      "loss": 1.6135,
      "theoretical_loss": 3.3983075178139597,
      "tokens_seen": 2241462272
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003263232323232324,
      "loss": 1.3254,
      "theoretical_loss": 3.3982994180138064,
      "tokens_seen": 2241527808
    },
    {
      "epoch": 0.35,
      "objective/train/aligned_token_count": 0.4692082107067108,
      "objective/train/avg_token_score": 0.020252209156751633,
      "objective/train/docs_used": 1270246,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5658974647521973,
      "objective/train/original_loss": 4.491634845733643,
      "objective/train/theoretical_loss": 3.3982913185167707,
      "objective/train/tokens_used": 600452576,
      "objective/train/unlikelihood_loss": 0.056142356246709824,
      "theoretical_loss": 3.3982913185167707,
      "tokens_seen": 2241593344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032630303030303033,
      "loss": 1.3362,
      "theoretical_loss": 3.3982913185167707,
      "tokens_seen": 2241593344
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003262828282828283,
      "loss": 1.4996,
      "theoretical_loss": 3.3982832193228334,
      "tokens_seen": 2241658880
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032626262626262624,
      "loss": 1.4219,
      "theoretical_loss": 3.3982751204319737,
      "tokens_seen": 2241724416
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032624242424242426,
      "loss": 1.6197,
      "theoretical_loss": 3.3982670218441715,
      "tokens_seen": 2241789952
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032622222222222227,
      "loss": 1.4509,
      "theoretical_loss": 3.398258923559407,
      "tokens_seen": 2241855488
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003262020202020202,
      "loss": 1.242,
      "theoretical_loss": 3.398250825577659,
      "tokens_seen": 2241921024
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003261818181818182,
      "loss": 1.5361,
      "theoretical_loss": 3.398242727898908,
      "tokens_seen": 2241986560
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032616161616161614,
      "loss": 1.3786,
      "theoretical_loss": 3.3982346305231346,
      "tokens_seen": 2242052096
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032614141414141415,
      "loss": 1.3969,
      "theoretical_loss": 3.398226533450317,
      "tokens_seen": 2242117632
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032612121212121216,
      "loss": 1.4295,
      "theoretical_loss": 3.3982184366804358,
      "tokens_seen": 2242183168
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003261010101010101,
      "loss": 1.2883,
      "theoretical_loss": 3.3982103402134713,
      "tokens_seen": 2242248704
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003260808080808081,
      "loss": 1.3568,
      "theoretical_loss": 3.3982022440494024,
      "tokens_seen": 2242314240
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003260606060606061,
      "loss": 1.3326,
      "theoretical_loss": 3.3981941481882094,
      "tokens_seen": 2242379776
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032604040404040404,
      "loss": 1.2553,
      "theoretical_loss": 3.398186052629872,
      "tokens_seen": 2242445312
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032602020202020205,
      "loss": 1.4345,
      "theoretical_loss": 3.39817795737437,
      "tokens_seen": 2242510848
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000326,
      "loss": 1.4429,
      "theoretical_loss": 3.3981698624216836,
      "tokens_seen": 2242576384
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032597979797979797,
      "loss": 1.2501,
      "theoretical_loss": 3.398161767771792,
      "tokens_seen": 2242641920
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.000325959595959596,
      "loss": 1.2652,
      "theoretical_loss": 3.3981536734246753,
      "tokens_seen": 2242707456
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032593939393939393,
      "loss": 1.3615,
      "theoretical_loss": 3.3981455793803135,
      "tokens_seen": 2242772992
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032591919191919195,
      "loss": 1.0111,
      "theoretical_loss": 3.3981374856386863,
      "tokens_seen": 2242838528
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003258989898989899,
      "loss": 1.4132,
      "theoretical_loss": 3.3981293921997735,
      "tokens_seen": 2242904064
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032587878787878786,
      "loss": 1.6166,
      "theoretical_loss": 3.398121299063555,
      "tokens_seen": 2242969600
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032585858585858587,
      "loss": 1.3921,
      "theoretical_loss": 3.3981132062300103,
      "tokens_seen": 2243035136
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003258383838383838,
      "loss": 1.3607,
      "theoretical_loss": 3.3981051136991196,
      "tokens_seen": 2243100672
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00032581818181818184,
      "loss": 1.2826,
      "theoretical_loss": 3.398097021470863,
      "tokens_seen": 2243166208
    },
    {
      "epoch": 0.35,
      "objective/train/aligned_token_count": 0.4957844614982605,
      "objective/train/avg_token_score": 0.006642804015427828,
      "objective/train/docs_used": 1270875,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5897886753082275,
      "objective/train/original_loss": 4.250694751739502,
      "objective/train/theoretical_loss": 3.39808892954522,
      "objective/train/tokens_used": 602090976,
      "objective/train/unlikelihood_loss": 0.054933879524469376,
      "theoretical_loss": 3.39808892954522,
      "tokens_seen": 2243231744
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0003257979797979798,
      "loss": 1.5069,
      "theoretical_loss": 3.39808892954522,
      "tokens_seen": 2243231744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003257777777777778,
      "loss": 1.36,
      "theoretical_loss": 3.39808083792217,
      "tokens_seen": 2243297280
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032575757575757576,
      "loss": 1.4108,
      "theoretical_loss": 3.3980727466016933,
      "tokens_seen": 2243362816
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003257373737373737,
      "loss": 1.3559,
      "theoretical_loss": 3.39806465558377,
      "tokens_seen": 2243428352
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032571717171717173,
      "loss": 1.3636,
      "theoretical_loss": 3.39805656486838,
      "tokens_seen": 2243493888
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256969696969697,
      "loss": 1.513,
      "theoretical_loss": 3.3980484744555026,
      "tokens_seen": 2243559424
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256767676767677,
      "loss": 1.3986,
      "theoretical_loss": 3.3980403843451175,
      "tokens_seen": 2243624960
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032565656565656566,
      "loss": 1.4283,
      "theoretical_loss": 3.3980322945372055,
      "tokens_seen": 2243690496
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256363636363636,
      "loss": 1.2542,
      "theoretical_loss": 3.3980242050317457,
      "tokens_seen": 2243756032
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003256161616161616,
      "loss": 1.4042,
      "theoretical_loss": 3.3980161158287183,
      "tokens_seen": 2243821568
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032559595959595963,
      "loss": 1.5923,
      "theoretical_loss": 3.3980080269281028,
      "tokens_seen": 2243887104
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003255757575757576,
      "loss": 1.2859,
      "theoretical_loss": 3.397999938329879,
      "tokens_seen": 2243952640
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032555555555555555,
      "loss": 1.4697,
      "theoretical_loss": 3.3979918500340274,
      "tokens_seen": 2244018176
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003255353535353535,
      "loss": 1.3444,
      "theoretical_loss": 3.3979837620405275,
      "tokens_seen": 2244083712
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003255151515151515,
      "loss": 1.4123,
      "theoretical_loss": 3.397975674349359,
      "tokens_seen": 2244149248
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032549494949494953,
      "loss": 1.6241,
      "theoretical_loss": 3.3979675869605024,
      "tokens_seen": 2244214784
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003254747474747475,
      "loss": 1.3032,
      "theoretical_loss": 3.397959499873937,
      "tokens_seen": 2244280320
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032545454545454544,
      "loss": 1.4131,
      "theoretical_loss": 3.3979514130896424,
      "tokens_seen": 2244345856
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003254343434343434,
      "loss": 1.4801,
      "theoretical_loss": 3.3979433266075993,
      "tokens_seen": 2244411392
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032541414141414146,
      "loss": 1.5695,
      "theoretical_loss": 3.3979352404277874,
      "tokens_seen": 2244476928
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003253939393939394,
      "loss": 1.5619,
      "theoretical_loss": 3.3979271545501857,
      "tokens_seen": 2244542464
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003253737373737374,
      "loss": 1.3973,
      "theoretical_loss": 3.397919068974775,
      "tokens_seen": 2244608000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032535353535353533,
      "loss": 1.2732,
      "theoretical_loss": 3.397910983701535,
      "tokens_seen": 2244673536
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003253333333333333,
      "loss": 1.313,
      "theoretical_loss": 3.3979028987304454,
      "tokens_seen": 2244739072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032531313131313136,
      "loss": 1.3015,
      "theoretical_loss": 3.397894814061486,
      "tokens_seen": 2244804608
    },
    {
      "epoch": 0.36,
      "objective/train/aligned_token_count": 0.5045821070671082,
      "objective/train/avg_token_score": 0.014392111450433731,
      "objective/train/docs_used": 1272103,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4862433671951294,
      "objective/train/original_loss": 4.055210113525391,
      "objective/train/theoretical_loss": 3.397886729694637,
      "objective/train/tokens_used": 603729376,
      "objective/train/unlikelihood_loss": 0.05690760910511017,
      "theoretical_loss": 3.397886729694637,
      "tokens_seen": 2244870144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003252929292929293,
      "loss": 1.3554,
      "theoretical_loss": 3.397886729694637,
      "tokens_seen": 2244870144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032527272727272727,
      "loss": 1.2141,
      "theoretical_loss": 3.397878645629878,
      "tokens_seen": 2244935680
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003252525252525252,
      "loss": 1.5109,
      "theoretical_loss": 3.3978705618671894,
      "tokens_seen": 2245001216
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032523232323232324,
      "loss": 1.3885,
      "theoretical_loss": 3.3978624784065508,
      "tokens_seen": 2245066752
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032521212121212125,
      "loss": 1.3561,
      "theoretical_loss": 3.3978543952479416,
      "tokens_seen": 2245132288
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003251919191919192,
      "loss": 1.5132,
      "theoretical_loss": 3.3978463123913425,
      "tokens_seen": 2245197824
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032517171717171716,
      "loss": 1.4369,
      "theoretical_loss": 3.397838229836733,
      "tokens_seen": 2245263360
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003251515151515151,
      "loss": 1.443,
      "theoretical_loss": 3.397830147584093,
      "tokens_seen": 2245328896
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032513131313131313,
      "loss": 1.6424,
      "theoretical_loss": 3.3978220656334024,
      "tokens_seen": 2245394432
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032511111111111114,
      "loss": 1.4265,
      "theoretical_loss": 3.3978139839846406,
      "tokens_seen": 2245459968
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003250909090909091,
      "loss": 1.4644,
      "theoretical_loss": 3.397805902637789,
      "tokens_seen": 2245525504
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032507070707070705,
      "loss": 1.2572,
      "theoretical_loss": 3.397797821592826,
      "tokens_seen": 2245591040
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032505050505050507,
      "loss": 1.2394,
      "theoretical_loss": 3.397789740849732,
      "tokens_seen": 2245656576
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000325030303030303,
      "loss": 1.191,
      "theoretical_loss": 3.3977816604084876,
      "tokens_seen": 2245722112
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032501010101010103,
      "loss": 1.5256,
      "theoretical_loss": 3.397773580269072,
      "tokens_seen": 2245787648
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000324989898989899,
      "loss": 1.5632,
      "theoretical_loss": 3.397765500431465,
      "tokens_seen": 2245853184
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032496969696969695,
      "loss": 1.5225,
      "theoretical_loss": 3.3977574208956467,
      "tokens_seen": 2245918720
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032494949494949496,
      "loss": 1.6012,
      "theoretical_loss": 3.3977493416615974,
      "tokens_seen": 2245984256
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032492929292929297,
      "loss": 1.5263,
      "theoretical_loss": 3.3977412627292964,
      "tokens_seen": 2246049792
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003249090909090909,
      "loss": 1.5229,
      "theoretical_loss": 3.397733184098724,
      "tokens_seen": 2246115328
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003248888888888889,
      "loss": 1.4393,
      "theoretical_loss": 3.39772510576986,
      "tokens_seen": 2246180864
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003248686868686869,
      "loss": 1.534,
      "theoretical_loss": 3.3977170277426847,
      "tokens_seen": 2246246400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032484848484848485,
      "loss": 1.4826,
      "theoretical_loss": 3.3977089500171775,
      "tokens_seen": 2246311936
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032482828282828286,
      "loss": 1.4662,
      "theoretical_loss": 3.3977008725933184,
      "tokens_seen": 2246377472
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003248080808080808,
      "loss": 1.257,
      "theoretical_loss": 3.3976927954710874,
      "tokens_seen": 2246443008
    },
    {
      "epoch": 0.36,
      "objective/train/aligned_token_count": 0.45869991183280945,
      "objective/train/avg_token_score": 0.009318995289504528,
      "objective/train/docs_used": 1272482,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4994103908538818,
      "objective/train/original_loss": 4.357715606689453,
      "objective/train/theoretical_loss": 3.397684718650465,
      "objective/train/tokens_used": 605367776,
      "objective/train/unlikelihood_loss": 0.05522417649626732,
      "theoretical_loss": 3.397684718650465,
      "tokens_seen": 2246508544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003247878787878788,
      "loss": 1.3811,
      "theoretical_loss": 3.397684718650465,
      "tokens_seen": 2246508544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003247676767676768,
      "loss": 1.5185,
      "theoretical_loss": 3.3976766421314304,
      "tokens_seen": 2246574080
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032474747474747474,
      "loss": 1.7004,
      "theoretical_loss": 3.397668565913964,
      "tokens_seen": 2246639616
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032472727272727276,
      "loss": 1.2445,
      "theoretical_loss": 3.3976604899980454,
      "tokens_seen": 2246705152
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003247070707070707,
      "loss": 1.5241,
      "theoretical_loss": 3.3976524143836544,
      "tokens_seen": 2246770688
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003246868686868687,
      "loss": 1.543,
      "theoretical_loss": 3.3976443390707716,
      "tokens_seen": 2246836224
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003246666666666667,
      "loss": 1.5563,
      "theoretical_loss": 3.3976362640593765,
      "tokens_seen": 2246901760
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032464646464646464,
      "loss": 1.3351,
      "theoretical_loss": 3.397628189349449,
      "tokens_seen": 2246967296
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032462626262626265,
      "loss": 1.4788,
      "theoretical_loss": 3.3976201149409695,
      "tokens_seen": 2247032832
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003246060606060606,
      "loss": 1.5545,
      "theoretical_loss": 3.3976120408339177,
      "tokens_seen": 2247098368
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003245858585858586,
      "loss": 1.3752,
      "theoretical_loss": 3.3976039670282736,
      "tokens_seen": 2247163904
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032456565656565657,
      "loss": 1.2239,
      "theoretical_loss": 3.3975958935240165,
      "tokens_seen": 2247229440
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032454545454545453,
      "loss": 1.2426,
      "theoretical_loss": 3.397587820321127,
      "tokens_seen": 2247294976
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032452525252525254,
      "loss": 1.5343,
      "theoretical_loss": 3.3975797474195857,
      "tokens_seen": 2247360512
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032450505050505055,
      "loss": 1.4623,
      "theoretical_loss": 3.397571674819371,
      "tokens_seen": 2247426048
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003244848484848485,
      "loss": 1.3089,
      "theoretical_loss": 3.3975636025204645,
      "tokens_seen": 2247491584
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032446464646464647,
      "loss": 1.7218,
      "theoretical_loss": 3.397555530522845,
      "tokens_seen": 2247557120
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003244444444444444,
      "loss": 1.3702,
      "theoretical_loss": 3.3975474588264927,
      "tokens_seen": 2247622656
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032442424242424243,
      "loss": 1.543,
      "theoretical_loss": 3.397539387431388,
      "tokens_seen": 2247688192
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032440404040404044,
      "loss": 1.3016,
      "theoretical_loss": 3.3975313163375107,
      "tokens_seen": 2247753728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003243838383838384,
      "loss": 1.5147,
      "theoretical_loss": 3.39752324554484,
      "tokens_seen": 2247819264
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032436363636363636,
      "loss": 1.2847,
      "theoretical_loss": 3.3975151750533574,
      "tokens_seen": 2247884800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003243434343434343,
      "loss": 1.2951,
      "theoretical_loss": 3.397507104863042,
      "tokens_seen": 2247950336
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003243232323232324,
      "loss": 1.5789,
      "theoretical_loss": 3.397499034973873,
      "tokens_seen": 2248015872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032430303030303034,
      "loss": 1.3023,
      "theoretical_loss": 3.397490965385832,
      "tokens_seen": 2248081408
    },
    {
      "epoch": 0.36,
      "objective/train/aligned_token_count": 0.4081744849681854,
      "objective/train/avg_token_score": 0.022701717913150787,
      "objective/train/docs_used": 1273740,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1919264793395996,
      "objective/train/original_loss": 4.763843536376953,
      "objective/train/theoretical_loss": 3.397482896098898,
      "objective/train/tokens_used": 607006176,
      "objective/train/unlikelihood_loss": 0.03907008469104767,
      "theoretical_loss": 3.397482896098898,
      "tokens_seen": 2248146944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003242828282828283,
      "loss": 1.4822,
      "theoretical_loss": 3.397482896098898,
      "tokens_seen": 2248146944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032426262626262625,
      "loss": 1.3441,
      "theoretical_loss": 3.397474827113051,
      "tokens_seen": 2248212480
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003242424242424242,
      "loss": 1.3524,
      "theoretical_loss": 3.397466758428271,
      "tokens_seen": 2248278016
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032422222222222227,
      "loss": 1.4582,
      "theoretical_loss": 3.3974586900445383,
      "tokens_seen": 2248343552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032420202020202023,
      "loss": 1.6156,
      "theoretical_loss": 3.3974506219618332,
      "tokens_seen": 2248409088
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003241818181818182,
      "loss": 1.4954,
      "theoretical_loss": 3.3974425541801345,
      "tokens_seen": 2248474624
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032416161616161614,
      "loss": 1.4916,
      "theoretical_loss": 3.3974344866994235,
      "tokens_seen": 2248540160
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003241414141414141,
      "loss": 1.4001,
      "theoretical_loss": 3.3974264195196797,
      "tokens_seen": 2248605696
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032412121212121217,
      "loss": 1.4004,
      "theoretical_loss": 3.3974183526408828,
      "tokens_seen": 2248671232
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003241010101010101,
      "loss": 1.455,
      "theoretical_loss": 3.397410286063013,
      "tokens_seen": 2248736768
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003240808080808081,
      "loss": 1.526,
      "theoretical_loss": 3.3974022197860503,
      "tokens_seen": 2248802304
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032406060606060604,
      "loss": 1.6362,
      "theoretical_loss": 3.3973941538099752,
      "tokens_seen": 2248867840
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032404040404040405,
      "loss": 1.4484,
      "theoretical_loss": 3.397386088134767,
      "tokens_seen": 2248933376
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032402020202020206,
      "loss": 1.2939,
      "theoretical_loss": 3.397378022760406,
      "tokens_seen": 2248998912
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000324,
      "loss": 1.6365,
      "theoretical_loss": 3.397369957686872,
      "tokens_seen": 2249064448
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032397979797979797,
      "loss": 1.5143,
      "theoretical_loss": 3.3973618929141454,
      "tokens_seen": 2249129984
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032395959595959593,
      "loss": 1.6437,
      "theoretical_loss": 3.397353828442206,
      "tokens_seen": 2249195520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032393939393939394,
      "loss": 1.5832,
      "theoretical_loss": 3.3973457642710336,
      "tokens_seen": 2249261056
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032391919191919195,
      "loss": 1.6569,
      "theoretical_loss": 3.397337700400609,
      "tokens_seen": 2249326592
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003238989898989899,
      "loss": 1.4805,
      "theoretical_loss": 3.397329636830911,
      "tokens_seen": 2249392128
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032387878787878786,
      "loss": 1.3956,
      "theoretical_loss": 3.397321573561921,
      "tokens_seen": 2249457664
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003238585858585859,
      "loss": 1.3033,
      "theoretical_loss": 3.3973135105936176,
      "tokens_seen": 2249523200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032383838383838383,
      "loss": 1.3286,
      "theoretical_loss": 3.397305447925982,
      "tokens_seen": 2249588736
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032381818181818184,
      "loss": 1.4236,
      "theoretical_loss": 3.3972973855589936,
      "tokens_seen": 2249654272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003237979797979798,
      "loss": 1.2614,
      "theoretical_loss": 3.397289323492633,
      "tokens_seen": 2249719808
    },
    {
      "epoch": 0.36,
      "objective/train/aligned_token_count": 0.4414100646972656,
      "objective/train/avg_token_score": 0.010622130706906319,
      "objective/train/docs_used": 1274460,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5247832536697388,
      "objective/train/original_loss": 4.059118747711182,
      "objective/train/theoretical_loss": 3.3972812617268797,
      "objective/train/tokens_used": 608644576,
      "objective/train/unlikelihood_loss": 0.07704553753137589,
      "theoretical_loss": 3.3972812617268797,
      "tokens_seen": 2249785344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032377777777777776,
      "loss": 1.4216,
      "theoretical_loss": 3.3972812617268797,
      "tokens_seen": 2249785344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032375757575757577,
      "loss": 1.4023,
      "theoretical_loss": 3.397273200261713,
      "tokens_seen": 2249850880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003237373737373737,
      "loss": 1.429,
      "theoretical_loss": 3.397265139097115,
      "tokens_seen": 2249916416
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032371717171717174,
      "loss": 1.5281,
      "theoretical_loss": 3.397257078233064,
      "tokens_seen": 2249981952
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003236969696969697,
      "loss": 1.2629,
      "theoretical_loss": 3.397249017669541,
      "tokens_seen": 2250047488
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003236767676767677,
      "loss": 1.4384,
      "theoretical_loss": 3.397240957406525,
      "tokens_seen": 2250113024
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032365656565656566,
      "loss": 1.4441,
      "theoretical_loss": 3.397232897443997,
      "tokens_seen": 2250178560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003236363636363636,
      "loss": 1.3831,
      "theoretical_loss": 3.397224837781937,
      "tokens_seen": 2250244096
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032361616161616163,
      "loss": 1.4635,
      "theoretical_loss": 3.397216778420325,
      "tokens_seen": 2250309632
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003235959595959596,
      "loss": 1.4113,
      "theoretical_loss": 3.3972087193591403,
      "tokens_seen": 2250375168
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003235757575757576,
      "loss": 1.6446,
      "theoretical_loss": 3.3972006605983633,
      "tokens_seen": 2250440704
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032355555555555555,
      "loss": 1.4168,
      "theoretical_loss": 3.397192602137975,
      "tokens_seen": 2250506240
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003235353535353535,
      "loss": 1.4173,
      "theoretical_loss": 3.397184543977954,
      "tokens_seen": 2250571776
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003235151515151515,
      "loss": 1.4192,
      "theoretical_loss": 3.3971764861182816,
      "tokens_seen": 2250637312
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032349494949494953,
      "loss": 1.547,
      "theoretical_loss": 3.397168428558937,
      "tokens_seen": 2250702848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234747474747475,
      "loss": 1.5355,
      "theoretical_loss": 3.3971603712999006,
      "tokens_seen": 2250768384
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032345454545454545,
      "loss": 1.2612,
      "theoretical_loss": 3.3971523143411524,
      "tokens_seen": 2250833920
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234343434343434,
      "loss": 1.5066,
      "theoretical_loss": 3.3971442576826725,
      "tokens_seen": 2250899456
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003234141414141414,
      "loss": 1.5522,
      "theoretical_loss": 3.3971362013244413,
      "tokens_seen": 2250964992
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003233939393939394,
      "loss": 1.5889,
      "theoretical_loss": 3.3971281452664384,
      "tokens_seen": 2251030528
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003233737373737374,
      "loss": 1.2159,
      "theoretical_loss": 3.3971200895086437,
      "tokens_seen": 2251096064
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032335353535353534,
      "loss": 1.5796,
      "theoretical_loss": 3.3971120340510375,
      "tokens_seen": 2251161600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003233333333333333,
      "loss": 1.679,
      "theoretical_loss": 3.3971039788936004,
      "tokens_seen": 2251227136
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032331313131313136,
      "loss": 1.5734,
      "theoretical_loss": 3.3970959240363117,
      "tokens_seen": 2251292672
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003232929292929293,
      "loss": 1.5368,
      "theoretical_loss": 3.397087869479152,
      "tokens_seen": 2251358208
    },
    {
      "epoch": 0.36,
      "objective/train/aligned_token_count": 0.2765762507915497,
      "objective/train/avg_token_score": 0.01770329289138317,
      "objective/train/docs_used": 1275535,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8958865404129028,
      "objective/train/original_loss": 4.806623458862305,
      "objective/train/theoretical_loss": 3.3970798152221016,
      "objective/train/tokens_used": 610282976,
      "objective/train/unlikelihood_loss": 0.04785655066370964,
      "theoretical_loss": 3.3970798152221016,
      "tokens_seen": 2251423744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003232727272727273,
      "loss": 1.2709,
      "theoretical_loss": 3.3970798152221016,
      "tokens_seen": 2251423744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032325252525252523,
      "loss": 1.453,
      "theoretical_loss": 3.3970717612651398,
      "tokens_seen": 2251489280
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032323232323232324,
      "loss": 1.2435,
      "theoretical_loss": 3.397063707608247,
      "tokens_seen": 2251554816
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032321212121212125,
      "loss": 1.3073,
      "theoretical_loss": 3.397055654251403,
      "tokens_seen": 2251620352
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003231919191919192,
      "loss": 1.2233,
      "theoretical_loss": 3.397047601194589,
      "tokens_seen": 2251685888
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032317171717171717,
      "loss": 1.371,
      "theoretical_loss": 3.397039548437784,
      "tokens_seen": 2251751424
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003231515151515151,
      "loss": 1.1985,
      "theoretical_loss": 3.3970314959809684,
      "tokens_seen": 2251816960
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003231313131313132,
      "loss": 1.1467,
      "theoretical_loss": 3.3970234438241222,
      "tokens_seen": 2251882496
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032311111111111115,
      "loss": 1.5857,
      "theoretical_loss": 3.397015391967226,
      "tokens_seen": 2251948032
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003230909090909091,
      "loss": 1.6056,
      "theoretical_loss": 3.3970073404102594,
      "tokens_seen": 2252013568
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032307070707070706,
      "loss": 1.6018,
      "theoretical_loss": 3.3969992891532024,
      "tokens_seen": 2252079104
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000323050505050505,
      "loss": 1.3144,
      "theoretical_loss": 3.3969912381960357,
      "tokens_seen": 2252144640
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003230303030303031,
      "loss": 1.0216,
      "theoretical_loss": 3.396983187538739,
      "tokens_seen": 2252210176
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032301010101010104,
      "loss": 1.2711,
      "theoretical_loss": 3.3969751371812915,
      "tokens_seen": 2252275712
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322989898989899,
      "loss": 1.372,
      "theoretical_loss": 3.396967087123675,
      "tokens_seen": 2252341248
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032296969696969695,
      "loss": 1.3176,
      "theoretical_loss": 3.396959037365869,
      "tokens_seen": 2252406784
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032294949494949496,
      "loss": 1.3419,
      "theoretical_loss": 3.3969509879078528,
      "tokens_seen": 2252472320
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322929292929293,
      "loss": 1.3391,
      "theoretical_loss": 3.396942938749608,
      "tokens_seen": 2252537856
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032290909090909093,
      "loss": 1.5176,
      "theoretical_loss": 3.396934889891113,
      "tokens_seen": 2252603392
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003228888888888889,
      "loss": 1.4952,
      "theoretical_loss": 3.3969268413323497,
      "tokens_seen": 2252668928
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032286868686868685,
      "loss": 1.3817,
      "theoretical_loss": 3.3969187930732967,
      "tokens_seen": 2252734464
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032284848484848486,
      "loss": 1.43,
      "theoretical_loss": 3.396910745113935,
      "tokens_seen": 2252800000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032282828282828287,
      "loss": 1.2426,
      "theoretical_loss": 3.3969026974542444,
      "tokens_seen": 2252865536
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003228080808080808,
      "loss": 1.5881,
      "theoretical_loss": 3.3968946500942048,
      "tokens_seen": 2252931072
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003227878787878788,
      "loss": 1.3793,
      "theoretical_loss": 3.396886603033797,
      "tokens_seen": 2252996608
    },
    {
      "epoch": 0.36,
      "objective/train/aligned_token_count": 0.3980327546596527,
      "objective/train/avg_token_score": 0.0226760171353817,
      "objective/train/docs_used": 1276002,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1978646516799927,
      "objective/train/original_loss": 4.280903339385986,
      "objective/train/theoretical_loss": 3.3968785562730006,
      "objective/train/tokens_used": 611921376,
      "objective/train/unlikelihood_loss": 0.05233257636427879,
      "theoretical_loss": 3.3968785562730006,
      "tokens_seen": 2253062144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032276767676767674,
      "loss": 1.3978,
      "theoretical_loss": 3.3968785562730006,
      "tokens_seen": 2253062144
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032274747474747475,
      "loss": 1.265,
      "theoretical_loss": 3.396870509811796,
      "tokens_seen": 2253127680
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032272727272727276,
      "loss": 1.3247,
      "theoretical_loss": 3.396862463650163,
      "tokens_seen": 2253193216
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003227070707070707,
      "loss": 1.4131,
      "theoretical_loss": 3.3968544177880817,
      "tokens_seen": 2253258752
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003226868686868687,
      "loss": 1.373,
      "theoretical_loss": 3.396846372225533,
      "tokens_seen": 2253324288
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003226666666666667,
      "loss": 1.3615,
      "theoretical_loss": 3.3968383269624964,
      "tokens_seen": 2253389824
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032264646464646464,
      "loss": 1.3092,
      "theoretical_loss": 3.396830281998952,
      "tokens_seen": 2253455360
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032262626262626265,
      "loss": 1.195,
      "theoretical_loss": 3.3968222373348804,
      "tokens_seen": 2253520896
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003226060606060606,
      "loss": 1.4153,
      "theoretical_loss": 3.396814192970261,
      "tokens_seen": 2253586432
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032258585858585857,
      "loss": 1.3903,
      "theoretical_loss": 3.3968061489050747,
      "tokens_seen": 2253651968
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003225656565656566,
      "loss": 1.5762,
      "theoretical_loss": 3.3967981051393012,
      "tokens_seen": 2253717504
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032254545454545453,
      "loss": 1.2702,
      "theoretical_loss": 3.3967900616729207,
      "tokens_seen": 2253783040
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032252525252525255,
      "loss": 1.3156,
      "theoretical_loss": 3.3967820185059137,
      "tokens_seen": 2253848576
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003225050505050505,
      "loss": 1.2227,
      "theoretical_loss": 3.3967739756382596,
      "tokens_seen": 2253914112
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003224848484848485,
      "loss": 1.6847,
      "theoretical_loss": 3.3967659330699393,
      "tokens_seen": 2253979648
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032246464646464647,
      "loss": 1.2627,
      "theoretical_loss": 3.396757890800933,
      "tokens_seen": 2254045184
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032244444444444443,
      "loss": 1.5797,
      "theoretical_loss": 3.39674984883122,
      "tokens_seen": 2254110720
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032242424242424244,
      "loss": 1.4668,
      "theoretical_loss": 3.3967418071607813,
      "tokens_seen": 2254176256
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003224040404040404,
      "loss": 1.3943,
      "theoretical_loss": 3.396733765789597,
      "tokens_seen": 2254241792
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003223838383838384,
      "loss": 1.439,
      "theoretical_loss": 3.3967257247176468,
      "tokens_seen": 2254307328
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032236363636363636,
      "loss": 1.3674,
      "theoretical_loss": 3.396717683944911,
      "tokens_seen": 2254372864
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003223434343434343,
      "loss": 1.4913,
      "theoretical_loss": 3.39670964347137,
      "tokens_seen": 2254438400
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032232323232323233,
      "loss": 1.6169,
      "theoretical_loss": 3.3967016032970037,
      "tokens_seen": 2254503936
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032230303030303034,
      "loss": 1.368,
      "theoretical_loss": 3.3966935634217927,
      "tokens_seen": 2254569472
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003222828282828283,
      "loss": 1.4036,
      "theoretical_loss": 3.3966855238457168,
      "tokens_seen": 2254635008
    },
    {
      "epoch": 0.36,
      "objective/train/aligned_token_count": 0.4541178047657013,
      "objective/train/avg_token_score": 0.00984613224864006,
      "objective/train/docs_used": 1277060,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.490248203277588,
      "objective/train/original_loss": 4.291781425476074,
      "objective/train/theoretical_loss": 3.396677484568756,
      "objective/train/tokens_used": 613559776,
      "objective/train/unlikelihood_loss": 0.06463318318128586,
      "theoretical_loss": 3.396677484568756,
      "tokens_seen": 2254700544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032226262626262626,
      "loss": 1.5803,
      "theoretical_loss": 3.396677484568756,
      "tokens_seen": 2254700544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003222424242424242,
      "loss": 1.3619,
      "theoretical_loss": 3.396669445590891,
      "tokens_seen": 2254766080
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003222222222222222,
      "loss": 1.3075,
      "theoretical_loss": 3.3966614069121017,
      "tokens_seen": 2254831616
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032220202020202024,
      "loss": 1.4748,
      "theoretical_loss": 3.3966533685323683,
      "tokens_seen": 2254897152
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003221818181818182,
      "loss": 1.6975,
      "theoretical_loss": 3.3966453304516713,
      "tokens_seen": 2254962688
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032216161616161615,
      "loss": 1.3707,
      "theoretical_loss": 3.39663729266999,
      "tokens_seen": 2255028224
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003221414141414141,
      "loss": 1.6489,
      "theoretical_loss": 3.3966292551873054,
      "tokens_seen": 2255093760
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032212121212121217,
      "loss": 1.4528,
      "theoretical_loss": 3.3966212180035975,
      "tokens_seen": 2255159296
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032210101010101013,
      "loss": 1.3579,
      "theoretical_loss": 3.3966131811188465,
      "tokens_seen": 2255224832
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003220808080808081,
      "loss": 1.3839,
      "theoretical_loss": 3.3966051445330323,
      "tokens_seen": 2255290368
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032206060606060604,
      "loss": 1.5057,
      "theoretical_loss": 3.3965971082461355,
      "tokens_seen": 2255355904
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322040404040404,
      "loss": 1.4487,
      "theoretical_loss": 3.396589072258136,
      "tokens_seen": 2255421440
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032202020202020206,
      "loss": 1.4962,
      "theoretical_loss": 3.3965810365690143,
      "tokens_seen": 2255486976
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000322,
      "loss": 1.3446,
      "theoretical_loss": 3.3965730011787505,
      "tokens_seen": 2255552512
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321979797979798,
      "loss": 1.5295,
      "theoretical_loss": 3.3965649660873245,
      "tokens_seen": 2255618048
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032195959595959593,
      "loss": 1.2575,
      "theoretical_loss": 3.3965569312947164,
      "tokens_seen": 2255683584
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032193939393939395,
      "loss": 1.5213,
      "theoretical_loss": 3.3965488968009074,
      "tokens_seen": 2255749120
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032191919191919196,
      "loss": 1.2468,
      "theoretical_loss": 3.3965408626058764,
      "tokens_seen": 2255814656
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003218989898989899,
      "loss": 1.4316,
      "theoretical_loss": 3.3965328287096046,
      "tokens_seen": 2255880192
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032187878787878787,
      "loss": 1.3831,
      "theoretical_loss": 3.396524795112072,
      "tokens_seen": 2255945728
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003218585858585858,
      "loss": 1.4654,
      "theoretical_loss": 3.396516761813259,
      "tokens_seen": 2256011264
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032183838383838384,
      "loss": 1.2699,
      "theoretical_loss": 3.396508728813145,
      "tokens_seen": 2256076800
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032181818181818185,
      "loss": 1.4817,
      "theoretical_loss": 3.3965006961117106,
      "tokens_seen": 2256142336
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003217979797979798,
      "loss": 1.3603,
      "theoretical_loss": 3.396492663708936,
      "tokens_seen": 2256207872
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032177777777777776,
      "loss": 1.4273,
      "theoretical_loss": 3.396484631604802,
      "tokens_seen": 2256273408
    },
    {
      "epoch": 0.36,
      "objective/train/aligned_token_count": 0.38404202461242676,
      "objective/train/avg_token_score": 0.020460879430174828,
      "objective/train/docs_used": 1277665,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3331819772720337,
      "objective/train/original_loss": 4.696850299835205,
      "objective/train/theoretical_loss": 3.3964765997992883,
      "objective/train/tokens_used": 615198176,
      "objective/train/unlikelihood_loss": 0.07580170780420303,
      "theoretical_loss": 3.3964765997992883,
      "tokens_seen": 2256338944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003217575757575758,
      "loss": 1.6563,
      "theoretical_loss": 3.3964765997992883,
      "tokens_seen": 2256338944
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032173737373737373,
      "loss": 1.5042,
      "theoretical_loss": 3.396468568292375,
      "tokens_seen": 2256404480
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032171717171717174,
      "loss": 1.5725,
      "theoretical_loss": 3.3964605370840424,
      "tokens_seen": 2256470016
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003216969696969697,
      "loss": 1.4661,
      "theoretical_loss": 3.3964525061742714,
      "tokens_seen": 2256535552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032167676767676766,
      "loss": 1.4444,
      "theoretical_loss": 3.396444475563041,
      "tokens_seen": 2256601088
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032165656565656567,
      "loss": 1.5849,
      "theoretical_loss": 3.3964364452503326,
      "tokens_seen": 2256666624
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003216363636363636,
      "loss": 1.4091,
      "theoretical_loss": 3.3964284152361257,
      "tokens_seen": 2256732160
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032161616161616163,
      "loss": 1.1892,
      "theoretical_loss": 3.396420385520401,
      "tokens_seen": 2256797696
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003215959595959596,
      "loss": 1.3864,
      "theoretical_loss": 3.3964123561031383,
      "tokens_seen": 2256863232
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003215757575757576,
      "loss": 1.2063,
      "theoretical_loss": 3.3964043269843183,
      "tokens_seen": 2256928768
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032155555555555556,
      "loss": 1.3192,
      "theoretical_loss": 3.396396298163921,
      "tokens_seen": 2256994304
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032153535353535357,
      "loss": 1.6945,
      "theoretical_loss": 3.3963882696419265,
      "tokens_seen": 2257059840
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032151515151515153,
      "loss": 1.3612,
      "theoretical_loss": 3.3963802414183153,
      "tokens_seen": 2257125376
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003214949494949495,
      "loss": 1.6943,
      "theoretical_loss": 3.396372213493067,
      "tokens_seen": 2257190912
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003214747474747475,
      "loss": 1.3488,
      "theoretical_loss": 3.396364185866163,
      "tokens_seen": 2257256448
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032145454545454545,
      "loss": 1.8195,
      "theoretical_loss": 3.396356158537583,
      "tokens_seen": 2257321984
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032143434343434346,
      "loss": 1.41,
      "theoretical_loss": 3.396348131507307,
      "tokens_seen": 2257387520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003214141414141414,
      "loss": 1.2774,
      "theoretical_loss": 3.3963401047753154,
      "tokens_seen": 2257453056
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032139393939393943,
      "loss": 1.2855,
      "theoretical_loss": 3.3963320783415885,
      "tokens_seen": 2257518592
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003213737373737374,
      "loss": 1.5862,
      "theoretical_loss": 3.3963240522061064,
      "tokens_seen": 2257584128
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032135353535353534,
      "loss": 1.2316,
      "theoretical_loss": 3.39631602636885,
      "tokens_seen": 2257649664
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032133333333333336,
      "loss": 1.4759,
      "theoretical_loss": 3.396308000829799,
      "tokens_seen": 2257715200
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003213131313131313,
      "loss": 1.3337,
      "theoretical_loss": 3.396299975588933,
      "tokens_seen": 2257780736
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003212929292929293,
      "loss": 1.4783,
      "theoretical_loss": 3.396291950646234,
      "tokens_seen": 2257846272
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003212727272727273,
      "loss": 1.4414,
      "theoretical_loss": 3.396283926001681,
      "tokens_seen": 2257911808
    },
    {
      "epoch": 0.36,
      "objective/train/aligned_token_count": 0.5074535608291626,
      "objective/train/avg_token_score": 0.006834826432168484,
      "objective/train/docs_used": 1279074,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7631499767303467,
      "objective/train/original_loss": 4.250295162200928,
      "objective/train/theoretical_loss": 3.396275901655254,
      "objective/train/tokens_used": 616836576,
      "objective/train/unlikelihood_loss": 0.06310959160327911,
      "theoretical_loss": 3.396275901655254,
      "tokens_seen": 2257977344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032125252525252524,
      "loss": 1.6222,
      "theoretical_loss": 3.396275901655254,
      "tokens_seen": 2257977344
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032123232323232325,
      "loss": 1.3458,
      "theoretical_loss": 3.3962678776069346,
      "tokens_seen": 2258042880
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032121212121212126,
      "loss": 1.3431,
      "theoretical_loss": 3.396259853856702,
      "tokens_seen": 2258108416
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003211919191919192,
      "loss": 1.2378,
      "theoretical_loss": 3.3962518304045366,
      "tokens_seen": 2258173952
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003211717171717172,
      "loss": 1.2636,
      "theoretical_loss": 3.396243807250419,
      "tokens_seen": 2258239488
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032115151515151513,
      "loss": 1.5049,
      "theoretical_loss": 3.3962357843943294,
      "tokens_seen": 2258305024
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032113131313131314,
      "loss": 1.3001,
      "theoretical_loss": 3.396227761836248,
      "tokens_seen": 2258370560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032111111111111115,
      "loss": 1.4809,
      "theoretical_loss": 3.396219739576155,
      "tokens_seen": 2258436096
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003210909090909091,
      "loss": 1.4123,
      "theoretical_loss": 3.3962117176140305,
      "tokens_seen": 2258501632
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032107070707070707,
      "loss": 1.3161,
      "theoretical_loss": 3.396203695949856,
      "tokens_seen": 2258567168
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000321050505050505,
      "loss": 1.2127,
      "theoretical_loss": 3.39619567458361,
      "tokens_seen": 2258632704
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003210303030303031,
      "loss": 1.2252,
      "theoretical_loss": 3.396187653515274,
      "tokens_seen": 2258698240
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032101010101010104,
      "loss": 1.3407,
      "theoretical_loss": 3.3961796327448277,
      "tokens_seen": 2258763776
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000320989898989899,
      "loss": 1.4325,
      "theoretical_loss": 3.396171612272252,
      "tokens_seen": 2258829312
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032096969696969696,
      "loss": 1.5917,
      "theoretical_loss": 3.3961635920975266,
      "tokens_seen": 2258894848
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003209494949494949,
      "loss": 1.4004,
      "theoretical_loss": 3.396155572220632,
      "tokens_seen": 2258960384
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000320929292929293,
      "loss": 1.5727,
      "theoretical_loss": 3.3961475526415486,
      "tokens_seen": 2259025920
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032090909090909094,
      "loss": 1.3955,
      "theoretical_loss": 3.396139533360256,
      "tokens_seen": 2259091456
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003208888888888889,
      "loss": 1.3422,
      "theoretical_loss": 3.3961315143767363,
      "tokens_seen": 2259156992
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032086868686868685,
      "loss": 1.3714,
      "theoretical_loss": 3.396123495690968,
      "tokens_seen": 2259222528
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003208484848484848,
      "loss": 1.2652,
      "theoretical_loss": 3.3961154773029323,
      "tokens_seen": 2259288064
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003208282828282829,
      "loss": 1.4912,
      "theoretical_loss": 3.396107459212609,
      "tokens_seen": 2259353600
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032080808080808083,
      "loss": 1.325,
      "theoretical_loss": 3.3960994414199783,
      "tokens_seen": 2259419136
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0003207878787878788,
      "loss": 1.4257,
      "theoretical_loss": 3.3960914239250215,
      "tokens_seen": 2259484672
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032076767676767674,
      "loss": 1.196,
      "theoretical_loss": 3.396083406727718,
      "tokens_seen": 2259550208
    },
    {
      "epoch": 0.36,
      "objective/train/aligned_token_count": 0.4451368451118469,
      "objective/train/avg_token_score": 0.005456716753542423,
      "objective/train/docs_used": 1279794,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5101059675216675,
      "objective/train/original_loss": 4.784608364105225,
      "objective/train/theoretical_loss": 3.3960753898280487,
      "objective/train/tokens_used": 618474976,
      "objective/train/unlikelihood_loss": 0.04287702962756157,
      "theoretical_loss": 3.3960753898280487,
      "tokens_seen": 2259615744
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00032074747474747476,
      "loss": 1.4806,
      "theoretical_loss": 3.3960753898280487,
      "tokens_seen": 2259615744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032072727272727277,
      "loss": 1.4367,
      "theoretical_loss": 3.3960673732259936,
      "tokens_seen": 2259681280
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003207070707070707,
      "loss": 1.4071,
      "theoretical_loss": 3.396059356921533,
      "tokens_seen": 2259746816
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003206868686868687,
      "loss": 1.2678,
      "theoretical_loss": 3.396051340914647,
      "tokens_seen": 2259812352
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032066666666666664,
      "loss": 1.2934,
      "theoretical_loss": 3.3960433252053166,
      "tokens_seen": 2259877888
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032064646464646465,
      "loss": 1.4741,
      "theoretical_loss": 3.3960353097935214,
      "tokens_seen": 2259943424
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032062626262626266,
      "loss": 1.3219,
      "theoretical_loss": 3.396027294679242,
      "tokens_seen": 2260008960
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003206060606060606,
      "loss": 1.2943,
      "theoretical_loss": 3.3960192798624593,
      "tokens_seen": 2260074496
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032058585858585857,
      "loss": 1.5051,
      "theoretical_loss": 3.3960112653431525,
      "tokens_seen": 2260140032
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003205656565656566,
      "loss": 1.3488,
      "theoretical_loss": 3.396003251121303,
      "tokens_seen": 2260205568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032054545454545454,
      "loss": 1.4134,
      "theoretical_loss": 3.3959952371968907,
      "tokens_seen": 2260271104
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032052525252525255,
      "loss": 1.2049,
      "theoretical_loss": 3.3959872235698954,
      "tokens_seen": 2260336640
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003205050505050505,
      "loss": 1.4153,
      "theoretical_loss": 3.3959792102402986,
      "tokens_seen": 2260402176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032048484848484847,
      "loss": 1.3853,
      "theoretical_loss": 3.3959711972080795,
      "tokens_seen": 2260467712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003204646464646465,
      "loss": 1.3351,
      "theoretical_loss": 3.395963184473219,
      "tokens_seen": 2260533248
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032044444444444443,
      "loss": 1.3555,
      "theoretical_loss": 3.3959551720356975,
      "tokens_seen": 2260598784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032042424242424244,
      "loss": 1.5299,
      "theoretical_loss": 3.3959471598954956,
      "tokens_seen": 2260664320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003204040404040404,
      "loss": 1.4875,
      "theoretical_loss": 3.3959391480525927,
      "tokens_seen": 2260729856
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203838383838384,
      "loss": 1.5177,
      "theoretical_loss": 3.39593113650697,
      "tokens_seen": 2260795392
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032036363636363637,
      "loss": 1.1689,
      "theoretical_loss": 3.3959231252586077,
      "tokens_seen": 2260860928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203434343434343,
      "loss": 1.5594,
      "theoretical_loss": 3.3959151143074857,
      "tokens_seen": 2260926464
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032032323232323234,
      "loss": 1.4139,
      "theoretical_loss": 3.395907103653585,
      "tokens_seen": 2260992000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003203030303030303,
      "loss": 1.316,
      "theoretical_loss": 3.3958990932968858,
      "tokens_seen": 2261057536
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202828282828283,
      "loss": 1.4796,
      "theoretical_loss": 3.395891083237368,
      "tokens_seen": 2261123072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032026262626262626,
      "loss": 1.5276,
      "theoretical_loss": 3.3958830734750123,
      "tokens_seen": 2261188608
    },
    {
      "epoch": 0.37,
      "objective/train/aligned_token_count": 0.5158846378326416,
      "objective/train/avg_token_score": 0.014068498276174068,
      "objective/train/docs_used": 1281070,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5966434478759766,
      "objective/train/original_loss": 3.950713872909546,
      "objective/train/theoretical_loss": 3.3958750640097994,
      "objective/train/tokens_used": 620113376,
      "objective/train/unlikelihood_loss": 0.051892075687646866,
      "theoretical_loss": 3.3958750640097994,
      "tokens_seen": 2261254144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003202424242424242,
      "loss": 1.6564,
      "theoretical_loss": 3.3958750640097994,
      "tokens_seen": 2261254144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032022222222222223,
      "loss": 1.5736,
      "theoretical_loss": 3.395867054841709,
      "tokens_seen": 2261319680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032020202020202024,
      "loss": 1.4138,
      "theoretical_loss": 3.3958590459707216,
      "tokens_seen": 2261385216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003201818181818182,
      "loss": 1.3033,
      "theoretical_loss": 3.3958510373968176,
      "tokens_seen": 2261450752
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032016161616161615,
      "loss": 1.4888,
      "theoretical_loss": 3.395843029119978,
      "tokens_seen": 2261516288
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003201414141414141,
      "loss": 1.6294,
      "theoretical_loss": 3.395835021140183,
      "tokens_seen": 2261581824
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003201212121212121,
      "loss": 1.5358,
      "theoretical_loss": 3.395827013457412,
      "tokens_seen": 2261647360
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032010101010101013,
      "loss": 1.5031,
      "theoretical_loss": 3.395819006071646,
      "tokens_seen": 2261712896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003200808080808081,
      "loss": 1.3777,
      "theoretical_loss": 3.3958109989828658,
      "tokens_seen": 2261778432
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032006060606060605,
      "loss": 1.2551,
      "theoretical_loss": 3.3958029921910513,
      "tokens_seen": 2261843968
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000320040404040404,
      "loss": 1.1802,
      "theoretical_loss": 3.395794985696183,
      "tokens_seen": 2261909504
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032002020202020207,
      "loss": 1.4263,
      "theoretical_loss": 3.395786979498241,
      "tokens_seen": 2261975040
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00032,
      "loss": 1.2827,
      "theoretical_loss": 3.3957789735972064,
      "tokens_seen": 2262040576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000319979797979798,
      "loss": 1.2766,
      "theoretical_loss": 3.3957709679930588,
      "tokens_seen": 2262106112
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031995959595959594,
      "loss": 1.4148,
      "theoretical_loss": 3.395762962685779,
      "tokens_seen": 2262171648
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003199393939393939,
      "loss": 1.3941,
      "theoretical_loss": 3.395754957675347,
      "tokens_seen": 2262237184
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031991919191919196,
      "loss": 1.6381,
      "theoretical_loss": 3.395746952961744,
      "tokens_seen": 2262302720
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003198989898989899,
      "loss": 1.448,
      "theoretical_loss": 3.39573894854495,
      "tokens_seen": 2262368256
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003198787878787879,
      "loss": 1.4175,
      "theoretical_loss": 3.395730944424945,
      "tokens_seen": 2262433792
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031985858585858583,
      "loss": 1.6616,
      "theoretical_loss": 3.3957229406017095,
      "tokens_seen": 2262499328
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031983838383838384,
      "loss": 1.3908,
      "theoretical_loss": 3.3957149370752244,
      "tokens_seen": 2262564864
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031981818181818185,
      "loss": 1.2474,
      "theoretical_loss": 3.39570693384547,
      "tokens_seen": 2262630400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197979797979798,
      "loss": 1.3953,
      "theoretical_loss": 3.3956989309124257,
      "tokens_seen": 2262695936
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031977777777777777,
      "loss": 1.3852,
      "theoretical_loss": 3.3956909282760734,
      "tokens_seen": 2262761472
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197575757575757,
      "loss": 1.2847,
      "theoretical_loss": 3.3956829259363928,
      "tokens_seen": 2262827008
    },
    {
      "epoch": 0.37,
      "objective/train/aligned_token_count": 0.4456256031990051,
      "objective/train/avg_token_score": 0.009554273448884487,
      "objective/train/docs_used": 1281765,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4664385318756104,
      "objective/train/original_loss": 4.559828758239746,
      "objective/train/theoretical_loss": 3.395674923893364,
      "objective/train/tokens_used": 621751776,
      "objective/train/unlikelihood_loss": 0.05876954644918442,
      "theoretical_loss": 3.395674923893364,
      "tokens_seen": 2262892544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003197373737373738,
      "loss": 1.635,
      "theoretical_loss": 3.395674923893364,
      "tokens_seen": 2262892544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031971717171717175,
      "loss": 1.2494,
      "theoretical_loss": 3.395666922146968,
      "tokens_seen": 2262958080
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003196969696969697,
      "loss": 1.4844,
      "theoretical_loss": 3.3956589206971852,
      "tokens_seen": 2263023616
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031967676767676766,
      "loss": 1.2211,
      "theoretical_loss": 3.3956509195439954,
      "tokens_seen": 2263089152
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031965656565656567,
      "loss": 1.3926,
      "theoretical_loss": 3.3956429186873796,
      "tokens_seen": 2263154688
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003196363636363637,
      "loss": 1.4106,
      "theoretical_loss": 3.395634918127318,
      "tokens_seen": 2263220224
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031961616161616164,
      "loss": 1.6201,
      "theoretical_loss": 3.3956269178637912,
      "tokens_seen": 2263285760
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003195959595959596,
      "loss": 1.4154,
      "theoretical_loss": 3.395618917896779,
      "tokens_seen": 2263351296
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031957575757575755,
      "loss": 1.4153,
      "theoretical_loss": 3.3956109182262626,
      "tokens_seen": 2263416832
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031955555555555556,
      "loss": 1.451,
      "theoretical_loss": 3.395602918852222,
      "tokens_seen": 2263482368
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003195353535353536,
      "loss": 1.401,
      "theoretical_loss": 3.395594919774638,
      "tokens_seen": 2263547904
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031951515151515153,
      "loss": 1.3213,
      "theoretical_loss": 3.39558692099349,
      "tokens_seen": 2263613440
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003194949494949495,
      "loss": 1.2835,
      "theoretical_loss": 3.39557892250876,
      "tokens_seen": 2263678976
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031947474747474745,
      "loss": 1.4184,
      "theoretical_loss": 3.3955709243204275,
      "tokens_seen": 2263744512
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031945454545454546,
      "loss": 1.3815,
      "theoretical_loss": 3.3955629264284726,
      "tokens_seen": 2263810048
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031943434343434347,
      "loss": 1.4012,
      "theoretical_loss": 3.3955549288328766,
      "tokens_seen": 2263875584
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003194141414141414,
      "loss": 1.5602,
      "theoretical_loss": 3.3955469315336195,
      "tokens_seen": 2263941120
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003193939393939394,
      "loss": 1.6686,
      "theoretical_loss": 3.395538934530682,
      "tokens_seen": 2264006656
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003193737373737374,
      "loss": 1.5422,
      "theoretical_loss": 3.395530937824044,
      "tokens_seen": 2264072192
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031935353535353535,
      "loss": 1.4225,
      "theoretical_loss": 3.3955229414136863,
      "tokens_seen": 2264137728
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031933333333333336,
      "loss": 1.5509,
      "theoretical_loss": 3.3955149452995896,
      "tokens_seen": 2264203264
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003193131313131313,
      "loss": 1.4872,
      "theoretical_loss": 3.395506949481734,
      "tokens_seen": 2264268800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192929292929293,
      "loss": 1.4178,
      "theoretical_loss": 3.3954989539601,
      "tokens_seen": 2264334336
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192727272727273,
      "loss": 1.4654,
      "theoretical_loss": 3.395490958734668,
      "tokens_seen": 2264399872
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031925252525252524,
      "loss": 1.3288,
      "theoretical_loss": 3.3954829638054185,
      "tokens_seen": 2264465408
    },
    {
      "epoch": 0.37,
      "objective/train/aligned_token_count": 0.3728005886077881,
      "objective/train/avg_token_score": 0.025287169963121414,
      "objective/train/docs_used": 1283066,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1364474296569824,
      "objective/train/original_loss": 4.377612113952637,
      "objective/train/theoretical_loss": 3.395474969172332,
      "objective/train/tokens_used": 623390176,
      "objective/train/unlikelihood_loss": 0.09229307621717453,
      "theoretical_loss": 3.395474969172332,
      "tokens_seen": 2264530944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031923232323232325,
      "loss": 1.3492,
      "theoretical_loss": 3.395474969172332,
      "tokens_seen": 2264530944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003192121212121212,
      "loss": 1.1642,
      "theoretical_loss": 3.395466974835389,
      "tokens_seen": 2264596480
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003191919191919192,
      "loss": 1.4769,
      "theoretical_loss": 3.39545898079457,
      "tokens_seen": 2264662016
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003191717171717172,
      "loss": 1.1757,
      "theoretical_loss": 3.3954509870498555,
      "tokens_seen": 2264727552
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031915151515151514,
      "loss": 1.504,
      "theoretical_loss": 3.3954429936012254,
      "tokens_seen": 2264793088
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031913131313131315,
      "loss": 1.3923,
      "theoretical_loss": 3.3954350004486606,
      "tokens_seen": 2264858624
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003191111111111111,
      "loss": 1.5972,
      "theoretical_loss": 3.3954270075921418,
      "tokens_seen": 2264924160
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003190909090909091,
      "loss": 1.5492,
      "theoretical_loss": 3.3954190150316492,
      "tokens_seen": 2264989696
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031907070707070707,
      "loss": 1.4975,
      "theoretical_loss": 3.3954110227671634,
      "tokens_seen": 2265055232
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031905050505050503,
      "loss": 1.4191,
      "theoretical_loss": 3.395403030798665,
      "tokens_seen": 2265120768
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031903030303030304,
      "loss": 1.3479,
      "theoretical_loss": 3.395395039126134,
      "tokens_seen": 2265186304
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031901010101010105,
      "loss": 1.3622,
      "theoretical_loss": 3.3953870477495514,
      "tokens_seen": 2265251840
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000318989898989899,
      "loss": 1.4282,
      "theoretical_loss": 3.395379056668897,
      "tokens_seen": 2265317376
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031896969696969696,
      "loss": 1.4593,
      "theoretical_loss": 3.395371065884152,
      "tokens_seen": 2265382912
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003189494949494949,
      "loss": 1.6206,
      "theoretical_loss": 3.395363075395296,
      "tokens_seen": 2265448448
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031892929292929293,
      "loss": 1.5248,
      "theoretical_loss": 3.395355085202311,
      "tokens_seen": 2265513984
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031890909090909094,
      "loss": 1.4146,
      "theoretical_loss": 3.395347095305176,
      "tokens_seen": 2265579520
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003188888888888889,
      "loss": 1.2943,
      "theoretical_loss": 3.3953391057038718,
      "tokens_seen": 2265645056
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031886868686868686,
      "loss": 1.5171,
      "theoretical_loss": 3.3953311163983795,
      "tokens_seen": 2265710592
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003188484848484848,
      "loss": 1.4393,
      "theoretical_loss": 3.395323127388679,
      "tokens_seen": 2265776128
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003188282828282829,
      "loss": 1.6842,
      "theoretical_loss": 3.3953151386747513,
      "tokens_seen": 2265841664
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031880808080808084,
      "loss": 1.3941,
      "theoretical_loss": 3.3953071502565764,
      "tokens_seen": 2265907200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003187878787878788,
      "loss": 1.2615,
      "theoretical_loss": 3.395299162134135,
      "tokens_seen": 2265972736
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031876767676767675,
      "loss": 1.5124,
      "theoretical_loss": 3.395291174307408,
      "tokens_seen": 2266038272
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003187474747474747,
      "loss": 1.1646,
      "theoretical_loss": 3.395283186776375,
      "tokens_seen": 2266103808
    },
    {
      "epoch": 0.37,
      "objective/train/aligned_token_count": 0.3621090054512024,
      "objective/train/avg_token_score": 0.02179253101348877,
      "objective/train/docs_used": 1283537,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2294130325317383,
      "objective/train/original_loss": 4.760871410369873,
      "objective/train/theoretical_loss": 3.3952751995410173,
      "objective/train/tokens_used": 625028576,
      "objective/train/unlikelihood_loss": 0.05112767964601517,
      "theoretical_loss": 3.3952751995410173,
      "tokens_seen": 2266169344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031872727272727277,
      "loss": 1.4634,
      "theoretical_loss": 3.3952751995410173,
      "tokens_seen": 2266169344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031870707070707073,
      "loss": 1.4078,
      "theoretical_loss": 3.395267212601315,
      "tokens_seen": 2266234880
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003186868686868687,
      "loss": 1.3478,
      "theoretical_loss": 3.3952592259572487,
      "tokens_seen": 2266300416
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031866666666666664,
      "loss": 1.115,
      "theoretical_loss": 3.395251239608799,
      "tokens_seen": 2266365952
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031864646464646465,
      "loss": 1.3535,
      "theoretical_loss": 3.3952432535559463,
      "tokens_seen": 2266431488
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031862626262626266,
      "loss": 1.6043,
      "theoretical_loss": 3.3952352677986712,
      "tokens_seen": 2266497024
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003186060606060606,
      "loss": 1.4207,
      "theoretical_loss": 3.395227282336954,
      "tokens_seen": 2266562560
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003185858585858586,
      "loss": 1.3105,
      "theoretical_loss": 3.3952192971707755,
      "tokens_seen": 2266628096
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031856565656565654,
      "loss": 1.6179,
      "theoretical_loss": 3.395211312300116,
      "tokens_seen": 2266693632
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031854545454545455,
      "loss": 1.553,
      "theoretical_loss": 3.3952033277249565,
      "tokens_seen": 2266759168
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031852525252525256,
      "loss": 1.145,
      "theoretical_loss": 3.395195343445277,
      "tokens_seen": 2266824704
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003185050505050505,
      "loss": 1.5126,
      "theoretical_loss": 3.3951873594610578,
      "tokens_seen": 2266890240
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031848484848484847,
      "loss": 1.534,
      "theoretical_loss": 3.3951793757722797,
      "tokens_seen": 2266955776
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003184646464646465,
      "loss": 1.2158,
      "theoretical_loss": 3.395171392378924,
      "tokens_seen": 2267021312
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031844444444444444,
      "loss": 1.3969,
      "theoretical_loss": 3.39516340928097,
      "tokens_seen": 2267086848
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031842424242424245,
      "loss": 1.5186,
      "theoretical_loss": 3.395155426478399,
      "tokens_seen": 2267152384
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003184040404040404,
      "loss": 1.4782,
      "theoretical_loss": 3.395147443971191,
      "tokens_seen": 2267217920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031838383838383836,
      "loss": 1.2187,
      "theoretical_loss": 3.395139461759327,
      "tokens_seen": 2267283456
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003183636363636364,
      "loss": 1.4741,
      "theoretical_loss": 3.395131479842788,
      "tokens_seen": 2267348992
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031834343434343433,
      "loss": 1.5372,
      "theoretical_loss": 3.3951234982215532,
      "tokens_seen": 2267414528
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031832323232323234,
      "loss": 1.2835,
      "theoretical_loss": 3.395115516895604,
      "tokens_seen": 2267480064
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003183030303030303,
      "loss": 1.4496,
      "theoretical_loss": 3.3951075358649208,
      "tokens_seen": 2267545600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182828282828283,
      "loss": 1.571,
      "theoretical_loss": 3.3950995551294842,
      "tokens_seen": 2267611136
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031826262626262627,
      "loss": 1.3076,
      "theoretical_loss": 3.3950915746892747,
      "tokens_seen": 2267676672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182424242424242,
      "loss": 1.3446,
      "theoretical_loss": 3.395083594544273,
      "tokens_seen": 2267742208
    },
    {
      "epoch": 0.37,
      "objective/train/aligned_token_count": 0.3189149498939514,
      "objective/train/avg_token_score": 0.01421320904046297,
      "objective/train/docs_used": 1284224,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1245386600494385,
      "objective/train/original_loss": 4.361290454864502,
      "objective/train/theoretical_loss": 3.395075614694459,
      "objective/train/tokens_used": 626666976,
      "objective/train/unlikelihood_loss": 0.06221999600529671,
      "theoretical_loss": 3.395075614694459,
      "tokens_seen": 2267807744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031822222222222224,
      "loss": 1.2747,
      "theoretical_loss": 3.395075614694459,
      "tokens_seen": 2267807744
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003182020202020202,
      "loss": 1.3077,
      "theoretical_loss": 3.395067635139814,
      "tokens_seen": 2267873280
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003181818181818182,
      "loss": 1.372,
      "theoretical_loss": 3.3950596558803183,
      "tokens_seen": 2267938816
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031816161616161616,
      "loss": 1.3845,
      "theoretical_loss": 3.3950516769159527,
      "tokens_seen": 2268004352
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003181414141414141,
      "loss": 1.3713,
      "theoretical_loss": 3.395043698246697,
      "tokens_seen": 2268069888
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031812121212121213,
      "loss": 1.4241,
      "theoretical_loss": 3.3950357198725327,
      "tokens_seen": 2268135424
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031810101010101014,
      "loss": 1.3318,
      "theoretical_loss": 3.39502774179344,
      "tokens_seen": 2268200960
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003180808080808081,
      "loss": 1.5411,
      "theoretical_loss": 3.395019764009399,
      "tokens_seen": 2268266496
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031806060606060605,
      "loss": 1.405,
      "theoretical_loss": 3.3950117865203904,
      "tokens_seen": 2268332032
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031804040404040406,
      "loss": 1.2797,
      "theoretical_loss": 3.3950038093263952,
      "tokens_seen": 2268397568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000318020202020202,
      "loss": 1.4282,
      "theoretical_loss": 3.3949958324273943,
      "tokens_seen": 2268463104
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031800000000000003,
      "loss": 1.382,
      "theoretical_loss": 3.3949878558233673,
      "tokens_seen": 2268528640
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317979797979798,
      "loss": 1.2363,
      "theoretical_loss": 3.3949798795142954,
      "tokens_seen": 2268594176
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031795959595959595,
      "loss": 1.3318,
      "theoretical_loss": 3.3949719035001586,
      "tokens_seen": 2268659712
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031793939393939396,
      "loss": 1.4382,
      "theoretical_loss": 3.394963927780938,
      "tokens_seen": 2268725248
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031791919191919197,
      "loss": 1.3401,
      "theoretical_loss": 3.3949559523566144,
      "tokens_seen": 2268790784
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003178989898989899,
      "loss": 1.294,
      "theoretical_loss": 3.394947977227168,
      "tokens_seen": 2268856320
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003178787878787879,
      "loss": 1.3724,
      "theoretical_loss": 3.3949400023925786,
      "tokens_seen": 2268921856
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031785858585858584,
      "loss": 1.3862,
      "theoretical_loss": 3.3949320278528283,
      "tokens_seen": 2268987392
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031783838383838385,
      "loss": 1.3396,
      "theoretical_loss": 3.394924053607897,
      "tokens_seen": 2269052928
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031781818181818186,
      "loss": 1.5495,
      "theoretical_loss": 3.394916079657765,
      "tokens_seen": 2269118464
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003177979797979798,
      "loss": 1.4592,
      "theoretical_loss": 3.394908106002413,
      "tokens_seen": 2269184000
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003177777777777778,
      "loss": 1.3121,
      "theoretical_loss": 3.394900132641822,
      "tokens_seen": 2269249536
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031775757575757573,
      "loss": 1.3915,
      "theoretical_loss": 3.3948921595759725,
      "tokens_seen": 2269315072
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003177373737373738,
      "loss": 1.1071,
      "theoretical_loss": 3.3948841868048447,
      "tokens_seen": 2269380608
    },
    {
      "epoch": 0.37,
      "objective/train/aligned_token_count": 0.4431818127632141,
      "objective/train/avg_token_score": 0.015034126117825508,
      "objective/train/docs_used": 1285503,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5824625492095947,
      "objective/train/original_loss": 4.151632308959961,
      "objective/train/theoretical_loss": 3.3948762143284195,
      "objective/train/tokens_used": 628305376,
      "objective/train/unlikelihood_loss": 0.25489455461502075,
      "theoretical_loss": 3.3948762143284195,
      "tokens_seen": 2269446144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031771717171717175,
      "loss": 1.6489,
      "theoretical_loss": 3.3948762143284195,
      "tokens_seen": 2269446144
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003176969696969697,
      "loss": 1.5055,
      "theoretical_loss": 3.394868242146677,
      "tokens_seen": 2269511680
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031767676767676767,
      "loss": 1.4638,
      "theoretical_loss": 3.3948602702595987,
      "tokens_seen": 2269577216
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003176565656565656,
      "loss": 1.5088,
      "theoretical_loss": 3.3948522986671645,
      "tokens_seen": 2269642752
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003176363636363637,
      "loss": 1.6326,
      "theoretical_loss": 3.394844327369355,
      "tokens_seen": 2269708288
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031761616161616165,
      "loss": 1.2584,
      "theoretical_loss": 3.3948363563661514,
      "tokens_seen": 2269773824
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003175959595959596,
      "loss": 1.2405,
      "theoretical_loss": 3.394828385657534,
      "tokens_seen": 2269839360
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031757575757575756,
      "loss": 1.5263,
      "theoretical_loss": 3.394820415243483,
      "tokens_seen": 2269904896
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003175555555555555,
      "loss": 1.2463,
      "theoretical_loss": 3.394812445123979,
      "tokens_seen": 2269970432
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003175353535353536,
      "loss": 1.2457,
      "theoretical_loss": 3.3948044752990034,
      "tokens_seen": 2270035968
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031751515151515154,
      "loss": 1.1539,
      "theoretical_loss": 3.3947965057685363,
      "tokens_seen": 2270101504
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003174949494949495,
      "loss": 1.4541,
      "theoretical_loss": 3.3947885365325585,
      "tokens_seen": 2270167040
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031747474747474745,
      "loss": 1.3777,
      "theoretical_loss": 3.3947805675910505,
      "tokens_seen": 2270232576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031745454545454546,
      "loss": 1.3803,
      "theoretical_loss": 3.3947725989439927,
      "tokens_seen": 2270298112
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003174343434343435,
      "loss": 1.3213,
      "theoretical_loss": 3.3947646305913657,
      "tokens_seen": 2270363648
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031741414141414143,
      "loss": 1.3852,
      "theoretical_loss": 3.394756662533151,
      "tokens_seen": 2270429184
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003173939393939394,
      "loss": 1.561,
      "theoretical_loss": 3.394748694769328,
      "tokens_seen": 2270494720
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031737373737373734,
      "loss": 1.284,
      "theoretical_loss": 3.394740727299878,
      "tokens_seen": 2270560256
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031735353535353536,
      "loss": 1.3187,
      "theoretical_loss": 3.394732760124782,
      "tokens_seen": 2270625792
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031733333333333337,
      "loss": 1.3845,
      "theoretical_loss": 3.39472479324402,
      "tokens_seen": 2270691328
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003173131313131313,
      "loss": 1.2897,
      "theoretical_loss": 3.3947168266575725,
      "tokens_seen": 2270756864
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003172929292929293,
      "loss": 1.7036,
      "theoretical_loss": 3.3947088603654203,
      "tokens_seen": 2270822400
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003172727272727273,
      "loss": 1.4416,
      "theoretical_loss": 3.3947008943675447,
      "tokens_seen": 2270887936
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031725252525252525,
      "loss": 1.2755,
      "theoretical_loss": 3.3946929286639254,
      "tokens_seen": 2270953472
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031723232323232326,
      "loss": 1.1962,
      "theoretical_loss": 3.3946849632545435,
      "tokens_seen": 2271019008
    },
    {
      "epoch": 0.37,
      "objective/train/aligned_token_count": 0.4025537669658661,
      "objective/train/avg_token_score": 0.04109661653637886,
      "objective/train/docs_used": 1286076,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.316990613937378,
      "objective/train/original_loss": 4.398165702819824,
      "objective/train/theoretical_loss": 3.3946769981393796,
      "objective/train/tokens_used": 629943776,
      "objective/train/unlikelihood_loss": 0.0520949587225914,
      "theoretical_loss": 3.3946769981393796,
      "tokens_seen": 2271084544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003172121212121212,
      "loss": 1.4544,
      "theoretical_loss": 3.3946769981393796,
      "tokens_seen": 2271084544
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003171919191919192,
      "loss": 1.4455,
      "theoretical_loss": 3.394669033318414,
      "tokens_seen": 2271150080
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003171717171717172,
      "loss": 1.6033,
      "theoretical_loss": 3.3946610687916285,
      "tokens_seen": 2271215616
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031715151515151514,
      "loss": 1.394,
      "theoretical_loss": 3.3946531045590023,
      "tokens_seen": 2271281152
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031713131313131315,
      "loss": 1.5442,
      "theoretical_loss": 3.3946451406205167,
      "tokens_seen": 2271346688
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003171111111111111,
      "loss": 1.4789,
      "theoretical_loss": 3.3946371769761527,
      "tokens_seen": 2271412224
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170909090909091,
      "loss": 1.397,
      "theoretical_loss": 3.3946292136258904,
      "tokens_seen": 2271477760
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003170707070707071,
      "loss": 1.2331,
      "theoretical_loss": 3.3946212505697106,
      "tokens_seen": 2271543296
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031705050505050503,
      "loss": 1.1416,
      "theoretical_loss": 3.3946132878075934,
      "tokens_seen": 2271608832
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031703030303030305,
      "loss": 1.3627,
      "theoretical_loss": 3.394605325339521,
      "tokens_seen": 2271674368
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000317010101010101,
      "loss": 1.3751,
      "theoretical_loss": 3.3945973631654724,
      "tokens_seen": 2271739904
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316989898989899,
      "loss": 1.4025,
      "theoretical_loss": 3.394589401285429,
      "tokens_seen": 2271805440
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031696969696969697,
      "loss": 1.3738,
      "theoretical_loss": 3.3945814396993716,
      "tokens_seen": 2271870976
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003169494949494949,
      "loss": 1.3204,
      "theoretical_loss": 3.394573478407281,
      "tokens_seen": 2271936512
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031692929292929294,
      "loss": 1.5831,
      "theoretical_loss": 3.394565517409137,
      "tokens_seen": 2272002048
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031690909090909095,
      "loss": 1.3489,
      "theoretical_loss": 3.394557556704921,
      "tokens_seen": 2272067584
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003168888888888889,
      "loss": 1.3825,
      "theoretical_loss": 3.3945495962946137,
      "tokens_seen": 2272133120
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031686868686868686,
      "loss": 1.0641,
      "theoretical_loss": 3.3945416361781953,
      "tokens_seen": 2272198656
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003168484848484848,
      "loss": 1.3117,
      "theoretical_loss": 3.394533676355647,
      "tokens_seen": 2272264192
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031682828282828283,
      "loss": 1.4181,
      "theoretical_loss": 3.3945257168269487,
      "tokens_seen": 2272329728
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031680808080808084,
      "loss": 1.26,
      "theoretical_loss": 3.394517757592082,
      "tokens_seen": 2272395264
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003167878787878788,
      "loss": 1.3754,
      "theoretical_loss": 3.3945097986510273,
      "tokens_seen": 2272460800
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031676767676767676,
      "loss": 1.2889,
      "theoretical_loss": 3.3945018400037648,
      "tokens_seen": 2272526336
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003167474747474747,
      "loss": 1.2485,
      "theoretical_loss": 3.3944938816502757,
      "tokens_seen": 2272591872
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003167272727272728,
      "loss": 1.4146,
      "theoretical_loss": 3.3944859235905405,
      "tokens_seen": 2272657408
    },
    {
      "epoch": 0.37,
      "objective/train/aligned_token_count": 0.4742179811000824,
      "objective/train/avg_token_score": 0.02209853194653988,
      "objective/train/docs_used": 1287335,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5478516817092896,
      "objective/train/original_loss": 4.039405822753906,
      "objective/train/theoretical_loss": 3.3944779658245396,
      "objective/train/tokens_used": 631582176,
      "objective/train/unlikelihood_loss": 0.05548189580440521,
      "theoretical_loss": 3.3944779658245396,
      "tokens_seen": 2272722944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031670707070707073,
      "loss": 1.4016,
      "theoretical_loss": 3.3944779658245396,
      "tokens_seen": 2272722944
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003166868686868687,
      "loss": 1.3712,
      "theoretical_loss": 3.394470008352254,
      "tokens_seen": 2272788480
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031666666666666665,
      "loss": 1.5329,
      "theoretical_loss": 3.394462051173665,
      "tokens_seen": 2272854016
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003166464646464646,
      "loss": 1.4236,
      "theoretical_loss": 3.3944540942887524,
      "tokens_seen": 2272919552
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031662626262626267,
      "loss": 1.2527,
      "theoretical_loss": 3.394446137697497,
      "tokens_seen": 2272985088
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003166060606060606,
      "loss": 1.2481,
      "theoretical_loss": 3.3944381813998796,
      "tokens_seen": 2273050624
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165858585858586,
      "loss": 1.1891,
      "theoretical_loss": 3.3944302253958814,
      "tokens_seen": 2273116160
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031656565656565654,
      "loss": 1.3889,
      "theoretical_loss": 3.394422269685482,
      "tokens_seen": 2273181696
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031654545454545455,
      "loss": 1.5707,
      "theoretical_loss": 3.394414314268663,
      "tokens_seen": 2273247232
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031652525252525256,
      "loss": 1.3709,
      "theoretical_loss": 3.394406359145405,
      "tokens_seen": 2273312768
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003165050505050505,
      "loss": 1.1651,
      "theoretical_loss": 3.3943984043156887,
      "tokens_seen": 2273378304
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003164848484848485,
      "loss": 1.2353,
      "theoretical_loss": 3.3943904497794946,
      "tokens_seen": 2273443840
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031646464646464643,
      "loss": 1.2444,
      "theoretical_loss": 3.3943824955368034,
      "tokens_seen": 2273509376
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031644444444444444,
      "loss": 1.6812,
      "theoretical_loss": 3.3943745415875957,
      "tokens_seen": 2273574912
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031642424242424246,
      "loss": 1.1759,
      "theoretical_loss": 3.3943665879318528,
      "tokens_seen": 2273640448
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003164040404040404,
      "loss": 1.4825,
      "theoretical_loss": 3.3943586345695547,
      "tokens_seen": 2273705984
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031638383838383837,
      "loss": 1.2999,
      "theoretical_loss": 3.3943506815006823,
      "tokens_seen": 2273771520
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003163636363636364,
      "loss": 1.4277,
      "theoretical_loss": 3.394342728725217,
      "tokens_seen": 2273837056
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003163434343434344,
      "loss": 1.3259,
      "theoretical_loss": 3.3943347762431384,
      "tokens_seen": 2273902592
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031632323232323235,
      "loss": 1.2956,
      "theoretical_loss": 3.394326824054428,
      "tokens_seen": 2273968128
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003163030303030303,
      "loss": 1.3248,
      "theoretical_loss": 3.3943188721590665,
      "tokens_seen": 2274033664
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031628282828282826,
      "loss": 1.4917,
      "theoretical_loss": 3.394310920557034,
      "tokens_seen": 2274099200
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003162626262626263,
      "loss": 1.3093,
      "theoretical_loss": 3.3943029692483124,
      "tokens_seen": 2274164736
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003162424242424243,
      "loss": 1.2935,
      "theoretical_loss": 3.394295018232881,
      "tokens_seen": 2274230272
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031622222222222224,
      "loss": 1.3317,
      "theoretical_loss": 3.3942870675107217,
      "tokens_seen": 2274295808
    },
    {
      "epoch": 0.37,
      "objective/train/aligned_token_count": 0.42173752188682556,
      "objective/train/avg_token_score": 0.009197893552482128,
      "objective/train/docs_used": 1288139,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2617578506469727,
      "objective/train/original_loss": 4.145997524261475,
      "objective/train/theoretical_loss": 3.3942791170818145,
      "objective/train/tokens_used": 633220576,
      "objective/train/unlikelihood_loss": 0.05465617775917053,
      "theoretical_loss": 3.3942791170818145,
      "tokens_seen": 2274361344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003162020202020202,
      "loss": 1.5026,
      "theoretical_loss": 3.3942791170818145,
      "tokens_seen": 2274361344
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031618181818181815,
      "loss": 1.3074,
      "theoretical_loss": 3.3942711669461407,
      "tokens_seen": 2274426880
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031616161616161617,
      "loss": 1.5164,
      "theoretical_loss": 3.3942632171036804,
      "tokens_seen": 2274492416
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003161414141414142,
      "loss": 1.4911,
      "theoretical_loss": 3.394255267554415,
      "tokens_seen": 2274557952
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031612121212121213,
      "loss": 1.3888,
      "theoretical_loss": 3.3942473182983246,
      "tokens_seen": 2274623488
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003161010101010101,
      "loss": 1.445,
      "theoretical_loss": 3.3942393693353905,
      "tokens_seen": 2274689024
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003160808080808081,
      "loss": 1.4263,
      "theoretical_loss": 3.3942314206655926,
      "tokens_seen": 2274754560
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031606060606060606,
      "loss": 1.1743,
      "theoretical_loss": 3.3942234722889126,
      "tokens_seen": 2274820096
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031604040404040407,
      "loss": 1.2941,
      "theoretical_loss": 3.394215524205331,
      "tokens_seen": 2274885632
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316020202020202,
      "loss": 1.4709,
      "theoretical_loss": 3.3942075764148285,
      "tokens_seen": 2274951168
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000316,
      "loss": 1.4272,
      "theoretical_loss": 3.3941996289173857,
      "tokens_seen": 2275016704
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.000315979797979798,
      "loss": 1.4151,
      "theoretical_loss": 3.3941916817129836,
      "tokens_seen": 2275082240
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031595959595959595,
      "loss": 1.456,
      "theoretical_loss": 3.3941837348016026,
      "tokens_seen": 2275147776
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031593939393939396,
      "loss": 1.2801,
      "theoretical_loss": 3.3941757881832233,
      "tokens_seen": 2275213312
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003159191919191919,
      "loss": 1.5407,
      "theoretical_loss": 3.3941678418578274,
      "tokens_seen": 2275278848
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031589898989898993,
      "loss": 1.5454,
      "theoretical_loss": 3.394159895825395,
      "tokens_seen": 2275344384
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003158787878787879,
      "loss": 1.3868,
      "theoretical_loss": 3.3941519500859063,
      "tokens_seen": 2275409920
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031585858585858584,
      "loss": 1.3965,
      "theoretical_loss": 3.3941440046393434,
      "tokens_seen": 2275475456
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031583838383838385,
      "loss": 1.3458,
      "theoretical_loss": 3.394136059485686,
      "tokens_seen": 2275540992
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003158181818181818,
      "loss": 1.0141,
      "theoretical_loss": 3.394128114624915,
      "tokens_seen": 2275606528
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003157979797979798,
      "loss": 1.3246,
      "theoretical_loss": 3.394120170057012,
      "tokens_seen": 2275672064
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0003157777777777778,
      "loss": 1.3736,
      "theoretical_loss": 3.394112225781957,
      "tokens_seen": 2275737600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031575757575757574,
      "loss": 1.5331,
      "theoretical_loss": 3.3941042817997307,
      "tokens_seen": 2275803136
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031573737373737375,
      "loss": 1.4423,
      "theoretical_loss": 3.394096338110314,
      "tokens_seen": 2275868672
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00031571717171717176,
      "loss": 1.5677,
      "theoretical_loss": 3.394088394713688,
      "tokens_seen": 2275934208
    },
    {
      "epoch": 0.37,
      "objective/train/aligned_token_count": 0.3288734257221222,
      "objective/train/avg_token_score": 0.016442084684967995,
      "objective/train/docs_used": 1289239,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0548652410507202,
      "objective/train/original_loss": 4.7988996505737305,
      "objective/train/theoretical_loss": 3.3940804516098333,
      "objective/train/tokens_used": 634858976,
      "objective/train/unlikelihood_loss": 0.04558462277054787,
      "theoretical_loss": 3.3940804516098333,
      "tokens_seen": 2275999744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003156969696969697,
      "loss": 1.1909,
      "theoretical_loss": 3.3940804516098333,
      "tokens_seen": 2275999744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031567676767676767,
      "loss": 1.2658,
      "theoretical_loss": 3.3940725087987307,
      "tokens_seen": 2276065280
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031565656565656563,
      "loss": 1.379,
      "theoretical_loss": 3.3940645662803606,
      "tokens_seen": 2276130816
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031563636363636364,
      "loss": 1.5352,
      "theoretical_loss": 3.3940566240547043,
      "tokens_seen": 2276196352
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031561616161616165,
      "loss": 1.3373,
      "theoretical_loss": 3.3940486821217424,
      "tokens_seen": 2276261888
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155959595959596,
      "loss": 1.2692,
      "theoretical_loss": 3.3940407404814557,
      "tokens_seen": 2276327424
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031557575757575757,
      "loss": 1.3048,
      "theoretical_loss": 3.394032799133825,
      "tokens_seen": 2276392960
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155555555555555,
      "loss": 1.4148,
      "theoretical_loss": 3.394024858078831,
      "tokens_seen": 2276458496
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003155353535353536,
      "loss": 1.4721,
      "theoretical_loss": 3.3940169173164545,
      "tokens_seen": 2276524032
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031551515151515154,
      "loss": 1.3577,
      "theoretical_loss": 3.394008976846676,
      "tokens_seen": 2276589568
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154949494949495,
      "loss": 1.4034,
      "theoretical_loss": 3.394001036669477,
      "tokens_seen": 2276655104
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031547474747474746,
      "loss": 1.2548,
      "theoretical_loss": 3.393993096784838,
      "tokens_seen": 2276720640
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154545454545454,
      "loss": 1.5906,
      "theoretical_loss": 3.3939851571927395,
      "tokens_seen": 2276786176
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003154343434343435,
      "loss": 1.1606,
      "theoretical_loss": 3.393977217893163,
      "tokens_seen": 2276851712
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031541414141414144,
      "loss": 1.5313,
      "theoretical_loss": 3.3939692788860887,
      "tokens_seen": 2276917248
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003153939393939394,
      "loss": 1.2417,
      "theoretical_loss": 3.393961340171497,
      "tokens_seen": 2276982784
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031537373737373735,
      "loss": 1.241,
      "theoretical_loss": 3.39395340174937,
      "tokens_seen": 2277048320
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031535353535353536,
      "loss": 1.4482,
      "theoretical_loss": 3.393945463619687,
      "tokens_seen": 2277113856
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031533333333333337,
      "loss": 1.3599,
      "theoretical_loss": 3.3939375257824302,
      "tokens_seen": 2277179392
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031531313131313133,
      "loss": 1.446,
      "theoretical_loss": 3.3939295882375795,
      "tokens_seen": 2277244928
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003152929292929293,
      "loss": 1.4587,
      "theoretical_loss": 3.3939216509851162,
      "tokens_seen": 2277310464
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031527272727272724,
      "loss": 1.4674,
      "theoretical_loss": 3.393913714025021,
      "tokens_seen": 2277376000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031525252525252525,
      "loss": 1.4673,
      "theoretical_loss": 3.393905777357274,
      "tokens_seen": 2277441536
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031523232323232327,
      "loss": 1.3774,
      "theoretical_loss": 3.393897840981857,
      "tokens_seen": 2277507072
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003152121212121212,
      "loss": 1.3097,
      "theoretical_loss": 3.3938899048987508,
      "tokens_seen": 2277572608
    },
    {
      "epoch": 0.38,
      "objective/train/aligned_token_count": 0.3794599175453186,
      "objective/train/avg_token_score": 0.012644610367715359,
      "objective/train/docs_used": 1289790,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3388731479644775,
      "objective/train/original_loss": 4.512577056884766,
      "objective/train/theoretical_loss": 3.3938819691079356,
      "objective/train/tokens_used": 636497376,
      "objective/train/unlikelihood_loss": 0.06975681334733963,
      "theoretical_loss": 3.3938819691079356,
      "tokens_seen": 2277638144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003151919191919192,
      "loss": 1.1216,
      "theoretical_loss": 3.3938819691079356,
      "tokens_seen": 2277638144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003151717171717172,
      "loss": 1.1285,
      "theoretical_loss": 3.393874033609392,
      "tokens_seen": 2277703680
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031515151515151515,
      "loss": 1.3093,
      "theoretical_loss": 3.393866098403102,
      "tokens_seen": 2277769216
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031513131313131316,
      "loss": 1.2332,
      "theoretical_loss": 3.3938581634890457,
      "tokens_seen": 2277834752
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003151111111111111,
      "loss": 1.4241,
      "theoretical_loss": 3.393850228867204,
      "tokens_seen": 2277900288
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031509090909090907,
      "loss": 1.2733,
      "theoretical_loss": 3.3938422945375577,
      "tokens_seen": 2277965824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003150707070707071,
      "loss": 1.2691,
      "theoretical_loss": 3.3938343605000876,
      "tokens_seen": 2278031360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031505050505050504,
      "loss": 1.5493,
      "theoretical_loss": 3.3938264267547744,
      "tokens_seen": 2278096896
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031503030303030305,
      "loss": 1.4782,
      "theoretical_loss": 3.3938184933015996,
      "tokens_seen": 2278162432
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000315010101010101,
      "loss": 1.4181,
      "theoretical_loss": 3.3938105601405435,
      "tokens_seen": 2278227968
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000314989898989899,
      "loss": 1.3255,
      "theoretical_loss": 3.3938026272715867,
      "tokens_seen": 2278293504
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000314969696969697,
      "loss": 1.3174,
      "theoretical_loss": 3.3937946946947104,
      "tokens_seen": 2278359040
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031494949494949493,
      "loss": 1.2296,
      "theoretical_loss": 3.393786762409896,
      "tokens_seen": 2278424576
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031492929292929294,
      "loss": 1.5981,
      "theoretical_loss": 3.3937788304171232,
      "tokens_seen": 2278490112
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003149090909090909,
      "loss": 1.2728,
      "theoretical_loss": 3.3937708987163733,
      "tokens_seen": 2278555648
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003148888888888889,
      "loss": 1.3194,
      "theoretical_loss": 3.393762967307628,
      "tokens_seen": 2278621184
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031486868686868687,
      "loss": 1.2689,
      "theoretical_loss": 3.393755036190867,
      "tokens_seen": 2278686720
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003148484848484848,
      "loss": 1.2996,
      "theoretical_loss": 3.393747105366071,
      "tokens_seen": 2278752256
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031482828282828284,
      "loss": 1.0475,
      "theoretical_loss": 3.393739174833222,
      "tokens_seen": 2278817792
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031480808080808085,
      "loss": 1.3342,
      "theoretical_loss": 3.3937312445923,
      "tokens_seen": 2278883328
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003147878787878788,
      "loss": 1.218,
      "theoretical_loss": 3.393723314643286,
      "tokens_seen": 2278948864
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031476767676767676,
      "loss": 1.3452,
      "theoretical_loss": 3.393715384986162,
      "tokens_seen": 2279014400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003147474747474747,
      "loss": 1.2913,
      "theoretical_loss": 3.393707455620907,
      "tokens_seen": 2279079936
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031472727272727273,
      "loss": 1.2618,
      "theoretical_loss": 3.3936995265475027,
      "tokens_seen": 2279145472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031470707070707074,
      "loss": 1.4637,
      "theoretical_loss": 3.3936915977659305,
      "tokens_seen": 2279211008
    },
    {
      "epoch": 0.38,
      "objective/train/aligned_token_count": 0.37243402004241943,
      "objective/train/avg_token_score": 0.01346712838858366,
      "objective/train/docs_used": 1290983,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3259166479110718,
      "objective/train/original_loss": 4.845729827880859,
      "objective/train/theoretical_loss": 3.39368366927617,
      "objective/train/tokens_used": 638135776,
      "objective/train/unlikelihood_loss": 0.042649686336517334,
      "theoretical_loss": 3.39368366927617,
      "tokens_seen": 2279276544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003146868686868687,
      "loss": 1.5732,
      "theoretical_loss": 3.39368366927617,
      "tokens_seen": 2279276544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031466666666666665,
      "loss": 1.5071,
      "theoretical_loss": 3.3936757410782032,
      "tokens_seen": 2279342080
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031464646464646466,
      "loss": 1.6084,
      "theoretical_loss": 3.393667813172011,
      "tokens_seen": 2279407616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003146262626262627,
      "loss": 1.2402,
      "theoretical_loss": 3.3936598855575735,
      "tokens_seen": 2279473152
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031460606060606063,
      "loss": 1.4012,
      "theoretical_loss": 3.393651958234872,
      "tokens_seen": 2279538688
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003145858585858586,
      "loss": 1.3239,
      "theoretical_loss": 3.393644031203887,
      "tokens_seen": 2279604224
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031456565656565655,
      "loss": 1.349,
      "theoretical_loss": 3.3936361044646,
      "tokens_seen": 2279669760
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031454545454545456,
      "loss": 1.2972,
      "theoretical_loss": 3.393628178016992,
      "tokens_seen": 2279735296
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031452525252525257,
      "loss": 1.6934,
      "theoretical_loss": 3.3936202518610425,
      "tokens_seen": 2279800832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003145050505050505,
      "loss": 1.2754,
      "theoretical_loss": 3.393612325996734,
      "tokens_seen": 2279866368
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003144848484848485,
      "loss": 1.2905,
      "theoretical_loss": 3.3936044004240467,
      "tokens_seen": 2279931904
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031446464646464644,
      "loss": 1.5761,
      "theoretical_loss": 3.3935964751429615,
      "tokens_seen": 2279997440
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003144444444444445,
      "loss": 1.3882,
      "theoretical_loss": 3.393588550153459,
      "tokens_seen": 2280062976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031442424242424246,
      "loss": 1.3124,
      "theoretical_loss": 3.393580625455521,
      "tokens_seen": 2280128512
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003144040404040404,
      "loss": 1.1929,
      "theoretical_loss": 3.3935727010491274,
      "tokens_seen": 2280194048
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003143838383838384,
      "loss": 1.5104,
      "theoretical_loss": 3.3935647769342596,
      "tokens_seen": 2280259584
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031436363636363633,
      "loss": 1.2717,
      "theoretical_loss": 3.3935568531108986,
      "tokens_seen": 2280325120
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003143434343434344,
      "loss": 1.208,
      "theoretical_loss": 3.393548929579025,
      "tokens_seen": 2280390656
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031432323232323235,
      "loss": 1.381,
      "theoretical_loss": 3.3935410063386193,
      "tokens_seen": 2280456192
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003143030303030303,
      "loss": 1.3861,
      "theoretical_loss": 3.3935330833896638,
      "tokens_seen": 2280521728
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031428282828282827,
      "loss": 1.2196,
      "theoretical_loss": 3.3935251607321377,
      "tokens_seen": 2280587264
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003142626262626262,
      "loss": 1.419,
      "theoretical_loss": 3.393517238366023,
      "tokens_seen": 2280652800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003142424242424243,
      "loss": 1.1766,
      "theoretical_loss": 3.3935093162913,
      "tokens_seen": 2280718336
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031422222222222225,
      "loss": 1.4375,
      "theoretical_loss": 3.3935013945079504,
      "tokens_seen": 2280783872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003142020202020202,
      "loss": 1.5103,
      "theoretical_loss": 3.3934934730159547,
      "tokens_seen": 2280849408
    },
    {
      "epoch": 0.38,
      "objective/train/aligned_token_count": 0.3643084168434143,
      "objective/train/avg_token_score": 0.0073792519979178905,
      "objective/train/docs_used": 1291619,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1794244050979614,
      "objective/train/original_loss": 4.9936370849609375,
      "objective/train/theoretical_loss": 3.3934855518152935,
      "objective/train/tokens_used": 639774176,
      "objective/train/unlikelihood_loss": 0.04381928592920303,
      "theoretical_loss": 3.3934855518152935,
      "tokens_seen": 2280914944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031418181818181816,
      "loss": 1.302,
      "theoretical_loss": 3.3934855518152935,
      "tokens_seen": 2280914944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031416161616161617,
      "loss": 1.3466,
      "theoretical_loss": 3.393477630905948,
      "tokens_seen": 2280980480
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141414141414142,
      "loss": 1.3189,
      "theoretical_loss": 3.393469710287899,
      "tokens_seen": 2281046016
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031412121212121214,
      "loss": 1.2833,
      "theoretical_loss": 3.3934617899611275,
      "tokens_seen": 2281111552
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003141010101010101,
      "loss": 1.4686,
      "theoretical_loss": 3.3934538699256143,
      "tokens_seen": 2281177088
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031408080808080805,
      "loss": 1.1074,
      "theoretical_loss": 3.393445950181341,
      "tokens_seen": 2281242624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031406060606060606,
      "loss": 1.4979,
      "theoretical_loss": 3.393438030728287,
      "tokens_seen": 2281308160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003140404040404041,
      "loss": 1.4589,
      "theoretical_loss": 3.393430111566435,
      "tokens_seen": 2281373696
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031402020202020203,
      "loss": 1.5319,
      "theoretical_loss": 3.3934221926957653,
      "tokens_seen": 2281439232
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000314,
      "loss": 1.4017,
      "theoretical_loss": 3.393414274116258,
      "tokens_seen": 2281504768
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000313979797979798,
      "loss": 1.5045,
      "theoretical_loss": 3.393406355827895,
      "tokens_seen": 2281570304
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031395959595959596,
      "loss": 1.4332,
      "theoretical_loss": 3.393398437830657,
      "tokens_seen": 2281635840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031393939393939397,
      "loss": 1.3022,
      "theoretical_loss": 3.3933905201245245,
      "tokens_seen": 2281701376
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003139191919191919,
      "loss": 1.3412,
      "theoretical_loss": 3.393382602709479,
      "tokens_seen": 2281766912
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138989898989899,
      "loss": 1.3812,
      "theoretical_loss": 3.3933746855855014,
      "tokens_seen": 2281832448
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138787878787879,
      "loss": 1.3659,
      "theoretical_loss": 3.393366768752572,
      "tokens_seen": 2281897984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031385858585858585,
      "loss": 1.2596,
      "theoretical_loss": 3.393358852210673,
      "tokens_seen": 2281963520
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031383838383838386,
      "loss": 1.5235,
      "theoretical_loss": 3.3933509359597838,
      "tokens_seen": 2282029056
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003138181818181818,
      "loss": 1.2453,
      "theoretical_loss": 3.393343019999886,
      "tokens_seen": 2282094592
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031379797979797983,
      "loss": 1.3648,
      "theoretical_loss": 3.393335104330961,
      "tokens_seen": 2282160128
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003137777777777778,
      "loss": 1.4178,
      "theoretical_loss": 3.3933271889529895,
      "tokens_seen": 2282225664
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031375757575757574,
      "loss": 1.2549,
      "theoretical_loss": 3.393319273865952,
      "tokens_seen": 2282291200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031373737373737375,
      "loss": 1.505,
      "theoretical_loss": 3.3933113590698305,
      "tokens_seen": 2282356736
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003137171717171717,
      "loss": 1.4863,
      "theoretical_loss": 3.3933034445646046,
      "tokens_seen": 2282422272
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003136969696969697,
      "loss": 1.4561,
      "theoretical_loss": 3.393295530350256,
      "tokens_seen": 2282487808
    },
    {
      "epoch": 0.38,
      "objective/train/aligned_token_count": 0.45967742800712585,
      "objective/train/avg_token_score": 0.004498833324760199,
      "objective/train/docs_used": 1292123,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3870453834533691,
      "objective/train/original_loss": 4.213576316833496,
      "objective/train/theoretical_loss": 3.3932876164267656,
      "objective/train/tokens_used": 641412576,
      "objective/train/unlikelihood_loss": 0.060575220733881,
      "theoretical_loss": 3.3932876164267656,
      "tokens_seen": 2282553344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003136767676767677,
      "loss": 1.3214,
      "theoretical_loss": 3.3932876164267656,
      "tokens_seen": 2282553344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031365656565656563,
      "loss": 1.3172,
      "theoretical_loss": 3.3932797027941146,
      "tokens_seen": 2282618880
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031363636363636365,
      "loss": 1.2125,
      "theoretical_loss": 3.3932717894522835,
      "tokens_seen": 2282684416
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031361616161616166,
      "loss": 1.2945,
      "theoretical_loss": 3.393263876401253,
      "tokens_seen": 2282749952
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003135959595959596,
      "loss": 1.5167,
      "theoretical_loss": 3.393255963641005,
      "tokens_seen": 2282815488
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031357575757575757,
      "loss": 1.2527,
      "theoretical_loss": 3.39324805117152,
      "tokens_seen": 2282881024
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031355555555555553,
      "loss": 1.4873,
      "theoretical_loss": 3.393240138992779,
      "tokens_seen": 2282946560
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031353535353535354,
      "loss": 1.2162,
      "theoretical_loss": 3.393232227104763,
      "tokens_seen": 2283012096
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031351515151515155,
      "loss": 1.3422,
      "theoretical_loss": 3.393224315507452,
      "tokens_seen": 2283077632
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003134949494949495,
      "loss": 1.3181,
      "theoretical_loss": 3.393216404200829,
      "tokens_seen": 2283143168
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031347474747474746,
      "loss": 1.4229,
      "theoretical_loss": 3.3932084931848734,
      "tokens_seen": 2283208704
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003134545454545454,
      "loss": 1.1308,
      "theoretical_loss": 3.3932005824595666,
      "tokens_seen": 2283274240
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003134343434343435,
      "loss": 1.1585,
      "theoretical_loss": 3.3931926720248895,
      "tokens_seen": 2283339776
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031341414141414144,
      "loss": 1.3696,
      "theoretical_loss": 3.3931847618808235,
      "tokens_seen": 2283405312
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133939393939394,
      "loss": 1.2045,
      "theoretical_loss": 3.393176852027349,
      "tokens_seen": 2283470848
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031337373737373736,
      "loss": 1.513,
      "theoretical_loss": 3.393168942464447,
      "tokens_seen": 2283536384
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133535353535353,
      "loss": 1.3988,
      "theoretical_loss": 3.3931610331920994,
      "tokens_seen": 2283601920
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003133333333333334,
      "loss": 1.3704,
      "theoretical_loss": 3.3931531242102864,
      "tokens_seen": 2283667456
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031331313131313134,
      "loss": 1.4667,
      "theoretical_loss": 3.393145215518989,
      "tokens_seen": 2283732992
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003132929292929293,
      "loss": 1.4136,
      "theoretical_loss": 3.393137307118188,
      "tokens_seen": 2283798528
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031327272727272725,
      "loss": 1.4526,
      "theoretical_loss": 3.393129399007865,
      "tokens_seen": 2283864064
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031325252525252526,
      "loss": 1.3586,
      "theoretical_loss": 3.3931214911880008,
      "tokens_seen": 2283929600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031323232323232327,
      "loss": 1.4388,
      "theoretical_loss": 3.393113583658576,
      "tokens_seen": 2283995136
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031321212121212123,
      "loss": 1.173,
      "theoretical_loss": 3.393105676419572,
      "tokens_seen": 2284060672
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003131919191919192,
      "loss": 1.2863,
      "theoretical_loss": 3.3930977694709696,
      "tokens_seen": 2284126208
    },
    {
      "epoch": 0.38,
      "objective/train/aligned_token_count": 0.4897971749305725,
      "objective/train/avg_token_score": 0.005325449630618095,
      "objective/train/docs_used": 1293247,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.586478590965271,
      "objective/train/original_loss": 3.9623570442199707,
      "objective/train/theoretical_loss": 3.3930898628127504,
      "objective/train/tokens_used": 643050976,
      "objective/train/unlikelihood_loss": 0.06217525526881218,
      "theoretical_loss": 3.3930898628127504,
      "tokens_seen": 2284191744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031317171717171714,
      "loss": 1.4252,
      "theoretical_loss": 3.3930898628127504,
      "tokens_seen": 2284191744
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031315151515151515,
      "loss": 1.5599,
      "theoretical_loss": 3.3930819564448944,
      "tokens_seen": 2284257280
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031313131313131316,
      "loss": 1.3327,
      "theoretical_loss": 3.3930740503673835,
      "tokens_seen": 2284322816
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003131111111111111,
      "loss": 1.2814,
      "theoretical_loss": 3.3930661445801977,
      "tokens_seen": 2284388352
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003130909090909091,
      "loss": 1.3333,
      "theoretical_loss": 3.3930582390833193,
      "tokens_seen": 2284453888
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003130707070707071,
      "loss": 1.4219,
      "theoretical_loss": 3.3930503338767286,
      "tokens_seen": 2284519424
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031305050505050505,
      "loss": 1.3005,
      "theoretical_loss": 3.393042428960406,
      "tokens_seen": 2284584960
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031303030303030306,
      "loss": 1.3976,
      "theoretical_loss": 3.393034524334334,
      "tokens_seen": 2284650496
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000313010101010101,
      "loss": 1.4101,
      "theoretical_loss": 3.3930266199984924,
      "tokens_seen": 2284716032
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031298989898989897,
      "loss": 1.3199,
      "theoretical_loss": 3.3930187159528624,
      "tokens_seen": 2284781568
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000312969696969697,
      "loss": 1.2353,
      "theoretical_loss": 3.393010812197425,
      "tokens_seen": 2284847104
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031294949494949494,
      "loss": 1.2993,
      "theoretical_loss": 3.3930029087321625,
      "tokens_seen": 2284912640
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031292929292929295,
      "loss": 1.4537,
      "theoretical_loss": 3.3929950055570544,
      "tokens_seen": 2284978176
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003129090909090909,
      "loss": 1.4237,
      "theoretical_loss": 3.3929871026720817,
      "tokens_seen": 2285043712
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031288888888888886,
      "loss": 1.4025,
      "theoretical_loss": 3.3929792000772263,
      "tokens_seen": 2285109248
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128686868686869,
      "loss": 1.3459,
      "theoretical_loss": 3.392971297772469,
      "tokens_seen": 2285174784
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128484848484849,
      "loss": 1.269,
      "theoretical_loss": 3.3929633957577905,
      "tokens_seen": 2285240320
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031282828282828284,
      "loss": 1.4989,
      "theoretical_loss": 3.3929554940331723,
      "tokens_seen": 2285305856
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003128080808080808,
      "loss": 1.4134,
      "theoretical_loss": 3.392947592598595,
      "tokens_seen": 2285371392
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127878787878788,
      "loss": 1.4434,
      "theoretical_loss": 3.3929396914540395,
      "tokens_seen": 2285436928
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031276767676767677,
      "loss": 1.3411,
      "theoretical_loss": 3.3929317905994876,
      "tokens_seen": 2285502464
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127474747474748,
      "loss": 1.2009,
      "theoretical_loss": 3.3929238900349197,
      "tokens_seen": 2285568000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031272727272727273,
      "loss": 1.4602,
      "theoretical_loss": 3.3929159897603167,
      "tokens_seen": 2285633536
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003127070707070707,
      "loss": 1.1351,
      "theoretical_loss": 3.3929080897756605,
      "tokens_seen": 2285699072
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003126868686868687,
      "loss": 1.586,
      "theoretical_loss": 3.3929001900809315,
      "tokens_seen": 2285764608
    },
    {
      "epoch": 0.38,
      "objective/train/aligned_token_count": 0.3095674514770508,
      "objective/train/avg_token_score": 0.030379366129636765,
      "objective/train/docs_used": 1293760,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0177284479141235,
      "objective/train/original_loss": 4.727108001708984,
      "objective/train/theoretical_loss": 3.3928922906761105,
      "objective/train/tokens_used": 644689376,
      "objective/train/unlikelihood_loss": 0.07617577165365219,
      "theoretical_loss": 3.3928922906761105,
      "tokens_seen": 2285830144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031266666666666666,
      "loss": 1.2254,
      "theoretical_loss": 3.3928922906761105,
      "tokens_seen": 2285830144
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031264646464646467,
      "loss": 1.5294,
      "theoretical_loss": 3.392884391561179,
      "tokens_seen": 2285895680
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003126262626262626,
      "loss": 1.3094,
      "theoretical_loss": 3.3928764927361184,
      "tokens_seen": 2285961216
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031260606060606064,
      "loss": 1.1948,
      "theoretical_loss": 3.392868594200909,
      "tokens_seen": 2286026752
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003125858585858586,
      "loss": 1.3352,
      "theoretical_loss": 3.392860695955532,
      "tokens_seen": 2286092288
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031256565656565655,
      "loss": 1.4472,
      "theoretical_loss": 3.3928527979999688,
      "tokens_seen": 2286157824
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031254545454545456,
      "loss": 1.5826,
      "theoretical_loss": 3.3928449003342003,
      "tokens_seen": 2286223360
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003125252525252525,
      "loss": 1.2689,
      "theoretical_loss": 3.3928370029582076,
      "tokens_seen": 2286288896
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031250505050505053,
      "loss": 1.2539,
      "theoretical_loss": 3.3928291058719715,
      "tokens_seen": 2286354432
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003124848484848485,
      "loss": 1.3971,
      "theoretical_loss": 3.3928212090754735,
      "tokens_seen": 2286419968
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031246464646464644,
      "loss": 1.4861,
      "theoretical_loss": 3.3928133125686943,
      "tokens_seen": 2286485504
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031244444444444446,
      "loss": 1.3704,
      "theoretical_loss": 3.3928054163516155,
      "tokens_seen": 2286551040
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031242424242424247,
      "loss": 1.4998,
      "theoretical_loss": 3.392797520424217,
      "tokens_seen": 2286616576
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003124040404040404,
      "loss": 1.4594,
      "theoretical_loss": 3.3927896247864813,
      "tokens_seen": 2286682112
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123838383838384,
      "loss": 1.1811,
      "theoretical_loss": 3.392781729438388,
      "tokens_seen": 2286747648
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031236363636363634,
      "loss": 1.3905,
      "theoretical_loss": 3.39277383437992,
      "tokens_seen": 2286813184
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031234343434343435,
      "loss": 1.4983,
      "theoretical_loss": 3.3927659396110568,
      "tokens_seen": 2286878720
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031232323232323236,
      "loss": 1.4324,
      "theoretical_loss": 3.39275804513178,
      "tokens_seen": 2286944256
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003123030303030303,
      "loss": 1.3279,
      "theoretical_loss": 3.3927501509420708,
      "tokens_seen": 2287009792
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003122828282828283,
      "loss": 1.4272,
      "theoretical_loss": 3.3927422570419106,
      "tokens_seen": 2287075328
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031226262626262623,
      "loss": 1.1917,
      "theoretical_loss": 3.3927343634312797,
      "tokens_seen": 2287140864
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003122424242424243,
      "loss": 1.5244,
      "theoretical_loss": 3.3927264701101594,
      "tokens_seen": 2287206400
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031222222222222225,
      "loss": 1.178,
      "theoretical_loss": 3.392718577078531,
      "tokens_seen": 2287271936
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003122020202020202,
      "loss": 1.4584,
      "theoretical_loss": 3.392710684336376,
      "tokens_seen": 2287337472
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031218181818181817,
      "loss": 1.1846,
      "theoretical_loss": 3.3927027918836745,
      "tokens_seen": 2287403008
    },
    {
      "epoch": 0.38,
      "objective/train/aligned_token_count": 0.3194037079811096,
      "objective/train/avg_token_score": 0.02210118994116783,
      "objective/train/docs_used": 1294930,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0830011367797852,
      "objective/train/original_loss": 4.479400157928467,
      "objective/train/theoretical_loss": 3.3926948997204085,
      "objective/train/tokens_used": 646327776,
      "objective/train/unlikelihood_loss": 0.10477761924266815,
      "theoretical_loss": 3.3926948997204085,
      "tokens_seen": 2287468544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003121616161616161,
      "loss": 1.2138,
      "theoretical_loss": 3.3926948997204085,
      "tokens_seen": 2287468544
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003121414141414142,
      "loss": 1.2854,
      "theoretical_loss": 3.3926870078465585,
      "tokens_seen": 2287534080
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031212121212121214,
      "loss": 1.2827,
      "theoretical_loss": 3.392679116262106,
      "tokens_seen": 2287599616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003121010101010101,
      "loss": 1.3801,
      "theoretical_loss": 3.3926712249670317,
      "tokens_seen": 2287665152
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031208080808080806,
      "loss": 1.578,
      "theoretical_loss": 3.392663333961317,
      "tokens_seen": 2287730688
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031206060606060607,
      "loss": 1.5125,
      "theoretical_loss": 3.392655443244943,
      "tokens_seen": 2287796224
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003120404040404041,
      "loss": 1.3125,
      "theoretical_loss": 3.3926475528178908,
      "tokens_seen": 2287861760
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031202020202020204,
      "loss": 1.405,
      "theoretical_loss": 3.392639662680141,
      "tokens_seen": 2287927296
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000312,
      "loss": 1.2856,
      "theoretical_loss": 3.3926317728316757,
      "tokens_seen": 2287992832
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031197979797979795,
      "loss": 1.4319,
      "theoretical_loss": 3.392623883272475,
      "tokens_seen": 2288058368
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031195959595959596,
      "loss": 1.3176,
      "theoretical_loss": 3.3926159940025205,
      "tokens_seen": 2288123904
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000311939393939394,
      "loss": 1.4789,
      "theoretical_loss": 3.3926081050217936,
      "tokens_seen": 2288189440
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031191919191919193,
      "loss": 1.4594,
      "theoretical_loss": 3.392600216330275,
      "tokens_seen": 2288254976
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003118989898989899,
      "loss": 1.4054,
      "theoretical_loss": 3.3925923279279453,
      "tokens_seen": 2288320512
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003118787878787879,
      "loss": 1.1994,
      "theoretical_loss": 3.3925844398147866,
      "tokens_seen": 2288386048
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031185858585858585,
      "loss": 1.3663,
      "theoretical_loss": 3.39257655199078,
      "tokens_seen": 2288451584
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031183838383838387,
      "loss": 1.1636,
      "theoretical_loss": 3.3925686644559057,
      "tokens_seen": 2288517120
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003118181818181818,
      "loss": 1.4278,
      "theoretical_loss": 3.3925607772101456,
      "tokens_seen": 2288582656
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117979797979798,
      "loss": 1.1649,
      "theoretical_loss": 3.39255289025348,
      "tokens_seen": 2288648192
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117777777777778,
      "loss": 1.5184,
      "theoretical_loss": 3.3925450035858913,
      "tokens_seen": 2288713728
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031175757575757575,
      "loss": 1.354,
      "theoretical_loss": 3.3925371172073597,
      "tokens_seen": 2288779264
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031173737373737376,
      "loss": 1.5778,
      "theoretical_loss": 3.3925292311178668,
      "tokens_seen": 2288844800
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003117171717171717,
      "loss": 1.1849,
      "theoretical_loss": 3.3925213453173932,
      "tokens_seen": 2288910336
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003116969696969697,
      "loss": 1.6056,
      "theoretical_loss": 3.3925134598059206,
      "tokens_seen": 2288975872
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003116767676767677,
      "loss": 1.58,
      "theoretical_loss": 3.3925055745834296,
      "tokens_seen": 2289041408
    },
    {
      "epoch": 0.38,
      "objective/train/aligned_token_count": 0.43163490295410156,
      "objective/train/avg_token_score": 0.027810361236333847,
      "objective/train/docs_used": 1295606,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5732510089874268,
      "objective/train/original_loss": 5.168575763702393,
      "objective/train/theoretical_loss": 3.392497689649902,
      "objective/train/tokens_used": 647966176,
      "objective/train/unlikelihood_loss": 0.03606986626982689,
      "theoretical_loss": 3.392497689649902,
      "tokens_seen": 2289106944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031165656565656564,
      "loss": 1.5815,
      "theoretical_loss": 3.392497689649902,
      "tokens_seen": 2289106944
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031163636363636365,
      "loss": 1.5668,
      "theoretical_loss": 3.3924898050053183,
      "tokens_seen": 2289172480
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003116161616161616,
      "loss": 1.36,
      "theoretical_loss": 3.3924819206496597,
      "tokens_seen": 2289238016
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003115959595959596,
      "loss": 1.2606,
      "theoretical_loss": 3.392474036582908,
      "tokens_seen": 2289303552
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003115757575757576,
      "loss": 1.3646,
      "theoretical_loss": 3.3924661528050435,
      "tokens_seen": 2289369088
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031155555555555553,
      "loss": 1.4083,
      "theoretical_loss": 3.3924582693160477,
      "tokens_seen": 2289434624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031153535353535354,
      "loss": 1.229,
      "theoretical_loss": 3.3924503861159017,
      "tokens_seen": 2289500160
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031151515151515156,
      "loss": 1.5474,
      "theoretical_loss": 3.392442503204587,
      "tokens_seen": 2289565696
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003114949494949495,
      "loss": 1.3648,
      "theoretical_loss": 3.392434620582084,
      "tokens_seen": 2289631232
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031147474747474747,
      "loss": 1.3014,
      "theoretical_loss": 3.392426738248375,
      "tokens_seen": 2289696768
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003114545454545454,
      "loss": 1.2736,
      "theoretical_loss": 3.39241885620344,
      "tokens_seen": 2289762304
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031143434343434344,
      "loss": 1.4334,
      "theoretical_loss": 3.3924109744472606,
      "tokens_seen": 2289827840
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031141414141414145,
      "loss": 1.4529,
      "theoretical_loss": 3.3924030929798183,
      "tokens_seen": 2289893376
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113939393939394,
      "loss": 1.2591,
      "theoretical_loss": 3.3923952118010936,
      "tokens_seen": 2289958912
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031137373737373736,
      "loss": 1.3507,
      "theoretical_loss": 3.392387330911068,
      "tokens_seen": 2290024448
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113535353535353,
      "loss": 1.3067,
      "theoretical_loss": 3.392379450309723,
      "tokens_seen": 2290089984
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003113333333333334,
      "loss": 1.3348,
      "theoretical_loss": 3.392371569997039,
      "tokens_seen": 2290155520
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031131313131313134,
      "loss": 1.4791,
      "theoretical_loss": 3.392363689972998,
      "tokens_seen": 2290221056
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112929292929293,
      "loss": 1.4525,
      "theoretical_loss": 3.3923558102375804,
      "tokens_seen": 2290286592
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031127272727272725,
      "loss": 1.467,
      "theoretical_loss": 3.392347930790768,
      "tokens_seen": 2290352128
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112525252525252,
      "loss": 1.4376,
      "theoretical_loss": 3.3923400516325417,
      "tokens_seen": 2290417664
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003112323232323233,
      "loss": 1.429,
      "theoretical_loss": 3.392332172762883,
      "tokens_seen": 2290483200
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031121212121212123,
      "loss": 1.5374,
      "theoretical_loss": 3.3923242941817717,
      "tokens_seen": 2290548736
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003111919191919192,
      "loss": 1.2073,
      "theoretical_loss": 3.3923164158891908,
      "tokens_seen": 2290614272
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031117171717171715,
      "loss": 1.1913,
      "theoretical_loss": 3.3923085378851208,
      "tokens_seen": 2290679808
    },
    {
      "epoch": 0.38,
      "objective/train/aligned_token_count": 0.3137218952178955,
      "objective/train/avg_token_score": 0.017470920458436012,
      "objective/train/docs_used": 1296233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9987515211105347,
      "objective/train/original_loss": 5.429305076599121,
      "objective/train/theoretical_loss": 3.3923006601695427,
      "objective/train/tokens_used": 649604576,
      "objective/train/unlikelihood_loss": 0.04133667051792145,
      "theoretical_loss": 3.3923006601695427,
      "tokens_seen": 2290745344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003111515151515152,
      "loss": 1.3272,
      "theoretical_loss": 3.3923006601695427,
      "tokens_seen": 2290745344
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031113131313131317,
      "loss": 1.5075,
      "theoretical_loss": 3.3922927827424374,
      "tokens_seen": 2290810880
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003111111111111111,
      "loss": 1.326,
      "theoretical_loss": 3.3922849056037867,
      "tokens_seen": 2290876416
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003110909090909091,
      "loss": 1.5044,
      "theoretical_loss": 3.392277028753572,
      "tokens_seen": 2290941952
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031107070707070704,
      "loss": 1.3687,
      "theoretical_loss": 3.3922691521917736,
      "tokens_seen": 2291007488
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003110505050505051,
      "loss": 1.0987,
      "theoretical_loss": 3.392261275918373,
      "tokens_seen": 2291073024
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031103030303030306,
      "loss": 1.3858,
      "theoretical_loss": 3.3922533999333515,
      "tokens_seen": 2291138560
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000311010101010101,
      "loss": 1.3893,
      "theoretical_loss": 3.3922455242366905,
      "tokens_seen": 2291204096
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000310989898989899,
      "loss": 1.2965,
      "theoretical_loss": 3.392237648828371,
      "tokens_seen": 2291269632
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031096969696969693,
      "loss": 1.5097,
      "theoretical_loss": 3.392229773708374,
      "tokens_seen": 2291335168
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000310949494949495,
      "loss": 1.5006,
      "theoretical_loss": 3.3922218988766812,
      "tokens_seen": 2291400704
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031092929292929295,
      "loss": 1.5,
      "theoretical_loss": 3.3922140243332732,
      "tokens_seen": 2291466240
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003109090909090909,
      "loss": 1.4117,
      "theoretical_loss": 3.3922061500781315,
      "tokens_seen": 2291531776
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031088888888888887,
      "loss": 1.4451,
      "theoretical_loss": 3.3921982761112375,
      "tokens_seen": 2291597312
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003108686868686869,
      "loss": 1.4606,
      "theoretical_loss": 3.392190402432572,
      "tokens_seen": 2291662848
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003108484848484849,
      "loss": 1.1391,
      "theoretical_loss": 3.392182529042117,
      "tokens_seen": 2291728384
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031082828282828285,
      "loss": 1.2425,
      "theoretical_loss": 3.3921746559398525,
      "tokens_seen": 2291793920
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003108080808080808,
      "loss": 1.5378,
      "theoretical_loss": 3.3921667831257607,
      "tokens_seen": 2291859456
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031078787878787876,
      "loss": 1.4458,
      "theoretical_loss": 3.392158910599822,
      "tokens_seen": 2291924992
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031076767676767677,
      "loss": 1.3178,
      "theoretical_loss": 3.3921510383620186,
      "tokens_seen": 2291990528
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003107474747474748,
      "loss": 1.4154,
      "theoretical_loss": 3.3921431664123305,
      "tokens_seen": 2292056064
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031072727272727274,
      "loss": 1.6187,
      "theoretical_loss": 3.39213529475074,
      "tokens_seen": 2292121600
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003107070707070707,
      "loss": 1.2322,
      "theoretical_loss": 3.392127423377228,
      "tokens_seen": 2292187136
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0003106868686868687,
      "loss": 1.2154,
      "theoretical_loss": 3.3921195522917755,
      "tokens_seen": 2292252672
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00031066666666666666,
      "loss": 1.315,
      "theoretical_loss": 3.392111681494364,
      "tokens_seen": 2292318208
    },
    {
      "epoch": 0.38,
      "objective/train/aligned_token_count": 0.5306695699691772,
      "objective/train/avg_token_score": 0.02912568300962448,
      "objective/train/docs_used": 1296233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5920521020889282,
      "objective/train/original_loss": 4.0167317390441895,
      "objective/train/theoretical_loss": 3.3921038109849744,
      "objective/train/tokens_used": 651242976,
      "objective/train/unlikelihood_loss": 0.07039138674736023,
      "theoretical_loss": 3.3921038109849744,
      "tokens_seen": 2292383744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003106464646464647,
      "loss": 1.4715,
      "theoretical_loss": 3.3921038109849744,
      "tokens_seen": 2292383744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031062626262626263,
      "loss": 1.6125,
      "theoretical_loss": 3.3920959407635882,
      "tokens_seen": 2292449280
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003106060606060606,
      "loss": 1.4429,
      "theoretical_loss": 3.3920880708301864,
      "tokens_seen": 2292514816
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003105858585858586,
      "loss": 1.3733,
      "theoretical_loss": 3.3920802011847506,
      "tokens_seen": 2292580352
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031056565656565656,
      "loss": 1.3472,
      "theoretical_loss": 3.3920723318272614,
      "tokens_seen": 2292645888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031054545454545457,
      "loss": 1.5162,
      "theoretical_loss": 3.392064462757701,
      "tokens_seen": 2292711424
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003105252525252525,
      "loss": 1.5003,
      "theoretical_loss": 3.39205659397605,
      "tokens_seen": 2292776960
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031050505050505054,
      "loss": 1.3211,
      "theoretical_loss": 3.3920487254822893,
      "tokens_seen": 2292842496
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003104848484848485,
      "loss": 1.5112,
      "theoretical_loss": 3.3920408572764007,
      "tokens_seen": 2292908032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031046464646464645,
      "loss": 1.4869,
      "theoretical_loss": 3.392032989358365,
      "tokens_seen": 2292973568
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031044444444444446,
      "loss": 1.4208,
      "theoretical_loss": 3.392025121728164,
      "tokens_seen": 2293039104
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003104242424242424,
      "loss": 1.372,
      "theoretical_loss": 3.3920172543857787,
      "tokens_seen": 2293104640
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031040404040404043,
      "loss": 1.4671,
      "theoretical_loss": 3.39200938733119,
      "tokens_seen": 2293170176
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003103838383838384,
      "loss": 1.5395,
      "theoretical_loss": 3.39200152056438,
      "tokens_seen": 2293235712
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031036363636363634,
      "loss": 1.2623,
      "theoretical_loss": 3.3919936540853293,
      "tokens_seen": 2293301248
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031034343434343435,
      "loss": 1.2479,
      "theoretical_loss": 3.3919857878940194,
      "tokens_seen": 2293366784
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031032323232323236,
      "loss": 1.2277,
      "theoretical_loss": 3.391977921990431,
      "tokens_seen": 2293432320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003103030303030303,
      "loss": 1.552,
      "theoretical_loss": 3.3919700563745456,
      "tokens_seen": 2293497856
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003102828282828283,
      "loss": 1.4732,
      "theoretical_loss": 3.3919621910463453,
      "tokens_seen": 2293563392
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031026262626262624,
      "loss": 1.4607,
      "theoretical_loss": 3.39195432600581,
      "tokens_seen": 2293628928
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031024242424242425,
      "loss": 1.4059,
      "theoretical_loss": 3.391946461252922,
      "tokens_seen": 2293694464
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031022222222222226,
      "loss": 1.5332,
      "theoretical_loss": 3.3919385967876625,
      "tokens_seen": 2293760000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003102020202020202,
      "loss": 1.2758,
      "theoretical_loss": 3.391930732610012,
      "tokens_seen": 2293825536
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031018181818181817,
      "loss": 1.4209,
      "theoretical_loss": 3.3919228687199525,
      "tokens_seen": 2293891072
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031016161616161613,
      "loss": 1.3323,
      "theoretical_loss": 3.3919150051174647,
      "tokens_seen": 2293956608
    },
    {
      "epoch": 0.39,
      "objective/train/aligned_token_count": 0.3692570924758911,
      "objective/train/avg_token_score": 0.019013360142707825,
      "objective/train/docs_used": 1296233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.219751000404358,
      "objective/train/original_loss": 4.565107345581055,
      "objective/train/theoretical_loss": 3.3919071418025304,
      "objective/train/tokens_used": 652881376,
      "objective/train/unlikelihood_loss": 0.062217194586992264,
      "theoretical_loss": 3.3919071418025304,
      "tokens_seen": 2294022144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003101414141414142,
      "loss": 1.4039,
      "theoretical_loss": 3.3919071418025304,
      "tokens_seen": 2294022144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031012121212121215,
      "loss": 1.2014,
      "theoretical_loss": 3.3918992787751305,
      "tokens_seen": 2294087680
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003101010101010101,
      "loss": 1.318,
      "theoretical_loss": 3.391891416035247,
      "tokens_seen": 2294153216
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031008080808080806,
      "loss": 1.3963,
      "theoretical_loss": 3.39188355358286,
      "tokens_seen": 2294218752
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000310060606060606,
      "loss": 1.3553,
      "theoretical_loss": 3.3918756914179515,
      "tokens_seen": 2294284288
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003100404040404041,
      "loss": 1.1823,
      "theoretical_loss": 3.3918678295405025,
      "tokens_seen": 2294349824
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031002020202020204,
      "loss": 1.2776,
      "theoretical_loss": 3.3918599679504946,
      "tokens_seen": 2294415360
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00031,
      "loss": 1.3436,
      "theoretical_loss": 3.391852106647909,
      "tokens_seen": 2294480896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030997979797979796,
      "loss": 1.3604,
      "theoretical_loss": 3.3918442456327265,
      "tokens_seen": 2294546432
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030995959595959597,
      "loss": 1.4994,
      "theoretical_loss": 3.391836384904929,
      "tokens_seen": 2294611968
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000309939393939394,
      "loss": 1.4167,
      "theoretical_loss": 3.3918285244644975,
      "tokens_seen": 2294677504
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030991919191919194,
      "loss": 1.4183,
      "theoretical_loss": 3.3918206643114135,
      "tokens_seen": 2294743040
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003098989898989899,
      "loss": 1.187,
      "theoretical_loss": 3.391812804445658,
      "tokens_seen": 2294808576
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030987878787878785,
      "loss": 1.465,
      "theoretical_loss": 3.3918049448672125,
      "tokens_seen": 2294874112
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030985858585858586,
      "loss": 1.4075,
      "theoretical_loss": 3.391797085576058,
      "tokens_seen": 2294939648
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030983838383838387,
      "loss": 1.6403,
      "theoretical_loss": 3.3917892265721763,
      "tokens_seen": 2295005184
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030981818181818183,
      "loss": 1.5449,
      "theoretical_loss": 3.391781367855548,
      "tokens_seen": 2295070720
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003097979797979798,
      "loss": 1.2429,
      "theoretical_loss": 3.391773509426155,
      "tokens_seen": 2295136256
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003097777777777778,
      "loss": 1.531,
      "theoretical_loss": 3.3917656512839787,
      "tokens_seen": 2295201792
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030975757575757575,
      "loss": 1.5201,
      "theoretical_loss": 3.391757793429,
      "tokens_seen": 2295267328
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030973737373737376,
      "loss": 1.4554,
      "theoretical_loss": 3.3917499358612,
      "tokens_seen": 2295332864
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003097171717171717,
      "loss": 1.3799,
      "theoretical_loss": 3.391742078580561,
      "tokens_seen": 2295398400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003096969696969697,
      "loss": 1.2612,
      "theoretical_loss": 3.391734221587063,
      "tokens_seen": 2295463936
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003096767676767677,
      "loss": 1.4497,
      "theoretical_loss": 3.391726364880688,
      "tokens_seen": 2295529472
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030965656565656565,
      "loss": 1.4073,
      "theoretical_loss": 3.3917185084614174,
      "tokens_seen": 2295595008
    },
    {
      "epoch": 0.39,
      "objective/train/aligned_token_count": 0.3643695116043091,
      "objective/train/avg_token_score": 0.015814119949936867,
      "objective/train/docs_used": 1296233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1693059206008911,
      "objective/train/original_loss": 4.68144416809082,
      "objective/train/theoretical_loss": 3.391710652329232,
      "objective/train/tokens_used": 654519776,
      "objective/train/unlikelihood_loss": 0.04476901888847351,
      "theoretical_loss": 3.391710652329232,
      "tokens_seen": 2295660544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030963636363636366,
      "loss": 1.5385,
      "theoretical_loss": 3.391710652329232,
      "tokens_seen": 2295660544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003096161616161616,
      "loss": 1.5498,
      "theoretical_loss": 3.391702796484114,
      "tokens_seen": 2295726080
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030959595959595957,
      "loss": 1.3705,
      "theoretical_loss": 3.391694940926044,
      "tokens_seen": 2295791616
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003095757575757576,
      "loss": 1.2588,
      "theoretical_loss": 3.3916870856550037,
      "tokens_seen": 2295857152
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030955555555555554,
      "loss": 1.401,
      "theoretical_loss": 3.3916792306709738,
      "tokens_seen": 2295922688
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030953535353535355,
      "loss": 1.3987,
      "theoretical_loss": 3.391671375973936,
      "tokens_seen": 2295988224
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003095151515151515,
      "loss": 1.3471,
      "theoretical_loss": 3.3916635215638724,
      "tokens_seen": 2296053760
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003094949494949495,
      "loss": 1.5478,
      "theoretical_loss": 3.391655667440763,
      "tokens_seen": 2296119296
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003094747474747475,
      "loss": 1.241,
      "theoretical_loss": 3.39164781360459,
      "tokens_seen": 2296184832
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003094545454545455,
      "loss": 1.5106,
      "theoretical_loss": 3.391639960055334,
      "tokens_seen": 2296250368
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030943434343434344,
      "loss": 1.3473,
      "theoretical_loss": 3.391632106792977,
      "tokens_seen": 2296315904
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003094141414141414,
      "loss": 1.458,
      "theoretical_loss": 3.3916242538175005,
      "tokens_seen": 2296381440
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003093939393939394,
      "loss": 1.6435,
      "theoretical_loss": 3.3916164011288847,
      "tokens_seen": 2296446976
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030937373737373737,
      "loss": 1.4213,
      "theoretical_loss": 3.3916085487271124,
      "tokens_seen": 2296512512
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003093535353535354,
      "loss": 1.394,
      "theoretical_loss": 3.391600696612164,
      "tokens_seen": 2296578048
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030933333333333334,
      "loss": 1.4394,
      "theoretical_loss": 3.3915928447840207,
      "tokens_seen": 2296643584
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030931313131313135,
      "loss": 1.3552,
      "theoretical_loss": 3.3915849932426645,
      "tokens_seen": 2296709120
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003092929292929293,
      "loss": 1.4329,
      "theoretical_loss": 3.3915771419880762,
      "tokens_seen": 2296774656
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030927272727272726,
      "loss": 1.345,
      "theoretical_loss": 3.3915692910202377,
      "tokens_seen": 2296840192
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030925252525252527,
      "loss": 1.3558,
      "theoretical_loss": 3.39156144033913,
      "tokens_seen": 2296905728
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030923232323232323,
      "loss": 1.2189,
      "theoretical_loss": 3.3915535899447344,
      "tokens_seen": 2296971264
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030921212121212124,
      "loss": 1.2527,
      "theoretical_loss": 3.3915457398370323,
      "tokens_seen": 2297036800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003091919191919192,
      "loss": 1.3738,
      "theoretical_loss": 3.391537890016005,
      "tokens_seen": 2297102336
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030917171717171715,
      "loss": 1.448,
      "theoretical_loss": 3.3915300404816344,
      "tokens_seen": 2297167872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030915151515151516,
      "loss": 1.1597,
      "theoretical_loss": 3.391522191233901,
      "tokens_seen": 2297233408
    },
    {
      "epoch": 0.39,
      "objective/train/aligned_token_count": 0.4037145674228668,
      "objective/train/avg_token_score": 0.005951590836048126,
      "objective/train/docs_used": 1296233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4598697423934937,
      "objective/train/original_loss": 4.722467422485352,
      "objective/train/theoretical_loss": 3.391514342272786,
      "objective/train/tokens_used": 656158176,
      "objective/train/unlikelihood_loss": 0.038441527634859085,
      "theoretical_loss": 3.391514342272786,
      "tokens_seen": 2297298944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003091313131313132,
      "loss": 1.4466,
      "theoretical_loss": 3.391514342272786,
      "tokens_seen": 2297298944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030911111111111113,
      "loss": 1.3704,
      "theoretical_loss": 3.391506493598272,
      "tokens_seen": 2297364480
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003090909090909091,
      "loss": 1.2262,
      "theoretical_loss": 3.3914986452103397,
      "tokens_seen": 2297430016
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030907070707070705,
      "loss": 1.3732,
      "theoretical_loss": 3.39149079710897,
      "tokens_seen": 2297495552
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030905050505050506,
      "loss": 1.2574,
      "theoretical_loss": 3.391482949294145,
      "tokens_seen": 2297561088
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030903030303030307,
      "loss": 1.3173,
      "theoretical_loss": 3.3914751017658458,
      "tokens_seen": 2297626624
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000309010101010101,
      "loss": 1.3587,
      "theoretical_loss": 3.3914672545240534,
      "tokens_seen": 2297692160
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000308989898989899,
      "loss": 1.3344,
      "theoretical_loss": 3.3914594075687496,
      "tokens_seen": 2297757696
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030896969696969694,
      "loss": 1.3116,
      "theoretical_loss": 3.391451560899916,
      "tokens_seen": 2297823232
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000308949494949495,
      "loss": 1.359,
      "theoretical_loss": 3.391443714517533,
      "tokens_seen": 2297888768
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030892929292929296,
      "loss": 1.0282,
      "theoretical_loss": 3.391435868421583,
      "tokens_seen": 2297954304
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003089090909090909,
      "loss": 1.321,
      "theoretical_loss": 3.391428022612047,
      "tokens_seen": 2298019840
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003088888888888889,
      "loss": 1.3739,
      "theoretical_loss": 3.391420177088906,
      "tokens_seen": 2298085376
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030886868686868683,
      "loss": 1.2092,
      "theoretical_loss": 3.391412331852142,
      "tokens_seen": 2298150912
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003088484848484849,
      "loss": 1.1849,
      "theoretical_loss": 3.391404486901736,
      "tokens_seen": 2298216448
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030882828282828285,
      "loss": 1.3004,
      "theoretical_loss": 3.3913966422376696,
      "tokens_seen": 2298281984
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003088080808080808,
      "loss": 1.2498,
      "theoretical_loss": 3.391388797859924,
      "tokens_seen": 2298347520
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030878787878787877,
      "loss": 1.2658,
      "theoretical_loss": 3.3913809537684805,
      "tokens_seen": 2298413056
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003087676767676768,
      "loss": 1.4311,
      "theoretical_loss": 3.391373109963321,
      "tokens_seen": 2298478592
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003087474747474748,
      "loss": 1.1184,
      "theoretical_loss": 3.391365266444426,
      "tokens_seen": 2298544128
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030872727272727275,
      "loss": 1.3735,
      "theoretical_loss": 3.3913574232117782,
      "tokens_seen": 2298609664
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003087070707070707,
      "loss": 1.2726,
      "theoretical_loss": 3.3913495802653575,
      "tokens_seen": 2298675200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030868686868686866,
      "loss": 1.2891,
      "theoretical_loss": 3.3913417376051465,
      "tokens_seen": 2298740736
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030866666666666667,
      "loss": 1.0164,
      "theoretical_loss": 3.391333895231126,
      "tokens_seen": 2298806272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003086464646464647,
      "loss": 1.5099,
      "theoretical_loss": 3.3913260531432776,
      "tokens_seen": 2298871808
    },
    {
      "epoch": 0.39,
      "objective/train/aligned_token_count": 0.42130985856056213,
      "objective/train/avg_token_score": 0.01315027941018343,
      "objective/train/docs_used": 1296233,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.473658800125122,
      "objective/train/original_loss": 4.4738264083862305,
      "objective/train/theoretical_loss": 3.3913182113415825,
      "objective/train/tokens_used": 657796576,
      "objective/train/unlikelihood_loss": 0.06518010050058365,
      "theoretical_loss": 3.3913182113415825,
      "tokens_seen": 2298937344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030862626262626264,
      "loss": 1.4101,
      "theoretical_loss": 3.3913182113415825,
      "tokens_seen": 2298937344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003086060606060606,
      "loss": 1.1034,
      "theoretical_loss": 3.391310369826022,
      "tokens_seen": 2299002880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003085858585858586,
      "loss": 1.6341,
      "theoretical_loss": 3.3913025285965777,
      "tokens_seen": 2299068416
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030856565656565656,
      "loss": 1.145,
      "theoretical_loss": 3.3912946876532315,
      "tokens_seen": 2299133952
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003085454545454546,
      "loss": 1.0626,
      "theoretical_loss": 3.391286846995964,
      "tokens_seen": 2299199488
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030852525252525253,
      "loss": 1.1242,
      "theoretical_loss": 3.3912790066247567,
      "tokens_seen": 2299265024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003085050505050505,
      "loss": 1.0286,
      "theoretical_loss": 3.3912711665395916,
      "tokens_seen": 2299330560
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003084848484848485,
      "loss": 1.2981,
      "theoretical_loss": 3.3912633267404497,
      "tokens_seen": 2299396096
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030846464646464646,
      "loss": 1.0791,
      "theoretical_loss": 3.391255487227313,
      "tokens_seen": 2299461632
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030844444444444447,
      "loss": 0.9205,
      "theoretical_loss": 3.3912476480001614,
      "tokens_seen": 2299527168
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003084242424242424,
      "loss": 1.3229,
      "theoretical_loss": 3.3912398090589777,
      "tokens_seen": 2299592704
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030840404040404043,
      "loss": 1.2715,
      "theoretical_loss": 3.391231970403743,
      "tokens_seen": 2299658240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003083838383838384,
      "loss": 1.2823,
      "theoretical_loss": 3.3912241320344387,
      "tokens_seen": 2299723776
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030836363636363635,
      "loss": 1.2244,
      "theoretical_loss": 3.3912162939510457,
      "tokens_seen": 2299789312
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030834343434343436,
      "loss": 1.2593,
      "theoretical_loss": 3.3912084561535467,
      "tokens_seen": 2299854848
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003083232323232323,
      "loss": 1.3573,
      "theoretical_loss": 3.3912006186419217,
      "tokens_seen": 2299920384
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030830303030303033,
      "loss": 1.0534,
      "theoretical_loss": 3.391192781416153,
      "tokens_seen": 2299985920
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003082828282828283,
      "loss": 1.1519,
      "theoretical_loss": 3.3911849444762217,
      "tokens_seen": 2300051456
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030826262626262624,
      "loss": 1.293,
      "theoretical_loss": 3.3911771078221093,
      "tokens_seen": 2300116992
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030824242424242425,
      "loss": 1.3531,
      "theoretical_loss": 3.391169271453797,
      "tokens_seen": 2300182528
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030822222222222226,
      "loss": 1.2815,
      "theoretical_loss": 3.3911614353712665,
      "tokens_seen": 2300248064
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003082020202020202,
      "loss": 1.2409,
      "theoretical_loss": 3.3911535995744995,
      "tokens_seen": 2300313600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003081818181818182,
      "loss": 1.0938,
      "theoretical_loss": 3.3911457640634772,
      "tokens_seen": 2300379136
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030816161616161613,
      "loss": 1.0029,
      "theoretical_loss": 3.3911379288381807,
      "tokens_seen": 2300444672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030814141414141414,
      "loss": 1.1595,
      "theoretical_loss": 3.391130093898592,
      "tokens_seen": 2300510208
    },
    {
      "epoch": 0.39,
      "objective/train/aligned_token_count": 0.3150048851966858,
      "objective/train/avg_token_score": 0.009434150531888008,
      "objective/train/docs_used": 1296929,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0510516166687012,
      "objective/train/original_loss": 4.600020408630371,
      "objective/train/theoretical_loss": 3.391122259244692,
      "objective/train/tokens_used": 659434976,
      "objective/train/unlikelihood_loss": 0.06388087570667267,
      "theoretical_loss": 3.391122259244692,
      "tokens_seen": 2300575744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030812121212121216,
      "loss": 1.2873,
      "theoretical_loss": 3.391122259244692,
      "tokens_seen": 2300575744
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003081010101010101,
      "loss": 1.2507,
      "theoretical_loss": 3.3911144248764624,
      "tokens_seen": 2300641280
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030808080808080807,
      "loss": 1.2046,
      "theoretical_loss": 3.3911065907938847,
      "tokens_seen": 2300706816
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000308060606060606,
      "loss": 1.27,
      "theoretical_loss": 3.3910987569969406,
      "tokens_seen": 2300772352
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003080404040404041,
      "loss": 1.2936,
      "theoretical_loss": 3.391090923485611,
      "tokens_seen": 2300837888
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030802020202020205,
      "loss": 1.0466,
      "theoretical_loss": 3.391083090259878,
      "tokens_seen": 2300903424
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000308,
      "loss": 1.2548,
      "theoretical_loss": 3.391075257319722,
      "tokens_seen": 2300968960
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030797979797979796,
      "loss": 1.1929,
      "theoretical_loss": 3.391067424665126,
      "tokens_seen": 2301034496
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003079595959595959,
      "loss": 1.3436,
      "theoretical_loss": 3.3910595922960693,
      "tokens_seen": 2301100032
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000307939393939394,
      "loss": 1.1891,
      "theoretical_loss": 3.3910517602125356,
      "tokens_seen": 2301165568
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030791919191919194,
      "loss": 1.4236,
      "theoretical_loss": 3.391043928414505,
      "tokens_seen": 2301231104
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003078989898989899,
      "loss": 1.3843,
      "theoretical_loss": 3.3910360969019595,
      "tokens_seen": 2301296640
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030787878787878786,
      "loss": 1.1209,
      "theoretical_loss": 3.3910282656748807,
      "tokens_seen": 2301362176
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003078585858585858,
      "loss": 1.0857,
      "theoretical_loss": 3.391020434733249,
      "tokens_seen": 2301427712
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003078383838383839,
      "loss": 1.2391,
      "theoretical_loss": 3.3910126040770474,
      "tokens_seen": 2301493248
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030781818181818183,
      "loss": 1.2651,
      "theoretical_loss": 3.3910047737062565,
      "tokens_seen": 2301558784
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003077979797979798,
      "loss": 1.2353,
      "theoretical_loss": 3.390996943620858,
      "tokens_seen": 2301624320
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030777777777777775,
      "loss": 1.1283,
      "theoretical_loss": 3.3909891138208326,
      "tokens_seen": 2301689856
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003077575757575758,
      "loss": 1.2395,
      "theoretical_loss": 3.390981284306163,
      "tokens_seen": 2301755392
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030773737373737377,
      "loss": 1.3752,
      "theoretical_loss": 3.3909734550768302,
      "tokens_seen": 2301820928
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003077171717171717,
      "loss": 1.1609,
      "theoretical_loss": 3.3909656261328154,
      "tokens_seen": 2301886464
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076969696969697,
      "loss": 1.1484,
      "theoretical_loss": 3.3909577974741003,
      "tokens_seen": 2301952000
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030767676767676764,
      "loss": 1.073,
      "theoretical_loss": 3.390949969100666,
      "tokens_seen": 2302017536
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076565656565657,
      "loss": 1.2327,
      "theoretical_loss": 3.390942141012495,
      "tokens_seen": 2302083072
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030763636363636366,
      "loss": 1.1147,
      "theoretical_loss": 3.3909343132095673,
      "tokens_seen": 2302148608
    },
    {
      "epoch": 0.39,
      "objective/train/aligned_token_count": 0.42057672142982483,
      "objective/train/avg_token_score": 0.027334054931998253,
      "objective/train/docs_used": 1297386,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.418977975845337,
      "objective/train/original_loss": 4.8929924964904785,
      "objective/train/theoretical_loss": 3.3909264856918657,
      "objective/train/tokens_used": 661073376,
      "objective/train/unlikelihood_loss": 0.03315281867980957,
      "theoretical_loss": 3.3909264856918657,
      "tokens_seen": 2302214144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003076161616161616,
      "loss": 1.355,
      "theoretical_loss": 3.3909264856918657,
      "tokens_seen": 2302214144
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075959595959596,
      "loss": 1.2428,
      "theoretical_loss": 3.3909186584593716,
      "tokens_seen": 2302279680
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075757575757576,
      "loss": 1.0449,
      "theoretical_loss": 3.3909108315120653,
      "tokens_seen": 2302345216
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075555555555556,
      "loss": 1.3174,
      "theoretical_loss": 3.3909030048499296,
      "tokens_seen": 2302410752
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030753535353535356,
      "loss": 1.2735,
      "theoretical_loss": 3.3908951784729453,
      "tokens_seen": 2302476288
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003075151515151515,
      "loss": 1.3767,
      "theoretical_loss": 3.3908873523810943,
      "tokens_seen": 2302541824
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030749494949494947,
      "loss": 1.2539,
      "theoretical_loss": 3.3908795265743574,
      "tokens_seen": 2302607360
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074747474747475,
      "loss": 1.3308,
      "theoretical_loss": 3.390871701052717,
      "tokens_seen": 2302672896
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074545454545455,
      "loss": 1.4026,
      "theoretical_loss": 3.3908638758161542,
      "tokens_seen": 2302738432
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030743434343434345,
      "loss": 1.1769,
      "theoretical_loss": 3.3908560508646497,
      "tokens_seen": 2302803968
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003074141414141414,
      "loss": 1.3659,
      "theoretical_loss": 3.3908482261981865,
      "tokens_seen": 2302869504
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003073939393939394,
      "loss": 1.2387,
      "theoretical_loss": 3.390840401816745,
      "tokens_seen": 2302935040
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030737373737373737,
      "loss": 1.2153,
      "theoretical_loss": 3.3908325777203077,
      "tokens_seen": 2303000576
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003073535353535354,
      "loss": 1.2136,
      "theoretical_loss": 3.3908247539088547,
      "tokens_seen": 2303066112
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030733333333333334,
      "loss": 1.1142,
      "theoretical_loss": 3.3908169303823685,
      "tokens_seen": 2303131648
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003073131313131313,
      "loss": 1.2083,
      "theoretical_loss": 3.390809107140831,
      "tokens_seen": 2303197184
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003072929292929293,
      "loss": 0.961,
      "theoretical_loss": 3.3908012841842226,
      "tokens_seen": 2303262720
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030727272727272727,
      "loss": 1.1029,
      "theoretical_loss": 3.3907934615125255,
      "tokens_seen": 2303328256
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003072525252525253,
      "loss": 1.0681,
      "theoretical_loss": 3.390785639125721,
      "tokens_seen": 2303393792
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030723232323232323,
      "loss": 1.206,
      "theoretical_loss": 3.3907778170237908,
      "tokens_seen": 2303459328
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030721212121212124,
      "loss": 1.3033,
      "theoretical_loss": 3.390769995206716,
      "tokens_seen": 2303524864
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003071919191919192,
      "loss": 1.0684,
      "theoretical_loss": 3.390762173674479,
      "tokens_seen": 2303590400
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030717171717171716,
      "loss": 1.4778,
      "theoretical_loss": 3.3907543524270602,
      "tokens_seen": 2303655936
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030715151515151517,
      "loss": 1.2259,
      "theoretical_loss": 3.3907465314644414,
      "tokens_seen": 2303721472
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003071313131313131,
      "loss": 1.3325,
      "theoretical_loss": 3.3907387107866054,
      "tokens_seen": 2303787008
    },
    {
      "epoch": 0.39,
      "objective/train/aligned_token_count": 0.3527003824710846,
      "objective/train/avg_token_score": 0.010025923140347004,
      "objective/train/docs_used": 1298736,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2935250997543335,
      "objective/train/original_loss": 4.880310535430908,
      "objective/train/theoretical_loss": 3.390730890393532,
      "objective/train/tokens_used": 662711776,
      "objective/train/unlikelihood_loss": 0.05143946781754494,
      "theoretical_loss": 3.390730890393532,
      "tokens_seen": 2303852544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030711111111111114,
      "loss": 1.1842,
      "theoretical_loss": 3.390730890393532,
      "tokens_seen": 2303852544
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003070909090909091,
      "loss": 1.2259,
      "theoretical_loss": 3.3907230702852034,
      "tokens_seen": 2303918080
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030707070707070705,
      "loss": 1.47,
      "theoretical_loss": 3.3907152504616014,
      "tokens_seen": 2303983616
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030705050505050506,
      "loss": 1.5371,
      "theoretical_loss": 3.390707430922707,
      "tokens_seen": 2304049152
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003070303030303031,
      "loss": 1.3781,
      "theoretical_loss": 3.390699611668503,
      "tokens_seen": 2304114688
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030701010101010103,
      "loss": 1.3398,
      "theoretical_loss": 3.3906917926989695,
      "tokens_seen": 2304180224
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000306989898989899,
      "loss": 1.2217,
      "theoretical_loss": 3.390683974014088,
      "tokens_seen": 2304245760
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030696969696969694,
      "loss": 1.1557,
      "theoretical_loss": 3.3906761556138414,
      "tokens_seen": 2304311296
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030694949494949495,
      "loss": 1.1502,
      "theoretical_loss": 3.39066833749821,
      "tokens_seen": 2304376832
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030692929292929297,
      "loss": 1.2739,
      "theoretical_loss": 3.390660519667176,
      "tokens_seen": 2304442368
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003069090909090909,
      "loss": 1.3691,
      "theoretical_loss": 3.3906527021207205,
      "tokens_seen": 2304507904
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003068888888888889,
      "loss": 1.0615,
      "theoretical_loss": 3.3906448848588253,
      "tokens_seen": 2304573440
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030686868686868684,
      "loss": 1.3709,
      "theoretical_loss": 3.390637067881472,
      "tokens_seen": 2304638976
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003068484848484849,
      "loss": 1.4468,
      "theoretical_loss": 3.390629251188642,
      "tokens_seen": 2304704512
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030682828282828286,
      "loss": 1.2041,
      "theoretical_loss": 3.390621434780317,
      "tokens_seen": 2304770048
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003068080808080808,
      "loss": 1.3067,
      "theoretical_loss": 3.3906136186564786,
      "tokens_seen": 2304835584
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030678787878787877,
      "loss": 1.1629,
      "theoretical_loss": 3.3906058028171078,
      "tokens_seen": 2304901120
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030676767676767673,
      "loss": 1.3646,
      "theoretical_loss": 3.390597987262187,
      "tokens_seen": 2304966656
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003067474747474748,
      "loss": 1.0836,
      "theoretical_loss": 3.3905901719916973,
      "tokens_seen": 2305032192
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030672727272727275,
      "loss": 1.3749,
      "theoretical_loss": 3.3905823570056204,
      "tokens_seen": 2305097728
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003067070707070707,
      "loss": 1.0669,
      "theoretical_loss": 3.3905745423039377,
      "tokens_seen": 2305163264
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030668686868686866,
      "loss": 1.265,
      "theoretical_loss": 3.3905667278866307,
      "tokens_seen": 2305228800
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003066666666666667,
      "loss": 1.0965,
      "theoretical_loss": 3.390558913753681,
      "tokens_seen": 2305294336
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003066464646464647,
      "loss": 1.3394,
      "theoretical_loss": 3.3905510999050708,
      "tokens_seen": 2305359872
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030662626262626264,
      "loss": 1.1764,
      "theoretical_loss": 3.3905432863407805,
      "tokens_seen": 2305425408
    },
    {
      "epoch": 0.39,
      "objective/train/aligned_token_count": 0.46841397881507874,
      "objective/train/avg_token_score": 0.009612136520445347,
      "objective/train/docs_used": 1299212,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.58378267288208,
      "objective/train/original_loss": 3.9019057750701904,
      "objective/train/theoretical_loss": 3.390535473060793,
      "objective/train/tokens_used": 664350176,
      "objective/train/unlikelihood_loss": 0.10278970748186111,
      "theoretical_loss": 3.390535473060793,
      "tokens_seen": 2305490944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003066060606060606,
      "loss": 1.402,
      "theoretical_loss": 3.390535473060793,
      "tokens_seen": 2305490944
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030658585858585856,
      "loss": 1.2497,
      "theoretical_loss": 3.390527660065089,
      "tokens_seen": 2305556480
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030656565656565657,
      "loss": 1.2126,
      "theoretical_loss": 3.3905198473536498,
      "tokens_seen": 2305622016
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003065454545454546,
      "loss": 1.3574,
      "theoretical_loss": 3.3905120349264584,
      "tokens_seen": 2305687552
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030652525252525254,
      "loss": 1.2106,
      "theoretical_loss": 3.3905042227834947,
      "tokens_seen": 2305753088
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003065050505050505,
      "loss": 1.1532,
      "theoretical_loss": 3.390496410924741,
      "tokens_seen": 2305818624
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003064848484848485,
      "loss": 1.1413,
      "theoretical_loss": 3.390488599350179,
      "tokens_seen": 2305884160
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030646464646464646,
      "loss": 1.2259,
      "theoretical_loss": 3.3904807880597905,
      "tokens_seen": 2305949696
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030644444444444447,
      "loss": 1.2263,
      "theoretical_loss": 3.390472977053556,
      "tokens_seen": 2306015232
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030642424242424243,
      "loss": 1.1926,
      "theoretical_loss": 3.3904651663314587,
      "tokens_seen": 2306080768
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003064040404040404,
      "loss": 1.292,
      "theoretical_loss": 3.390457355893479,
      "tokens_seen": 2306146304
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063838383838384,
      "loss": 1.2367,
      "theoretical_loss": 3.390449545739599,
      "tokens_seen": 2306211840
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030636363636363635,
      "loss": 1.2084,
      "theoretical_loss": 3.3904417358698002,
      "tokens_seen": 2306277376
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030634343434343437,
      "loss": 1.364,
      "theoretical_loss": 3.390433926284064,
      "tokens_seen": 2306342912
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063232323232323,
      "loss": 1.1916,
      "theoretical_loss": 3.390426116982372,
      "tokens_seen": 2306408448
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003063030303030303,
      "loss": 1.5234,
      "theoretical_loss": 3.390418307964706,
      "tokens_seen": 2306473984
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062828282828283,
      "loss": 1.2173,
      "theoretical_loss": 3.3904104992310478,
      "tokens_seen": 2306539520
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030626262626262625,
      "loss": 1.2636,
      "theoretical_loss": 3.3904026907813787,
      "tokens_seen": 2306605056
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030624242424242426,
      "loss": 1.1215,
      "theoretical_loss": 3.39039488261568,
      "tokens_seen": 2306670592
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062222222222222,
      "loss": 1.1888,
      "theoretical_loss": 3.3903870747339333,
      "tokens_seen": 2306736128
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003062020202020202,
      "loss": 1.2218,
      "theoretical_loss": 3.3903792671361215,
      "tokens_seen": 2306801664
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061818181818182,
      "loss": 1.4681,
      "theoretical_loss": 3.3903714598222243,
      "tokens_seen": 2306867200
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030616161616161614,
      "loss": 1.3683,
      "theoretical_loss": 3.390363652792225,
      "tokens_seen": 2306932736
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030614141414141415,
      "loss": 1.3316,
      "theoretical_loss": 3.390355846046104,
      "tokens_seen": 2306998272
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061212121212121,
      "loss": 1.3096,
      "theoretical_loss": 3.3903480395838437,
      "tokens_seen": 2307063808
    },
    {
      "epoch": 0.39,
      "objective/train/aligned_token_count": 0.37829911708831787,
      "objective/train/avg_token_score": 0.009323440492153168,
      "objective/train/docs_used": 1299979,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.326233983039856,
      "objective/train/original_loss": 4.481205463409424,
      "objective/train/theoretical_loss": 3.3903402334054253,
      "objective/train/tokens_used": 665988576,
      "objective/train/unlikelihood_loss": 0.06780549138784409,
      "theoretical_loss": 3.3903402334054253,
      "tokens_seen": 2307129344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003061010101010101,
      "loss": 1.2669,
      "theoretical_loss": 3.3903402334054253,
      "tokens_seen": 2307129344
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003060808080808081,
      "loss": 1.4712,
      "theoretical_loss": 3.39033242751083,
      "tokens_seen": 2307194880
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030606060606060603,
      "loss": 1.2825,
      "theoretical_loss": 3.390324621900041,
      "tokens_seen": 2307260416
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030604040404040404,
      "loss": 1.4692,
      "theoretical_loss": 3.390316816573038,
      "tokens_seen": 2307325952
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030602020202020205,
      "loss": 1.2649,
      "theoretical_loss": 3.390309011529804,
      "tokens_seen": 2307391488
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000306,
      "loss": 1.2396,
      "theoretical_loss": 3.39030120677032,
      "tokens_seen": 2307457024
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030597979797979797,
      "loss": 1.1179,
      "theoretical_loss": 3.390293402294567,
      "tokens_seen": 2307522560
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.000305959595959596,
      "loss": 1.2811,
      "theoretical_loss": 3.3902855981025284,
      "tokens_seen": 2307588096
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030593939393939394,
      "loss": 1.3667,
      "theoretical_loss": 3.3902777941941844,
      "tokens_seen": 2307653632
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030591919191919195,
      "loss": 1.0795,
      "theoretical_loss": 3.390269990569517,
      "tokens_seen": 2307719168
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058989898989899,
      "loss": 1.3023,
      "theoretical_loss": 3.3902621872285077,
      "tokens_seen": 2307784704
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030587878787878786,
      "loss": 1.0989,
      "theoretical_loss": 3.3902543841711386,
      "tokens_seen": 2307850240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030585858585858587,
      "loss": 1.0674,
      "theoretical_loss": 3.3902465813973905,
      "tokens_seen": 2307915776
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003058383838383839,
      "loss": 1.2032,
      "theoretical_loss": 3.390238778907246,
      "tokens_seen": 2307981312
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030581818181818184,
      "loss": 1.3444,
      "theoretical_loss": 3.3902309767006864,
      "tokens_seen": 2308046848
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003057979797979798,
      "loss": 1.2769,
      "theoretical_loss": 3.390223174777693,
      "tokens_seen": 2308112384
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030577777777777775,
      "loss": 1.3699,
      "theoretical_loss": 3.390215373138248,
      "tokens_seen": 2308177920
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030575757575757576,
      "loss": 1.1034,
      "theoretical_loss": 3.390207571782332,
      "tokens_seen": 2308243456
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003057373737373738,
      "loss": 1.4012,
      "theoretical_loss": 3.390199770709928,
      "tokens_seen": 2308308992
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030571717171717173,
      "loss": 1.211,
      "theoretical_loss": 3.390191969921017,
      "tokens_seen": 2308374528
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003056969696969697,
      "loss": 1.2935,
      "theoretical_loss": 3.3901841694155803,
      "tokens_seen": 2308440064
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030567676767676765,
      "loss": 1.2698,
      "theoretical_loss": 3.3901763691936,
      "tokens_seen": 2308505600
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003056565656565657,
      "loss": 1.2488,
      "theoretical_loss": 3.390168569255058,
      "tokens_seen": 2308571136
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.00030563636363636367,
      "loss": 1.3065,
      "theoretical_loss": 3.3901607695999356,
      "tokens_seen": 2308636672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0003056161616161616,
      "loss": 1.2432,
      "theoretical_loss": 3.390152970228214,
      "tokens_seen": 2308702208
    },
    {
      "epoch": 0.39,
      "objective/train/aligned_token_count": 0.35997068881988525,
      "objective/train/avg_token_score": 0.012620855122804642,
      "objective/train/docs_used": 1301383,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2265704870224,
      "objective/train/original_loss": 4.811745643615723,
      "objective/train/theoretical_loss": 3.390145171139876,
      "objective/train/tokens_used": 667626976,
      "objective/train/unlikelihood_loss": 0.04698840156197548,
      "theoretical_loss": 3.390145171139876,
      "tokens_seen": 2308767744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003055959595959596,
      "loss": 1.3292,
      "theoretical_loss": 3.390145171139876,
      "tokens_seen": 2308767744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030557575757575754,
      "loss": 1.2977,
      "theoretical_loss": 3.390137372334902,
      "tokens_seen": 2308833280
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003055555555555556,
      "loss": 1.2004,
      "theoretical_loss": 3.3901295738132746,
      "tokens_seen": 2308898816
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030553535353535356,
      "loss": 1.3133,
      "theoretical_loss": 3.390121775574975,
      "tokens_seen": 2308964352
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003055151515151515,
      "loss": 1.4018,
      "theoretical_loss": 3.3901139776199853,
      "tokens_seen": 2309029888
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054949494949495,
      "loss": 1.4115,
      "theoretical_loss": 3.390106179948287,
      "tokens_seen": 2309095424
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054747474747475,
      "loss": 1.2831,
      "theoretical_loss": 3.390098382559861,
      "tokens_seen": 2309160960
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054545454545455,
      "loss": 1.2944,
      "theoretical_loss": 3.39009058545469,
      "tokens_seen": 2309226496
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030543434343434345,
      "loss": 1.2277,
      "theoretical_loss": 3.390082788632755,
      "tokens_seen": 2309292032
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003054141414141414,
      "loss": 1.3132,
      "theoretical_loss": 3.3900749920940383,
      "tokens_seen": 2309357568
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030539393939393937,
      "loss": 1.4092,
      "theoretical_loss": 3.3900671958385207,
      "tokens_seen": 2309423104
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053737373737374,
      "loss": 1.1246,
      "theoretical_loss": 3.390059399866185,
      "tokens_seen": 2309488640
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053535353535354,
      "loss": 1.3691,
      "theoretical_loss": 3.390051604177012,
      "tokens_seen": 2309554176
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030533333333333335,
      "loss": 1.3168,
      "theoretical_loss": 3.3900438087709834,
      "tokens_seen": 2309619712
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003053131313131313,
      "loss": 1.3891,
      "theoretical_loss": 3.3900360136480816,
      "tokens_seen": 2309685248
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003052929292929293,
      "loss": 1.2372,
      "theoretical_loss": 3.3900282188082875,
      "tokens_seen": 2309750784
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030527272727272727,
      "loss": 1.271,
      "theoretical_loss": 3.3900204242515835,
      "tokens_seen": 2309816320
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003052525252525253,
      "loss": 1.393,
      "theoretical_loss": 3.3900126299779503,
      "tokens_seen": 2309881856
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030523232323232324,
      "loss": 1.4966,
      "theoretical_loss": 3.3900048359873707,
      "tokens_seen": 2309947392
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003052121212121212,
      "loss": 1.2359,
      "theoretical_loss": 3.3899970422798256,
      "tokens_seen": 2310012928
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003051919191919192,
      "loss": 1.3159,
      "theoretical_loss": 3.389989248855297,
      "tokens_seen": 2310078464
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030517171717171716,
      "loss": 1.3298,
      "theoretical_loss": 3.389981455713767,
      "tokens_seen": 2310144000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003051515151515152,
      "loss": 1.3,
      "theoretical_loss": 3.3899736628552164,
      "tokens_seen": 2310209536
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030513131313131313,
      "loss": 1.4629,
      "theoretical_loss": 3.389965870279627,
      "tokens_seen": 2310275072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030511111111111114,
      "loss": 1.3711,
      "theoretical_loss": 3.389958077986982,
      "tokens_seen": 2310340608
    },
    {
      "epoch": 0.4,
      "objective/train/aligned_token_count": 0.3856304883956909,
      "objective/train/avg_token_score": 0.024019302800297737,
      "objective/train/docs_used": 1301708,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.323414921760559,
      "objective/train/original_loss": 4.8101348876953125,
      "objective/train/theoretical_loss": 3.3899502859772612,
      "objective/train/tokens_used": 669265376,
      "objective/train/unlikelihood_loss": 0.039997730404138565,
      "theoretical_loss": 3.3899502859772612,
      "tokens_seen": 2310406144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003050909090909091,
      "loss": 1.268,
      "theoretical_loss": 3.3899502859772612,
      "tokens_seen": 2310406144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030507070707070706,
      "loss": 1.4568,
      "theoretical_loss": 3.389942494250447,
      "tokens_seen": 2310471680
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030505050505050507,
      "loss": 1.046,
      "theoretical_loss": 3.389934702806521,
      "tokens_seen": 2310537216
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000305030303030303,
      "loss": 1.2006,
      "theoretical_loss": 3.389926911645466,
      "tokens_seen": 2310602752
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030501010101010104,
      "loss": 1.4522,
      "theoretical_loss": 3.389919120767262,
      "tokens_seen": 2310668288
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000304989898989899,
      "loss": 1.4433,
      "theoretical_loss": 3.3899113301718913,
      "tokens_seen": 2310733824
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030496969696969695,
      "loss": 1.2675,
      "theoretical_loss": 3.3899035398593362,
      "tokens_seen": 2310799360
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030494949494949496,
      "loss": 1.0547,
      "theoretical_loss": 3.389895749829578,
      "tokens_seen": 2310864896
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030492929292929297,
      "loss": 1.315,
      "theoretical_loss": 3.389887960082598,
      "tokens_seen": 2310930432
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030490909090909093,
      "loss": 1.3819,
      "theoretical_loss": 3.3898801706183788,
      "tokens_seen": 2310995968
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003048888888888889,
      "loss": 1.4238,
      "theoretical_loss": 3.3898723814369016,
      "tokens_seen": 2311061504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030486868686868684,
      "loss": 1.3426,
      "theoretical_loss": 3.389864592538148,
      "tokens_seen": 2311127040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030484848484848485,
      "loss": 1.3422,
      "theoretical_loss": 3.3898568039221004,
      "tokens_seen": 2311192576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030482828282828286,
      "loss": 1.528,
      "theoretical_loss": 3.3898490155887395,
      "tokens_seen": 2311258112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003048080808080808,
      "loss": 1.5163,
      "theoretical_loss": 3.3898412275380476,
      "tokens_seen": 2311323648
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003047878787878788,
      "loss": 1.4546,
      "theoretical_loss": 3.3898334397700065,
      "tokens_seen": 2311389184
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030476767676767673,
      "loss": 1.2827,
      "theoretical_loss": 3.389825652284598,
      "tokens_seen": 2311454720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003047474747474748,
      "loss": 1.2328,
      "theoretical_loss": 3.3898178650818034,
      "tokens_seen": 2311520256
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030472727272727276,
      "loss": 0.8864,
      "theoretical_loss": 3.389810078161605,
      "tokens_seen": 2311585792
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003047070707070707,
      "loss": 1.2343,
      "theoretical_loss": 3.3898022915239836,
      "tokens_seen": 2311651328
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030468686868686867,
      "loss": 1.474,
      "theoretical_loss": 3.389794505168922,
      "tokens_seen": 2311716864
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030466666666666663,
      "loss": 1.2737,
      "theoretical_loss": 3.389786719096401,
      "tokens_seen": 2311782400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003046464646464647,
      "loss": 1.2406,
      "theoretical_loss": 3.389778933306403,
      "tokens_seen": 2311847936
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030462626262626265,
      "loss": 1.2341,
      "theoretical_loss": 3.3897711477989096,
      "tokens_seen": 2311913472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003046060606060606,
      "loss": 1.1053,
      "theoretical_loss": 3.3897633625739028,
      "tokens_seen": 2311979008
    },
    {
      "epoch": 0.4,
      "objective/train/aligned_token_count": 0.30816227197647095,
      "objective/train/avg_token_score": 0.013038579374551773,
      "objective/train/docs_used": 1302992,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1209722757339478,
      "objective/train/original_loss": 5.080594062805176,
      "objective/train/theoretical_loss": 3.3897555776313637,
      "objective/train/tokens_used": 670903776,
      "objective/train/unlikelihood_loss": 0.06270943582057953,
      "theoretical_loss": 3.3897555776313637,
      "tokens_seen": 2312044544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030458585858585856,
      "loss": 1.3195,
      "theoretical_loss": 3.3897555776313637,
      "tokens_seen": 2312044544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003045656565656565,
      "loss": 1.3964,
      "theoretical_loss": 3.3897477929712743,
      "tokens_seen": 2312110080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003045454545454546,
      "loss": 1.3246,
      "theoretical_loss": 3.3897400085936167,
      "tokens_seen": 2312175616
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030452525252525254,
      "loss": 1.3529,
      "theoretical_loss": 3.3897322244983723,
      "tokens_seen": 2312241152
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003045050505050505,
      "loss": 1.3689,
      "theoretical_loss": 3.389724440685523,
      "tokens_seen": 2312306688
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030448484848484846,
      "loss": 1.2891,
      "theoretical_loss": 3.3897166571550503,
      "tokens_seen": 2312372224
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030446464646464647,
      "loss": 1.4361,
      "theoretical_loss": 3.389708873906936,
      "tokens_seen": 2312437760
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003044444444444445,
      "loss": 1.1947,
      "theoretical_loss": 3.3897010909411622,
      "tokens_seen": 2312503296
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030442424242424243,
      "loss": 1.3051,
      "theoretical_loss": 3.3896933082577103,
      "tokens_seen": 2312568832
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003044040404040404,
      "loss": 1.0764,
      "theoretical_loss": 3.3896855258565624,
      "tokens_seen": 2312634368
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030438383838383835,
      "loss": 1.2603,
      "theoretical_loss": 3.3896777437376997,
      "tokens_seen": 2312699904
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030436363636363636,
      "loss": 1.3683,
      "theoretical_loss": 3.3896699619011046,
      "tokens_seen": 2312765440
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030434343434343437,
      "loss": 1.297,
      "theoretical_loss": 3.389662180346759,
      "tokens_seen": 2312830976
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030432323232323233,
      "loss": 1.559,
      "theoretical_loss": 3.3896543990746437,
      "tokens_seen": 2312896512
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003043030303030303,
      "loss": 1.1921,
      "theoretical_loss": 3.3896466180847407,
      "tokens_seen": 2312962048
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042828282828283,
      "loss": 1.1718,
      "theoretical_loss": 3.3896388373770328,
      "tokens_seen": 2313027584
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042626262626263,
      "loss": 1.285,
      "theoretical_loss": 3.389631056951501,
      "tokens_seen": 2313093120
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030424242424242426,
      "loss": 1.4129,
      "theoretical_loss": 3.389623276808127,
      "tokens_seen": 2313158656
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042222222222222,
      "loss": 1.1192,
      "theoretical_loss": 3.389615496946892,
      "tokens_seen": 2313224192
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003042020202020202,
      "loss": 1.3893,
      "theoretical_loss": 3.389607717367779,
      "tokens_seen": 2313289728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003041818181818182,
      "loss": 1.0889,
      "theoretical_loss": 3.3895999380707695,
      "tokens_seen": 2313355264
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003041616161616162,
      "loss": 1.5004,
      "theoretical_loss": 3.389592159055845,
      "tokens_seen": 2313420800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030414141414141416,
      "loss": 1.1498,
      "theoretical_loss": 3.389584380322987,
      "tokens_seen": 2313486336
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003041212121212121,
      "loss": 1.3367,
      "theoretical_loss": 3.389576601872178,
      "tokens_seen": 2313551872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003041010101010101,
      "loss": 1.1432,
      "theoretical_loss": 3.389568823703399,
      "tokens_seen": 2313617408
    },
    {
      "epoch": 0.4,
      "objective/train/aligned_token_count": 0.29899805784225464,
      "objective/train/avg_token_score": 0.03673693910241127,
      "objective/train/docs_used": 1303654,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9932365417480469,
      "objective/train/original_loss": 4.882896900177002,
      "objective/train/theoretical_loss": 3.3895610458166323,
      "objective/train/tokens_used": 672542176,
      "objective/train/unlikelihood_loss": 0.057898398488759995,
      "theoretical_loss": 3.3895610458166323,
      "tokens_seen": 2313682944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003040808080808081,
      "loss": 1.2048,
      "theoretical_loss": 3.3895610458166323,
      "tokens_seen": 2313682944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003040606060606061,
      "loss": 1.1649,
      "theoretical_loss": 3.3895532682118596,
      "tokens_seen": 2313748480
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030404040404040405,
      "loss": 1.1116,
      "theoretical_loss": 3.3895454908890628,
      "tokens_seen": 2313814016
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000304020202020202,
      "loss": 1.4856,
      "theoretical_loss": 3.389537713848223,
      "tokens_seen": 2313879552
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000304,
      "loss": 1.6482,
      "theoretical_loss": 3.389529937089323,
      "tokens_seen": 2313945088
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303979797979798,
      "loss": 1.3507,
      "theoretical_loss": 3.3895221606123442,
      "tokens_seen": 2314010624
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303959595959596,
      "loss": 1.2936,
      "theoretical_loss": 3.389514384417268,
      "tokens_seen": 2314076160
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030393939393939394,
      "loss": 1.3648,
      "theoretical_loss": 3.389506608504077,
      "tokens_seen": 2314141696
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030391919191919195,
      "loss": 1.1018,
      "theoretical_loss": 3.3894988328727518,
      "tokens_seen": 2314207232
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003038989898989899,
      "loss": 1.382,
      "theoretical_loss": 3.3894910575232755,
      "tokens_seen": 2314272768
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030387878787878787,
      "loss": 1.2937,
      "theoretical_loss": 3.389483282455629,
      "tokens_seen": 2314338304
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003038585858585859,
      "loss": 1.2498,
      "theoretical_loss": 3.389475507669794,
      "tokens_seen": 2314403840
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030383838383838383,
      "loss": 1.3394,
      "theoretical_loss": 3.3894677331657537,
      "tokens_seen": 2314469376
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030381818181818185,
      "loss": 1.4034,
      "theoretical_loss": 3.3894599589434886,
      "tokens_seen": 2314534912
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003037979797979798,
      "loss": 1.3533,
      "theoretical_loss": 3.3894521850029804,
      "tokens_seen": 2314600448
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030377777777777776,
      "loss": 1.4532,
      "theoretical_loss": 3.3894444113442117,
      "tokens_seen": 2314665984
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030375757575757577,
      "loss": 1.1857,
      "theoretical_loss": 3.389436637967164,
      "tokens_seen": 2314731520
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003037373737373738,
      "loss": 1.4857,
      "theoretical_loss": 3.389428864871819,
      "tokens_seen": 2314797056
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030371717171717174,
      "loss": 1.377,
      "theoretical_loss": 3.389421092058159,
      "tokens_seen": 2314862592
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003036969696969697,
      "loss": 1.4042,
      "theoretical_loss": 3.389413319526165,
      "tokens_seen": 2314928128
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030367676767676765,
      "loss": 1.5411,
      "theoretical_loss": 3.389405547275819,
      "tokens_seen": 2314993664
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030365656565656566,
      "loss": 1.3192,
      "theoretical_loss": 3.3893977753071036,
      "tokens_seen": 2315059200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003036363636363637,
      "loss": 1.4209,
      "theoretical_loss": 3.3893900036199995,
      "tokens_seen": 2315124736
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030361616161616163,
      "loss": 1.3064,
      "theoretical_loss": 3.3893822322144898,
      "tokens_seen": 2315190272
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003035959595959596,
      "loss": 1.4298,
      "theoretical_loss": 3.389374461090555,
      "tokens_seen": 2315255808
    },
    {
      "epoch": 0.4,
      "objective/train/aligned_token_count": 0.47629520297050476,
      "objective/train/avg_token_score": 0.007926389575004578,
      "objective/train/docs_used": 1304892,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.557091236114502,
      "objective/train/original_loss": 4.0050249099731445,
      "objective/train/theoretical_loss": 3.3893666902481776,
      "objective/train/tokens_used": 674180576,
      "objective/train/unlikelihood_loss": 0.06536270678043365,
      "theoretical_loss": 3.3893666902481776,
      "tokens_seen": 2315321344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030357575757575754,
      "loss": 1.3845,
      "theoretical_loss": 3.3893666902481776,
      "tokens_seen": 2315321344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003035555555555556,
      "loss": 1.2308,
      "theoretical_loss": 3.38935891968734,
      "tokens_seen": 2315386880
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030353535353535357,
      "loss": 1.2302,
      "theoretical_loss": 3.389351149408023,
      "tokens_seen": 2315452416
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003035151515151515,
      "loss": 1.4393,
      "theoretical_loss": 3.389343379410209,
      "tokens_seen": 2315517952
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034949494949495,
      "loss": 1.4026,
      "theoretical_loss": 3.3893356096938794,
      "tokens_seen": 2315583488
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030347474747474744,
      "loss": 1.3512,
      "theoretical_loss": 3.389327840259017,
      "tokens_seen": 2315649024
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034545454545455,
      "loss": 1.3894,
      "theoretical_loss": 3.389320071105602,
      "tokens_seen": 2315714560
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030343434343434346,
      "loss": 1.1371,
      "theoretical_loss": 3.3893123022336176,
      "tokens_seen": 2315780096
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003034141414141414,
      "loss": 1.1908,
      "theoretical_loss": 3.3893045336430454,
      "tokens_seen": 2315845632
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003033939393939394,
      "loss": 1.473,
      "theoretical_loss": 3.389296765333867,
      "tokens_seen": 2315911168
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003033737373737374,
      "loss": 1.4724,
      "theoretical_loss": 3.3892889973060645,
      "tokens_seen": 2315976704
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003033535353535354,
      "loss": 1.1666,
      "theoretical_loss": 3.3892812295596193,
      "tokens_seen": 2316042240
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030333333333333335,
      "loss": 1.3499,
      "theoretical_loss": 3.3892734620945135,
      "tokens_seen": 2316107776
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003033131313131313,
      "loss": 1.2771,
      "theoretical_loss": 3.3892656949107294,
      "tokens_seen": 2316173312
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030329292929292927,
      "loss": 1.5259,
      "theoretical_loss": 3.3892579280082478,
      "tokens_seen": 2316238848
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003032727272727273,
      "loss": 1.3074,
      "theoretical_loss": 3.389250161387052,
      "tokens_seen": 2316304384
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003032525252525253,
      "loss": 1.5571,
      "theoretical_loss": 3.3892423950471224,
      "tokens_seen": 2316369920
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030323232323232324,
      "loss": 1.3977,
      "theoretical_loss": 3.389234628988442,
      "tokens_seen": 2316435456
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003032121212121212,
      "loss": 1.3522,
      "theoretical_loss": 3.3892268632109914,
      "tokens_seen": 2316500992
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003031919191919192,
      "loss": 1.6511,
      "theoretical_loss": 3.389219097714754,
      "tokens_seen": 2316566528
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030317171717171717,
      "loss": 1.3572,
      "theoretical_loss": 3.38921133249971,
      "tokens_seen": 2316632064
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003031515151515152,
      "loss": 1.0567,
      "theoretical_loss": 3.389203567565843,
      "tokens_seen": 2316697600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030313131313131314,
      "loss": 1.3032,
      "theoretical_loss": 3.3891958029131333,
      "tokens_seen": 2316763136
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003031111111111111,
      "loss": 1.3153,
      "theoretical_loss": 3.389188038541564,
      "tokens_seen": 2316828672
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003030909090909091,
      "loss": 1.5591,
      "theoretical_loss": 3.3891802744511166,
      "tokens_seen": 2316894208
    },
    {
      "debugging/Self-BLEU-5": 0.15659319952366813,
      "debugging/distinct-1-grams": 0.7962866038742501,
      "debugging/distinct-2-grams": 0.9550000000000001,
      "debugging/entropy-1-grams": 4.929609510498728,
      "debugging/entropy-2-grams": 5.4030103897236375,
      "debugging/length": 474.3333333333333,
      "debugging/num_segments": 3,
      "epoch": 0.4,
      "objective/train/aligned_token_count": 0.43438416719436646,
      "objective/train/avg_token_score": 0.018595926463603973,
      "objective/train/docs_used": 1305637,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3358829021453857,
      "objective/train/original_loss": 4.557432174682617,
      "objective/train/theoretical_loss": 3.3891725106417723,
      "objective/train/tokens_used": 675818976,
      "objective/train/unlikelihood_loss": 0.04952618479728699,
      "theoretical_loss": 3.3891725106417723,
      "tokens_seen": 2316959744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030307070707070706,
      "loss": 1.5101,
      "theoretical_loss": 3.3891725106417723,
      "tokens_seen": 2316959744
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003030505050505051,
      "loss": 1.6241,
      "theoretical_loss": 3.3891647471135133,
      "tokens_seen": 2317025280
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030303030303030303,
      "loss": 1.3985,
      "theoretical_loss": 3.3891569838663225,
      "tokens_seen": 2317090816
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000303010101010101,
      "loss": 1.5132,
      "theoretical_loss": 3.38914922090018,
      "tokens_seen": 2317156352
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302989898989899,
      "loss": 1.4192,
      "theoretical_loss": 3.389141458215069,
      "tokens_seen": 2317221888
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030296969696969695,
      "loss": 1.3171,
      "theoretical_loss": 3.3891336958109712,
      "tokens_seen": 2317287424
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030294949494949497,
      "loss": 1.2823,
      "theoretical_loss": 3.3891259336878683,
      "tokens_seen": 2317352960
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003029292929292929,
      "loss": 1.1762,
      "theoretical_loss": 3.389118171845742,
      "tokens_seen": 2317418496
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030290909090909093,
      "loss": 1.3718,
      "theoretical_loss": 3.389110410284574,
      "tokens_seen": 2317484032
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003028888888888889,
      "loss": 1.2652,
      "theoretical_loss": 3.3891026490043465,
      "tokens_seen": 2317549568
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030286868686868685,
      "loss": 1.3804,
      "theoretical_loss": 3.389094888005042,
      "tokens_seen": 2317615104
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030284848484848486,
      "loss": 1.371,
      "theoretical_loss": 3.3890871272866416,
      "tokens_seen": 2317680640
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003028282828282828,
      "loss": 1.3334,
      "theoretical_loss": 3.389079366849127,
      "tokens_seen": 2317746176
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003028080808080808,
      "loss": 1.3739,
      "theoretical_loss": 3.3890716066924806,
      "tokens_seen": 2317811712
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003027878787878788,
      "loss": 1.2034,
      "theoretical_loss": 3.3890638468166845,
      "tokens_seen": 2317877248
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030276767676767674,
      "loss": 1.5391,
      "theoretical_loss": 3.38905608722172,
      "tokens_seen": 2317942784
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030274747474747475,
      "loss": 1.2926,
      "theoretical_loss": 3.3890483279075694,
      "tokens_seen": 2318008320
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030272727272727276,
      "loss": 1.2363,
      "theoretical_loss": 3.3890405688742145,
      "tokens_seen": 2318073856
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003027070707070707,
      "loss": 1.3786,
      "theoretical_loss": 3.3890328101216367,
      "tokens_seen": 2318139392
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003026868686868687,
      "loss": 1.1363,
      "theoretical_loss": 3.389025051649819,
      "tokens_seen": 2318204928
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030266666666666663,
      "loss": 1.3674,
      "theoretical_loss": 3.389017293458742,
      "tokens_seen": 2318270464
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030264646464646464,
      "loss": 1.1345,
      "theoretical_loss": 3.389009535548389,
      "tokens_seen": 2318336000
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030262626262626266,
      "loss": 1.4569,
      "theoretical_loss": 3.3890017779187405,
      "tokens_seen": 2318401536
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003026060606060606,
      "loss": 1.1793,
      "theoretical_loss": 3.3889940205697795,
      "tokens_seen": 2318467072
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030258585858585857,
      "loss": 1.3625,
      "theoretical_loss": 3.3889862635014874,
      "tokens_seen": 2318532608
    },
    {
      "epoch": 0.4,
      "objective/train/aligned_token_count": 0.3646138906478882,
      "objective/train/avg_token_score": 0.055039163678884506,
      "objective/train/docs_used": 1307149,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.19109046459198,
      "objective/train/original_loss": 4.393966197967529,
      "objective/train/theoretical_loss": 3.3889785067138463,
      "objective/train/tokens_used": 677457376,
      "objective/train/unlikelihood_loss": 0.06055895984172821,
      "theoretical_loss": 3.3889785067138463,
      "tokens_seen": 2318598144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003025656565656566,
      "loss": 1.4126,
      "theoretical_loss": 3.3889785067138463,
      "tokens_seen": 2318598144
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003025454545454546,
      "loss": 1.1879,
      "theoretical_loss": 3.3889707502068376,
      "tokens_seen": 2318663680
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030252525252525255,
      "loss": 1.3418,
      "theoretical_loss": 3.388962993980444,
      "tokens_seen": 2318729216
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003025050505050505,
      "loss": 1.4377,
      "theoretical_loss": 3.3889552380346473,
      "tokens_seen": 2318794752
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030248484848484846,
      "loss": 1.2935,
      "theoretical_loss": 3.3889474823694288,
      "tokens_seen": 2318860288
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030246464646464647,
      "loss": 1.3236,
      "theoretical_loss": 3.3889397269847708,
      "tokens_seen": 2318925824
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003024444444444445,
      "loss": 1.4837,
      "theoretical_loss": 3.3889319718806554,
      "tokens_seen": 2318991360
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030242424242424244,
      "loss": 1.1537,
      "theoretical_loss": 3.388924217057064,
      "tokens_seen": 2319056896
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003024040404040404,
      "loss": 1.3011,
      "theoretical_loss": 3.388916462513979,
      "tokens_seen": 2319122432
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030238383838383835,
      "loss": 1.327,
      "theoretical_loss": 3.3889087082513827,
      "tokens_seen": 2319187968
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003023636363636364,
      "loss": 1.2941,
      "theoretical_loss": 3.388900954269256,
      "tokens_seen": 2319253504
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003023434343434344,
      "loss": 1.3944,
      "theoretical_loss": 3.388893200567581,
      "tokens_seen": 2319319040
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030232323232323233,
      "loss": 1.4892,
      "theoretical_loss": 3.3888854471463405,
      "tokens_seen": 2319384576
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003023030303030303,
      "loss": 1.2106,
      "theoretical_loss": 3.388877694005516,
      "tokens_seen": 2319450112
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030228282828282825,
      "loss": 1.5681,
      "theoretical_loss": 3.388869941145089,
      "tokens_seen": 2319515648
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003022626262626263,
      "loss": 1.2976,
      "theoretical_loss": 3.3888621885650423,
      "tokens_seen": 2319581184
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030224242424242427,
      "loss": 1.1376,
      "theoretical_loss": 3.388854436265357,
      "tokens_seen": 2319646720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003022222222222222,
      "loss": 1.5385,
      "theoretical_loss": 3.3888466842460154,
      "tokens_seen": 2319712256
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003022020202020202,
      "loss": 1.5192,
      "theoretical_loss": 3.3888389325069994,
      "tokens_seen": 2319777792
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021818181818182,
      "loss": 1.2542,
      "theoretical_loss": 3.3888311810482907,
      "tokens_seen": 2319843328
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021616161616162,
      "loss": 1.334,
      "theoretical_loss": 3.388823429869872,
      "tokens_seen": 2319908864
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030214141414141416,
      "loss": 1.1614,
      "theoretical_loss": 3.3888156789717243,
      "tokens_seen": 2319974400
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021212121212121,
      "loss": 1.2211,
      "theoretical_loss": 3.38880792835383,
      "tokens_seen": 2320039936
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003021010101010101,
      "loss": 1.6699,
      "theoretical_loss": 3.3888001780161714,
      "tokens_seen": 2320105472
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003020808080808081,
      "loss": 1.2965,
      "theoretical_loss": 3.3887924279587303,
      "tokens_seen": 2320171008
    },
    {
      "epoch": 0.4,
      "objective/train/aligned_token_count": 0.37616080045700073,
      "objective/train/avg_token_score": 0.058650314807891846,
      "objective/train/docs_used": 1307847,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2421090602874756,
      "objective/train/original_loss": 4.878607749938965,
      "objective/train/theoretical_loss": 3.388784678181488,
      "objective/train/tokens_used": 679095776,
      "objective/train/unlikelihood_loss": 0.04421050846576691,
      "theoretical_loss": 3.388784678181488,
      "tokens_seen": 2320236544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003020606060606061,
      "loss": 1.2704,
      "theoretical_loss": 3.388784678181488,
      "tokens_seen": 2320236544
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030204040404040405,
      "loss": 1.5955,
      "theoretical_loss": 3.388776928684427,
      "tokens_seen": 2320302080
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302020202020202,
      "loss": 1.4401,
      "theoretical_loss": 3.388769179467529,
      "tokens_seen": 2320367616
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000302,
      "loss": 1.2776,
      "theoretical_loss": 3.3887614305307765,
      "tokens_seen": 2320433152
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301979797979798,
      "loss": 1.2772,
      "theoretical_loss": 3.388753681874151,
      "tokens_seen": 2320498688
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301959595959596,
      "loss": 1.2832,
      "theoretical_loss": 3.3887459334976344,
      "tokens_seen": 2320564224
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030193939393939395,
      "loss": 1.2607,
      "theoretical_loss": 3.3887381854012086,
      "tokens_seen": 2320629760
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003019191919191919,
      "loss": 1.5267,
      "theoretical_loss": 3.388730437584856,
      "tokens_seen": 2320695296
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003018989898989899,
      "loss": 1.3252,
      "theoretical_loss": 3.3887226900485583,
      "tokens_seen": 2320760832
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030187878787878787,
      "loss": 1.3564,
      "theoretical_loss": 3.388714942792298,
      "tokens_seen": 2320826368
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003018585858585859,
      "loss": 1.4979,
      "theoretical_loss": 3.388707195816056,
      "tokens_seen": 2320891904
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030183838383838384,
      "loss": 1.2916,
      "theoretical_loss": 3.388699449119815,
      "tokens_seen": 2320957440
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030181818181818185,
      "loss": 1.4677,
      "theoretical_loss": 3.3886917027035572,
      "tokens_seen": 2321022976
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003017979797979798,
      "loss": 1.4057,
      "theoretical_loss": 3.3886839565672635,
      "tokens_seen": 2321088512
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030177777777777776,
      "loss": 1.1926,
      "theoretical_loss": 3.3886762107109174,
      "tokens_seen": 2321154048
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003017575757575758,
      "loss": 1.5298,
      "theoretical_loss": 3.3886684651344994,
      "tokens_seen": 2321219584
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030173737373737373,
      "loss": 1.2734,
      "theoretical_loss": 3.3886607198379926,
      "tokens_seen": 2321285120
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030171717171717174,
      "loss": 1.2796,
      "theoretical_loss": 3.388652974821378,
      "tokens_seen": 2321350656
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003016969696969697,
      "loss": 1.4198,
      "theoretical_loss": 3.3886452300846384,
      "tokens_seen": 2321416192
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030167676767676766,
      "loss": 1.4449,
      "theoretical_loss": 3.3886374856277555,
      "tokens_seen": 2321481728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030165656565656567,
      "loss": 1.1509,
      "theoretical_loss": 3.3886297414507114,
      "tokens_seen": 2321547264
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003016363636363637,
      "loss": 1.2739,
      "theoretical_loss": 3.388621997553488,
      "tokens_seen": 2321612800
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030161616161616164,
      "loss": 1.4117,
      "theoretical_loss": 3.388614253936067,
      "tokens_seen": 2321678336
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003015959595959596,
      "loss": 1.5221,
      "theoretical_loss": 3.3886065105984304,
      "tokens_seen": 2321743872
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030157575757575755,
      "loss": 1.357,
      "theoretical_loss": 3.3885987675405613,
      "tokens_seen": 2321809408
    },
    {
      "epoch": 0.4,
      "objective/train/aligned_token_count": 0.39320626854896545,
      "objective/train/avg_token_score": 0.014931839890778065,
      "objective/train/docs_used": 1309222,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2071703672409058,
      "objective/train/original_loss": 4.620835304260254,
      "objective/train/theoretical_loss": 3.38859102476244,
      "objective/train/tokens_used": 680734176,
      "objective/train/unlikelihood_loss": 0.048910029232501984,
      "theoretical_loss": 3.38859102476244,
      "tokens_seen": 2321874944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030155555555555556,
      "loss": 1.1984,
      "theoretical_loss": 3.38859102476244,
      "tokens_seen": 2321874944
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030153535353535357,
      "loss": 1.4257,
      "theoretical_loss": 3.38858328226405,
      "tokens_seen": 2321940480
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030151515151515153,
      "loss": 1.5666,
      "theoretical_loss": 3.3885755400453723,
      "tokens_seen": 2322006016
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003014949494949495,
      "loss": 1.4227,
      "theoretical_loss": 3.388567798106389,
      "tokens_seen": 2322071552
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030147474747474744,
      "loss": 1.4813,
      "theoretical_loss": 3.388560056447083,
      "tokens_seen": 2322137088
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003014545454545455,
      "loss": 1.6086,
      "theoretical_loss": 3.3885523150674355,
      "tokens_seen": 2322202624
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030143434343434346,
      "loss": 1.363,
      "theoretical_loss": 3.3885445739674283,
      "tokens_seen": 2322268160
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003014141414141414,
      "loss": 1.4199,
      "theoretical_loss": 3.388536833147044,
      "tokens_seen": 2322333696
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003013939393939394,
      "loss": 1.364,
      "theoretical_loss": 3.388529092606264,
      "tokens_seen": 2322399232
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030137373737373734,
      "loss": 1.4825,
      "theoretical_loss": 3.3885213523450712,
      "tokens_seen": 2322464768
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003013535353535354,
      "loss": 1.1633,
      "theoretical_loss": 3.3885136123634467,
      "tokens_seen": 2322530304
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030133333333333336,
      "loss": 1.3236,
      "theoretical_loss": 3.3885058726613733,
      "tokens_seen": 2322595840
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003013131313131313,
      "loss": 1.4534,
      "theoretical_loss": 3.388498133238832,
      "tokens_seen": 2322661376
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030129292929292927,
      "loss": 1.3583,
      "theoretical_loss": 3.388490394095806,
      "tokens_seen": 2322726912
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030127272727272723,
      "loss": 1.4128,
      "theoretical_loss": 3.3884826552322767,
      "tokens_seen": 2322792448
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003012525252525253,
      "loss": 1.2525,
      "theoretical_loss": 3.388474916648226,
      "tokens_seen": 2322857984
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030123232323232325,
      "loss": 1.3705,
      "theoretical_loss": 3.388467178343636,
      "tokens_seen": 2322923520
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003012121212121212,
      "loss": 1.326,
      "theoretical_loss": 3.3884594403184893,
      "tokens_seen": 2322989056
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030119191919191916,
      "loss": 1.4207,
      "theoretical_loss": 3.3884517025727674,
      "tokens_seen": 2323054592
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003011717171717172,
      "loss": 1.3393,
      "theoretical_loss": 3.388443965106452,
      "tokens_seen": 2323120128
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003011515151515152,
      "loss": 1.45,
      "theoretical_loss": 3.3884362279195255,
      "tokens_seen": 2323185664
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030113131313131314,
      "loss": 1.3483,
      "theoretical_loss": 3.3884284910119704,
      "tokens_seen": 2323251200
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003011111111111111,
      "loss": 1.3011,
      "theoretical_loss": 3.388420754383768,
      "tokens_seen": 2323316736
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030109090909090906,
      "loss": 1.4264,
      "theoretical_loss": 3.3884130180349006,
      "tokens_seen": 2323382272
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030107070707070707,
      "loss": 1.3491,
      "theoretical_loss": 3.3884052819653503,
      "tokens_seen": 2323447808
    },
    {
      "epoch": 0.4,
      "objective/train/aligned_token_count": 0.35599949955940247,
      "objective/train/avg_token_score": 0.013976586051285267,
      "objective/train/docs_used": 1309762,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2380194664001465,
      "objective/train/original_loss": 4.7337870597839355,
      "objective/train/theoretical_loss": 3.388397546175099,
      "objective/train/tokens_used": 682372576,
      "objective/train/unlikelihood_loss": 0.06674149632453918,
      "theoretical_loss": 3.388397546175099,
      "tokens_seen": 2323513344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003010505050505051,
      "loss": 1.4418,
      "theoretical_loss": 3.388397546175099,
      "tokens_seen": 2323513344
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030103030303030304,
      "loss": 1.2086,
      "theoretical_loss": 3.388389810664129,
      "tokens_seen": 2323578880
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000301010101010101,
      "loss": 1.3186,
      "theoretical_loss": 3.3883820754324216,
      "tokens_seen": 2323644416
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000300989898989899,
      "loss": 1.4295,
      "theoretical_loss": 3.3883743404799596,
      "tokens_seen": 2323709952
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030096969696969696,
      "loss": 1.4046,
      "theoretical_loss": 3.388366605806725,
      "tokens_seen": 2323775488
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030094949494949497,
      "loss": 1.3986,
      "theoretical_loss": 3.3883588714127,
      "tokens_seen": 2323841024
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030092929292929293,
      "loss": 1.404,
      "theoretical_loss": 3.388351137297866,
      "tokens_seen": 2323906560
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003009090909090909,
      "loss": 1.3683,
      "theoretical_loss": 3.3883434034622057,
      "tokens_seen": 2323972096
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003008888888888889,
      "loss": 1.2672,
      "theoretical_loss": 3.3883356699057003,
      "tokens_seen": 2324037632
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030086868686868685,
      "loss": 1.2514,
      "theoretical_loss": 3.3883279366283325,
      "tokens_seen": 2324103168
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030084848484848486,
      "loss": 1.4696,
      "theoretical_loss": 3.3883202036300846,
      "tokens_seen": 2324168704
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003008282828282828,
      "loss": 1.2462,
      "theoretical_loss": 3.3883124709109382,
      "tokens_seen": 2324234240
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030080808080808083,
      "loss": 1.5372,
      "theoretical_loss": 3.3883047384708753,
      "tokens_seen": 2324299776
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007878787878788,
      "loss": 1.2044,
      "theoretical_loss": 3.3882970063098785,
      "tokens_seen": 2324365312
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007676767676768,
      "loss": 1.1372,
      "theoretical_loss": 3.3882892744279287,
      "tokens_seen": 2324430848
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030074747474747476,
      "loss": 1.5126,
      "theoretical_loss": 3.3882815428250095,
      "tokens_seen": 2324496384
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007272727272727,
      "loss": 1.2965,
      "theoretical_loss": 3.3882738115011017,
      "tokens_seen": 2324561920
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003007070707070707,
      "loss": 1.4344,
      "theoretical_loss": 3.388266080456188,
      "tokens_seen": 2324627456
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003006868686868687,
      "loss": 1.328,
      "theoretical_loss": 3.388258349690251,
      "tokens_seen": 2324692992
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003006666666666667,
      "loss": 1.4662,
      "theoretical_loss": 3.388250619203271,
      "tokens_seen": 2324758528
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030064646464646465,
      "loss": 1.311,
      "theoretical_loss": 3.388242888995232,
      "tokens_seen": 2324824064
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.00030062626262626266,
      "loss": 1.1502,
      "theoretical_loss": 3.388235159066115,
      "tokens_seen": 2324889600
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003006060606060606,
      "loss": 1.2941,
      "theoretical_loss": 3.388227429415902,
      "tokens_seen": 2324955136
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003005858585858586,
      "loss": 1.1959,
      "theoretical_loss": 3.3882197000445755,
      "tokens_seen": 2325020672
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0003005656565656566,
      "loss": 1.1962,
      "theoretical_loss": 3.388211970952118,
      "tokens_seen": 2325086208
    },
    {
      "epoch": 0.4,
      "objective/train/aligned_token_count": 0.513318657875061,
      "objective/train/avg_token_score": 0.009959396906197071,
      "objective/train/docs_used": 1311230,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5987526178359985,
      "objective/train/original_loss": 3.8751535415649414,
      "objective/train/theoretical_loss": 3.3882042421385106,
      "objective/train/tokens_used": 684010976,
      "objective/train/unlikelihood_loss": 0.0738629624247551,
      "theoretical_loss": 3.3882042421385106,
      "tokens_seen": 2325151744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030054545454545454,
      "loss": 1.3623,
      "theoretical_loss": 3.3882042421385106,
      "tokens_seen": 2325151744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030052525252525255,
      "loss": 1.5836,
      "theoretical_loss": 3.388196513603736,
      "tokens_seen": 2325217280
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003005050505050505,
      "loss": 1.2247,
      "theoretical_loss": 3.388188785347776,
      "tokens_seen": 2325282816
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030048484848484847,
      "loss": 1.3711,
      "theoretical_loss": 3.3881810573706126,
      "tokens_seen": 2325348352
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003004646464646465,
      "loss": 1.4734,
      "theoretical_loss": 3.3881733296722283,
      "tokens_seen": 2325413888
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003004444444444445,
      "loss": 1.3643,
      "theoretical_loss": 3.388165602252605,
      "tokens_seen": 2325479424
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030042424242424245,
      "loss": 1.5469,
      "theoretical_loss": 3.388157875111725,
      "tokens_seen": 2325544960
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003004040404040404,
      "loss": 1.233,
      "theoretical_loss": 3.38815014824957,
      "tokens_seen": 2325610496
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030038383838383836,
      "loss": 1.331,
      "theoretical_loss": 3.3881424216661222,
      "tokens_seen": 2325676032
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030036363636363637,
      "loss": 1.297,
      "theoretical_loss": 3.3881346953613636,
      "tokens_seen": 2325741568
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003003434343434344,
      "loss": 1.351,
      "theoretical_loss": 3.3881269693352762,
      "tokens_seen": 2325807104
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030032323232323234,
      "loss": 1.4152,
      "theoretical_loss": 3.388119243587843,
      "tokens_seen": 2325872640
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003003030303030303,
      "loss": 1.4547,
      "theoretical_loss": 3.388111518119045,
      "tokens_seen": 2325938176
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030028282828282825,
      "loss": 1.4602,
      "theoretical_loss": 3.388103792928865,
      "tokens_seen": 2326003712
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002626262626263,
      "loss": 1.45,
      "theoretical_loss": 3.3880960680172842,
      "tokens_seen": 2326069248
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002424242424243,
      "loss": 1.2813,
      "theoretical_loss": 3.388088343384286,
      "tokens_seen": 2326134784
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030022222222222223,
      "loss": 1.5594,
      "theoretical_loss": 3.3880806190298514,
      "tokens_seen": 2326200320
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003002020202020202,
      "loss": 1.3237,
      "theoretical_loss": 3.388072894953963,
      "tokens_seen": 2326265856
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030018181818181815,
      "loss": 1.3153,
      "theoretical_loss": 3.388065171156603,
      "tokens_seen": 2326331392
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003001616161616162,
      "loss": 1.548,
      "theoretical_loss": 3.3880574476377534,
      "tokens_seen": 2326396928
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030014141414141417,
      "loss": 1.5302,
      "theoretical_loss": 3.3880497243973964,
      "tokens_seen": 2326462464
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003001212121212121,
      "loss": 1.2711,
      "theoretical_loss": 3.3880420014355135,
      "tokens_seen": 2326528000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003001010101010101,
      "loss": 1.471,
      "theoretical_loss": 3.388034278752088,
      "tokens_seen": 2326593536
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003000808080808081,
      "loss": 1.2223,
      "theoretical_loss": 3.388026556347101,
      "tokens_seen": 2326659072
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003000606060606061,
      "loss": 1.0905,
      "theoretical_loss": 3.3880188342205346,
      "tokens_seen": 2326724608
    },
    {
      "epoch": 0.41,
      "objective/train/aligned_token_count": 0.4186827838420868,
      "objective/train/avg_token_score": 0.03725993633270264,
      "objective/train/docs_used": 1311908,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.323853850364685,
      "objective/train/original_loss": 4.203466892242432,
      "objective/train/theoretical_loss": 3.3880111123723715,
      "objective/train/tokens_used": 685649376,
      "objective/train/unlikelihood_loss": 0.06094466894865036,
      "theoretical_loss": 3.3880111123723715,
      "tokens_seen": 2326790144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00030004040404040406,
      "loss": 1.403,
      "theoretical_loss": 3.3880111123723715,
      "tokens_seen": 2326790144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000300020202020202,
      "loss": 1.4474,
      "theoretical_loss": 3.3880033908025937,
      "tokens_seen": 2326855680
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0003,
      "loss": 1.3732,
      "theoretical_loss": 3.387995669511183,
      "tokens_seen": 2326921216
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000299979797979798,
      "loss": 1.4023,
      "theoretical_loss": 3.3879879484981217,
      "tokens_seen": 2326986752
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000299959595959596,
      "loss": 1.3328,
      "theoretical_loss": 3.3879802277633924,
      "tokens_seen": 2327052288
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029993939393939395,
      "loss": 1.5088,
      "theoretical_loss": 3.3879725073069764,
      "tokens_seen": 2327117824
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002999191919191919,
      "loss": 1.3227,
      "theoretical_loss": 3.3879647871288565,
      "tokens_seen": 2327183360
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029989898989898987,
      "loss": 1.2824,
      "theoretical_loss": 3.3879570672290145,
      "tokens_seen": 2327248896
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002998787878787879,
      "loss": 1.2443,
      "theoretical_loss": 3.387949347607432,
      "tokens_seen": 2327314432
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002998585858585859,
      "loss": 1.3132,
      "theoretical_loss": 3.3879416282640924,
      "tokens_seen": 2327379968
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029983838383838385,
      "loss": 1.5057,
      "theoretical_loss": 3.387933909198977,
      "tokens_seen": 2327445504
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002998181818181818,
      "loss": 1.3784,
      "theoretical_loss": 3.3879261904120677,
      "tokens_seen": 2327511040
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002997979797979798,
      "loss": 1.371,
      "theoretical_loss": 3.3879184719033475,
      "tokens_seen": 2327576576
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029977777777777777,
      "loss": 1.308,
      "theoretical_loss": 3.3879107536727977,
      "tokens_seen": 2327642112
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002997575757575758,
      "loss": 1.3715,
      "theoretical_loss": 3.3879030357204014,
      "tokens_seen": 2327707648
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029973737373737374,
      "loss": 1.3225,
      "theoretical_loss": 3.3878953180461395,
      "tokens_seen": 2327773184
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002997171717171717,
      "loss": 1.292,
      "theoretical_loss": 3.387887600649995,
      "tokens_seen": 2327838720
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002996969696969697,
      "loss": 1.4853,
      "theoretical_loss": 3.38787988353195,
      "tokens_seen": 2327904256
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029967676767676766,
      "loss": 1.458,
      "theoretical_loss": 3.387872166691986,
      "tokens_seen": 2327969792
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002996565656565657,
      "loss": 1.5755,
      "theoretical_loss": 3.3878644501300865,
      "tokens_seen": 2328035328
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029963636363636363,
      "loss": 1.4686,
      "theoretical_loss": 3.3878567338462324,
      "tokens_seen": 2328100864
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029961616161616164,
      "loss": 1.4439,
      "theoretical_loss": 3.387849017840406,
      "tokens_seen": 2328166400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002995959595959596,
      "loss": 1.3398,
      "theoretical_loss": 3.3878413021125904,
      "tokens_seen": 2328231936
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029957575757575756,
      "loss": 1.2501,
      "theoretical_loss": 3.3878335866627665,
      "tokens_seen": 2328297472
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029955555555555557,
      "loss": 1.3139,
      "theoretical_loss": 3.387825871490917,
      "tokens_seen": 2328363008
    },
    {
      "epoch": 0.41,
      "objective/train/aligned_token_count": 0.3224584460258484,
      "objective/train/avg_token_score": 0.013725139200687408,
      "objective/train/docs_used": 1312576,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0300214290618896,
      "objective/train/original_loss": 5.195791244506836,
      "objective/train/theoretical_loss": 3.3878181565970245,
      "objective/train/tokens_used": 687287776,
      "objective/train/unlikelihood_loss": 0.051812510937452316,
      "theoretical_loss": 3.3878181565970245,
      "tokens_seen": 2328428544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002995353535353535,
      "loss": 1.2663,
      "theoretical_loss": 3.3878181565970245,
      "tokens_seen": 2328428544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029951515151515153,
      "loss": 1.3875,
      "theoretical_loss": 3.387810441981071,
      "tokens_seen": 2328494080
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002994949494949495,
      "loss": 1.433,
      "theoretical_loss": 3.3878027276430376,
      "tokens_seen": 2328559616
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029947474747474745,
      "loss": 1.3504,
      "theoretical_loss": 3.387795013582908,
      "tokens_seen": 2328625152
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029945454545454546,
      "loss": 1.4285,
      "theoretical_loss": 3.387787299800663,
      "tokens_seen": 2328690688
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029943434343434347,
      "loss": 1.3871,
      "theoretical_loss": 3.387779586296286,
      "tokens_seen": 2328756224
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029941414141414143,
      "loss": 1.5346,
      "theoretical_loss": 3.387771873069758,
      "tokens_seen": 2328821760
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002993939393939394,
      "loss": 1.3443,
      "theoretical_loss": 3.3877641601210624,
      "tokens_seen": 2328887296
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029937373737373734,
      "loss": 1.3198,
      "theoretical_loss": 3.3877564474501805,
      "tokens_seen": 2328952832
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029935353535353535,
      "loss": 1.5291,
      "theoretical_loss": 3.3877487350570945,
      "tokens_seen": 2329018368
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029933333333333336,
      "loss": 1.3841,
      "theoretical_loss": 3.3877410229417873,
      "tokens_seen": 2329083904
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002993131313131313,
      "loss": 1.2398,
      "theoretical_loss": 3.3877333111042405,
      "tokens_seen": 2329149440
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992929292929293,
      "loss": 1.0373,
      "theoretical_loss": 3.387725599544436,
      "tokens_seen": 2329214976
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029927272727272723,
      "loss": 1.6582,
      "theoretical_loss": 3.3877178882623564,
      "tokens_seen": 2329280512
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992525252525253,
      "loss": 1.4018,
      "theoretical_loss": 3.387710177257984,
      "tokens_seen": 2329346048
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029923232323232326,
      "loss": 1.1924,
      "theoretical_loss": 3.3877024665313007,
      "tokens_seen": 2329411584
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002992121212121212,
      "loss": 1.3124,
      "theoretical_loss": 3.387694756082289,
      "tokens_seen": 2329477120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029919191919191917,
      "loss": 1.2712,
      "theoretical_loss": 3.3876870459109307,
      "tokens_seen": 2329542656
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991717171717171,
      "loss": 1.3236,
      "theoretical_loss": 3.387679336017208,
      "tokens_seen": 2329608192
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991515151515152,
      "loss": 1.1823,
      "theoretical_loss": 3.387671626401104,
      "tokens_seen": 2329673728
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029913131313131315,
      "loss": 1.3918,
      "theoretical_loss": 3.3876639170626,
      "tokens_seen": 2329739264
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002991111111111111,
      "loss": 1.3946,
      "theoretical_loss": 3.387656208001678,
      "tokens_seen": 2329804800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029909090909090906,
      "loss": 1.3118,
      "theoretical_loss": 3.3876484992183205,
      "tokens_seen": 2329870336
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029907070707070713,
      "loss": 1.5544,
      "theoretical_loss": 3.38764079071251,
      "tokens_seen": 2329935872
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002990505050505051,
      "loss": 1.2841,
      "theoretical_loss": 3.3876330824842285,
      "tokens_seen": 2330001408
    },
    {
      "epoch": 0.41,
      "objective/train/aligned_token_count": 0.29771506786346436,
      "objective/train/avg_token_score": 0.010447874665260315,
      "objective/train/docs_used": 1313793,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9548224806785583,
      "objective/train/original_loss": 5.10629940032959,
      "objective/train/theoretical_loss": 3.387625374533458,
      "objective/train/tokens_used": 688926176,
      "objective/train/unlikelihood_loss": 0.039890166372060776,
      "theoretical_loss": 3.387625374533458,
      "tokens_seen": 2330066944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029903030303030304,
      "loss": 1.2233,
      "theoretical_loss": 3.387625374533458,
      "tokens_seen": 2330066944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000299010101010101,
      "loss": 1.3124,
      "theoretical_loss": 3.387617666860181,
      "tokens_seen": 2330132480
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029898989898989895,
      "loss": 1.3452,
      "theoretical_loss": 3.3876099594643794,
      "tokens_seen": 2330198016
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298969696969697,
      "loss": 1.3325,
      "theoretical_loss": 3.387602252346036,
      "tokens_seen": 2330263552
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298949494949495,
      "loss": 1.2539,
      "theoretical_loss": 3.3875945455051317,
      "tokens_seen": 2330329088
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029892929292929293,
      "loss": 1.2469,
      "theoretical_loss": 3.38758683894165,
      "tokens_seen": 2330394624
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002989090909090909,
      "loss": 1.2959,
      "theoretical_loss": 3.3875791326555733,
      "tokens_seen": 2330460160
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002988888888888889,
      "loss": 1.4881,
      "theoretical_loss": 3.3875714266468826,
      "tokens_seen": 2330525696
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002988686868686869,
      "loss": 1.1873,
      "theoretical_loss": 3.387563720915561,
      "tokens_seen": 2330591232
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029884848484848487,
      "loss": 1.3852,
      "theoretical_loss": 3.3875560154615907,
      "tokens_seen": 2330656768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002988282828282828,
      "loss": 1.1676,
      "theoretical_loss": 3.3875483102849535,
      "tokens_seen": 2330722304
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002988080808080808,
      "loss": 1.1926,
      "theoretical_loss": 3.387540605385632,
      "tokens_seen": 2330787840
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002987878787878788,
      "loss": 1.2662,
      "theoretical_loss": 3.387532900763608,
      "tokens_seen": 2330853376
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002987676767676768,
      "loss": 1.4213,
      "theoretical_loss": 3.3875251964188635,
      "tokens_seen": 2330918912
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029874747474747476,
      "loss": 1.3983,
      "theoretical_loss": 3.3875174923513818,
      "tokens_seen": 2330984448
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002987272727272727,
      "loss": 1.4192,
      "theoretical_loss": 3.387509788561144,
      "tokens_seen": 2331049984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029870707070707073,
      "loss": 1.4373,
      "theoretical_loss": 3.3875020850481334,
      "tokens_seen": 2331115520
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986868686868687,
      "loss": 1.5373,
      "theoretical_loss": 3.3874943818123313,
      "tokens_seen": 2331181056
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986666666666667,
      "loss": 1.2498,
      "theoretical_loss": 3.387486678853721,
      "tokens_seen": 2331246592
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029864646464646466,
      "loss": 1.3819,
      "theoretical_loss": 3.3874789761722828,
      "tokens_seen": 2331312128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986262626262626,
      "loss": 1.5267,
      "theoretical_loss": 3.387471273768001,
      "tokens_seen": 2331377664
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002986060606060606,
      "loss": 1.1649,
      "theoretical_loss": 3.3874635716408568,
      "tokens_seen": 2331443200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002985858585858586,
      "loss": 1.3822,
      "theoretical_loss": 3.3874558697908324,
      "tokens_seen": 2331508736
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002985656565656566,
      "loss": 1.5529,
      "theoretical_loss": 3.3874481682179107,
      "tokens_seen": 2331574272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029854545454545455,
      "loss": 1.3514,
      "theoretical_loss": 3.3874404669220732,
      "tokens_seen": 2331639808
    },
    {
      "epoch": 0.41,
      "objective/train/aligned_token_count": 0.3087732195854187,
      "objective/train/avg_token_score": 0.03237170726060867,
      "objective/train/docs_used": 1314556,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0826610326766968,
      "objective/train/original_loss": 4.7997612953186035,
      "objective/train/theoretical_loss": 3.3874327659033026,
      "objective/train/tokens_used": 690564576,
      "objective/train/unlikelihood_loss": 0.04036524519324303,
      "theoretical_loss": 3.3874327659033026,
      "tokens_seen": 2331705344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029852525252525256,
      "loss": 1.2415,
      "theoretical_loss": 3.3874327659033026,
      "tokens_seen": 2331705344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002985050505050505,
      "loss": 1.2645,
      "theoretical_loss": 3.3874250651615814,
      "tokens_seen": 2331770880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029848484848484847,
      "loss": 1.4787,
      "theoretical_loss": 3.3874173646968906,
      "tokens_seen": 2331836416
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984646464646465,
      "loss": 1.2571,
      "theoretical_loss": 3.3874096645092138,
      "tokens_seen": 2331901952
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029844444444444444,
      "loss": 1.2408,
      "theoretical_loss": 3.3874019645985327,
      "tokens_seen": 2331967488
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029842424242424245,
      "loss": 1.4366,
      "theoretical_loss": 3.3873942649648296,
      "tokens_seen": 2332033024
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002984040404040404,
      "loss": 1.3848,
      "theoretical_loss": 3.387386565608087,
      "tokens_seen": 2332098560
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029838383838383837,
      "loss": 1.6009,
      "theoretical_loss": 3.3873788665282865,
      "tokens_seen": 2332164096
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983636363636364,
      "loss": 1.4112,
      "theoretical_loss": 3.3873711677254112,
      "tokens_seen": 2332229632
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983434343434344,
      "loss": 1.4061,
      "theoretical_loss": 3.3873634691994425,
      "tokens_seen": 2332295168
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029832323232323234,
      "loss": 1.3012,
      "theoretical_loss": 3.3873557709503634,
      "tokens_seen": 2332360704
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002983030303030303,
      "loss": 1.3866,
      "theoretical_loss": 3.3873480729781558,
      "tokens_seen": 2332426240
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029828282828282826,
      "loss": 1.3447,
      "theoretical_loss": 3.387340375282802,
      "tokens_seen": 2332491776
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029826262626262627,
      "loss": 1.3566,
      "theoretical_loss": 3.3873326778642845,
      "tokens_seen": 2332557312
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002982424242424243,
      "loss": 1.4103,
      "theoretical_loss": 3.387324980722585,
      "tokens_seen": 2332622848
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029822222222222224,
      "loss": 1.3748,
      "theoretical_loss": 3.387317283857686,
      "tokens_seen": 2332688384
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002982020202020202,
      "loss": 1.3649,
      "theoretical_loss": 3.38730958726957,
      "tokens_seen": 2332753920
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029818181818181815,
      "loss": 1.3568,
      "theoretical_loss": 3.387301890958219,
      "tokens_seen": 2332819456
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002981616161616162,
      "loss": 1.4599,
      "theoretical_loss": 3.3872941949236157,
      "tokens_seen": 2332884992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002981414141414142,
      "loss": 1.5607,
      "theoretical_loss": 3.387286499165742,
      "tokens_seen": 2332950528
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029812121212121213,
      "loss": 1.3177,
      "theoretical_loss": 3.3872788036845805,
      "tokens_seen": 2333016064
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002981010101010101,
      "loss": 1.3511,
      "theoretical_loss": 3.387271108480113,
      "tokens_seen": 2333081600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029808080808080804,
      "loss": 1.3203,
      "theoretical_loss": 3.3872634135523225,
      "tokens_seen": 2333147136
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002980606060606061,
      "loss": 1.3498,
      "theoretical_loss": 3.3872557189011903,
      "tokens_seen": 2333212672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029804040404040407,
      "loss": 1.2351,
      "theoretical_loss": 3.3872480245266994,
      "tokens_seen": 2333278208
    },
    {
      "epoch": 0.41,
      "objective/train/aligned_token_count": 0.45839443802833557,
      "objective/train/avg_token_score": 0.03555627539753914,
      "objective/train/docs_used": 1315828,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4466710090637207,
      "objective/train/original_loss": 4.289207935333252,
      "objective/train/theoretical_loss": 3.3872403304288317,
      "objective/train/tokens_used": 692202976,
      "objective/train/unlikelihood_loss": 0.07773689180612564,
      "theoretical_loss": 3.3872403304288317,
      "tokens_seen": 2333343744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298020202020202,
      "loss": 1.3172,
      "theoretical_loss": 3.3872403304288317,
      "tokens_seen": 2333343744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000298,
      "loss": 1.3595,
      "theoretical_loss": 3.38723263660757,
      "tokens_seen": 2333409280
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029797979797979794,
      "loss": 1.2366,
      "theoretical_loss": 3.387224943062896,
      "tokens_seen": 2333474816
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000297959595959596,
      "loss": 1.1637,
      "theoretical_loss": 3.3872172497947926,
      "tokens_seen": 2333540352
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029793939393939396,
      "loss": 1.4065,
      "theoretical_loss": 3.387209556803241,
      "tokens_seen": 2333605888
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002979191919191919,
      "loss": 1.3971,
      "theoretical_loss": 3.3872018640882247,
      "tokens_seen": 2333671424
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029789898989898987,
      "loss": 1.1752,
      "theoretical_loss": 3.3871941716497256,
      "tokens_seen": 2333736960
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978787878787879,
      "loss": 1.3734,
      "theoretical_loss": 3.3871864794877258,
      "tokens_seen": 2333802496
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978585858585859,
      "loss": 1.4392,
      "theoretical_loss": 3.387178787602208,
      "tokens_seen": 2333868032
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029783838383838385,
      "loss": 1.4705,
      "theoretical_loss": 3.3871710959931534,
      "tokens_seen": 2333933568
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002978181818181818,
      "loss": 1.1946,
      "theoretical_loss": 3.387163404660546,
      "tokens_seen": 2333999104
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029779797979797976,
      "loss": 1.3705,
      "theoretical_loss": 3.3871557136043666,
      "tokens_seen": 2334064640
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002977777777777778,
      "loss": 1.0616,
      "theoretical_loss": 3.3871480228245985,
      "tokens_seen": 2334130176
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002977575757575758,
      "loss": 1.0673,
      "theoretical_loss": 3.3871403323212235,
      "tokens_seen": 2334195712
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029773737373737374,
      "loss": 1.453,
      "theoretical_loss": 3.3871326420942243,
      "tokens_seen": 2334261248
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002977171717171717,
      "loss": 1.27,
      "theoretical_loss": 3.3871249521435827,
      "tokens_seen": 2334326784
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002976969696969697,
      "loss": 1.5134,
      "theoretical_loss": 3.387117262469281,
      "tokens_seen": 2334392320
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029767676767676767,
      "loss": 1.4628,
      "theoretical_loss": 3.387109573071302,
      "tokens_seen": 2334457856
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002976565656565657,
      "loss": 1.2572,
      "theoretical_loss": 3.387101883949628,
      "tokens_seen": 2334523392
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029763636363636364,
      "loss": 1.2418,
      "theoretical_loss": 3.387094195104241,
      "tokens_seen": 2334588928
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002976161616161616,
      "loss": 1.1627,
      "theoretical_loss": 3.387086506535123,
      "tokens_seen": 2334654464
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002975959595959596,
      "loss": 1.2074,
      "theoretical_loss": 3.3870788182422573,
      "tokens_seen": 2334720000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029757575757575756,
      "loss": 1.4521,
      "theoretical_loss": 3.3870711302256256,
      "tokens_seen": 2334785536
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029755555555555557,
      "loss": 1.5053,
      "theoretical_loss": 3.3870634424852097,
      "tokens_seen": 2334851072
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029753535353535353,
      "loss": 1.4822,
      "theoretical_loss": 3.387055755020993,
      "tokens_seen": 2334916608
    },
    {
      "epoch": 0.41,
      "objective/train/aligned_token_count": 0.31836509704589844,
      "objective/train/avg_token_score": 0.017019353806972504,
      "objective/train/docs_used": 1316418,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0808972120285034,
      "objective/train/original_loss": 4.780982971191406,
      "objective/train/theoretical_loss": 3.387048067832957,
      "objective/train/tokens_used": 693841376,
      "objective/train/unlikelihood_loss": 0.05668047070503235,
      "theoretical_loss": 3.387048067832957,
      "tokens_seen": 2334982144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029751515151515154,
      "loss": 1.5536,
      "theoretical_loss": 3.387048067832957,
      "tokens_seen": 2334982144
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002974949494949495,
      "loss": 1.4085,
      "theoretical_loss": 3.3870403809210847,
      "tokens_seen": 2335047680
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029747474747474745,
      "loss": 1.2136,
      "theoretical_loss": 3.3870326942853577,
      "tokens_seen": 2335113216
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029745454545454546,
      "loss": 1.3145,
      "theoretical_loss": 3.387025007925759,
      "tokens_seen": 2335178752
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002974343434343434,
      "loss": 1.3836,
      "theoretical_loss": 3.387017321842271,
      "tokens_seen": 2335244288
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029741414141414143,
      "loss": 1.3291,
      "theoretical_loss": 3.387009636034875,
      "tokens_seen": 2335309824
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973939393939394,
      "loss": 1.3387,
      "theoretical_loss": 3.3870019505035542,
      "tokens_seen": 2335375360
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973737373737374,
      "loss": 1.42,
      "theoretical_loss": 3.386994265248291,
      "tokens_seen": 2335440896
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029735353535353536,
      "loss": 1.304,
      "theoretical_loss": 3.386986580269067,
      "tokens_seen": 2335506432
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029733333333333337,
      "loss": 1.2459,
      "theoretical_loss": 3.3869788955658655,
      "tokens_seen": 2335571968
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002973131313131313,
      "loss": 1.3319,
      "theoretical_loss": 3.3869712111386683,
      "tokens_seen": 2335637504
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002972929292929293,
      "loss": 1.3937,
      "theoretical_loss": 3.3869635269874574,
      "tokens_seen": 2335703040
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002972727272727273,
      "loss": 1.2894,
      "theoretical_loss": 3.386955843112216,
      "tokens_seen": 2335768576
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029725252525252525,
      "loss": 1.466,
      "theoretical_loss": 3.386948159512926,
      "tokens_seen": 2335834112
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029723232323232326,
      "loss": 1.2976,
      "theoretical_loss": 3.386940476189569,
      "tokens_seen": 2335899648
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002972121212121212,
      "loss": 1.5405,
      "theoretical_loss": 3.3869327931421287,
      "tokens_seen": 2335965184
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971919191919192,
      "loss": 1.265,
      "theoretical_loss": 3.3869251103705866,
      "tokens_seen": 2336030720
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971717171717172,
      "loss": 1.3659,
      "theoretical_loss": 3.3869174278749257,
      "tokens_seen": 2336096256
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971515151515152,
      "loss": 1.2269,
      "theoretical_loss": 3.386909745655128,
      "tokens_seen": 2336161792
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029713131313131315,
      "loss": 1.2583,
      "theoretical_loss": 3.386902063711175,
      "tokens_seen": 2336227328
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002971111111111111,
      "loss": 1.4692,
      "theoretical_loss": 3.386894382043051,
      "tokens_seen": 2336292864
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029709090909090907,
      "loss": 1.3863,
      "theoretical_loss": 3.386886700650736,
      "tokens_seen": 2336358400
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002970707070707071,
      "loss": 1.4846,
      "theoretical_loss": 3.3868790195342147,
      "tokens_seen": 2336423936
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002970505050505051,
      "loss": 1.4048,
      "theoretical_loss": 3.3868713386934677,
      "tokens_seen": 2336489472
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029703030303030305,
      "loss": 1.3043,
      "theoretical_loss": 3.386863658128478,
      "tokens_seen": 2336555008
    },
    {
      "epoch": 0.41,
      "objective/train/aligned_token_count": 0.45008552074432373,
      "objective/train/avg_token_score": 0.010576860047876835,
      "objective/train/docs_used": 1317715,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.429062008857727,
      "objective/train/original_loss": 4.223320960998535,
      "objective/train/theoretical_loss": 3.386855977839228,
      "objective/train/tokens_used": 695479776,
      "objective/train/unlikelihood_loss": 0.040715496987104416,
      "theoretical_loss": 3.386855977839228,
      "tokens_seen": 2336620544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000297010101010101,
      "loss": 1.4037,
      "theoretical_loss": 3.386855977839228,
      "tokens_seen": 2336620544
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029698989898989896,
      "loss": 1.5114,
      "theoretical_loss": 3.3868482978257006,
      "tokens_seen": 2336686080
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000296969696969697,
      "loss": 1.4268,
      "theoretical_loss": 3.386840618087877,
      "tokens_seen": 2336751616
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000296949494949495,
      "loss": 1.5413,
      "theoretical_loss": 3.3868329386257403,
      "tokens_seen": 2336817152
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029692929292929294,
      "loss": 1.5755,
      "theoretical_loss": 3.386825259439273,
      "tokens_seen": 2336882688
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002969090909090909,
      "loss": 1.3947,
      "theoretical_loss": 3.386817580528457,
      "tokens_seen": 2336948224
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029688888888888885,
      "loss": 1.3844,
      "theoretical_loss": 3.386809901893275,
      "tokens_seen": 2337013760
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002968686868686869,
      "loss": 1.3668,
      "theoretical_loss": 3.386802223533709,
      "tokens_seen": 2337079296
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002968484848484849,
      "loss": 1.2511,
      "theoretical_loss": 3.3867945454497423,
      "tokens_seen": 2337144832
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029682828282828283,
      "loss": 1.3388,
      "theoretical_loss": 3.386786867641356,
      "tokens_seen": 2337210368
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002968080808080808,
      "loss": 1.3557,
      "theoretical_loss": 3.3867791901085336,
      "tokens_seen": 2337275904
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002967878787878788,
      "loss": 1.3147,
      "theoretical_loss": 3.386771512851257,
      "tokens_seen": 2337341440
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002967676767676768,
      "loss": 1.4156,
      "theoretical_loss": 3.386763835869508,
      "tokens_seen": 2337406976
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029674747474747477,
      "loss": 1.5811,
      "theoretical_loss": 3.38675615916327,
      "tokens_seen": 2337472512
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002967272727272727,
      "loss": 1.3891,
      "theoretical_loss": 3.386748482732525,
      "tokens_seen": 2337538048
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002967070707070707,
      "loss": 1.174,
      "theoretical_loss": 3.3867408065772553,
      "tokens_seen": 2337603584
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002966868686868687,
      "loss": 1.2649,
      "theoretical_loss": 3.3867331306974435,
      "tokens_seen": 2337669120
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002966666666666667,
      "loss": 1.0837,
      "theoretical_loss": 3.3867254550930714,
      "tokens_seen": 2337734656
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029664646464646466,
      "loss": 1.1901,
      "theoretical_loss": 3.386717779764122,
      "tokens_seen": 2337800192
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002966262626262626,
      "loss": 1.3378,
      "theoretical_loss": 3.386710104710578,
      "tokens_seen": 2337865728
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002966060606060606,
      "loss": 1.4181,
      "theoretical_loss": 3.386702429932421,
      "tokens_seen": 2337931264
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965858585858586,
      "loss": 1.3675,
      "theoretical_loss": 3.386694755429634,
      "tokens_seen": 2337996800
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965656565656566,
      "loss": 1.4307,
      "theoretical_loss": 3.3866870812021985,
      "tokens_seen": 2338062336
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029654545454545455,
      "loss": 1.3351,
      "theoretical_loss": 3.3866794072500976,
      "tokens_seen": 2338127872
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965252525252525,
      "loss": 1.1742,
      "theoretical_loss": 3.3866717335733143,
      "tokens_seen": 2338193408
    },
    {
      "epoch": 0.41,
      "objective/train/aligned_token_count": 0.5104472041130066,
      "objective/train/avg_token_score": 0.017596814781427383,
      "objective/train/docs_used": 1318542,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5858395099639893,
      "objective/train/original_loss": 3.725752830505371,
      "objective/train/theoretical_loss": 3.38666406017183,
      "objective/train/tokens_used": 697118176,
      "objective/train/unlikelihood_loss": 0.08415797352790833,
      "theoretical_loss": 3.38666406017183,
      "tokens_seen": 2338258944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002965050505050505,
      "loss": 1.3267,
      "theoretical_loss": 3.38666406017183,
      "tokens_seen": 2338258944
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964848484848485,
      "loss": 1.3185,
      "theoretical_loss": 3.3866563870456274,
      "tokens_seen": 2338324480
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964646464646465,
      "loss": 1.4953,
      "theoretical_loss": 3.386648714194689,
      "tokens_seen": 2338390016
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029644444444444445,
      "loss": 1.2821,
      "theoretical_loss": 3.386641041618997,
      "tokens_seen": 2338455552
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964242424242424,
      "loss": 1.333,
      "theoretical_loss": 3.386633369318534,
      "tokens_seen": 2338521088
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002964040404040404,
      "loss": 1.2939,
      "theoretical_loss": 3.3866256972932827,
      "tokens_seen": 2338586624
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029638383838383837,
      "loss": 1.2444,
      "theoretical_loss": 3.386618025543225,
      "tokens_seen": 2338652160
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002963636363636364,
      "loss": 1.3476,
      "theoretical_loss": 3.3866103540683437,
      "tokens_seen": 2338717696
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029634343434343434,
      "loss": 1.5575,
      "theoretical_loss": 3.386602682868621,
      "tokens_seen": 2338783232
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029632323232323235,
      "loss": 1.4264,
      "theoretical_loss": 3.3865950119440393,
      "tokens_seen": 2338848768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002963030303030303,
      "loss": 1.3683,
      "theoretical_loss": 3.3865873412945815,
      "tokens_seen": 2338914304
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029628282828282826,
      "loss": 1.388,
      "theoretical_loss": 3.386579670920229,
      "tokens_seen": 2338979840
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002962626262626263,
      "loss": 1.3813,
      "theoretical_loss": 3.386572000820965,
      "tokens_seen": 2339045376
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029624242424242423,
      "loss": 1.4794,
      "theoretical_loss": 3.386564330996772,
      "tokens_seen": 2339110912
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029622222222222224,
      "loss": 1.36,
      "theoretical_loss": 3.3865566614476323,
      "tokens_seen": 2339176448
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002962020202020202,
      "loss": 1.094,
      "theoretical_loss": 3.3865489921735277,
      "tokens_seen": 2339241984
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029618181818181816,
      "loss": 1.47,
      "theoretical_loss": 3.3865413231744417,
      "tokens_seen": 2339307520
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029616161616161617,
      "loss": 1.3171,
      "theoretical_loss": 3.386533654450356,
      "tokens_seen": 2339373056
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002961414141414142,
      "loss": 1.4659,
      "theoretical_loss": 3.386525986001253,
      "tokens_seen": 2339438592
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029612121212121214,
      "loss": 1.4489,
      "theoretical_loss": 3.3865183178271154,
      "tokens_seen": 2339504128
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002961010101010101,
      "loss": 1.2219,
      "theoretical_loss": 3.386510649927926,
      "tokens_seen": 2339569664
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029608080808080805,
      "loss": 1.6557,
      "theoretical_loss": 3.386502982303667,
      "tokens_seen": 2339635200
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029606060606060606,
      "loss": 1.1883,
      "theoretical_loss": 3.38649531495432,
      "tokens_seen": 2339700736
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029604040404040407,
      "loss": 1.2398,
      "theoretical_loss": 3.3864876478798687,
      "tokens_seen": 2339766272
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029602020202020203,
      "loss": 1.1956,
      "theoretical_loss": 3.386479981080295,
      "tokens_seen": 2339831808
    },
    {
      "epoch": 0.41,
      "objective/train/aligned_token_count": 0.5395283699035645,
      "objective/train/avg_token_score": 0.02159624919295311,
      "objective/train/docs_used": 1319951,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7791123390197754,
      "objective/train/original_loss": 3.8667173385620117,
      "objective/train/theoretical_loss": 3.3864723145555806,
      "objective/train/tokens_used": 698756576,
      "objective/train/unlikelihood_loss": 0.05942104384303093,
      "theoretical_loss": 3.3864723145555806,
      "tokens_seen": 2339897344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000296,
      "loss": 1.66,
      "theoretical_loss": 3.3864723145555806,
      "tokens_seen": 2339897344
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029597979797979794,
      "loss": 1.4367,
      "theoretical_loss": 3.3864646483057093,
      "tokens_seen": 2339962880
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.000295959595959596,
      "loss": 1.2796,
      "theoretical_loss": 3.386456982330663,
      "tokens_seen": 2340028416
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029593939393939396,
      "loss": 1.3171,
      "theoretical_loss": 3.3864493166304235,
      "tokens_seen": 2340093952
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002959191919191919,
      "loss": 1.3064,
      "theoretical_loss": 3.3864416512049744,
      "tokens_seen": 2340159488
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958989898989899,
      "loss": 1.42,
      "theoretical_loss": 3.386433986054297,
      "tokens_seen": 2340225024
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029587878787878783,
      "loss": 1.3351,
      "theoretical_loss": 3.386426321178375,
      "tokens_seen": 2340290560
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958585858585859,
      "loss": 1.5767,
      "theoretical_loss": 3.38641865657719,
      "tokens_seen": 2340356096
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029583838383838386,
      "loss": 1.2766,
      "theoretical_loss": 3.3864109922507244,
      "tokens_seen": 2340421632
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002958181818181818,
      "loss": 1.1799,
      "theoretical_loss": 3.386403328198961,
      "tokens_seen": 2340487168
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029579797979797977,
      "loss": 1.4878,
      "theoretical_loss": 3.3863956644218822,
      "tokens_seen": 2340552704
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002957777777777778,
      "loss": 1.54,
      "theoretical_loss": 3.386388000919471,
      "tokens_seen": 2340618240
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002957575757575758,
      "loss": 1.3752,
      "theoretical_loss": 3.3863803376917083,
      "tokens_seen": 2340683776
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029573737373737375,
      "loss": 1.1428,
      "theoretical_loss": 3.3863726747385785,
      "tokens_seen": 2340749312
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002957171717171717,
      "loss": 1.3977,
      "theoretical_loss": 3.386365012060063,
      "tokens_seen": 2340814848
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029569696969696966,
      "loss": 1.2719,
      "theoretical_loss": 3.386357349656144,
      "tokens_seen": 2340880384
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029567676767676773,
      "loss": 1.426,
      "theoretical_loss": 3.3863496875268044,
      "tokens_seen": 2340945920
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002956565656565657,
      "loss": 1.2751,
      "theoretical_loss": 3.386342025672027,
      "tokens_seen": 2341011456
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029563636363636364,
      "loss": 1.4353,
      "theoretical_loss": 3.386334364091794,
      "tokens_seen": 2341076992
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002956161616161616,
      "loss": 1.3797,
      "theoretical_loss": 3.386326702786087,
      "tokens_seen": 2341142528
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002955959595959596,
      "loss": 1.1246,
      "theoretical_loss": 3.38631904175489,
      "tokens_seen": 2341208064
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002955757575757576,
      "loss": 1.3255,
      "theoretical_loss": 3.386311380998185,
      "tokens_seen": 2341273600
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002955555555555556,
      "loss": 1.2215,
      "theoretical_loss": 3.386303720515954,
      "tokens_seen": 2341339136
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00029553535353535353,
      "loss": 1.189,
      "theoretical_loss": 3.3862960603081795,
      "tokens_seen": 2341404672
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002955151515151515,
      "loss": 1.7435,
      "theoretical_loss": 3.3862884003748444,
      "tokens_seen": 2341470208
    },
    {
      "epoch": 0.41,
      "objective/train/aligned_token_count": 0.3271627426147461,
      "objective/train/avg_token_score": 0.020861048251390457,
      "objective/train/docs_used": 1320683,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9508933424949646,
      "objective/train/original_loss": 4.241996765136719,
      "objective/train/theoretical_loss": 3.386280740715931,
      "objective/train/tokens_used": 700394976,
      "objective/train/unlikelihood_loss": 0.09151821583509445,
      "theoretical_loss": 3.386280740715931,
      "tokens_seen": 2341535744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002954949494949495,
      "loss": 1.182,
      "theoretical_loss": 3.386280740715931,
      "tokens_seen": 2341535744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002954747474747475,
      "loss": 1.6398,
      "theoretical_loss": 3.3862730813314217,
      "tokens_seen": 2341601280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029545454545454547,
      "loss": 1.4839,
      "theoretical_loss": 3.3862654222212996,
      "tokens_seen": 2341666816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029543434343434343,
      "loss": 1.3762,
      "theoretical_loss": 3.386257763385546,
      "tokens_seen": 2341732352
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029541414141414144,
      "loss": 1.3949,
      "theoretical_loss": 3.3862501048241445,
      "tokens_seen": 2341797888
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002953939393939394,
      "loss": 1.3373,
      "theoretical_loss": 3.386242446537077,
      "tokens_seen": 2341863424
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002953737373737374,
      "loss": 1.3981,
      "theoretical_loss": 3.386234788524326,
      "tokens_seen": 2341928960
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029535353535353536,
      "loss": 1.3007,
      "theoretical_loss": 3.3862271307858745,
      "tokens_seen": 2341994496
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002953333333333333,
      "loss": 1.4548,
      "theoretical_loss": 3.3862194733217046,
      "tokens_seen": 2342060032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029531313131313133,
      "loss": 1.483,
      "theoretical_loss": 3.386211816131799,
      "tokens_seen": 2342125568
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002952929292929293,
      "loss": 1.4656,
      "theoretical_loss": 3.3862041592161396,
      "tokens_seen": 2342191104
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002952727272727273,
      "loss": 1.2616,
      "theoretical_loss": 3.38619650257471,
      "tokens_seen": 2342256640
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029525252525252526,
      "loss": 1.6351,
      "theoretical_loss": 3.3861888462074914,
      "tokens_seen": 2342322176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029523232323232327,
      "loss": 1.4237,
      "theoretical_loss": 3.3861811901144674,
      "tokens_seen": 2342387712
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002952121212121212,
      "loss": 1.3561,
      "theoretical_loss": 3.38617353429562,
      "tokens_seen": 2342453248
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002951919191919192,
      "loss": 1.3891,
      "theoretical_loss": 3.3861658787509317,
      "tokens_seen": 2342518784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002951717171717172,
      "loss": 1.2434,
      "theoretical_loss": 3.3861582234803853,
      "tokens_seen": 2342584320
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029515151515151515,
      "loss": 1.2424,
      "theoretical_loss": 3.3861505684839632,
      "tokens_seen": 2342649856
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029513131313131316,
      "loss": 1.3477,
      "theoretical_loss": 3.386142913761648,
      "tokens_seen": 2342715392
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002951111111111111,
      "loss": 1.1252,
      "theoretical_loss": 3.3861352593134217,
      "tokens_seen": 2342780928
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002950909090909091,
      "loss": 1.3671,
      "theoretical_loss": 3.3861276051392672,
      "tokens_seen": 2342846464
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002950707070707071,
      "loss": 1.1082,
      "theoretical_loss": 3.3861199512391673,
      "tokens_seen": 2342912000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002950505050505051,
      "loss": 1.3716,
      "theoretical_loss": 3.386112297613104,
      "tokens_seen": 2342977536
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029503030303030305,
      "loss": 1.469,
      "theoretical_loss": 3.38610464426106,
      "tokens_seen": 2343043072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000295010101010101,
      "loss": 1.4787,
      "theoretical_loss": 3.386096991183018,
      "tokens_seen": 2343108608
    },
    {
      "epoch": 0.42,
      "objective/train/aligned_token_count": 0.47140762209892273,
      "objective/train/avg_token_score": 0.01286020502448082,
      "objective/train/docs_used": 1321245,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.682588815689087,
      "objective/train/original_loss": 4.142392635345459,
      "objective/train/theoretical_loss": 3.3860893383789605,
      "objective/train/tokens_used": 702033376,
      "objective/train/unlikelihood_loss": 0.06471728533506393,
      "theoretical_loss": 3.3860893383789605,
      "tokens_seen": 2343174144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029498989898989897,
      "loss": 1.3862,
      "theoretical_loss": 3.3860893383789605,
      "tokens_seen": 2343174144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000294969696969697,
      "loss": 1.2825,
      "theoretical_loss": 3.38608168584887,
      "tokens_seen": 2343239680
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000294949494949495,
      "loss": 1.5629,
      "theoretical_loss": 3.386074033592729,
      "tokens_seen": 2343305216
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029492929292929295,
      "loss": 1.3939,
      "theoretical_loss": 3.3860663816105196,
      "tokens_seen": 2343370752
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002949090909090909,
      "loss": 1.3254,
      "theoretical_loss": 3.386058729902225,
      "tokens_seen": 2343436288
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029488888888888886,
      "loss": 1.514,
      "theoretical_loss": 3.3860510784678275,
      "tokens_seen": 2343501824
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002948686868686869,
      "loss": 1.574,
      "theoretical_loss": 3.3860434273073095,
      "tokens_seen": 2343567360
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002948484848484849,
      "loss": 1.1568,
      "theoretical_loss": 3.3860357764206537,
      "tokens_seen": 2343632896
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029482828282828284,
      "loss": 1.5747,
      "theoretical_loss": 3.3860281258078424,
      "tokens_seen": 2343698432
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002948080808080808,
      "loss": 1.3345,
      "theoretical_loss": 3.3860204754688583,
      "tokens_seen": 2343763968
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029478787878787875,
      "loss": 1.4053,
      "theoretical_loss": 3.3860128254036845,
      "tokens_seen": 2343829504
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002947676767676768,
      "loss": 1.4436,
      "theoretical_loss": 3.3860051756123024,
      "tokens_seen": 2343895040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002947474747474748,
      "loss": 1.2996,
      "theoretical_loss": 3.3859975260946955,
      "tokens_seen": 2343960576
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029472727272727273,
      "loss": 1.4964,
      "theoretical_loss": 3.385989876850846,
      "tokens_seen": 2344026112
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002947070707070707,
      "loss": 1.4539,
      "theoretical_loss": 3.3859822278807363,
      "tokens_seen": 2344091648
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029468686868686864,
      "loss": 1.5263,
      "theoretical_loss": 3.3859745791843494,
      "tokens_seen": 2344157184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002946666666666667,
      "loss": 1.3613,
      "theoretical_loss": 3.385966930761667,
      "tokens_seen": 2344222720
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029464646464646467,
      "loss": 1.3915,
      "theoretical_loss": 3.3859592826126725,
      "tokens_seen": 2344288256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002946262626262626,
      "loss": 1.4129,
      "theoretical_loss": 3.3859516347373484,
      "tokens_seen": 2344353792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002946060606060606,
      "loss": 1.4366,
      "theoretical_loss": 3.3859439871356765,
      "tokens_seen": 2344419328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002945858585858586,
      "loss": 1.2693,
      "theoretical_loss": 3.3859363398076407,
      "tokens_seen": 2344484864
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002945656565656566,
      "loss": 1.4416,
      "theoretical_loss": 3.3859286927532217,
      "tokens_seen": 2344550400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029454545454545456,
      "loss": 1.3292,
      "theoretical_loss": 3.385921045972404,
      "tokens_seen": 2344615936
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002945252525252525,
      "loss": 1.1381,
      "theoretical_loss": 3.385913399465169,
      "tokens_seen": 2344681472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029450505050505047,
      "loss": 1.2509,
      "theoretical_loss": 3.3859057532314996,
      "tokens_seen": 2344747008
    },
    {
      "epoch": 0.42,
      "objective/train/aligned_token_count": 0.4610826075077057,
      "objective/train/avg_token_score": 0.022878697142004967,
      "objective/train/docs_used": 1322388,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4502899646759033,
      "objective/train/original_loss": 4.400845050811768,
      "objective/train/theoretical_loss": 3.385898107271378,
      "objective/train/tokens_used": 703671776,
      "objective/train/unlikelihood_loss": 0.0546007975935936,
      "theoretical_loss": 3.385898107271378,
      "tokens_seen": 2344812544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002944848484848485,
      "loss": 1.6367,
      "theoretical_loss": 3.385898107271378,
      "tokens_seen": 2344812544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002944646464646465,
      "loss": 1.339,
      "theoretical_loss": 3.385890461584787,
      "tokens_seen": 2344878080
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029444444444444445,
      "loss": 1.2468,
      "theoretical_loss": 3.3858828161717094,
      "tokens_seen": 2344943616
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002944242424242424,
      "loss": 1.3644,
      "theoretical_loss": 3.385875171032128,
      "tokens_seen": 2345009152
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002944040404040404,
      "loss": 1.6762,
      "theoretical_loss": 3.3858675261660247,
      "tokens_seen": 2345074688
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943838383838384,
      "loss": 1.1652,
      "theoretical_loss": 3.3858598815733822,
      "tokens_seen": 2345140224
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943636363636364,
      "loss": 1.4899,
      "theoretical_loss": 3.3858522372541833,
      "tokens_seen": 2345205760
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029434343434343434,
      "loss": 1.111,
      "theoretical_loss": 3.385844593208411,
      "tokens_seen": 2345271296
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943232323232323,
      "loss": 1.5833,
      "theoretical_loss": 3.3858369494360465,
      "tokens_seen": 2345336832
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002943030303030303,
      "loss": 1.2551,
      "theoretical_loss": 3.385829305937074,
      "tokens_seen": 2345402368
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029428282828282827,
      "loss": 1.3052,
      "theoretical_loss": 3.385821662711475,
      "tokens_seen": 2345467904
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002942626262626263,
      "loss": 1.3932,
      "theoretical_loss": 3.385814019759233,
      "tokens_seen": 2345533440
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029424242424242424,
      "loss": 1.3276,
      "theoretical_loss": 3.3858063770803293,
      "tokens_seen": 2345598976
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029422222222222225,
      "loss": 1.2919,
      "theoretical_loss": 3.3857987346747476,
      "tokens_seen": 2345664512
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002942020202020202,
      "loss": 1.49,
      "theoretical_loss": 3.38579109254247,
      "tokens_seen": 2345730048
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029418181818181816,
      "loss": 1.4133,
      "theoretical_loss": 3.3857834506834794,
      "tokens_seen": 2345795584
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002941616161616162,
      "loss": 1.4346,
      "theoretical_loss": 3.385775809097758,
      "tokens_seen": 2345861120
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029414141414141413,
      "loss": 1.5293,
      "theoretical_loss": 3.3857681677852884,
      "tokens_seen": 2345926656
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029412121212121214,
      "loss": 1.5854,
      "theoretical_loss": 3.3857605267460538,
      "tokens_seen": 2345992192
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002941010101010101,
      "loss": 1.3585,
      "theoretical_loss": 3.385752885980036,
      "tokens_seen": 2346057728
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029408080808080805,
      "loss": 1.3399,
      "theoretical_loss": 3.385745245487218,
      "tokens_seen": 2346123264
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029406060606060607,
      "loss": 1.315,
      "theoretical_loss": 3.385737605267583,
      "tokens_seen": 2346188800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002940404040404041,
      "loss": 1.4681,
      "theoretical_loss": 3.3857299653211124,
      "tokens_seen": 2346254336
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029402020202020203,
      "loss": 1.4397,
      "theoretical_loss": 3.3857223256477895,
      "tokens_seen": 2346319872
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000294,
      "loss": 1.3611,
      "theoretical_loss": 3.3857146862475966,
      "tokens_seen": 2346385408
    },
    {
      "epoch": 0.42,
      "objective/train/aligned_token_count": 0.39888808131217957,
      "objective/train/avg_token_score": 0.01974659226834774,
      "objective/train/docs_used": 1323122,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3201321363449097,
      "objective/train/original_loss": 4.9252028465271,
      "objective/train/theoretical_loss": 3.385707047120517,
      "objective/train/tokens_used": 705310176,
      "objective/train/unlikelihood_loss": 0.04510696604847908,
      "theoretical_loss": 3.385707047120517,
      "tokens_seen": 2346450944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029397979797979795,
      "loss": 1.2576,
      "theoretical_loss": 3.385707047120517,
      "tokens_seen": 2346450944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029395959595959596,
      "loss": 1.3288,
      "theoretical_loss": 3.3856994082665324,
      "tokens_seen": 2346516480
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029393939393939397,
      "loss": 1.4952,
      "theoretical_loss": 3.385691769685626,
      "tokens_seen": 2346582016
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002939191919191919,
      "loss": 1.5096,
      "theoretical_loss": 3.38568413137778,
      "tokens_seen": 2346647552
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938989898989899,
      "loss": 1.2154,
      "theoretical_loss": 3.3856764933429773,
      "tokens_seen": 2346713088
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938787878787879,
      "loss": 1.161,
      "theoretical_loss": 3.3856688555812005,
      "tokens_seen": 2346778624
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938585858585859,
      "loss": 1.5185,
      "theoretical_loss": 3.3856612180924324,
      "tokens_seen": 2346844160
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029383838383838386,
      "loss": 1.5155,
      "theoretical_loss": 3.385653580876655,
      "tokens_seen": 2346909696
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002938181818181818,
      "loss": 1.3608,
      "theoretical_loss": 3.3856459439338513,
      "tokens_seen": 2346975232
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937979797979798,
      "loss": 1.3051,
      "theoretical_loss": 3.3856383072640046,
      "tokens_seen": 2347040768
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937777777777778,
      "loss": 1.3274,
      "theoretical_loss": 3.385630670867096,
      "tokens_seen": 2347106304
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937575757575758,
      "loss": 1.3074,
      "theoretical_loss": 3.3856230347431095,
      "tokens_seen": 2347171840
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029373737373737375,
      "loss": 1.4971,
      "theoretical_loss": 3.385615398892027,
      "tokens_seen": 2347237376
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002937171717171717,
      "loss": 1.2324,
      "theoretical_loss": 3.3856077633138315,
      "tokens_seen": 2347302912
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029369696969696967,
      "loss": 1.2736,
      "theoretical_loss": 3.3856001280085053,
      "tokens_seen": 2347368448
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029367676767676773,
      "loss": 1.4152,
      "theoretical_loss": 3.3855924929760315,
      "tokens_seen": 2347433984
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002936565656565657,
      "loss": 1.5543,
      "theoretical_loss": 3.385584858216392,
      "tokens_seen": 2347499520
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029363636363636365,
      "loss": 1.4467,
      "theoretical_loss": 3.38557722372957,
      "tokens_seen": 2347565056
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002936161616161616,
      "loss": 1.3345,
      "theoretical_loss": 3.3855695895155478,
      "tokens_seen": 2347630592
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029359595959595956,
      "loss": 1.5153,
      "theoretical_loss": 3.3855619555743086,
      "tokens_seen": 2347696128
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002935757575757576,
      "loss": 1.3653,
      "theoretical_loss": 3.3855543219058344,
      "tokens_seen": 2347761664
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002935555555555556,
      "loss": 1.4731,
      "theoretical_loss": 3.385546688510108,
      "tokens_seen": 2347827200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029353535353535354,
      "loss": 1.4075,
      "theoretical_loss": 3.3855390553871128,
      "tokens_seen": 2347892736
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002935151515151515,
      "loss": 1.4475,
      "theoretical_loss": 3.3855314225368303,
      "tokens_seen": 2347958272
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002934949494949495,
      "loss": 1.4245,
      "theoretical_loss": 3.385523789959244,
      "tokens_seen": 2348023808
    },
    {
      "epoch": 0.42,
      "objective/train/aligned_token_count": 0.3812316656112671,
      "objective/train/avg_token_score": 0.008906248025596142,
      "objective/train/docs_used": 1324638,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3218107223510742,
      "objective/train/original_loss": 5.085808277130127,
      "objective/train/theoretical_loss": 3.385516157654336,
      "objective/train/tokens_used": 706948576,
      "objective/train/unlikelihood_loss": 0.041298095136880875,
      "theoretical_loss": 3.385516157654336,
      "tokens_seen": 2348089344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002934747474747475,
      "loss": 1.4418,
      "theoretical_loss": 3.385516157654336,
      "tokens_seen": 2348089344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002934545454545455,
      "loss": 1.3803,
      "theoretical_loss": 3.3855085256220887,
      "tokens_seen": 2348154880
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029343434343434343,
      "loss": 1.3033,
      "theoretical_loss": 3.3855008938624858,
      "tokens_seen": 2348220416
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002934141414141414,
      "loss": 1.5552,
      "theoretical_loss": 3.3854932623755087,
      "tokens_seen": 2348285952
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933939393939394,
      "loss": 1.1522,
      "theoretical_loss": 3.385485631161141,
      "tokens_seen": 2348351488
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933737373737374,
      "loss": 1.2792,
      "theoretical_loss": 3.3854780002193654,
      "tokens_seen": 2348417024
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029335353535353537,
      "loss": 1.5459,
      "theoretical_loss": 3.3854703695501636,
      "tokens_seen": 2348482560
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933333333333333,
      "loss": 1.3472,
      "theoretical_loss": 3.3854627391535193,
      "tokens_seen": 2348548096
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002933131313131313,
      "loss": 1.5468,
      "theoretical_loss": 3.3854551090294143,
      "tokens_seen": 2348613632
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932929292929293,
      "loss": 1.3396,
      "theoretical_loss": 3.385447479177832,
      "tokens_seen": 2348679168
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932727272727273,
      "loss": 1.3714,
      "theoretical_loss": 3.3854398495987548,
      "tokens_seen": 2348744704
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029325252525252526,
      "loss": 1.2384,
      "theoretical_loss": 3.385432220292165,
      "tokens_seen": 2348810240
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002932323232323232,
      "loss": 1.1892,
      "theoretical_loss": 3.3854245912580456,
      "tokens_seen": 2348875776
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029321212121212123,
      "loss": 1.2717,
      "theoretical_loss": 3.3854169624963797,
      "tokens_seen": 2348941312
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002931919191919192,
      "loss": 1.1871,
      "theoretical_loss": 3.385409334007149,
      "tokens_seen": 2349006848
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002931717171717172,
      "loss": 1.115,
      "theoretical_loss": 3.385401705790337,
      "tokens_seen": 2349072384
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029315151515151515,
      "loss": 1.3643,
      "theoretical_loss": 3.385394077845926,
      "tokens_seen": 2349137920
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002931313131313131,
      "loss": 1.2296,
      "theoretical_loss": 3.3853864501738986,
      "tokens_seen": 2349203456
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002931111111111111,
      "loss": 1.2722,
      "theoretical_loss": 3.3853788227742374,
      "tokens_seen": 2349268992
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002930909090909091,
      "loss": 1.5436,
      "theoretical_loss": 3.385371195646926,
      "tokens_seen": 2349334528
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002930707070707071,
      "loss": 1.4991,
      "theoretical_loss": 3.3853635687919454,
      "tokens_seen": 2349400064
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029305050505050505,
      "loss": 1.5477,
      "theoretical_loss": 3.38535594220928,
      "tokens_seen": 2349465600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029303030303030306,
      "loss": 1.3506,
      "theoretical_loss": 3.3853483158989115,
      "tokens_seen": 2349531136
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000293010101010101,
      "loss": 1.3633,
      "theoretical_loss": 3.3853406898608225,
      "tokens_seen": 2349596672
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029298989898989897,
      "loss": 1.327,
      "theoretical_loss": 3.385333064094996,
      "tokens_seen": 2349662208
    },
    {
      "epoch": 0.42,
      "objective/train/aligned_token_count": 0.3013807535171509,
      "objective/train/avg_token_score": 0.019836248829960823,
      "objective/train/docs_used": 1325195,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0126372575759888,
      "objective/train/original_loss": 5.093283176422119,
      "objective/train/theoretical_loss": 3.385325438601415,
      "objective/train/tokens_used": 708586976,
      "objective/train/unlikelihood_loss": 0.03497455269098282,
      "theoretical_loss": 3.385325438601415,
      "tokens_seen": 2349727744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000292969696969697,
      "loss": 1.4783,
      "theoretical_loss": 3.385325438601415,
      "tokens_seen": 2349727744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029294949494949494,
      "loss": 1.3299,
      "theoretical_loss": 3.3853178133800617,
      "tokens_seen": 2349793280
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029292929292929295,
      "loss": 1.2168,
      "theoretical_loss": 3.385310188430919,
      "tokens_seen": 2349858816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002929090909090909,
      "loss": 1.5405,
      "theoretical_loss": 3.3853025637539695,
      "tokens_seen": 2349924352
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029288888888888886,
      "loss": 1.3766,
      "theoretical_loss": 3.385294939349196,
      "tokens_seen": 2349989888
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928686868686869,
      "loss": 0.9968,
      "theoretical_loss": 3.385287315216581,
      "tokens_seen": 2350055424
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928484848484849,
      "loss": 1.1878,
      "theoretical_loss": 3.385279691356107,
      "tokens_seen": 2350120960
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029282828282828284,
      "loss": 1.5326,
      "theoretical_loss": 3.3852720677677577,
      "tokens_seen": 2350186496
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002928080808080808,
      "loss": 1.3972,
      "theoretical_loss": 3.385264444451515,
      "tokens_seen": 2350252032
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029278787878787876,
      "loss": 1.3954,
      "theoretical_loss": 3.3852568214073617,
      "tokens_seen": 2350317568
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029276767676767677,
      "loss": 1.1285,
      "theoretical_loss": 3.38524919863528,
      "tokens_seen": 2350383104
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002927474747474748,
      "loss": 1.1983,
      "theoretical_loss": 3.3852415761352534,
      "tokens_seen": 2350448640
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029272727272727274,
      "loss": 1.6663,
      "theoretical_loss": 3.3852339539072647,
      "tokens_seen": 2350514176
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002927070707070707,
      "loss": 1.3656,
      "theoretical_loss": 3.3852263319512956,
      "tokens_seen": 2350579712
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029268686868686865,
      "loss": 1.2903,
      "theoretical_loss": 3.38521871026733,
      "tokens_seen": 2350645248
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002926666666666667,
      "loss": 1.332,
      "theoretical_loss": 3.3852110888553497,
      "tokens_seen": 2350710784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029264646464646467,
      "loss": 1.3226,
      "theoretical_loss": 3.385203467715338,
      "tokens_seen": 2350776320
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029262626262626263,
      "loss": 1.3683,
      "theoretical_loss": 3.385195846847277,
      "tokens_seen": 2350841856
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002926060606060606,
      "loss": 1.3889,
      "theoretical_loss": 3.38518822625115,
      "tokens_seen": 2350907392
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029258585858585854,
      "loss": 1.6004,
      "theoretical_loss": 3.38518060592694,
      "tokens_seen": 2350972928
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002925656565656566,
      "loss": 1.1249,
      "theoretical_loss": 3.3851729858746284,
      "tokens_seen": 2351038464
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029254545454545456,
      "loss": 1.2784,
      "theoretical_loss": 3.385165366094199,
      "tokens_seen": 2351104000
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002925252525252525,
      "loss": 1.3577,
      "theoretical_loss": 3.3851577465856346,
      "tokens_seen": 2351169536
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002925050505050505,
      "loss": 1.3408,
      "theoretical_loss": 3.3851501273489175,
      "tokens_seen": 2351235072
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924848484848485,
      "loss": 1.4983,
      "theoretical_loss": 3.38514250838403,
      "tokens_seen": 2351300608
    },
    {
      "epoch": 0.42,
      "objective/train/aligned_token_count": 0.49492913484573364,
      "objective/train/avg_token_score": 0.013261919841170311,
      "objective/train/docs_used": 1325921,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4051340818405151,
      "objective/train/original_loss": 3.6577250957489014,
      "objective/train/theoretical_loss": 3.3851348896909563,
      "objective/train/tokens_used": 710225376,
      "objective/train/unlikelihood_loss": 0.06033783406019211,
      "theoretical_loss": 3.3851348896909563,
      "tokens_seen": 2351366144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924646464646465,
      "loss": 1.4179,
      "theoretical_loss": 3.3851348896909563,
      "tokens_seen": 2351366144
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029244444444444446,
      "loss": 1.3944,
      "theoretical_loss": 3.3851272712696776,
      "tokens_seen": 2351431680
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002924242424242424,
      "loss": 1.2103,
      "theoretical_loss": 3.385119653120177,
      "tokens_seen": 2351497216
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029240404040404037,
      "loss": 1.3193,
      "theoretical_loss": 3.3851120352424378,
      "tokens_seen": 2351562752
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923838383838384,
      "loss": 1.6829,
      "theoretical_loss": 3.385104417636442,
      "tokens_seen": 2351628288
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923636363636364,
      "loss": 1.2676,
      "theoretical_loss": 3.385096800302173,
      "tokens_seen": 2351693824
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029234343434343435,
      "loss": 1.4469,
      "theoretical_loss": 3.385089183239613,
      "tokens_seen": 2351759360
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923232323232323,
      "loss": 1.404,
      "theoretical_loss": 3.3850815664487452,
      "tokens_seen": 2351824896
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002923030303030303,
      "loss": 1.3104,
      "theoretical_loss": 3.385073949929552,
      "tokens_seen": 2351890432
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922828282828283,
      "loss": 1.4883,
      "theoretical_loss": 3.3850663336820164,
      "tokens_seen": 2351955968
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922626262626263,
      "loss": 1.3124,
      "theoretical_loss": 3.3850587177061207,
      "tokens_seen": 2352021504
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029224242424242424,
      "loss": 1.3919,
      "theoretical_loss": 3.385051102001848,
      "tokens_seen": 2352087040
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922222222222222,
      "loss": 1.3929,
      "theoretical_loss": 3.385043486569181,
      "tokens_seen": 2352152576
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002922020202020202,
      "loss": 1.5922,
      "theoretical_loss": 3.3850358714081024,
      "tokens_seen": 2352218112
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002921818181818182,
      "loss": 1.3539,
      "theoretical_loss": 3.385028256518595,
      "tokens_seen": 2352283648
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002921616161616162,
      "loss": 1.3798,
      "theoretical_loss": 3.385020641900642,
      "tokens_seen": 2352349184
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029214141414141414,
      "loss": 1.4083,
      "theoretical_loss": 3.385013027554225,
      "tokens_seen": 2352414720
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029212121212121215,
      "loss": 1.6837,
      "theoretical_loss": 3.3850054134793277,
      "tokens_seen": 2352480256
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002921010101010101,
      "loss": 1.4031,
      "theoretical_loss": 3.3849977996759324,
      "tokens_seen": 2352545792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002920808080808081,
      "loss": 1.3732,
      "theoretical_loss": 3.384990186144022,
      "tokens_seen": 2352611328
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029206060606060607,
      "loss": 1.2938,
      "theoretical_loss": 3.3849825728835796,
      "tokens_seen": 2352676864
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029204040404040403,
      "loss": 1.4738,
      "theoretical_loss": 3.3849749598945875,
      "tokens_seen": 2352742400
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029202020202020204,
      "loss": 1.1486,
      "theoretical_loss": 3.3849673471770285,
      "tokens_seen": 2352807936
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000292,
      "loss": 1.4867,
      "theoretical_loss": 3.3849597347308853,
      "tokens_seen": 2352873472
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000291979797979798,
      "loss": 1.3583,
      "theoretical_loss": 3.3849521225561414,
      "tokens_seen": 2352939008
    },
    {
      "epoch": 0.42,
      "objective/train/aligned_token_count": 0.40890762209892273,
      "objective/train/avg_token_score": 0.022630106657743454,
      "objective/train/docs_used": 1327039,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3502763509750366,
      "objective/train/original_loss": 4.7145185470581055,
      "objective/train/theoretical_loss": 3.3849445106527787,
      "objective/train/tokens_used": 711863776,
      "objective/train/unlikelihood_loss": 0.04300157353281975,
      "theoretical_loss": 3.3849445106527787,
      "tokens_seen": 2353004544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029195959595959596,
      "loss": 1.5215,
      "theoretical_loss": 3.3849445106527787,
      "tokens_seen": 2353004544
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000291939393939394,
      "loss": 1.2127,
      "theoretical_loss": 3.38493689902078,
      "tokens_seen": 2353070080
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029191919191919193,
      "loss": 1.3912,
      "theoretical_loss": 3.3849292876601287,
      "tokens_seen": 2353135616
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918989898989899,
      "loss": 1.2113,
      "theoretical_loss": 3.384921676570807,
      "tokens_seen": 2353201152
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918787878787879,
      "loss": 1.2496,
      "theoretical_loss": 3.384914065752798,
      "tokens_seen": 2353266688
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029185858585858586,
      "loss": 1.477,
      "theoretical_loss": 3.3849064552060844,
      "tokens_seen": 2353332224
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029183838383838387,
      "loss": 1.2464,
      "theoretical_loss": 3.3848988449306487,
      "tokens_seen": 2353397760
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002918181818181818,
      "loss": 1.3349,
      "theoretical_loss": 3.3848912349264744,
      "tokens_seen": 2353463296
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917979797979798,
      "loss": 1.2989,
      "theoretical_loss": 3.384883625193543,
      "tokens_seen": 2353528832
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917777777777778,
      "loss": 1.6739,
      "theoretical_loss": 3.3848760157318387,
      "tokens_seen": 2353594368
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917575757575758,
      "loss": 1.3725,
      "theoretical_loss": 3.3848684065413437,
      "tokens_seen": 2353659904
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029173737373737376,
      "loss": 1.3346,
      "theoretical_loss": 3.38486079762204,
      "tokens_seen": 2353725440
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002917171717171717,
      "loss": 1.4585,
      "theoretical_loss": 3.3848531889739117,
      "tokens_seen": 2353790976
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916969696969697,
      "loss": 1.5088,
      "theoretical_loss": 3.3848455805969406,
      "tokens_seen": 2353856512
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916767676767677,
      "loss": 1.3138,
      "theoretical_loss": 3.3848379724911104,
      "tokens_seen": 2353922048
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916565656565657,
      "loss": 1.4471,
      "theoretical_loss": 3.384830364656403,
      "tokens_seen": 2353987584
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029163636363636365,
      "loss": 1.2704,
      "theoretical_loss": 3.3848227570928016,
      "tokens_seen": 2354053120
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002916161616161616,
      "loss": 1.2343,
      "theoretical_loss": 3.384815149800289,
      "tokens_seen": 2354118656
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029159595959595957,
      "loss": 1.4497,
      "theoretical_loss": 3.384807542778848,
      "tokens_seen": 2354184192
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029157575757575763,
      "loss": 1.2666,
      "theoretical_loss": 3.384799936028461,
      "tokens_seen": 2354249728
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915555555555556,
      "loss": 1.4356,
      "theoretical_loss": 3.3847923295491116,
      "tokens_seen": 2354315264
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029153535353535355,
      "loss": 1.4451,
      "theoretical_loss": 3.384784723340782,
      "tokens_seen": 2354380800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002915151515151515,
      "loss": 1.3591,
      "theoretical_loss": 3.384777117403455,
      "tokens_seen": 2354446336
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029149494949494946,
      "loss": 1.1864,
      "theoretical_loss": 3.3847695117371135,
      "tokens_seen": 2354511872
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914747474747475,
      "loss": 1.6164,
      "theoretical_loss": 3.384761906341741,
      "tokens_seen": 2354577408
    },
    {
      "epoch": 0.42,
      "objective/train/aligned_token_count": 0.455278605222702,
      "objective/train/avg_token_score": 0.014150572940707207,
      "objective/train/docs_used": 1327759,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.452375888824463,
      "objective/train/original_loss": 4.775256633758545,
      "objective/train/theoretical_loss": 3.3847543012173187,
      "objective/train/tokens_used": 713502176,
      "objective/train/unlikelihood_loss": 0.05769788846373558,
      "theoretical_loss": 3.3847543012173187,
      "tokens_seen": 2354642944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914545454545455,
      "loss": 1.5235,
      "theoretical_loss": 3.3847543012173187,
      "tokens_seen": 2354642944
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029143434343434344,
      "loss": 1.1975,
      "theoretical_loss": 3.384746696363831,
      "tokens_seen": 2354708480
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002914141414141414,
      "loss": 1.2643,
      "theoretical_loss": 3.3847390917812596,
      "tokens_seen": 2354774016
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029139393939393935,
      "loss": 1.5832,
      "theoretical_loss": 3.3847314874695877,
      "tokens_seen": 2354839552
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002913737373737374,
      "loss": 1.3005,
      "theoretical_loss": 3.3847238834287987,
      "tokens_seen": 2354905088
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002913535353535354,
      "loss": 1.2276,
      "theoretical_loss": 3.384716279658875,
      "tokens_seen": 2354970624
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029133333333333333,
      "loss": 1.3797,
      "theoretical_loss": 3.3847086761597986,
      "tokens_seen": 2355036160
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002913131313131313,
      "loss": 1.2732,
      "theoretical_loss": 3.3847010729315534,
      "tokens_seen": 2355101696
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912929292929293,
      "loss": 1.4515,
      "theoretical_loss": 3.384693469974122,
      "tokens_seen": 2355167232
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912727272727273,
      "loss": 1.4209,
      "theoretical_loss": 3.3846858672874864,
      "tokens_seen": 2355232768
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029125252525252527,
      "loss": 1.2338,
      "theoretical_loss": 3.384678264871631,
      "tokens_seen": 2355298304
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912323232323232,
      "loss": 1.4176,
      "theoretical_loss": 3.384670662726537,
      "tokens_seen": 2355363840
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002912121212121212,
      "loss": 1.3808,
      "theoretical_loss": 3.3846630608521884,
      "tokens_seen": 2355429376
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002911919191919192,
      "loss": 1.5465,
      "theoretical_loss": 3.384655459248567,
      "tokens_seen": 2355494912
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002911717171717172,
      "loss": 1.2696,
      "theoretical_loss": 3.3846478579156565,
      "tokens_seen": 2355560448
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029115151515151516,
      "loss": 1.443,
      "theoretical_loss": 3.3846402568534395,
      "tokens_seen": 2355625984
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002911313131313131,
      "loss": 1.4776,
      "theoretical_loss": 3.3846326560618984,
      "tokens_seen": 2355691520
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029111111111111113,
      "loss": 1.3422,
      "theoretical_loss": 3.384625055541017,
      "tokens_seen": 2355757056
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002910909090909091,
      "loss": 1.3967,
      "theoretical_loss": 3.3846174552907766,
      "tokens_seen": 2355822592
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002910707070707071,
      "loss": 1.3379,
      "theoretical_loss": 3.3846098553111617,
      "tokens_seen": 2355888128
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029105050505050505,
      "loss": 1.504,
      "theoretical_loss": 3.3846022556021538,
      "tokens_seen": 2355953664
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000291030303030303,
      "loss": 1.1635,
      "theoretical_loss": 3.3845946561637366,
      "tokens_seen": 2356019200
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000291010101010101,
      "loss": 1.2227,
      "theoretical_loss": 3.3845870569958927,
      "tokens_seen": 2356084736
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000290989898989899,
      "loss": 1.5219,
      "theoretical_loss": 3.384579458098605,
      "tokens_seen": 2356150272
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.000290969696969697,
      "loss": 1.3421,
      "theoretical_loss": 3.3845718594718557,
      "tokens_seen": 2356215808
    },
    {
      "epoch": 0.42,
      "objective/train/aligned_token_count": 0.3460410535335541,
      "objective/train/avg_token_score": 0.027831710875034332,
      "objective/train/docs_used": 1329159,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1032332181930542,
      "objective/train/original_loss": 4.591887474060059,
      "objective/train/theoretical_loss": 3.3845642611156284,
      "objective/train/tokens_used": 715140576,
      "objective/train/unlikelihood_loss": 0.04450482130050659,
      "theoretical_loss": 3.3845642611156284,
      "tokens_seen": 2356281344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029094949494949495,
      "loss": 1.2214,
      "theoretical_loss": 3.3845642611156284,
      "tokens_seen": 2356281344
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029092929292929296,
      "loss": 1.3379,
      "theoretical_loss": 3.3845566630299055,
      "tokens_seen": 2356346880
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002909090909090909,
      "loss": 1.4552,
      "theoretical_loss": 3.384549065214671,
      "tokens_seen": 2356412416
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029088888888888887,
      "loss": 1.339,
      "theoretical_loss": 3.384541467669906,
      "tokens_seen": 2356477952
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002908686868686869,
      "loss": 1.444,
      "theoretical_loss": 3.384533870395594,
      "tokens_seen": 2356543488
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029084848484848484,
      "loss": 1.2578,
      "theoretical_loss": 3.3845262733917183,
      "tokens_seen": 2356609024
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029082828282828285,
      "loss": 1.2881,
      "theoretical_loss": 3.384518676658262,
      "tokens_seen": 2356674560
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002908080808080808,
      "loss": 1.0789,
      "theoretical_loss": 3.3845110801952063,
      "tokens_seen": 2356740096
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029078787878787876,
      "loss": 1.5228,
      "theoretical_loss": 3.384503484002536,
      "tokens_seen": 2356805632
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002907676767676768,
      "loss": 1.5573,
      "theoretical_loss": 3.3844958880802327,
      "tokens_seen": 2356871168
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002907474747474748,
      "loss": 1.471,
      "theoretical_loss": 3.38448829242828,
      "tokens_seen": 2356936704
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029072727272727274,
      "loss": 1.1686,
      "theoretical_loss": 3.3844806970466603,
      "tokens_seen": 2357002240
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002907070707070707,
      "loss": 1.669,
      "theoretical_loss": 3.384473101935357,
      "tokens_seen": 2357067776
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029068686868686866,
      "loss": 1.4405,
      "theoretical_loss": 3.384465507094352,
      "tokens_seen": 2357133312
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029066666666666667,
      "loss": 1.4692,
      "theoretical_loss": 3.384457912523629,
      "tokens_seen": 2357198848
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002906464646464647,
      "loss": 1.4492,
      "theoretical_loss": 3.3844503182231707,
      "tokens_seen": 2357264384
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029062626262626263,
      "loss": 1.4208,
      "theoretical_loss": 3.3844427241929598,
      "tokens_seen": 2357329920
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002906060606060606,
      "loss": 1.4766,
      "theoretical_loss": 3.384435130432979,
      "tokens_seen": 2357395456
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029058585858585855,
      "loss": 1.4459,
      "theoretical_loss": 3.3844275369432113,
      "tokens_seen": 2357460992
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002905656565656566,
      "loss": 1.3856,
      "theoretical_loss": 3.3844199437236404,
      "tokens_seen": 2357526528
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029054545454545457,
      "loss": 1.491,
      "theoretical_loss": 3.384412350774248,
      "tokens_seen": 2357592064
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029052525252525253,
      "loss": 1.0693,
      "theoretical_loss": 3.384404758095018,
      "tokens_seen": 2357657600
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002905050505050505,
      "loss": 1.2916,
      "theoretical_loss": 3.3843971656859315,
      "tokens_seen": 2357723136
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002904848484848485,
      "loss": 1.5648,
      "theoretical_loss": 3.384389573546974,
      "tokens_seen": 2357788672
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002904646464646465,
      "loss": 1.2684,
      "theoretical_loss": 3.384381981678126,
      "tokens_seen": 2357854208
    },
    {
      "epoch": 0.42,
      "objective/train/aligned_token_count": 0.48154935240745544,
      "objective/train/avg_token_score": 0.005818034056574106,
      "objective/train/docs_used": 1329871,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4948874711990356,
      "objective/train/original_loss": 4.238726615905762,
      "objective/train/theoretical_loss": 3.3843743900793717,
      "objective/train/tokens_used": 716778976,
      "objective/train/unlikelihood_loss": 0.054188963025808334,
      "theoretical_loss": 3.3843743900793717,
      "tokens_seen": 2357919744
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00029044444444444446,
      "loss": 1.6344,
      "theoretical_loss": 3.3843743900793717,
      "tokens_seen": 2357919744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002904242424242424,
      "loss": 1.4277,
      "theoretical_loss": 3.384366798750694,
      "tokens_seen": 2357985280
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002904040404040404,
      "loss": 1.5767,
      "theoretical_loss": 3.384359207692075,
      "tokens_seen": 2358050816
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029038383838383844,
      "loss": 1.4814,
      "theoretical_loss": 3.384351616903498,
      "tokens_seen": 2358116352
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002903636363636364,
      "loss": 1.4248,
      "theoretical_loss": 3.3843440263849462,
      "tokens_seen": 2358181888
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029034343434343436,
      "loss": 1.4056,
      "theoretical_loss": 3.3843364361364023,
      "tokens_seen": 2358247424
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002903232323232323,
      "loss": 1.6208,
      "theoretical_loss": 3.3843288461578487,
      "tokens_seen": 2358312960
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029030303030303027,
      "loss": 1.5078,
      "theoretical_loss": 3.384321256449269,
      "tokens_seen": 2358378496
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029028282828282833,
      "loss": 1.3485,
      "theoretical_loss": 3.3843136670106455,
      "tokens_seen": 2358444032
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002902626262626263,
      "loss": 1.3757,
      "theoretical_loss": 3.384306077841962,
      "tokens_seen": 2358509568
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029024242424242425,
      "loss": 1.4029,
      "theoretical_loss": 3.3842984889432,
      "tokens_seen": 2358575104
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002902222222222222,
      "loss": 1.5263,
      "theoretical_loss": 3.3842909003143435,
      "tokens_seen": 2358640640
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002902020202020202,
      "loss": 1.2675,
      "theoretical_loss": 3.3842833119553752,
      "tokens_seen": 2358706176
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029018181818181823,
      "loss": 1.3893,
      "theoretical_loss": 3.3842757238662777,
      "tokens_seen": 2358771712
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002901616161616162,
      "loss": 1.447,
      "theoretical_loss": 3.384268136047034,
      "tokens_seen": 2358837248
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029014141414141414,
      "loss": 1.2949,
      "theoretical_loss": 3.3842605484976276,
      "tokens_seen": 2358902784
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002901212121212121,
      "loss": 1.3848,
      "theoretical_loss": 3.3842529612180408,
      "tokens_seen": 2358968320
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002901010101010101,
      "loss": 1.3274,
      "theoretical_loss": 3.3842453742082563,
      "tokens_seen": 2359033856
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900808080808081,
      "loss": 1.4611,
      "theoretical_loss": 3.3842377874682574,
      "tokens_seen": 2359099392
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002900606060606061,
      "loss": 1.2985,
      "theoretical_loss": 3.3842302009980267,
      "tokens_seen": 2359164928
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029004040404040403,
      "loss": 1.4262,
      "theoretical_loss": 3.384222614797548,
      "tokens_seen": 2359230464
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000290020202020202,
      "loss": 1.2718,
      "theoretical_loss": 3.384215028866803,
      "tokens_seen": 2359296000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00029,
      "loss": 1.6012,
      "theoretical_loss": 3.3842074432057756,
      "tokens_seen": 2359361536
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000289979797979798,
      "loss": 1.4847,
      "theoretical_loss": 3.384199857814448,
      "tokens_seen": 2359427072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028995959595959597,
      "loss": 1.479,
      "theoretical_loss": 3.3841922726928035,
      "tokens_seen": 2359492608
    },
    {
      "epoch": 0.43,
      "objective/train/aligned_token_count": 0.4097629487514496,
      "objective/train/avg_token_score": 0.010148090310394764,
      "objective/train/docs_used": 1331053,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2300500869750977,
      "objective/train/original_loss": 4.669225215911865,
      "objective/train/theoretical_loss": 3.384184687840825,
      "objective/train/tokens_used": 718417376,
      "objective/train/unlikelihood_loss": 0.050324972718954086,
      "theoretical_loss": 3.384184687840825,
      "tokens_seen": 2359558144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002899393939393939,
      "loss": 1.3844,
      "theoretical_loss": 3.384184687840825,
      "tokens_seen": 2359558144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028991919191919194,
      "loss": 1.2041,
      "theoretical_loss": 3.3841771032584953,
      "tokens_seen": 2359623680
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002898989898989899,
      "loss": 1.5068,
      "theoretical_loss": 3.3841695189457974,
      "tokens_seen": 2359689216
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002898787878787879,
      "loss": 1.5442,
      "theoretical_loss": 3.384161934902714,
      "tokens_seen": 2359754752
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028985858585858586,
      "loss": 1.3691,
      "theoretical_loss": 3.384154351129229,
      "tokens_seen": 2359820288
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002898383838383838,
      "loss": 1.5031,
      "theoretical_loss": 3.384146767625324,
      "tokens_seen": 2359885824
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028981818181818183,
      "loss": 1.4352,
      "theoretical_loss": 3.3841391843909827,
      "tokens_seen": 2359951360
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002897979797979798,
      "loss": 1.4476,
      "theoretical_loss": 3.384131601426188,
      "tokens_seen": 2360016896
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002897777777777778,
      "loss": 1.434,
      "theoretical_loss": 3.384124018730922,
      "tokens_seen": 2360082432
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028975757575757576,
      "loss": 1.3902,
      "theoretical_loss": 3.3841164363051686,
      "tokens_seen": 2360147968
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028973737373737377,
      "loss": 1.5177,
      "theoretical_loss": 3.384108854148911,
      "tokens_seen": 2360213504
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002897171717171717,
      "loss": 1.3624,
      "theoretical_loss": 3.3841012722621313,
      "tokens_seen": 2360279040
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002896969696969697,
      "loss": 1.4619,
      "theoretical_loss": 3.3840936906448125,
      "tokens_seen": 2360344576
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002896767676767677,
      "loss": 1.2784,
      "theoretical_loss": 3.3840861092969377,
      "tokens_seen": 2360410112
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028965656565656565,
      "loss": 1.4228,
      "theoretical_loss": 3.38407852821849,
      "tokens_seen": 2360475648
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028963636363636366,
      "loss": 1.2324,
      "theoretical_loss": 3.3840709474094526,
      "tokens_seen": 2360541184
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002896161616161616,
      "loss": 1.3592,
      "theoretical_loss": 3.3840633668698077,
      "tokens_seen": 2360606720
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028959595959595957,
      "loss": 1.5428,
      "theoretical_loss": 3.384055786599539,
      "tokens_seen": 2360672256
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895757575757576,
      "loss": 1.3871,
      "theoretical_loss": 3.384048206598629,
      "tokens_seen": 2360737792
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895555555555556,
      "loss": 1.5465,
      "theoretical_loss": 3.3840406268670606,
      "tokens_seen": 2360803328
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028953535353535355,
      "loss": 1.3346,
      "theoretical_loss": 3.384033047404817,
      "tokens_seen": 2360868864
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002895151515151515,
      "loss": 1.1777,
      "theoretical_loss": 3.384025468211881,
      "tokens_seen": 2360934400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028949494949494947,
      "loss": 1.3053,
      "theoretical_loss": 3.384017889288236,
      "tokens_seen": 2360999936
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894747474747475,
      "loss": 1.3116,
      "theoretical_loss": 3.384010310633864,
      "tokens_seen": 2361065472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894545454545455,
      "loss": 1.5858,
      "theoretical_loss": 3.384002732248749,
      "tokens_seen": 2361131008
    },
    {
      "epoch": 0.43,
      "objective/train/aligned_token_count": 0.43194037675857544,
      "objective/train/avg_token_score": 0.022422827780246735,
      "objective/train/docs_used": 1331743,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3715589046478271,
      "objective/train/original_loss": 3.8805229663848877,
      "objective/train/theoretical_loss": 3.383995154132873,
      "objective/train/tokens_used": 720055776,
      "objective/train/unlikelihood_loss": 0.09763460606336594,
      "theoretical_loss": 3.383995154132873,
      "tokens_seen": 2361196544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028943434343434344,
      "loss": 1.3238,
      "theoretical_loss": 3.383995154132873,
      "tokens_seen": 2361196544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002894141414141414,
      "loss": 1.3955,
      "theoretical_loss": 3.38398757628622,
      "tokens_seen": 2361262080
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028939393939393936,
      "loss": 1.5074,
      "theoretical_loss": 3.383979998708772,
      "tokens_seen": 2361327616
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002893737373737374,
      "loss": 1.4602,
      "theoretical_loss": 3.3839724214005127,
      "tokens_seen": 2361393152
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002893535353535354,
      "loss": 1.5743,
      "theoretical_loss": 3.3839648443614245,
      "tokens_seen": 2361458688
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028933333333333334,
      "loss": 1.383,
      "theoretical_loss": 3.383957267591491,
      "tokens_seen": 2361524224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002893131313131313,
      "loss": 1.362,
      "theoretical_loss": 3.3839496910906943,
      "tokens_seen": 2361589760
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028929292929292925,
      "loss": 1.6377,
      "theoretical_loss": 3.3839421148590185,
      "tokens_seen": 2361655296
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002892727272727273,
      "loss": 1.3465,
      "theoretical_loss": 3.3839345388964457,
      "tokens_seen": 2361720832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028925252525252527,
      "loss": 1.3312,
      "theoretical_loss": 3.383926963202959,
      "tokens_seen": 2361786368
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028923232323232323,
      "loss": 1.4537,
      "theoretical_loss": 3.3839193877785414,
      "tokens_seen": 2361851904
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002892121212121212,
      "loss": 1.5915,
      "theoretical_loss": 3.3839118126231758,
      "tokens_seen": 2361917440
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002891919191919192,
      "loss": 1.4522,
      "theoretical_loss": 3.3839042377368456,
      "tokens_seen": 2361982976
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002891717171717172,
      "loss": 1.4073,
      "theoretical_loss": 3.3838966631195335,
      "tokens_seen": 2362048512
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028915151515151517,
      "loss": 1.3297,
      "theoretical_loss": 3.3838890887712223,
      "tokens_seen": 2362114048
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002891313131313131,
      "loss": 1.1539,
      "theoretical_loss": 3.383881514691896,
      "tokens_seen": 2362179584
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002891111111111111,
      "loss": 1.1587,
      "theoretical_loss": 3.3838739408815357,
      "tokens_seen": 2362245120
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002890909090909091,
      "loss": 1.2817,
      "theoretical_loss": 3.383866367340126,
      "tokens_seen": 2362310656
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002890707070707071,
      "loss": 1.1959,
      "theoretical_loss": 3.3838587940676494,
      "tokens_seen": 2362376192
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028905050505050506,
      "loss": 1.2907,
      "theoretical_loss": 3.3838512210640888,
      "tokens_seen": 2362441728
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000289030303030303,
      "loss": 1.106,
      "theoretical_loss": 3.383843648329427,
      "tokens_seen": 2362507264
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000289010101010101,
      "loss": 1.2883,
      "theoretical_loss": 3.3838360758636474,
      "tokens_seen": 2362572800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288989898989899,
      "loss": 1.2487,
      "theoretical_loss": 3.383828503666733,
      "tokens_seen": 2362638336
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288969696969697,
      "loss": 1.5694,
      "theoretical_loss": 3.3838209317386667,
      "tokens_seen": 2362703872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028894949494949495,
      "loss": 1.2995,
      "theoretical_loss": 3.383813360079431,
      "tokens_seen": 2362769408
    },
    {
      "epoch": 0.43,
      "objective/train/aligned_token_count": 0.28928396105766296,
      "objective/train/avg_token_score": 0.04601091518998146,
      "objective/train/docs_used": 1333010,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0358519554138184,
      "objective/train/original_loss": 5.88991641998291,
      "objective/train/theoretical_loss": 3.3838057886890094,
      "objective/train/tokens_used": 721694176,
      "objective/train/unlikelihood_loss": 0.04012729600071907,
      "theoretical_loss": 3.3838057886890094,
      "tokens_seen": 2362834944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002889292929292929,
      "loss": 1.3935,
      "theoretical_loss": 3.3838057886890094,
      "tokens_seen": 2362834944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002889090909090909,
      "loss": 1.5173,
      "theoretical_loss": 3.383798217567385,
      "tokens_seen": 2362900480
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002888888888888889,
      "loss": 1.3416,
      "theoretical_loss": 3.3837906467145404,
      "tokens_seen": 2362966016
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002888686868686869,
      "loss": 1.1838,
      "theoretical_loss": 3.383783076130459,
      "tokens_seen": 2363031552
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028884848484848484,
      "loss": 1.5555,
      "theoretical_loss": 3.383775505815124,
      "tokens_seen": 2363097088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028882828282828285,
      "loss": 1.4281,
      "theoretical_loss": 3.3837679357685175,
      "tokens_seen": 2363162624
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028882828282828285,
      "loss": 1.3372,
      "theoretical_loss": 3.3837603659906232,
      "tokens_seen": 2363228160
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002888080808080808,
      "loss": 1.3909,
      "theoretical_loss": 3.383752796481424,
      "tokens_seen": 2363293696
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002887878787878788,
      "loss": 1.2163,
      "theoretical_loss": 3.3837452272409028,
      "tokens_seen": 2363359232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002887676767676768,
      "loss": 1.4808,
      "theoretical_loss": 3.3837376582690424,
      "tokens_seen": 2363424768
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028874747474747474,
      "loss": 1.3635,
      "theoretical_loss": 3.3837300895658267,
      "tokens_seen": 2363490304
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028872727272727275,
      "loss": 1.542,
      "theoretical_loss": 3.3837225211312374,
      "tokens_seen": 2363555840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002887070707070707,
      "loss": 1.5443,
      "theoretical_loss": 3.3837149529652586,
      "tokens_seen": 2363621376
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002886868686868687,
      "loss": 1.3479,
      "theoretical_loss": 3.383707385067873,
      "tokens_seen": 2363686912
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028866666666666667,
      "loss": 1.3904,
      "theoretical_loss": 3.383699817439064,
      "tokens_seen": 2363752448
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002886464646464647,
      "loss": 1.4806,
      "theoretical_loss": 3.3836922500788136,
      "tokens_seen": 2363817984
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028862626262626264,
      "loss": 1.4541,
      "theoretical_loss": 3.3836846829871057,
      "tokens_seen": 2363883520
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002886060606060606,
      "loss": 1.498,
      "theoretical_loss": 3.3836771161639225,
      "tokens_seen": 2363949056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002885858585858586,
      "loss": 1.412,
      "theoretical_loss": 3.383669549609248,
      "tokens_seen": 2364014592
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028856565656565656,
      "loss": 1.3955,
      "theoretical_loss": 3.383661983323065,
      "tokens_seen": 2364080128
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002885454545454546,
      "loss": 1.2436,
      "theoretical_loss": 3.383654417305356,
      "tokens_seen": 2364145664
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028852525252525253,
      "loss": 1.3896,
      "theoretical_loss": 3.3836468515561045,
      "tokens_seen": 2364211200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002885050505050505,
      "loss": 1.1,
      "theoretical_loss": 3.3836392860752933,
      "tokens_seen": 2364276736
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884848484848485,
      "loss": 1.5186,
      "theoretical_loss": 3.3836317208629056,
      "tokens_seen": 2364342272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884646464646465,
      "loss": 1.2568,
      "theoretical_loss": 3.383624155918924,
      "tokens_seen": 2364407808
    },
    {
      "epoch": 0.43,
      "objective/train/aligned_token_count": 0.3367546498775482,
      "objective/train/avg_token_score": 0.03062385879456997,
      "objective/train/docs_used": 1333580,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1720012426376343,
      "objective/train/original_loss": 5.483919143676758,
      "objective/train/theoretical_loss": 3.3836165912433325,
      "objective/train/tokens_used": 723332576,
      "objective/train/unlikelihood_loss": 0.0514683723449707,
      "theoretical_loss": 3.3836165912433325,
      "tokens_seen": 2364473344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028844444444444447,
      "loss": 1.4126,
      "theoretical_loss": 3.3836165912433325,
      "tokens_seen": 2364473344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884242424242424,
      "loss": 1.299,
      "theoretical_loss": 3.3836090268361136,
      "tokens_seen": 2364538880
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002884040404040404,
      "loss": 1.4568,
      "theoretical_loss": 3.38360146269725,
      "tokens_seen": 2364604416
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883838383838384,
      "loss": 1.2094,
      "theoretical_loss": 3.383593898826725,
      "tokens_seen": 2364669952
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883636363636364,
      "loss": 1.1791,
      "theoretical_loss": 3.3835863352245217,
      "tokens_seen": 2364735488
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028834343434343436,
      "loss": 1.3631,
      "theoretical_loss": 3.383578771890623,
      "tokens_seen": 2364801024
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883232323232323,
      "loss": 1.1566,
      "theoretical_loss": 3.3835712088250123,
      "tokens_seen": 2364866560
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002883030303030303,
      "loss": 1.5564,
      "theoretical_loss": 3.383563646027672,
      "tokens_seen": 2364932096
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028828282828282834,
      "loss": 1.4775,
      "theoretical_loss": 3.383556083498586,
      "tokens_seen": 2364997632
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002882626262626263,
      "loss": 1.1822,
      "theoretical_loss": 3.3835485212377368,
      "tokens_seen": 2365063168
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028824242424242425,
      "loss": 1.2825,
      "theoretical_loss": 3.3835409592451073,
      "tokens_seen": 2365128704
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002882222222222222,
      "loss": 1.6162,
      "theoretical_loss": 3.383533397520681,
      "tokens_seen": 2365194240
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028820202020202017,
      "loss": 1.3432,
      "theoretical_loss": 3.383525836064441,
      "tokens_seen": 2365259776
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028818181818181823,
      "loss": 1.3098,
      "theoretical_loss": 3.38351827487637,
      "tokens_seen": 2365325312
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002881616161616162,
      "loss": 1.5097,
      "theoretical_loss": 3.383510713956451,
      "tokens_seen": 2365390848
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028814141414141415,
      "loss": 1.4224,
      "theoretical_loss": 3.3835031533046678,
      "tokens_seen": 2365456384
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002881212121212121,
      "loss": 1.2661,
      "theoretical_loss": 3.3834955929210024,
      "tokens_seen": 2365521920
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028810101010101006,
      "loss": 1.4468,
      "theoretical_loss": 3.3834880328054386,
      "tokens_seen": 2365587456
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002880808080808081,
      "loss": 1.4258,
      "theoretical_loss": 3.383480472957959,
      "tokens_seen": 2365652992
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002880606060606061,
      "loss": 1.0788,
      "theoretical_loss": 3.3834729133785473,
      "tokens_seen": 2365718528
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028804040404040404,
      "loss": 1.4418,
      "theoretical_loss": 3.383465354067186,
      "tokens_seen": 2365784064
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288020202020202,
      "loss": 1.7686,
      "theoretical_loss": 3.3834577950238582,
      "tokens_seen": 2365849600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000288,
      "loss": 1.3001,
      "theoretical_loss": 3.3834502362485472,
      "tokens_seen": 2365915136
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000287979797979798,
      "loss": 1.1934,
      "theoretical_loss": 3.383442677741236,
      "tokens_seen": 2365980672
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000287959595959596,
      "loss": 1.5071,
      "theoretical_loss": 3.383435119501908,
      "tokens_seen": 2366046208
    },
    {
      "epoch": 0.43,
      "objective/train/aligned_token_count": 0.4615713655948639,
      "objective/train/avg_token_score": 0.01951819844543934,
      "objective/train/docs_used": 1334672,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.498369812965393,
      "objective/train/original_loss": 4.349247932434082,
      "objective/train/theoretical_loss": 3.383427561530546,
      "objective/train/tokens_used": 724970976,
      "objective/train/unlikelihood_loss": 0.04598849639296532,
      "theoretical_loss": 3.383427561530546,
      "tokens_seen": 2366111744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028793939393939393,
      "loss": 1.6664,
      "theoretical_loss": 3.383427561530546,
      "tokens_seen": 2366111744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002879191919191919,
      "loss": 1.5305,
      "theoretical_loss": 3.3834200038271325,
      "tokens_seen": 2366177280
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878989898989899,
      "loss": 1.3154,
      "theoretical_loss": 3.3834124463916515,
      "tokens_seen": 2366242816
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878787878787879,
      "loss": 1.4503,
      "theoretical_loss": 3.383404889224085,
      "tokens_seen": 2366308352
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028785858585858587,
      "loss": 1.266,
      "theoretical_loss": 3.3833973323244178,
      "tokens_seen": 2366373888
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002878383838383838,
      "loss": 1.5802,
      "theoretical_loss": 3.383389775692631,
      "tokens_seen": 2366439424
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028781818181818184,
      "loss": 1.2874,
      "theoretical_loss": 3.3833822193287095,
      "tokens_seen": 2366504960
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002877979797979798,
      "loss": 1.2666,
      "theoretical_loss": 3.383374663232635,
      "tokens_seen": 2366570496
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002877777777777778,
      "loss": 1.4265,
      "theoretical_loss": 3.3833671074043914,
      "tokens_seen": 2366636032
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028775757575757576,
      "loss": 1.6023,
      "theoretical_loss": 3.383359551843961,
      "tokens_seen": 2366701568
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002877373737373737,
      "loss": 1.5889,
      "theoretical_loss": 3.3833519965513283,
      "tokens_seen": 2366767104
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028771717171717173,
      "loss": 1.3125,
      "theoretical_loss": 3.383344441526475,
      "tokens_seen": 2366832640
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876969696969697,
      "loss": 1.3513,
      "theoretical_loss": 3.3833368867693845,
      "tokens_seen": 2366898176
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876767676767677,
      "loss": 1.3827,
      "theoretical_loss": 3.38332933228004,
      "tokens_seen": 2366963712
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028765656565656565,
      "loss": 1.2565,
      "theoretical_loss": 3.3833217780584253,
      "tokens_seen": 2367029248
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028763636363636366,
      "loss": 1.1851,
      "theoretical_loss": 3.3833142241045224,
      "tokens_seen": 2367094784
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002876161616161616,
      "loss": 1.398,
      "theoretical_loss": 3.3833066704183152,
      "tokens_seen": 2367160320
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002875959595959596,
      "loss": 1.4702,
      "theoretical_loss": 3.3832991169997864,
      "tokens_seen": 2367225856
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002875757575757576,
      "loss": 1.4969,
      "theoretical_loss": 3.383291563848919,
      "tokens_seen": 2367291392
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028755555555555555,
      "loss": 1.288,
      "theoretical_loss": 3.3832840109656965,
      "tokens_seen": 2367356928
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028753535353535356,
      "loss": 1.2385,
      "theoretical_loss": 3.383276458350102,
      "tokens_seen": 2367422464
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002875151515151515,
      "loss": 1.3774,
      "theoretical_loss": 3.383268906002118,
      "tokens_seen": 2367488000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028749494949494947,
      "loss": 1.4814,
      "theoretical_loss": 3.383261353921728,
      "tokens_seen": 2367553536
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002874747474747475,
      "loss": 1.4596,
      "theoretical_loss": 3.383253802108915,
      "tokens_seen": 2367619072
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002874545454545455,
      "loss": 1.5521,
      "theoretical_loss": 3.383246250563663,
      "tokens_seen": 2367684608
    },
    {
      "epoch": 0.43,
      "objective/train/aligned_token_count": 0.45448437333106995,
      "objective/train/avg_token_score": 0.006110110320150852,
      "objective/train/docs_used": 1335172,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.44041109085083,
      "objective/train/original_loss": 3.8892815113067627,
      "objective/train/theoretical_loss": 3.383238699285954,
      "objective/train/tokens_used": 726609376,
      "objective/train/unlikelihood_loss": 0.0706142708659172,
      "theoretical_loss": 3.383238699285954,
      "tokens_seen": 2367750144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028743434343434345,
      "loss": 1.4402,
      "theoretical_loss": 3.383238699285954,
      "tokens_seen": 2367750144
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002874141414141414,
      "loss": 1.4211,
      "theoretical_loss": 3.383231148275771,
      "tokens_seen": 2367815680
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028739393939393936,
      "loss": 1.4792,
      "theoretical_loss": 3.383223597533098,
      "tokens_seen": 2367881216
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002873737373737374,
      "loss": 1.4774,
      "theoretical_loss": 3.383216047057918,
      "tokens_seen": 2367946752
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002873535353535354,
      "loss": 1.5526,
      "theoretical_loss": 3.383208496850213,
      "tokens_seen": 2368012288
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028733333333333334,
      "loss": 1.1989,
      "theoretical_loss": 3.3832009469099678,
      "tokens_seen": 2368077824
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002873131313131313,
      "loss": 1.5552,
      "theoretical_loss": 3.3831933972371644,
      "tokens_seen": 2368143360
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028729292929292926,
      "loss": 1.6741,
      "theoretical_loss": 3.383185847831786,
      "tokens_seen": 2368208896
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002872727272727273,
      "loss": 1.3995,
      "theoretical_loss": 3.3831782986938164,
      "tokens_seen": 2368274432
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002872525252525253,
      "loss": 1.6226,
      "theoretical_loss": 3.383170749823238,
      "tokens_seen": 2368339968
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028723232323232324,
      "loss": 1.3627,
      "theoretical_loss": 3.383163201220034,
      "tokens_seen": 2368405504
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002872121212121212,
      "loss": 1.3269,
      "theoretical_loss": 3.3831556528841884,
      "tokens_seen": 2368471040
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028719191919191915,
      "loss": 1.377,
      "theoretical_loss": 3.383148104815683,
      "tokens_seen": 2368536576
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002871717171717172,
      "loss": 1.3432,
      "theoretical_loss": 3.383140557014502,
      "tokens_seen": 2368602112
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028715151515151517,
      "loss": 1.4068,
      "theoretical_loss": 3.383133009480628,
      "tokens_seen": 2368667648
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028713131313131313,
      "loss": 1.2287,
      "theoretical_loss": 3.383125462214044,
      "tokens_seen": 2368733184
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002871111111111111,
      "loss": 1.4338,
      "theoretical_loss": 3.3831179152147337,
      "tokens_seen": 2368798720
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002870909090909091,
      "loss": 1.442,
      "theoretical_loss": 3.3831103684826798,
      "tokens_seen": 2368864256
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002870707070707071,
      "loss": 1.3788,
      "theoretical_loss": 3.383102822017866,
      "tokens_seen": 2368929792
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028705050505050506,
      "loss": 1.4295,
      "theoretical_loss": 3.3830952758202746,
      "tokens_seen": 2368995328
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000287030303030303,
      "loss": 1.3666,
      "theoretical_loss": 3.383087729889889,
      "tokens_seen": 2369060864
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000287010101010101,
      "loss": 1.3447,
      "theoretical_loss": 3.383080184226693,
      "tokens_seen": 2369126400
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028698989898989904,
      "loss": 1.1664,
      "theoretical_loss": 3.383072638830669,
      "tokens_seen": 2369191936
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000286969696969697,
      "loss": 1.3989,
      "theoretical_loss": 3.3830650937018008,
      "tokens_seen": 2369257472
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028694949494949496,
      "loss": 1.4609,
      "theoretical_loss": 3.3830575488400707,
      "tokens_seen": 2369323008
    },
    {
      "epoch": 0.43,
      "objective/train/aligned_token_count": 0.4973118305206299,
      "objective/train/avg_token_score": 0.01196665596216917,
      "objective/train/docs_used": 1336475,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.605056881904602,
      "objective/train/original_loss": 4.532284259796143,
      "objective/train/theoretical_loss": 3.3830500042454625,
      "objective/train/tokens_used": 728247776,
      "objective/train/unlikelihood_loss": 0.04307335987687111,
      "theoretical_loss": 3.3830500042454625,
      "tokens_seen": 2369388544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002869292929292929,
      "loss": 1.271,
      "theoretical_loss": 3.3830500042454625,
      "tokens_seen": 2369388544
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002869090909090909,
      "loss": 1.4417,
      "theoretical_loss": 3.3830424599179594,
      "tokens_seen": 2369454080
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028688888888888894,
      "loss": 1.4232,
      "theoretical_loss": 3.383034915857544,
      "tokens_seen": 2369519616
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868686868686869,
      "loss": 1.2976,
      "theoretical_loss": 3.3830273720642,
      "tokens_seen": 2369585152
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028684848484848485,
      "loss": 1.251,
      "theoretical_loss": 3.3830198285379103,
      "tokens_seen": 2369650688
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868282828282828,
      "loss": 1.3399,
      "theoretical_loss": 3.3830122852786584,
      "tokens_seen": 2369716224
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002868080808080808,
      "loss": 1.3407,
      "theoretical_loss": 3.3830047422864267,
      "tokens_seen": 2369781760
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028678787878787883,
      "loss": 1.279,
      "theoretical_loss": 3.382997199561199,
      "tokens_seen": 2369847296
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002867676767676768,
      "loss": 1.1155,
      "theoretical_loss": 3.382989657102959,
      "tokens_seen": 2369912832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028674747474747474,
      "loss": 1.6392,
      "theoretical_loss": 3.3829821149116883,
      "tokens_seen": 2369978368
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002867272727272727,
      "loss": 1.5048,
      "theoretical_loss": 3.3829745729873713,
      "tokens_seen": 2370043904
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002867070707070707,
      "loss": 1.3851,
      "theoretical_loss": 3.382967031329991,
      "tokens_seen": 2370109440
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002866868686868687,
      "loss": 1.2604,
      "theoretical_loss": 3.38295948993953,
      "tokens_seen": 2370174976
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002866666666666667,
      "loss": 1.5078,
      "theoretical_loss": 3.382951948815972,
      "tokens_seen": 2370240512
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028664646464646463,
      "loss": 1.3849,
      "theoretical_loss": 3.3829444079593003,
      "tokens_seen": 2370306048
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028662626262626265,
      "loss": 1.2267,
      "theoretical_loss": 3.3829368673694975,
      "tokens_seen": 2370371584
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002866060606060606,
      "loss": 1.4573,
      "theoretical_loss": 3.382929327046547,
      "tokens_seen": 2370437120
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002865858585858586,
      "loss": 1.3225,
      "theoretical_loss": 3.3829217869904324,
      "tokens_seen": 2370502656
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028656565656565657,
      "loss": 1.5227,
      "theoretical_loss": 3.3829142472011364,
      "tokens_seen": 2370568192
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028654545454545453,
      "loss": 1.4416,
      "theoretical_loss": 3.382906707678642,
      "tokens_seen": 2370633728
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028652525252525254,
      "loss": 1.4845,
      "theoretical_loss": 3.382899168422933,
      "tokens_seen": 2370699264
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002865050505050505,
      "loss": 1.4616,
      "theoretical_loss": 3.382891629433993,
      "tokens_seen": 2370764800
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002864848484848485,
      "loss": 1.3842,
      "theoretical_loss": 3.382884090711803,
      "tokens_seen": 2370830336
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028646464646464646,
      "loss": 1.2925,
      "theoretical_loss": 3.3828765522563486,
      "tokens_seen": 2370895872
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002864444444444445,
      "loss": 1.596,
      "theoretical_loss": 3.3828690140676123,
      "tokens_seen": 2370961408
    },
    {
      "epoch": 0.43,
      "objective/train/aligned_token_count": 0.3946114480495453,
      "objective/train/avg_token_score": 0.014777363277971745,
      "objective/train/docs_used": 1336773,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3121557235717773,
      "objective/train/original_loss": 4.668807506561279,
      "objective/train/theoretical_loss": 3.382861476145577,
      "objective/train/tokens_used": 729886176,
      "objective/train/unlikelihood_loss": 0.045156847685575485,
      "theoretical_loss": 3.382861476145577,
      "tokens_seen": 2371026944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028642424242424243,
      "loss": 1.1919,
      "theoretical_loss": 3.382861476145577,
      "tokens_seen": 2371026944
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002864040404040404,
      "loss": 1.4444,
      "theoretical_loss": 3.3828539384902254,
      "tokens_seen": 2371092480
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863838383838384,
      "loss": 1.2072,
      "theoretical_loss": 3.3828464011015416,
      "tokens_seen": 2371158016
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028636363636363636,
      "loss": 1.1969,
      "theoretical_loss": 3.3828388639795084,
      "tokens_seen": 2371223552
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028634343434343437,
      "loss": 1.4388,
      "theoretical_loss": 3.382831327124109,
      "tokens_seen": 2371289088
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863232323232323,
      "loss": 1.4012,
      "theoretical_loss": 3.3828237905353267,
      "tokens_seen": 2371354624
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002863030303030303,
      "loss": 1.7014,
      "theoretical_loss": 3.382816254213145,
      "tokens_seen": 2371420160
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862828282828283,
      "loss": 1.2458,
      "theoretical_loss": 3.3828087181575457,
      "tokens_seen": 2371485696
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862626262626263,
      "loss": 1.7266,
      "theoretical_loss": 3.382801182368514,
      "tokens_seen": 2371551232
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028624242424242426,
      "loss": 1.5237,
      "theoretical_loss": 3.382793646846032,
      "tokens_seen": 2371616768
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862222222222222,
      "loss": 1.4128,
      "theoretical_loss": 3.3827861115900824,
      "tokens_seen": 2371682304
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002862020202020202,
      "loss": 1.2924,
      "theoretical_loss": 3.3827785766006495,
      "tokens_seen": 2371747840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002861818181818182,
      "loss": 1.1952,
      "theoretical_loss": 3.3827710418777164,
      "tokens_seen": 2371813376
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002861616161616162,
      "loss": 1.6068,
      "theoretical_loss": 3.3827635074212656,
      "tokens_seen": 2371878912
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028614141414141415,
      "loss": 1.4706,
      "theoretical_loss": 3.382755973231281,
      "tokens_seen": 2371944448
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002861212121212121,
      "loss": 1.486,
      "theoretical_loss": 3.382748439307745,
      "tokens_seen": 2372009984
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028610101010101007,
      "loss": 1.4444,
      "theoretical_loss": 3.3827409056506417,
      "tokens_seen": 2372075520
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028608080808080813,
      "loss": 1.2881,
      "theoretical_loss": 3.382733372259954,
      "tokens_seen": 2372141056
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002860606060606061,
      "loss": 1.54,
      "theoretical_loss": 3.3827258391356647,
      "tokens_seen": 2372206592
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028604040404040404,
      "loss": 1.5966,
      "theoretical_loss": 3.3827183062777575,
      "tokens_seen": 2372272128
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000286020202020202,
      "loss": 1.3591,
      "theoretical_loss": 3.382710773686216,
      "tokens_seen": 2372337664
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028599999999999996,
      "loss": 1.445,
      "theoretical_loss": 3.3827032413610225,
      "tokens_seen": 2372403200
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000285979797979798,
      "loss": 1.3746,
      "theoretical_loss": 3.3826957093021606,
      "tokens_seen": 2372468736
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.000285959595959596,
      "loss": 1.0061,
      "theoretical_loss": 3.3826881775096136,
      "tokens_seen": 2372534272
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028593939393939394,
      "loss": 1.2972,
      "theoretical_loss": 3.3826806459833647,
      "tokens_seen": 2372599808
    },
    {
      "epoch": 0.43,
      "objective/train/aligned_token_count": 0.41758307814598083,
      "objective/train/avg_token_score": 0.018000949174165726,
      "objective/train/docs_used": 1338329,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3701995611190796,
      "objective/train/original_loss": 4.445309162139893,
      "objective/train/theoretical_loss": 3.3826731147233975,
      "objective/train/tokens_used": 731524576,
      "objective/train/unlikelihood_loss": 0.05343393608927727,
      "theoretical_loss": 3.3826731147233975,
      "tokens_seen": 2372665344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002859191919191919,
      "loss": 1.3068,
      "theoretical_loss": 3.3826731147233975,
      "tokens_seen": 2372665344
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858989898989899,
      "loss": 1.4712,
      "theoretical_loss": 3.3826655837296946,
      "tokens_seen": 2372730880
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858787878787879,
      "loss": 1.43,
      "theoretical_loss": 3.382658053002239,
      "tokens_seen": 2372796416
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858585858585859,
      "loss": 1.1567,
      "theoretical_loss": 3.3826505225410153,
      "tokens_seen": 2372861952
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028583838383838383,
      "loss": 1.2357,
      "theoretical_loss": 3.3826429923460055,
      "tokens_seen": 2372927488
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002858181818181818,
      "loss": 1.3614,
      "theoretical_loss": 3.382635462417193,
      "tokens_seen": 2372993024
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857979797979798,
      "loss": 1.4572,
      "theoretical_loss": 3.382627932754562,
      "tokens_seen": 2373058560
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857777777777778,
      "loss": 1.4098,
      "theoretical_loss": 3.3826204033580938,
      "tokens_seen": 2373124096
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028575757575757577,
      "loss": 1.4758,
      "theoretical_loss": 3.3826128742277737,
      "tokens_seen": 2373189632
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002857373737373737,
      "loss": 1.3241,
      "theoretical_loss": 3.382605345363584,
      "tokens_seen": 2373255168
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028571717171717173,
      "loss": 1.3815,
      "theoretical_loss": 3.3825978167655077,
      "tokens_seen": 2373320704
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002856969696969697,
      "loss": 1.6823,
      "theoretical_loss": 3.382590288433528,
      "tokens_seen": 2373386240
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002856767676767677,
      "loss": 1.31,
      "theoretical_loss": 3.3825827603676295,
      "tokens_seen": 2373451776
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028565656565656566,
      "loss": 1.4413,
      "theoretical_loss": 3.3825752325677936,
      "tokens_seen": 2373517312
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002856363636363636,
      "loss": 1.2879,
      "theoretical_loss": 3.382567705034005,
      "tokens_seen": 2373582848
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002856161616161616,
      "loss": 1.5229,
      "theoretical_loss": 3.382560177766246,
      "tokens_seen": 2373648384
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002855959595959596,
      "loss": 1.6247,
      "theoretical_loss": 3.3825526507645005,
      "tokens_seen": 2373713920
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002855757575757576,
      "loss": 1.2641,
      "theoretical_loss": 3.382545124028751,
      "tokens_seen": 2373779456
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028555555555555555,
      "loss": 1.3307,
      "theoretical_loss": 3.3825375975589815,
      "tokens_seen": 2373844992
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028553535353535356,
      "loss": 1.3931,
      "theoretical_loss": 3.382530071355175,
      "tokens_seen": 2373910528
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002855151515151515,
      "loss": 1.3758,
      "theoretical_loss": 3.3825225454173147,
      "tokens_seen": 2373976064
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002854949494949495,
      "loss": 1.1934,
      "theoretical_loss": 3.382515019745384,
      "tokens_seen": 2374041600
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002854747474747475,
      "loss": 1.5804,
      "theoretical_loss": 3.3825074943393663,
      "tokens_seen": 2374107136
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028545454545454544,
      "loss": 1.3004,
      "theoretical_loss": 3.382499969199244,
      "tokens_seen": 2374172672
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00028543434343434346,
      "loss": 1.5092,
      "theoretical_loss": 3.3824924443250013,
      "tokens_seen": 2374238208
    },
    {
      "epoch": 0.43,
      "objective/train/aligned_token_count": 0.44379276037216187,
      "objective/train/avg_token_score": 0.013660687953233719,
      "objective/train/docs_used": 1338943,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5327610969543457,
      "objective/train/original_loss": 4.388190746307373,
      "objective/train/theoretical_loss": 3.3824849197166214,
      "objective/train/tokens_used": 733162976,
      "objective/train/unlikelihood_loss": 0.06412756443023682,
      "theoretical_loss": 3.3824849197166214,
      "tokens_seen": 2374303744
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002854141414141414,
      "loss": 1.3118,
      "theoretical_loss": 3.3824849197166214,
      "tokens_seen": 2374303744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028539393939393937,
      "loss": 1.5673,
      "theoretical_loss": 3.382477395374087,
      "tokens_seen": 2374369280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002853737373737374,
      "loss": 1.3696,
      "theoretical_loss": 3.3824698712973817,
      "tokens_seen": 2374434816
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002853535353535354,
      "loss": 1.1372,
      "theoretical_loss": 3.3824623474864888,
      "tokens_seen": 2374500352
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028533333333333335,
      "loss": 1.3919,
      "theoretical_loss": 3.3824548239413916,
      "tokens_seen": 2374565888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002853131313131313,
      "loss": 1.358,
      "theoretical_loss": 3.3824473006620734,
      "tokens_seen": 2374631424
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002852929292929293,
      "loss": 1.5123,
      "theoretical_loss": 3.3824397776485173,
      "tokens_seen": 2374696960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028527272727272727,
      "loss": 1.3465,
      "theoretical_loss": 3.382432254900707,
      "tokens_seen": 2374762496
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002852525252525253,
      "loss": 1.4702,
      "theoretical_loss": 3.382424732418625,
      "tokens_seen": 2374828032
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028523232323232324,
      "loss": 1.2749,
      "theoretical_loss": 3.382417210202255,
      "tokens_seen": 2374893568
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002852121212121212,
      "loss": 1.37,
      "theoretical_loss": 3.3824096882515806,
      "tokens_seen": 2374959104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002851919191919192,
      "loss": 1.6034,
      "theoretical_loss": 3.382402166566585,
      "tokens_seen": 2375024640
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002851717171717172,
      "loss": 1.4731,
      "theoretical_loss": 3.382394645147251,
      "tokens_seen": 2375090176
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002851515151515152,
      "loss": 1.3706,
      "theoretical_loss": 3.382387123993562,
      "tokens_seen": 2375155712
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028513131313131313,
      "loss": 1.5478,
      "theoretical_loss": 3.382379603105502,
      "tokens_seen": 2375221248
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002851111111111111,
      "loss": 1.5693,
      "theoretical_loss": 3.382372082483053,
      "tokens_seen": 2375286784
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002850909090909091,
      "loss": 1.3201,
      "theoretical_loss": 3.3823645621261997,
      "tokens_seen": 2375352320
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002850707070707071,
      "loss": 1.3534,
      "theoretical_loss": 3.3823570420349243,
      "tokens_seen": 2375417856
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028505050505050507,
      "loss": 1.4791,
      "theoretical_loss": 3.3823495222092106,
      "tokens_seen": 2375483392
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000285030303030303,
      "loss": 1.4434,
      "theoretical_loss": 3.3823420026490423,
      "tokens_seen": 2375548928
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000285010101010101,
      "loss": 1.2365,
      "theoretical_loss": 3.3823344833544016,
      "tokens_seen": 2375614464
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028498989898989905,
      "loss": 1.2356,
      "theoretical_loss": 3.382326964325273,
      "tokens_seen": 2375680000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000284969696969697,
      "loss": 1.1734,
      "theoretical_loss": 3.3823194455616385,
      "tokens_seen": 2375745536
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028494949494949496,
      "loss": 1.2798,
      "theoretical_loss": 3.3823119270634825,
      "tokens_seen": 2375811072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002849292929292929,
      "loss": 1.4012,
      "theoretical_loss": 3.3823044088307883,
      "tokens_seen": 2375876608
    },
    {
      "epoch": 0.44,
      "objective/train/aligned_token_count": 0.5108137726783752,
      "objective/train/avg_token_score": 0.008281651884317398,
      "objective/train/docs_used": 1339545,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5575460195541382,
      "objective/train/original_loss": 4.1155500411987305,
      "objective/train/theoretical_loss": 3.3822968908635382,
      "objective/train/tokens_used": 734801376,
      "objective/train/unlikelihood_loss": 0.048413120210170746,
      "theoretical_loss": 3.3822968908635382,
      "tokens_seen": 2375942144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002849090909090909,
      "loss": 1.3599,
      "theoretical_loss": 3.3822968908635382,
      "tokens_seen": 2375942144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028488888888888894,
      "loss": 1.2925,
      "theoretical_loss": 3.3822893731617167,
      "tokens_seen": 2376007680
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002848686868686869,
      "loss": 1.521,
      "theoretical_loss": 3.382281855725306,
      "tokens_seen": 2376073216
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028484848484848485,
      "loss": 1.4509,
      "theoretical_loss": 3.38227433855429,
      "tokens_seen": 2376138752
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002848282828282828,
      "loss": 1.1725,
      "theoretical_loss": 3.3822668216486527,
      "tokens_seen": 2376204288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028480808080808077,
      "loss": 1.3792,
      "theoretical_loss": 3.3822593050083762,
      "tokens_seen": 2376269824
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028478787878787883,
      "loss": 1.5095,
      "theoretical_loss": 3.382251788633444,
      "tokens_seen": 2376335360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847676767676768,
      "loss": 1.4708,
      "theoretical_loss": 3.38224427252384,
      "tokens_seen": 2376400896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028474747474747475,
      "loss": 1.3668,
      "theoretical_loss": 3.3822367566795473,
      "tokens_seen": 2376466432
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847272727272727,
      "loss": 1.2346,
      "theoretical_loss": 3.382229241100549,
      "tokens_seen": 2376531968
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002847070707070707,
      "loss": 1.6055,
      "theoretical_loss": 3.382221725786829,
      "tokens_seen": 2376597504
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846868686868687,
      "loss": 1.5056,
      "theoretical_loss": 3.3822142107383697,
      "tokens_seen": 2376663040
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846666666666667,
      "loss": 1.5082,
      "theoretical_loss": 3.382206695955155,
      "tokens_seen": 2376728576
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028464646464646464,
      "loss": 1.6026,
      "theoretical_loss": 3.3821991814371684,
      "tokens_seen": 2376794112
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846262626262626,
      "loss": 1.3587,
      "theoretical_loss": 3.3821916671843923,
      "tokens_seen": 2376859648
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002846060606060606,
      "loss": 1.2202,
      "theoretical_loss": 3.3821841531968113,
      "tokens_seen": 2376925184
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002845858585858586,
      "loss": 1.4576,
      "theoretical_loss": 3.382176639474408,
      "tokens_seen": 2376990720
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002845656565656566,
      "loss": 1.5177,
      "theoretical_loss": 3.382169126017166,
      "tokens_seen": 2377056256
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028454545454545453,
      "loss": 1.204,
      "theoretical_loss": 3.382161612825068,
      "tokens_seen": 2377121792
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028452525252525254,
      "loss": 1.4748,
      "theoretical_loss": 3.3821540998980986,
      "tokens_seen": 2377187328
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002845050505050505,
      "loss": 1.54,
      "theoretical_loss": 3.3821465872362397,
      "tokens_seen": 2377252864
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002844848484848485,
      "loss": 1.2059,
      "theoretical_loss": 3.3821390748394755,
      "tokens_seen": 2377318400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028446464646464647,
      "loss": 1.4906,
      "theoretical_loss": 3.382131562707789,
      "tokens_seen": 2377383936
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002844444444444444,
      "loss": 1.428,
      "theoretical_loss": 3.382124050841164,
      "tokens_seen": 2377449472
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028442424242424244,
      "loss": 1.3324,
      "theoretical_loss": 3.382116539239583,
      "tokens_seen": 2377515008
    },
    {
      "epoch": 0.44,
      "objective/train/aligned_token_count": 0.4179496467113495,
      "objective/train/avg_token_score": 0.006778097245842218,
      "objective/train/docs_used": 1340666,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3509007692337036,
      "objective/train/original_loss": 4.567722320556641,
      "objective/train/theoretical_loss": 3.38210902790303,
      "objective/train/tokens_used": 736439776,
      "objective/train/unlikelihood_loss": 0.03927852585911751,
      "theoretical_loss": 3.38210902790303,
      "tokens_seen": 2377580544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002844040404040404,
      "loss": 1.4267,
      "theoretical_loss": 3.38210902790303,
      "tokens_seen": 2377580544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002843838383838384,
      "loss": 1.5695,
      "theoretical_loss": 3.3821015168314883,
      "tokens_seen": 2377646080
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028436363636363636,
      "loss": 1.3673,
      "theoretical_loss": 3.3820940060249414,
      "tokens_seen": 2377711616
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028434343434343437,
      "loss": 1.3044,
      "theoretical_loss": 3.382086495483372,
      "tokens_seen": 2377777152
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028432323232323233,
      "loss": 1.6347,
      "theoretical_loss": 3.382078985206764,
      "tokens_seen": 2377842688
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002843030303030303,
      "loss": 1.4181,
      "theoretical_loss": 3.3820714751951004,
      "tokens_seen": 2377908224
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002842828282828283,
      "loss": 1.2901,
      "theoretical_loss": 3.3820639654483653,
      "tokens_seen": 2377973760
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028426262626262625,
      "loss": 1.6701,
      "theoretical_loss": 3.382056455966541,
      "tokens_seen": 2378039296
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028424242424242427,
      "loss": 1.5209,
      "theoretical_loss": 3.3820489467496113,
      "tokens_seen": 2378104832
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002842222222222222,
      "loss": 1.3022,
      "theoretical_loss": 3.38204143779756,
      "tokens_seen": 2378170368
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002842020202020202,
      "loss": 1.392,
      "theoretical_loss": 3.38203392911037,
      "tokens_seen": 2378235904
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841818181818182,
      "loss": 1.313,
      "theoretical_loss": 3.382026420688024,
      "tokens_seen": 2378301440
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841616161616162,
      "loss": 1.3319,
      "theoretical_loss": 3.3820189125305067,
      "tokens_seen": 2378366976
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028414141414141416,
      "loss": 1.3494,
      "theoretical_loss": 3.382011404637801,
      "tokens_seen": 2378432512
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002841212121212121,
      "loss": 1.2837,
      "theoretical_loss": 3.38200389700989,
      "tokens_seen": 2378498048
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028410101010101007,
      "loss": 1.4177,
      "theoretical_loss": 3.381996389646757,
      "tokens_seen": 2378563584
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002840808080808081,
      "loss": 1.4751,
      "theoretical_loss": 3.3819888825483853,
      "tokens_seen": 2378629120
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002840606060606061,
      "loss": 1.4126,
      "theoretical_loss": 3.381981375714759,
      "tokens_seen": 2378694656
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028404040404040405,
      "loss": 1.3667,
      "theoretical_loss": 3.3819738691458605,
      "tokens_seen": 2378760192
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000284020202020202,
      "loss": 1.5295,
      "theoretical_loss": 3.3819663628416743,
      "tokens_seen": 2378825728
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028399999999999996,
      "loss": 1.4211,
      "theoretical_loss": 3.3819588568021826,
      "tokens_seen": 2378891264
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028397979797979803,
      "loss": 1.5185,
      "theoretical_loss": 3.3819513510273698,
      "tokens_seen": 2378956800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000283959595959596,
      "loss": 1.4247,
      "theoretical_loss": 3.381943845517218,
      "tokens_seen": 2379022336
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028393939393939394,
      "loss": 1.2742,
      "theoretical_loss": 3.3819363402717117,
      "tokens_seen": 2379087872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002839191919191919,
      "loss": 1.1976,
      "theoretical_loss": 3.381928835290834,
      "tokens_seen": 2379153408
    },
    {
      "epoch": 0.44,
      "objective/train/aligned_token_count": 0.3921065628528595,
      "objective/train/avg_token_score": 0.011915666982531548,
      "objective/train/docs_used": 1341274,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2844879627227783,
      "objective/train/original_loss": 4.640810012817383,
      "objective/train/theoretical_loss": 3.3819213305745683,
      "objective/train/tokens_used": 738078176,
      "objective/train/unlikelihood_loss": 0.04996131733059883,
      "theoretical_loss": 3.3819213305745683,
      "tokens_seen": 2379218944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028389898989898986,
      "loss": 1.3874,
      "theoretical_loss": 3.3819213305745683,
      "tokens_seen": 2379218944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002838787878787879,
      "loss": 1.3481,
      "theoretical_loss": 3.381913826122898,
      "tokens_seen": 2379284480
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002838585858585859,
      "loss": 1.4152,
      "theoretical_loss": 3.3819063219358063,
      "tokens_seen": 2379350016
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028383838383838384,
      "loss": 1.5741,
      "theoretical_loss": 3.3818988180132763,
      "tokens_seen": 2379415552
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002838181818181818,
      "loss": 1.302,
      "theoretical_loss": 3.381891314355292,
      "tokens_seen": 2379481088
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837979797979798,
      "loss": 1.3789,
      "theoretical_loss": 3.3818838109618365,
      "tokens_seen": 2379546624
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837777777777778,
      "loss": 1.2682,
      "theoretical_loss": 3.3818763078328935,
      "tokens_seen": 2379612160
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028375757575757577,
      "loss": 1.3848,
      "theoretical_loss": 3.3818688049684456,
      "tokens_seen": 2379677696
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028373737373737373,
      "loss": 1.3637,
      "theoretical_loss": 3.381861302368477,
      "tokens_seen": 2379743232
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002837171717171717,
      "loss": 1.5685,
      "theoretical_loss": 3.381853800032971,
      "tokens_seen": 2379808768
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836969696969697,
      "loss": 1.4117,
      "theoretical_loss": 3.3818462979619106,
      "tokens_seen": 2379874304
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836767676767677,
      "loss": 1.6689,
      "theoretical_loss": 3.381838796155279,
      "tokens_seen": 2379939840
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028365656565656566,
      "loss": 1.4027,
      "theoretical_loss": 3.3818312946130606,
      "tokens_seen": 2380005376
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002836363636363636,
      "loss": 1.4279,
      "theoretical_loss": 3.381823793335238,
      "tokens_seen": 2380070912
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028361616161616163,
      "loss": 1.4208,
      "theoretical_loss": 3.381816292321795,
      "tokens_seen": 2380136448
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028359595959595964,
      "loss": 1.6876,
      "theoretical_loss": 3.3818087915727144,
      "tokens_seen": 2380201984
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002835757575757576,
      "loss": 1.3847,
      "theoretical_loss": 3.3818012910879798,
      "tokens_seen": 2380267520
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028355555555555556,
      "loss": 1.4627,
      "theoretical_loss": 3.3817937908675755,
      "tokens_seen": 2380333056
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002835353535353535,
      "loss": 1.5326,
      "theoretical_loss": 3.3817862909114837,
      "tokens_seen": 2380398592
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002835151515151515,
      "loss": 1.608,
      "theoretical_loss": 3.3817787912196886,
      "tokens_seen": 2380464128
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028349494949494954,
      "loss": 1.3762,
      "theoretical_loss": 3.381771291792173,
      "tokens_seen": 2380529664
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002834747474747475,
      "loss": 1.2198,
      "theoretical_loss": 3.381763792628921,
      "tokens_seen": 2380595200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028345454545454545,
      "loss": 1.3058,
      "theoretical_loss": 3.3817562937299153,
      "tokens_seen": 2380660736
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002834343434343434,
      "loss": 1.3922,
      "theoretical_loss": 3.38174879509514,
      "tokens_seen": 2380726272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002834141414141414,
      "loss": 1.3416,
      "theoretical_loss": 3.381741296724578,
      "tokens_seen": 2380791808
    },
    {
      "epoch": 0.44,
      "objective/train/aligned_token_count": 0.35202834010124207,
      "objective/train/avg_token_score": 0.028343692421913147,
      "objective/train/docs_used": 1342636,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1041020154953003,
      "objective/train/original_loss": 4.720988750457764,
      "objective/train/theoretical_loss": 3.381733798618213,
      "objective/train/tokens_used": 739716576,
      "objective/train/unlikelihood_loss": 0.032304927706718445,
      "theoretical_loss": 3.381733798618213,
      "tokens_seen": 2380857344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028339393939393943,
      "loss": 1.2181,
      "theoretical_loss": 3.381733798618213,
      "tokens_seen": 2380857344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002833737373737374,
      "loss": 1.4288,
      "theoretical_loss": 3.3817263007760285,
      "tokens_seen": 2380922880
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028335353535353534,
      "loss": 1.2034,
      "theoretical_loss": 3.3817188031980074,
      "tokens_seen": 2380988416
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028333333333333335,
      "loss": 1.4006,
      "theoretical_loss": 3.3817113058841337,
      "tokens_seen": 2381053952
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002833131313131313,
      "loss": 1.3987,
      "theoretical_loss": 3.3817038088343905,
      "tokens_seen": 2381119488
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002832929292929293,
      "loss": 1.3982,
      "theoretical_loss": 3.3816963120487613,
      "tokens_seen": 2381185024
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002832727272727273,
      "loss": 1.3613,
      "theoretical_loss": 3.3816888155272298,
      "tokens_seen": 2381250560
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028325252525252524,
      "loss": 1.4329,
      "theoretical_loss": 3.381681319269779,
      "tokens_seen": 2381316096
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028323232323232325,
      "loss": 1.2718,
      "theoretical_loss": 3.381673823276392,
      "tokens_seen": 2381381632
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002832121212121212,
      "loss": 1.323,
      "theoretical_loss": 3.3816663275470535,
      "tokens_seen": 2381447168
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831919191919192,
      "loss": 1.2278,
      "theoretical_loss": 3.3816588320817456,
      "tokens_seen": 2381512704
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028317171717171717,
      "loss": 1.4736,
      "theoretical_loss": 3.381651336880453,
      "tokens_seen": 2381578240
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831515151515152,
      "loss": 1.4174,
      "theoretical_loss": 3.381643841943158,
      "tokens_seen": 2381643776
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028313131313131314,
      "loss": 1.3346,
      "theoretical_loss": 3.3816363472698443,
      "tokens_seen": 2381709312
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002831111111111111,
      "loss": 1.4491,
      "theoretical_loss": 3.3816288528604956,
      "tokens_seen": 2381774848
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002830909090909091,
      "loss": 1.3748,
      "theoretical_loss": 3.3816213587150954,
      "tokens_seen": 2381840384
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028307070707070706,
      "loss": 1.4253,
      "theoretical_loss": 3.3816138648336267,
      "tokens_seen": 2381905920
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002830505050505051,
      "loss": 1.2331,
      "theoretical_loss": 3.3816063712160735,
      "tokens_seen": 2381971456
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028303030303030303,
      "loss": 1.1187,
      "theoretical_loss": 3.381598877862419,
      "tokens_seen": 2382036992
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000283010101010101,
      "loss": 1.185,
      "theoretical_loss": 3.381591384772647,
      "tokens_seen": 2382102528
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282989898989899,
      "loss": 1.4666,
      "theoretical_loss": 3.38158389194674,
      "tokens_seen": 2382168064
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282969696969697,
      "loss": 1.2409,
      "theoretical_loss": 3.381576399384682,
      "tokens_seen": 2382233600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028294949494949497,
      "loss": 1.305,
      "theoretical_loss": 3.3815689070864567,
      "tokens_seen": 2382299136
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002829292929292929,
      "loss": 1.601,
      "theoretical_loss": 3.3815614150520474,
      "tokens_seen": 2382364672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002829090909090909,
      "loss": 1.5077,
      "theoretical_loss": 3.3815539232814373,
      "tokens_seen": 2382430208
    },
    {
      "epoch": 0.44,
      "objective/train/aligned_token_count": 0.4714687168598175,
      "objective/train/avg_token_score": 0.010577002540230751,
      "objective/train/docs_used": 1343061,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5413758754730225,
      "objective/train/original_loss": 4.354613780975342,
      "objective/train/theoretical_loss": 3.3815464317746105,
      "objective/train/tokens_used": 741354976,
      "objective/train/unlikelihood_loss": 0.05472417175769806,
      "theoretical_loss": 3.3815464317746105,
      "tokens_seen": 2382495744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828888888888889,
      "loss": 1.2853,
      "theoretical_loss": 3.3815464317746105,
      "tokens_seen": 2382495744
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828686868686869,
      "loss": 1.4258,
      "theoretical_loss": 3.3815389405315495,
      "tokens_seen": 2382561280
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028284848484848486,
      "loss": 1.5563,
      "theoretical_loss": 3.3815314495522384,
      "tokens_seen": 2382626816
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828282828282828,
      "loss": 1.5105,
      "theoretical_loss": 3.3815239588366603,
      "tokens_seen": 2382692352
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002828080808080808,
      "loss": 1.5981,
      "theoretical_loss": 3.381516468384799,
      "tokens_seen": 2382757888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028278787878787884,
      "loss": 1.3008,
      "theoretical_loss": 3.381508978196638,
      "tokens_seen": 2382823424
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002827676767676768,
      "loss": 1.4973,
      "theoretical_loss": 3.3815014882721606,
      "tokens_seen": 2382888960
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028274747474747475,
      "loss": 1.7113,
      "theoretical_loss": 3.38149399861135,
      "tokens_seen": 2382954496
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002827272727272727,
      "loss": 1.2877,
      "theoretical_loss": 3.38148650921419,
      "tokens_seen": 2383020032
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028270707070707067,
      "loss": 1.3951,
      "theoretical_loss": 3.3814790200806644,
      "tokens_seen": 2383085568
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028268686868686873,
      "loss": 1.5716,
      "theoretical_loss": 3.381471531210756,
      "tokens_seen": 2383151104
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002826666666666667,
      "loss": 1.4062,
      "theoretical_loss": 3.3814640426044487,
      "tokens_seen": 2383216640
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028264646464646465,
      "loss": 1.383,
      "theoretical_loss": 3.3814565542617254,
      "tokens_seen": 2383282176
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002826262626262626,
      "loss": 1.6649,
      "theoretical_loss": 3.3814490661825705,
      "tokens_seen": 2383347712
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002826060606060606,
      "loss": 1.2902,
      "theoretical_loss": 3.3814415783669665,
      "tokens_seen": 2383413248
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825858585858586,
      "loss": 1.2342,
      "theoretical_loss": 3.3814340908148974,
      "tokens_seen": 2383478784
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825656565656566,
      "loss": 1.3031,
      "theoretical_loss": 3.381426603526347,
      "tokens_seen": 2383544320
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028254545454545454,
      "loss": 1.3312,
      "theoretical_loss": 3.381419116501298,
      "tokens_seen": 2383609856
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825252525252525,
      "loss": 1.3918,
      "theoretical_loss": 3.381411629739735,
      "tokens_seen": 2383675392
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002825050505050505,
      "loss": 1.603,
      "theoretical_loss": 3.3814041432416397,
      "tokens_seen": 2383740928
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002824848484848485,
      "loss": 1.5161,
      "theoretical_loss": 3.3813966570069973,
      "tokens_seen": 2383806464
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002824646464646465,
      "loss": 1.3009,
      "theoretical_loss": 3.381389171035791,
      "tokens_seen": 2383872000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028244444444444443,
      "loss": 1.3383,
      "theoretical_loss": 3.3813816853280034,
      "tokens_seen": 2383937536
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028242424242424244,
      "loss": 1.3306,
      "theoretical_loss": 3.381374199883618,
      "tokens_seen": 2384003072
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002824040404040404,
      "loss": 1.5509,
      "theoretical_loss": 3.3813667147026196,
      "tokens_seen": 2384068608
    },
    {
      "epoch": 0.44,
      "objective/train/aligned_token_count": 0.3167155385017395,
      "objective/train/avg_token_score": 0.03705340996384621,
      "objective/train/docs_used": 1344107,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.83567214012146,
      "objective/train/original_loss": 5.089759826660156,
      "objective/train/theoretical_loss": 3.381359229784991,
      "objective/train/tokens_used": 742993376,
      "objective/train/unlikelihood_loss": 0.040563907474279404,
      "theoretical_loss": 3.381359229784991,
      "tokens_seen": 2384134144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002823838383838384,
      "loss": 1.264,
      "theoretical_loss": 3.381359229784991,
      "tokens_seen": 2384134144
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028236363636363637,
      "loss": 1.4175,
      "theoretical_loss": 3.3813517451307153,
      "tokens_seen": 2384199680
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002823434343434343,
      "loss": 1.475,
      "theoretical_loss": 3.3813442607397763,
      "tokens_seen": 2384265216
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028232323232323233,
      "loss": 1.5295,
      "theoretical_loss": 3.381336776612158,
      "tokens_seen": 2384330752
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002823030303030303,
      "loss": 1.5294,
      "theoretical_loss": 3.3813292927478424,
      "tokens_seen": 2384396288
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002822828282828283,
      "loss": 1.4084,
      "theoretical_loss": 3.3813218091468147,
      "tokens_seen": 2384461824
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028226262626262626,
      "loss": 1.466,
      "theoretical_loss": 3.3813143258090577,
      "tokens_seen": 2384527360
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028224242424242427,
      "loss": 1.5126,
      "theoretical_loss": 3.3813068427345545,
      "tokens_seen": 2384592896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028222222222222223,
      "loss": 1.3749,
      "theoretical_loss": 3.3812993599232892,
      "tokens_seen": 2384658432
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002822020202020202,
      "loss": 1.502,
      "theoretical_loss": 3.381291877375245,
      "tokens_seen": 2384723968
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821818181818182,
      "loss": 1.4527,
      "theoretical_loss": 3.3812843950904057,
      "tokens_seen": 2384789504
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028216161616161615,
      "loss": 1.5023,
      "theoretical_loss": 3.3812769130687546,
      "tokens_seen": 2384855040
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028214141414141416,
      "loss": 1.5412,
      "theoretical_loss": 3.381269431310275,
      "tokens_seen": 2384920576
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821212121212121,
      "loss": 1.3337,
      "theoretical_loss": 3.3812619498149505,
      "tokens_seen": 2384986112
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002821010101010101,
      "loss": 1.424,
      "theoretical_loss": 3.3812544685827652,
      "tokens_seen": 2385051648
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002820808080808081,
      "loss": 1.4564,
      "theoretical_loss": 3.381246987613702,
      "tokens_seen": 2385117184
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002820606060606061,
      "loss": 1.4897,
      "theoretical_loss": 3.3812395069077446,
      "tokens_seen": 2385182720
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028204040404040406,
      "loss": 1.371,
      "theoretical_loss": 3.3812320264648767,
      "tokens_seen": 2385248256
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000282020202020202,
      "loss": 1.4244,
      "theoretical_loss": 3.381224546285081,
      "tokens_seen": 2385313792
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028199999999999997,
      "loss": 1.244,
      "theoretical_loss": 3.3812170663683423,
      "tokens_seen": 2385379328
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000281979797979798,
      "loss": 1.2754,
      "theoretical_loss": 3.3812095867146432,
      "tokens_seen": 2385444864
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000281959595959596,
      "loss": 1.5573,
      "theoretical_loss": 3.3812021073239675,
      "tokens_seen": 2385510400
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028193939393939395,
      "loss": 1.4114,
      "theoretical_loss": 3.381194628196299,
      "tokens_seen": 2385575936
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002819191919191919,
      "loss": 1.3219,
      "theoretical_loss": 3.3811871493316206,
      "tokens_seen": 2385641472
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028189898989898986,
      "loss": 1.4295,
      "theoretical_loss": 3.3811796707299164,
      "tokens_seen": 2385707008
    },
    {
      "epoch": 0.44,
      "objective/train/aligned_token_count": 0.43493402004241943,
      "objective/train/avg_token_score": 0.010483779944479465,
      "objective/train/docs_used": 1344624,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4551893472671509,
      "objective/train/original_loss": 4.279529571533203,
      "objective/train/theoretical_loss": 3.3811721923911695,
      "objective/train/tokens_used": 744631776,
      "objective/train/unlikelihood_loss": 0.07575725018978119,
      "theoretical_loss": 3.3811721923911695,
      "tokens_seen": 2385772544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028187878787878793,
      "loss": 1.4003,
      "theoretical_loss": 3.3811721923911695,
      "tokens_seen": 2385772544
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002818585858585859,
      "loss": 1.6059,
      "theoretical_loss": 3.381164714315364,
      "tokens_seen": 2385838080
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028183838383838384,
      "loss": 1.3359,
      "theoretical_loss": 3.3811572365024825,
      "tokens_seen": 2385903616
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002818181818181818,
      "loss": 1.4396,
      "theoretical_loss": 3.3811497589525095,
      "tokens_seen": 2385969152
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817979797979798,
      "loss": 1.2297,
      "theoretical_loss": 3.381142281665428,
      "tokens_seen": 2386034688
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817777777777778,
      "loss": 1.4438,
      "theoretical_loss": 3.3811348046412215,
      "tokens_seen": 2386100224
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817575757575758,
      "loss": 1.4777,
      "theoretical_loss": 3.3811273278798737,
      "tokens_seen": 2386165760
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028173737373737373,
      "loss": 1.5505,
      "theoretical_loss": 3.3811198513813685,
      "tokens_seen": 2386231296
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002817171717171717,
      "loss": 1.5478,
      "theoretical_loss": 3.3811123751456886,
      "tokens_seen": 2386296832
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028169696969696976,
      "loss": 1.3448,
      "theoretical_loss": 3.3811048991728185,
      "tokens_seen": 2386362368
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002816767676767677,
      "loss": 1.3288,
      "theoretical_loss": 3.3810974234627413,
      "tokens_seen": 2386427904
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028165656565656567,
      "loss": 1.6563,
      "theoretical_loss": 3.38108994801544,
      "tokens_seen": 2386493440
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002816363636363636,
      "loss": 1.3731,
      "theoretical_loss": 3.3810824728308995,
      "tokens_seen": 2386558976
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002816161616161616,
      "loss": 1.2648,
      "theoretical_loss": 3.381074997909102,
      "tokens_seen": 2386624512
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028159595959595965,
      "loss": 1.2388,
      "theoretical_loss": 3.3810675232500316,
      "tokens_seen": 2386690048
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815757575757576,
      "loss": 1.2634,
      "theoretical_loss": 3.3810600488536715,
      "tokens_seen": 2386755584
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028155555555555556,
      "loss": 0.9898,
      "theoretical_loss": 3.381052574720006,
      "tokens_seen": 2386821120
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815353535353535,
      "loss": 1.3161,
      "theoretical_loss": 3.3810451008490183,
      "tokens_seen": 2386886656
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002815151515151515,
      "loss": 1.361,
      "theoretical_loss": 3.3810376272406915,
      "tokens_seen": 2386952192
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028149494949494954,
      "loss": 1.2821,
      "theoretical_loss": 3.38103015389501,
      "tokens_seen": 2387017728
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002814747474747475,
      "loss": 1.4033,
      "theoretical_loss": 3.381022680811957,
      "tokens_seen": 2387083264
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028145454545454546,
      "loss": 1.311,
      "theoretical_loss": 3.3810152079915152,
      "tokens_seen": 2387148800
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002814343434343434,
      "loss": 1.4525,
      "theoretical_loss": 3.3810077354336694,
      "tokens_seen": 2387214336
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002814141414141414,
      "loss": 1.1911,
      "theoretical_loss": 3.381000263138403,
      "tokens_seen": 2387279872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028139393939393943,
      "loss": 1.4063,
      "theoretical_loss": 3.3809927911056987,
      "tokens_seen": 2387345408
    },
    {
      "epoch": 0.44,
      "objective/train/aligned_token_count": 0.5529081225395203,
      "objective/train/avg_token_score": 0.021423161029815674,
      "objective/train/docs_used": 1344624,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7752543687820435,
      "objective/train/original_loss": 4.1110334396362305,
      "objective/train/theoretical_loss": 3.380985319335541,
      "objective/train/tokens_used": 746270176,
      "objective/train/unlikelihood_loss": 0.07261250168085098,
      "theoretical_loss": 3.380985319335541,
      "tokens_seen": 2387410944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002813737373737374,
      "loss": 1.4646,
      "theoretical_loss": 3.380985319335541,
      "tokens_seen": 2387410944
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028135353535353535,
      "loss": 1.5991,
      "theoretical_loss": 3.380977847827913,
      "tokens_seen": 2387476480
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002813333333333333,
      "loss": 1.1626,
      "theoretical_loss": 3.3809703765827983,
      "tokens_seen": 2387542016
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002813131313131313,
      "loss": 1.4584,
      "theoretical_loss": 3.3809629056001804,
      "tokens_seen": 2387607552
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028129292929292933,
      "loss": 1.2904,
      "theoretical_loss": 3.3809554348800432,
      "tokens_seen": 2387673088
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002812727272727273,
      "loss": 1.3856,
      "theoretical_loss": 3.38094796442237,
      "tokens_seen": 2387738624
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028125252525252524,
      "loss": 1.2955,
      "theoretical_loss": 3.3809404942271444,
      "tokens_seen": 2387804160
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028123232323232325,
      "loss": 1.1989,
      "theoretical_loss": 3.38093302429435,
      "tokens_seen": 2387869696
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002812121212121212,
      "loss": 1.4256,
      "theoretical_loss": 3.3809255546239707,
      "tokens_seen": 2387935232
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002811919191919192,
      "loss": 1.2467,
      "theoretical_loss": 3.3809180852159892,
      "tokens_seen": 2388000768
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002811717171717172,
      "loss": 1.4979,
      "theoretical_loss": 3.3809106160703903,
      "tokens_seen": 2388066304
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028115151515151513,
      "loss": 1.4656,
      "theoretical_loss": 3.3809031471871567,
      "tokens_seen": 2388131840
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028113131313131314,
      "loss": 1.187,
      "theoretical_loss": 3.3808956785662723,
      "tokens_seen": 2388197376
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002811111111111111,
      "loss": 1.38,
      "theoretical_loss": 3.3808882102077207,
      "tokens_seen": 2388262912
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002810909090909091,
      "loss": 1.3937,
      "theoretical_loss": 3.3808807421114855,
      "tokens_seen": 2388328448
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028107070707070707,
      "loss": 1.2744,
      "theoretical_loss": 3.38087327427755,
      "tokens_seen": 2388393984
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002810505050505051,
      "loss": 1.2549,
      "theoretical_loss": 3.3808658067058976,
      "tokens_seen": 2388459520
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028103030303030304,
      "loss": 1.19,
      "theoretical_loss": 3.380858339396513,
      "tokens_seen": 2388525056
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000281010101010101,
      "loss": 1.3395,
      "theoretical_loss": 3.3808508723493786,
      "tokens_seen": 2388590592
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000280989898989899,
      "loss": 1.4702,
      "theoretical_loss": 3.3808434055644785,
      "tokens_seen": 2388656128
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028096969696969696,
      "loss": 1.4098,
      "theoretical_loss": 3.3808359390417966,
      "tokens_seen": 2388721664
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000280949494949495,
      "loss": 1.4335,
      "theoretical_loss": 3.380828472781316,
      "tokens_seen": 2388787200
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028092929292929293,
      "loss": 1.2851,
      "theoretical_loss": 3.3808210067830204,
      "tokens_seen": 2388852736
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002809090909090909,
      "loss": 1.4252,
      "theoretical_loss": 3.3808135410468934,
      "tokens_seen": 2388918272
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808888888888889,
      "loss": 1.3942,
      "theoretical_loss": 3.380806075572919,
      "tokens_seen": 2388983808
    },
    {
      "epoch": 0.44,
      "objective/train/aligned_token_count": 0.46419844031333923,
      "objective/train/avg_token_score": 0.020229149609804153,
      "objective/train/docs_used": 1345276,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4691895246505737,
      "objective/train/original_loss": 4.4459147453308105,
      "objective/train/theoretical_loss": 3.3807986103610803,
      "objective/train/tokens_used": 747908576,
      "objective/train/unlikelihood_loss": 0.05134518817067146,
      "theoretical_loss": 3.3807986103610803,
      "tokens_seen": 2389049344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808686868686869,
      "loss": 1.3653,
      "theoretical_loss": 3.3807986103610803,
      "tokens_seen": 2389049344
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028084848484848487,
      "loss": 1.2709,
      "theoretical_loss": 3.3807911454113606,
      "tokens_seen": 2389114880
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808282828282828,
      "loss": 1.3926,
      "theoretical_loss": 3.3807836807237446,
      "tokens_seen": 2389180416
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002808080808080808,
      "loss": 1.3716,
      "theoretical_loss": 3.3807762162982153,
      "tokens_seen": 2389245952
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002807878787878788,
      "loss": 1.5344,
      "theoretical_loss": 3.380768752134756,
      "tokens_seen": 2389311488
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002807676767676768,
      "loss": 1.1028,
      "theoretical_loss": 3.3807612882333506,
      "tokens_seen": 2389377024
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028074747474747476,
      "loss": 1.3541,
      "theoretical_loss": 3.3807538245939828,
      "tokens_seen": 2389442560
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002807272727272727,
      "loss": 1.3906,
      "theoretical_loss": 3.3807463612166364,
      "tokens_seen": 2389508096
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028070707070707067,
      "loss": 1.3479,
      "theoretical_loss": 3.3807388981012942,
      "tokens_seen": 2389573632
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028068686868686874,
      "loss": 1.379,
      "theoretical_loss": 3.3807314352479407,
      "tokens_seen": 2389639168
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002806666666666667,
      "loss": 1.2722,
      "theoretical_loss": 3.3807239726565594,
      "tokens_seen": 2389704704
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028064646464646465,
      "loss": 1.462,
      "theoretical_loss": 3.3807165103271335,
      "tokens_seen": 2389770240
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002806262626262626,
      "loss": 1.3133,
      "theoretical_loss": 3.380709048259647,
      "tokens_seen": 2389835776
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028060606060606057,
      "loss": 1.3268,
      "theoretical_loss": 3.380701586454083,
      "tokens_seen": 2389901312
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028058585858585863,
      "loss": 1.3673,
      "theoretical_loss": 3.3806941249104256,
      "tokens_seen": 2389966848
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002805656565656566,
      "loss": 1.6385,
      "theoretical_loss": 3.3806866636286586,
      "tokens_seen": 2390032384
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028054545454545454,
      "loss": 1.392,
      "theoretical_loss": 3.380679202608765,
      "tokens_seen": 2390097920
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002805252525252525,
      "loss": 1.4179,
      "theoretical_loss": 3.380671741850729,
      "tokens_seen": 2390163456
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002805050505050505,
      "loss": 1.5563,
      "theoretical_loss": 3.380664281354534,
      "tokens_seen": 2390228992
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002804848484848485,
      "loss": 1.4218,
      "theoretical_loss": 3.380656821120163,
      "tokens_seen": 2390294528
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002804646464646465,
      "loss": 1.4375,
      "theoretical_loss": 3.3806493611476007,
      "tokens_seen": 2390360064
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00028044444444444444,
      "loss": 1.3986,
      "theoretical_loss": 3.3806419014368303,
      "tokens_seen": 2390425600
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002804242424242424,
      "loss": 1.2133,
      "theoretical_loss": 3.3806344419878354,
      "tokens_seen": 2390491136
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002804040404040404,
      "loss": 1.1597,
      "theoretical_loss": 3.3806269828005995,
      "tokens_seen": 2390556672
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002803838383838384,
      "loss": 1.4783,
      "theoretical_loss": 3.3806195238751067,
      "tokens_seen": 2390622208
    },
    {
      "epoch": 0.44,
      "objective/train/aligned_token_count": 0.4629765450954437,
      "objective/train/avg_token_score": 0.0038181368727236986,
      "objective/train/docs_used": 1345792,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4926215410232544,
      "objective/train/original_loss": 3.676708698272705,
      "objective/train/theoretical_loss": 3.38061206521134,
      "objective/train/tokens_used": 749546976,
      "objective/train/unlikelihood_loss": 0.08295217901468277,
      "theoretical_loss": 3.38061206521134,
      "tokens_seen": 2390687744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028036363636363637,
      "loss": 1.7014,
      "theoretical_loss": 3.38061206521134,
      "tokens_seen": 2390687744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028034343434343433,
      "loss": 1.3941,
      "theoretical_loss": 3.3806046068092837,
      "tokens_seen": 2390753280
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002803232323232323,
      "loss": 1.4447,
      "theoretical_loss": 3.380597148668921,
      "tokens_seen": 2390818816
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002803030303030303,
      "loss": 1.341,
      "theoretical_loss": 3.380589690790236,
      "tokens_seen": 2390884352
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002802828282828283,
      "loss": 1.1669,
      "theoretical_loss": 3.380582233173212,
      "tokens_seen": 2390949888
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028026262626262627,
      "loss": 1.3318,
      "theoretical_loss": 3.380574775817832,
      "tokens_seen": 2391015424
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002802424242424242,
      "loss": 1.4913,
      "theoretical_loss": 3.3805673187240806,
      "tokens_seen": 2391080960
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028022222222222223,
      "loss": 1.2693,
      "theoretical_loss": 3.3805598618919412,
      "tokens_seen": 2391146496
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002802020202020202,
      "loss": 1.632,
      "theoretical_loss": 3.3805524053213976,
      "tokens_seen": 2391212032
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002801818181818182,
      "loss": 1.3351,
      "theoretical_loss": 3.3805449490124335,
      "tokens_seen": 2391277568
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028016161616161616,
      "loss": 1.6101,
      "theoretical_loss": 3.380537492965032,
      "tokens_seen": 2391343104
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002801414141414141,
      "loss": 1.3961,
      "theoretical_loss": 3.3805300371791764,
      "tokens_seen": 2391408640
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002801212121212121,
      "loss": 1.3157,
      "theoretical_loss": 3.3805225816548523,
      "tokens_seen": 2391474176
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028010101010101014,
      "loss": 1.3766,
      "theoretical_loss": 3.380515126392041,
      "tokens_seen": 2391539712
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002800808080808081,
      "loss": 1.6082,
      "theoretical_loss": 3.380507671390728,
      "tokens_seen": 2391605248
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028006060606060605,
      "loss": 1.5636,
      "theoretical_loss": 3.380500216650896,
      "tokens_seen": 2391670784
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028004040404040406,
      "loss": 1.4588,
      "theoretical_loss": 3.380492762172529,
      "tokens_seen": 2391736320
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000280020202020202,
      "loss": 1.2098,
      "theoretical_loss": 3.3804853079556105,
      "tokens_seen": 2391801856
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00028000000000000003,
      "loss": 1.5383,
      "theoretical_loss": 3.380477854000124,
      "tokens_seen": 2391867392
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000279979797979798,
      "loss": 1.381,
      "theoretical_loss": 3.3804704003060535,
      "tokens_seen": 2391932928
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027995959595959594,
      "loss": 1.4397,
      "theoretical_loss": 3.380462946873383,
      "tokens_seen": 2391998464
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027993939393939395,
      "loss": 1.4376,
      "theoretical_loss": 3.380455493702095,
      "tokens_seen": 2392064000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002799191919191919,
      "loss": 1.4103,
      "theoretical_loss": 3.3804480407921744,
      "tokens_seen": 2392129536
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798989898989899,
      "loss": 1.4325,
      "theoretical_loss": 3.380440588143604,
      "tokens_seen": 2392195072
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798787878787879,
      "loss": 1.3514,
      "theoretical_loss": 3.380433135756368,
      "tokens_seen": 2392260608
    },
    {
      "epoch": 0.45,
      "objective/train/aligned_token_count": 0.4523460268974304,
      "objective/train/avg_token_score": 0.00830022618174553,
      "objective/train/docs_used": 1347158,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.287163257598877,
      "objective/train/original_loss": 3.8665153980255127,
      "objective/train/theoretical_loss": 3.3804256836304503,
      "objective/train/tokens_used": 751185376,
      "objective/train/unlikelihood_loss": 0.0731152817606926,
      "theoretical_loss": 3.3804256836304503,
      "tokens_seen": 2392326144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798585858585859,
      "loss": 1.5072,
      "theoretical_loss": 3.3804256836304503,
      "tokens_seen": 2392326144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027983838383838385,
      "loss": 1.3035,
      "theoretical_loss": 3.380418231765834,
      "tokens_seen": 2392391680
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002798181818181818,
      "loss": 1.2277,
      "theoretical_loss": 3.380410780162503,
      "tokens_seen": 2392457216
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002797979797979798,
      "loss": 1.3406,
      "theoretical_loss": 3.3804033288204405,
      "tokens_seen": 2392522752
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027977777777777777,
      "loss": 1.3591,
      "theoretical_loss": 3.3803958777396312,
      "tokens_seen": 2392588288
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002797575757575758,
      "loss": 1.3303,
      "theoretical_loss": 3.380388426920058,
      "tokens_seen": 2392653824
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027973737373737374,
      "loss": 1.3684,
      "theoretical_loss": 3.380380976361705,
      "tokens_seen": 2392719360
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002797171717171717,
      "loss": 1.2951,
      "theoretical_loss": 3.3803735260645555,
      "tokens_seen": 2392784896
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002796969696969697,
      "loss": 1.4226,
      "theoretical_loss": 3.3803660760285936,
      "tokens_seen": 2392850432
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002796767676767677,
      "loss": 1.4524,
      "theoretical_loss": 3.3803586262538023,
      "tokens_seen": 2392915968
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002796565656565657,
      "loss": 1.3109,
      "theoretical_loss": 3.3803511767401666,
      "tokens_seen": 2392981504
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027963636363636363,
      "loss": 1.2903,
      "theoretical_loss": 3.3803437274876686,
      "tokens_seen": 2393047040
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002796161616161616,
      "loss": 1.3995,
      "theoretical_loss": 3.3803362784962934,
      "tokens_seen": 2393112576
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002795959595959596,
      "loss": 1.2145,
      "theoretical_loss": 3.380328829766024,
      "tokens_seen": 2393178112
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002795757575757576,
      "loss": 1.4327,
      "theoretical_loss": 3.380321381296844,
      "tokens_seen": 2393243648
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027955555555555557,
      "loss": 1.3489,
      "theoretical_loss": 3.380313933088737,
      "tokens_seen": 2393309184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002795353535353535,
      "loss": 1.3499,
      "theoretical_loss": 3.380306485141687,
      "tokens_seen": 2393374720
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002795151515151515,
      "loss": 1.3077,
      "theoretical_loss": 3.380299037455678,
      "tokens_seen": 2393440256
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027949494949494955,
      "loss": 1.2469,
      "theoretical_loss": 3.3802915900306933,
      "tokens_seen": 2393505792
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794747474747475,
      "loss": 1.2864,
      "theoretical_loss": 3.3802841428667163,
      "tokens_seen": 2393571328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027945454545454546,
      "loss": 1.3159,
      "theoretical_loss": 3.3802766959637314,
      "tokens_seen": 2393636864
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794343434343434,
      "loss": 1.3202,
      "theoretical_loss": 3.380269249321722,
      "tokens_seen": 2393702400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002794141414141414,
      "loss": 1.2987,
      "theoretical_loss": 3.3802618029406717,
      "tokens_seen": 2393767936
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027939393939393944,
      "loss": 1.3117,
      "theoretical_loss": 3.3802543568205645,
      "tokens_seen": 2393833472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002793737373737374,
      "loss": 1.3603,
      "theoretical_loss": 3.3802469109613837,
      "tokens_seen": 2393899008
    },
    {
      "epoch": 0.45,
      "objective/train/aligned_token_count": 0.49242424964904785,
      "objective/train/avg_token_score": 0.021008413285017014,
      "objective/train/docs_used": 1347750,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5664385557174683,
      "objective/train/original_loss": 3.8690271377563477,
      "objective/train/theoretical_loss": 3.3802394653631134,
      "objective/train/tokens_used": 752823776,
      "objective/train/unlikelihood_loss": 0.055746890604496,
      "theoretical_loss": 3.3802394653631134,
      "tokens_seen": 2393964544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027935353535353535,
      "loss": 1.5735,
      "theoretical_loss": 3.3802394653631134,
      "tokens_seen": 2393964544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002793333333333333,
      "loss": 1.3153,
      "theoretical_loss": 3.380232020025737,
      "tokens_seen": 2394030080
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002793131313131313,
      "loss": 1.1486,
      "theoretical_loss": 3.3802245749492386,
      "tokens_seen": 2394095616
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027929292929292933,
      "loss": 1.3841,
      "theoretical_loss": 3.3802171301336017,
      "tokens_seen": 2394161152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002792727272727273,
      "loss": 1.4639,
      "theoretical_loss": 3.38020968557881,
      "tokens_seen": 2394226688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027925252525252525,
      "loss": 1.4286,
      "theoretical_loss": 3.380202241284847,
      "tokens_seen": 2394292224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002792323232323232,
      "loss": 1.3575,
      "theoretical_loss": 3.3801947972516966,
      "tokens_seen": 2394357760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002792121212121212,
      "loss": 1.5849,
      "theoretical_loss": 3.3801873534793425,
      "tokens_seen": 2394423296
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002791919191919192,
      "loss": 1.3265,
      "theoretical_loss": 3.3801799099677687,
      "tokens_seen": 2394488832
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002791717171717172,
      "loss": 1.2447,
      "theoretical_loss": 3.380172466716959,
      "tokens_seen": 2394554368
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027915151515151514,
      "loss": 1.222,
      "theoretical_loss": 3.3801650237268968,
      "tokens_seen": 2394619904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027913131313131315,
      "loss": 1.1653,
      "theoretical_loss": 3.3801575809975652,
      "tokens_seen": 2394685440
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002791111111111111,
      "loss": 1.2802,
      "theoretical_loss": 3.380150138528949,
      "tokens_seen": 2394750976
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002790909090909091,
      "loss": 1.2354,
      "theoretical_loss": 3.380142696321032,
      "tokens_seen": 2394816512
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002790707070707071,
      "loss": 1.2586,
      "theoretical_loss": 3.3801352543737972,
      "tokens_seen": 2394882048
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027905050505050503,
      "loss": 1.3284,
      "theoretical_loss": 3.3801278126872285,
      "tokens_seen": 2394947584
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027903030303030304,
      "loss": 1.2358,
      "theoretical_loss": 3.3801203712613095,
      "tokens_seen": 2395013120
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000279010101010101,
      "loss": 1.3904,
      "theoretical_loss": 3.3801129300960246,
      "tokens_seen": 2395078656
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000278989898989899,
      "loss": 1.4965,
      "theoretical_loss": 3.380105489191357,
      "tokens_seen": 2395144192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027896969696969697,
      "loss": 1.5838,
      "theoretical_loss": 3.380098048547291,
      "tokens_seen": 2395209728
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000278949494949495,
      "loss": 1.2873,
      "theoretical_loss": 3.380090608163809,
      "tokens_seen": 2395275264
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027892929292929294,
      "loss": 1.3179,
      "theoretical_loss": 3.3800831680408963,
      "tokens_seen": 2395340800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002789090909090909,
      "loss": 1.3696,
      "theoretical_loss": 3.380075728178536,
      "tokens_seen": 2395406336
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002788888888888889,
      "loss": 1.6295,
      "theoretical_loss": 3.380068288576712,
      "tokens_seen": 2395471872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027886868686868686,
      "loss": 1.3536,
      "theoretical_loss": 3.3800608492354076,
      "tokens_seen": 2395537408
    },
    {
      "epoch": 0.45,
      "objective/train/aligned_token_count": 0.4375,
      "objective/train/avg_token_score": 0.010807149112224579,
      "objective/train/docs_used": 1348886,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3535983562469482,
      "objective/train/original_loss": 4.886846542358398,
      "objective/train/theoretical_loss": 3.3800534101546065,
      "objective/train/tokens_used": 754462176,
      "objective/train/unlikelihood_loss": 0.03203345090150833,
      "theoretical_loss": 3.3800534101546065,
      "tokens_seen": 2395602944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027884848484848487,
      "loss": 1.2346,
      "theoretical_loss": 3.3800534101546065,
      "tokens_seen": 2395602944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027882828282828283,
      "loss": 1.5895,
      "theoretical_loss": 3.3800459713342934,
      "tokens_seen": 2395668480
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002788080808080808,
      "loss": 1.3255,
      "theoretical_loss": 3.380038532774451,
      "tokens_seen": 2395734016
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787878787878788,
      "loss": 1.4681,
      "theoretical_loss": 3.380031094475064,
      "tokens_seen": 2395799552
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787676767676768,
      "loss": 1.3716,
      "theoretical_loss": 3.3800236564361157,
      "tokens_seen": 2395865088
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027874747474747476,
      "loss": 1.4271,
      "theoretical_loss": 3.3800162186575893,
      "tokens_seen": 2395930624
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787272727272727,
      "loss": 1.523,
      "theoretical_loss": 3.3800087811394697,
      "tokens_seen": 2395996160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002787070707070707,
      "loss": 1.284,
      "theoretical_loss": 3.3800013438817396,
      "tokens_seen": 2396061696
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786868686868687,
      "loss": 1.3963,
      "theoretical_loss": 3.3799939068843834,
      "tokens_seen": 2396127232
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786666666666667,
      "loss": 1.3333,
      "theoretical_loss": 3.3799864701473847,
      "tokens_seen": 2396192768
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027864646464646466,
      "loss": 1.2567,
      "theoretical_loss": 3.379979033670727,
      "tokens_seen": 2396258304
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002786262626262626,
      "loss": 1.3232,
      "theoretical_loss": 3.3799715974543947,
      "tokens_seen": 2396323840
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027860606060606057,
      "loss": 1.1253,
      "theoretical_loss": 3.379964161498371,
      "tokens_seen": 2396389376
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027858585858585864,
      "loss": 1.4086,
      "theoretical_loss": 3.3799567258026397,
      "tokens_seen": 2396454912
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002785656565656566,
      "loss": 1.3954,
      "theoretical_loss": 3.3799492903671844,
      "tokens_seen": 2396520448
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027854545454545455,
      "loss": 1.3583,
      "theoretical_loss": 3.3799418551919898,
      "tokens_seen": 2396585984
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002785252525252525,
      "loss": 1.2857,
      "theoretical_loss": 3.379934420277039,
      "tokens_seen": 2396651520
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027850505050505046,
      "loss": 1.42,
      "theoretical_loss": 3.3799269856223155,
      "tokens_seen": 2396717056
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027848484848484853,
      "loss": 1.4374,
      "theoretical_loss": 3.379919551227804,
      "tokens_seen": 2396782592
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784646464646465,
      "loss": 1.4401,
      "theoretical_loss": 3.379912117093487,
      "tokens_seen": 2396848128
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027844444444444444,
      "loss": 1.392,
      "theoretical_loss": 3.3799046832193493,
      "tokens_seen": 2396913664
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784242424242424,
      "loss": 1.1728,
      "theoretical_loss": 3.3798972496053743,
      "tokens_seen": 2396979200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002784040404040404,
      "loss": 1.2069,
      "theoretical_loss": 3.379889816251546,
      "tokens_seen": 2397044736
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002783838383838384,
      "loss": 1.1546,
      "theoretical_loss": 3.3798823831578475,
      "tokens_seen": 2397110272
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002783636363636364,
      "loss": 1.3609,
      "theoretical_loss": 3.3798749503242638,
      "tokens_seen": 2397175808
    },
    {
      "epoch": 0.45,
      "objective/train/aligned_token_count": 0.4404325485229492,
      "objective/train/avg_token_score": 0.01453663595020771,
      "objective/train/docs_used": 1349689,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5041776895523071,
      "objective/train/original_loss": 4.179414749145508,
      "objective/train/theoretical_loss": 3.3798675177507773,
      "objective/train/tokens_used": 756100576,
      "objective/train/unlikelihood_loss": 0.08225324749946594,
      "theoretical_loss": 3.3798675177507773,
      "tokens_seen": 2397241344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027834343434343434,
      "loss": 1.4081,
      "theoretical_loss": 3.3798675177507773,
      "tokens_seen": 2397241344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002783232323232323,
      "loss": 1.2035,
      "theoretical_loss": 3.379860085437373,
      "tokens_seen": 2397306880
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027830303030303036,
      "loss": 1.5473,
      "theoretical_loss": 3.3798526533840336,
      "tokens_seen": 2397372416
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002782828282828283,
      "loss": 1.5783,
      "theoretical_loss": 3.379845221590744,
      "tokens_seen": 2397437952
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027826262626262627,
      "loss": 1.3014,
      "theoretical_loss": 3.379837790057487,
      "tokens_seen": 2397503488
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027824242424242423,
      "loss": 1.4455,
      "theoretical_loss": 3.3798303587842473,
      "tokens_seen": 2397569024
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002782222222222222,
      "loss": 1.5219,
      "theoretical_loss": 3.3798229277710083,
      "tokens_seen": 2397634560
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027820202020202025,
      "loss": 1.2654,
      "theoretical_loss": 3.3798154970177534,
      "tokens_seen": 2397700096
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002781818181818182,
      "loss": 1.2638,
      "theoretical_loss": 3.3798080665244665,
      "tokens_seen": 2397765632
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027816161616161616,
      "loss": 1.242,
      "theoretical_loss": 3.379800636291132,
      "tokens_seen": 2397831168
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002781414141414141,
      "loss": 1.3277,
      "theoretical_loss": 3.379793206317734,
      "tokens_seen": 2397896704
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027812121212121213,
      "loss": 1.2848,
      "theoretical_loss": 3.3797857766042547,
      "tokens_seen": 2397962240
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027810101010101014,
      "loss": 1.3055,
      "theoretical_loss": 3.379778347150679,
      "tokens_seen": 2398027776
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002780808080808081,
      "loss": 1.5947,
      "theoretical_loss": 3.3797709179569906,
      "tokens_seen": 2398093312
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027806060606060606,
      "loss": 1.4486,
      "theoretical_loss": 3.379763489023173,
      "tokens_seen": 2398158848
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000278040404040404,
      "loss": 1.6998,
      "theoretical_loss": 3.3797560603492105,
      "tokens_seen": 2398224384
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000278020202020202,
      "loss": 1.4201,
      "theoretical_loss": 3.3797486319350867,
      "tokens_seen": 2398289920
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027800000000000004,
      "loss": 1.4088,
      "theoretical_loss": 3.3797412037807857,
      "tokens_seen": 2398355456
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277979797979798,
      "loss": 1.4876,
      "theoretical_loss": 3.3797337758862906,
      "tokens_seen": 2398420992
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027795959595959595,
      "loss": 1.5457,
      "theoretical_loss": 3.3797263482515856,
      "tokens_seen": 2398486528
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027793939393939396,
      "loss": 1.5658,
      "theoretical_loss": 3.3797189208766545,
      "tokens_seen": 2398552064
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002779191919191919,
      "loss": 1.2483,
      "theoretical_loss": 3.3797114937614814,
      "tokens_seen": 2398617600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027789898989898993,
      "loss": 1.574,
      "theoretical_loss": 3.37970406690605,
      "tokens_seen": 2398683136
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002778787878787879,
      "loss": 1.4241,
      "theoretical_loss": 3.3796966403103434,
      "tokens_seen": 2398748672
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027785858585858584,
      "loss": 0.9828,
      "theoretical_loss": 3.3796892139743466,
      "tokens_seen": 2398814208
    },
    {
      "epoch": 0.45,
      "objective/train/aligned_token_count": 0.42100438475608826,
      "objective/train/avg_token_score": 0.007781416177749634,
      "objective/train/docs_used": 1350650,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.393656849861145,
      "objective/train/original_loss": 4.208992958068848,
      "objective/train/theoretical_loss": 3.3796817878980425,
      "objective/train/tokens_used": 757738976,
      "objective/train/unlikelihood_loss": 0.06533461809158325,
      "theoretical_loss": 3.3796817878980425,
      "tokens_seen": 2398879744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027783838383838385,
      "loss": 1.1971,
      "theoretical_loss": 3.3796817878980425,
      "tokens_seen": 2398879744
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002778181818181818,
      "loss": 1.2063,
      "theoretical_loss": 3.379674362081415,
      "tokens_seen": 2398945280
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777979797979798,
      "loss": 1.2698,
      "theoretical_loss": 3.3796669365244485,
      "tokens_seen": 2399010816
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777777777777778,
      "loss": 1.4213,
      "theoretical_loss": 3.3796595112271266,
      "tokens_seen": 2399076352
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777575757575758,
      "loss": 1.2401,
      "theoretical_loss": 3.379652086189433,
      "tokens_seen": 2399141888
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027773737373737375,
      "loss": 1.365,
      "theoretical_loss": 3.3796446614113513,
      "tokens_seen": 2399207424
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002777171717171717,
      "loss": 1.384,
      "theoretical_loss": 3.379637236892866,
      "tokens_seen": 2399272960
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002776969696969697,
      "loss": 1.4138,
      "theoretical_loss": 3.37962981263396,
      "tokens_seen": 2399338496
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027767676767676767,
      "loss": 1.4686,
      "theoretical_loss": 3.379622388634618,
      "tokens_seen": 2399404032
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002776565656565657,
      "loss": 1.4807,
      "theoretical_loss": 3.3796149648948237,
      "tokens_seen": 2399469568
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027763636363636364,
      "loss": 1.5021,
      "theoretical_loss": 3.3796075414145603,
      "tokens_seen": 2399535104
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002776161616161616,
      "loss": 1.3843,
      "theoretical_loss": 3.3796001181938125,
      "tokens_seen": 2399600640
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002775959595959596,
      "loss": 1.5654,
      "theoretical_loss": 3.3795926952325637,
      "tokens_seen": 2399666176
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002775757575757576,
      "loss": 1.3517,
      "theoretical_loss": 3.3795852725307975,
      "tokens_seen": 2399731712
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002775555555555556,
      "loss": 1.463,
      "theoretical_loss": 3.3795778500884976,
      "tokens_seen": 2399797248
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027753535353535353,
      "loss": 1.4221,
      "theoretical_loss": 3.379570427905649,
      "tokens_seen": 2399862784
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002775151515151515,
      "loss": 1.4883,
      "theoretical_loss": 3.3795630059822344,
      "tokens_seen": 2399928320
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774949494949495,
      "loss": 1.5241,
      "theoretical_loss": 3.379555584318238,
      "tokens_seen": 2399993856
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774747474747475,
      "loss": 1.4847,
      "theoretical_loss": 3.379548162913644,
      "tokens_seen": 2400059392
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027745454545454547,
      "loss": 1.443,
      "theoretical_loss": 3.379540741768436,
      "tokens_seen": 2400124928
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774343434343434,
      "loss": 1.5061,
      "theoretical_loss": 3.3795333208825973,
      "tokens_seen": 2400190464
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002774141414141414,
      "loss": 1.4144,
      "theoretical_loss": 3.3795259002561124,
      "tokens_seen": 2400256000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027739393939393945,
      "loss": 1.3557,
      "theoretical_loss": 3.379518479888965,
      "tokens_seen": 2400321536
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002773737373737374,
      "loss": 1.2024,
      "theoretical_loss": 3.379511059781139,
      "tokens_seen": 2400387072
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027735353535353536,
      "loss": 1.2473,
      "theoretical_loss": 3.3795036399326186,
      "tokens_seen": 2400452608
    },
    {
      "epoch": 0.45,
      "objective/train/aligned_token_count": 0.3300953209400177,
      "objective/train/avg_token_score": 0.025813382118940353,
      "objective/train/docs_used": 1352096,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1708762645721436,
      "objective/train/original_loss": 5.187501907348633,
      "objective/train/theoretical_loss": 3.3794962203433867,
      "objective/train/tokens_used": 759377376,
      "objective/train/unlikelihood_loss": 0.03745998814702034,
      "theoretical_loss": 3.3794962203433867,
      "tokens_seen": 2400518144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002773333333333333,
      "loss": 1.4941,
      "theoretical_loss": 3.3794962203433867,
      "tokens_seen": 2400518144
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002773131313131313,
      "loss": 1.2409,
      "theoretical_loss": 3.3794888010134283,
      "tokens_seen": 2400583680
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027729292929292934,
      "loss": 1.3939,
      "theoretical_loss": 3.379481381942726,
      "tokens_seen": 2400649216
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002772727272727273,
      "loss": 1.5559,
      "theoretical_loss": 3.379473963131265,
      "tokens_seen": 2400714752
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027725252525252525,
      "loss": 1.2818,
      "theoretical_loss": 3.3794665445790284,
      "tokens_seen": 2400780288
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002772323232323232,
      "loss": 1.4539,
      "theoretical_loss": 3.379459126286,
      "tokens_seen": 2400845824
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002772121212121212,
      "loss": 1.3352,
      "theoretical_loss": 3.379451708252164,
      "tokens_seen": 2400911360
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027719191919191923,
      "loss": 1.4395,
      "theoretical_loss": 3.3794442904775037,
      "tokens_seen": 2400976896
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771717171717172,
      "loss": 1.4772,
      "theoretical_loss": 3.3794368729620037,
      "tokens_seen": 2401042432
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027715151515151514,
      "loss": 1.5502,
      "theoretical_loss": 3.3794294557056475,
      "tokens_seen": 2401107968
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771313131313131,
      "loss": 1.2814,
      "theoretical_loss": 3.3794220387084195,
      "tokens_seen": 2401173504
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002771111111111111,
      "loss": 1.3659,
      "theoretical_loss": 3.3794146219703025,
      "tokens_seen": 2401239040
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770909090909091,
      "loss": 1.3856,
      "theoretical_loss": 3.3794072054912814,
      "tokens_seen": 2401304576
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002770707070707071,
      "loss": 1.2235,
      "theoretical_loss": 3.379399789271339,
      "tokens_seen": 2401370112
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027705050505050504,
      "loss": 1.2366,
      "theoretical_loss": 3.379392373310461,
      "tokens_seen": 2401435648
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277030303030303,
      "loss": 1.4892,
      "theoretical_loss": 3.3793849576086297,
      "tokens_seen": 2401501184
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000277010101010101,
      "loss": 1.1678,
      "theoretical_loss": 3.379377542165829,
      "tokens_seen": 2401566720
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000276989898989899,
      "loss": 1.1829,
      "theoretical_loss": 3.3793701269820433,
      "tokens_seen": 2401632256
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000276969696969697,
      "loss": 1.2928,
      "theoretical_loss": 3.3793627120572567,
      "tokens_seen": 2401697792
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027694949494949493,
      "loss": 1.2661,
      "theoretical_loss": 3.3793552973914522,
      "tokens_seen": 2401763328
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027692929292929294,
      "loss": 1.3704,
      "theoretical_loss": 3.379347882984615,
      "tokens_seen": 2401828864
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002769090909090909,
      "loss": 1.3863,
      "theoretical_loss": 3.3793404688367277,
      "tokens_seen": 2401894400
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768888888888889,
      "loss": 1.4656,
      "theoretical_loss": 3.379333054947775,
      "tokens_seen": 2401959936
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027686868686868687,
      "loss": 1.3682,
      "theoretical_loss": 3.3793256413177404,
      "tokens_seen": 2402025472
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768484848484848,
      "loss": 1.4678,
      "theoretical_loss": 3.3793182279466083,
      "tokens_seen": 2402091008
    },
    {
      "epoch": 0.45,
      "objective/train/aligned_token_count": 0.48356547951698303,
      "objective/train/avg_token_score": 0.007656682748347521,
      "objective/train/docs_used": 1352871,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.642473578453064,
      "objective/train/original_loss": 4.2449469566345215,
      "objective/train/theoretical_loss": 3.379310814834362,
      "objective/train/tokens_used": 761015776,
      "objective/train/unlikelihood_loss": 0.06536135822534561,
      "theoretical_loss": 3.379310814834362,
      "tokens_seen": 2402156544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027682828282828283,
      "loss": 1.3587,
      "theoretical_loss": 3.379310814834362,
      "tokens_seen": 2402156544
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002768080808080808,
      "loss": 1.298,
      "theoretical_loss": 3.3793034019809856,
      "tokens_seen": 2402222080
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002767878787878788,
      "loss": 1.493,
      "theoretical_loss": 3.3792959893864625,
      "tokens_seen": 2402287616
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027676767676767676,
      "loss": 1.3794,
      "theoretical_loss": 3.3792885770507777,
      "tokens_seen": 2402353152
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027674747474747477,
      "loss": 1.5329,
      "theoretical_loss": 3.3792811649739143,
      "tokens_seen": 2402418688
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002767272727272727,
      "loss": 1.4254,
      "theoretical_loss": 3.3792737531558568,
      "tokens_seen": 2402484224
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027670707070707074,
      "loss": 1.4161,
      "theoretical_loss": 3.3792663415965882,
      "tokens_seen": 2402549760
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766868686868687,
      "loss": 1.3344,
      "theoretical_loss": 3.379258930296093,
      "tokens_seen": 2402615296
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027666666666666665,
      "loss": 1.2244,
      "theoretical_loss": 3.3792515192543555,
      "tokens_seen": 2402680832
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027664646464646466,
      "loss": 1.4277,
      "theoretical_loss": 3.379244108471359,
      "tokens_seen": 2402746368
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002766262626262626,
      "loss": 1.4814,
      "theoretical_loss": 3.3792366979470874,
      "tokens_seen": 2402811904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027660606060606063,
      "loss": 1.2447,
      "theoretical_loss": 3.3792292876815244,
      "tokens_seen": 2402877440
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765858585858586,
      "loss": 1.3277,
      "theoretical_loss": 3.379221877674655,
      "tokens_seen": 2402942976
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765656565656566,
      "loss": 1.1567,
      "theoretical_loss": 3.3792144679264617,
      "tokens_seen": 2403008512
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027654545454545456,
      "loss": 1.4681,
      "theoretical_loss": 3.3792070584369296,
      "tokens_seen": 2403074048
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765252525252525,
      "loss": 1.3071,
      "theoretical_loss": 3.3791996492060417,
      "tokens_seen": 2403139584
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002765050505050505,
      "loss": 1.5039,
      "theoretical_loss": 3.379192240233783,
      "tokens_seen": 2403205120
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002764848484848485,
      "loss": 1.3967,
      "theoretical_loss": 3.3791848315201363,
      "tokens_seen": 2403270656
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002764646464646465,
      "loss": 1.5303,
      "theoretical_loss": 3.379177423065086,
      "tokens_seen": 2403336192
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027644444444444445,
      "loss": 1.3719,
      "theoretical_loss": 3.379170014868616,
      "tokens_seen": 2403401728
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002764242424242424,
      "loss": 1.4207,
      "theoretical_loss": 3.37916260693071,
      "tokens_seen": 2403467264
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002764040404040404,
      "loss": 1.4786,
      "theoretical_loss": 3.3791551992513527,
      "tokens_seen": 2403532800
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763838383838384,
      "loss": 1.4421,
      "theoretical_loss": 3.3791477918305275,
      "tokens_seen": 2403598336
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763636363636364,
      "loss": 1.4136,
      "theoretical_loss": 3.3791403846682178,
      "tokens_seen": 2403663872
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027634343434343434,
      "loss": 1.5299,
      "theoretical_loss": 3.3791329777644084,
      "tokens_seen": 2403729408
    },
    {
      "epoch": 0.45,
      "objective/train/aligned_token_count": 0.3915567100048065,
      "objective/train/avg_token_score": 0.0073168897069990635,
      "objective/train/docs_used": 1354213,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4102452993392944,
      "objective/train/original_loss": 4.5515875816345215,
      "objective/train/theoretical_loss": 3.379125571119083,
      "objective/train/tokens_used": 762654176,
      "objective/train/unlikelihood_loss": 0.056296207010746,
      "theoretical_loss": 3.379125571119083,
      "tokens_seen": 2403794944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763232323232323,
      "loss": 1.288,
      "theoretical_loss": 3.379125571119083,
      "tokens_seen": 2403794944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002763030303030303,
      "loss": 1.3891,
      "theoretical_loss": 3.379118164732225,
      "tokens_seen": 2403860480
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002762828282828283,
      "loss": 1.1599,
      "theoretical_loss": 3.3791107586038187,
      "tokens_seen": 2403926016
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002762626262626263,
      "loss": 1.3495,
      "theoretical_loss": 3.379103352733848,
      "tokens_seen": 2403991552
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027624242424242423,
      "loss": 1.2897,
      "theoretical_loss": 3.3790959471222974,
      "tokens_seen": 2404057088
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002762222222222222,
      "loss": 1.4134,
      "theoretical_loss": 3.3790885417691503,
      "tokens_seen": 2404122624
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027620202020202026,
      "loss": 1.7981,
      "theoretical_loss": 3.3790811366743903,
      "tokens_seen": 2404188160
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761818181818182,
      "loss": 1.2719,
      "theoretical_loss": 3.379073731838002,
      "tokens_seen": 2404253696
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027616161616161617,
      "loss": 1.3692,
      "theoretical_loss": 3.3790663272599692,
      "tokens_seen": 2404319232
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761414141414141,
      "loss": 1.5045,
      "theoretical_loss": 3.3790589229402754,
      "tokens_seen": 2404384768
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002761212121212121,
      "loss": 1.2248,
      "theoretical_loss": 3.379051518878905,
      "tokens_seen": 2404450304
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027610101010101015,
      "loss": 1.3799,
      "theoretical_loss": 3.3790441150758417,
      "tokens_seen": 2404515840
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002760808080808081,
      "loss": 1.6212,
      "theoretical_loss": 3.3790367115310698,
      "tokens_seen": 2404581376
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027606060606060606,
      "loss": 1.2316,
      "theoretical_loss": 3.379029308244573,
      "tokens_seen": 2404646912
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000276040404040404,
      "loss": 1.5246,
      "theoretical_loss": 3.379021905216335,
      "tokens_seen": 2404712448
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027602020202020203,
      "loss": 1.3812,
      "theoretical_loss": 3.3790145024463403,
      "tokens_seen": 2404777984
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027600000000000004,
      "loss": 1.5349,
      "theoretical_loss": 3.3790070999345723,
      "tokens_seen": 2404843520
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000275979797979798,
      "loss": 1.2826,
      "theoretical_loss": 3.3789996976810155,
      "tokens_seen": 2404909056
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027595959595959595,
      "loss": 1.6458,
      "theoretical_loss": 3.378992295685653,
      "tokens_seen": 2404974592
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002759393939393939,
      "loss": 1.5943,
      "theoretical_loss": 3.37898489394847,
      "tokens_seen": 2405040128
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002759191919191919,
      "loss": 1.2717,
      "theoretical_loss": 3.3789774924694496,
      "tokens_seen": 2405105664
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027589898989898993,
      "loss": 1.447,
      "theoretical_loss": 3.378970091248576,
      "tokens_seen": 2405171200
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002758787878787879,
      "loss": 1.5976,
      "theoretical_loss": 3.378962690285833,
      "tokens_seen": 2405236736
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027585858585858585,
      "loss": 1.3752,
      "theoretical_loss": 3.3789552895812047,
      "tokens_seen": 2405302272
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027583838383838386,
      "loss": 1.5788,
      "theoretical_loss": 3.378947889134675,
      "tokens_seen": 2405367808
    },
    {
      "epoch": 0.45,
      "objective/train/aligned_token_count": 0.47928884625434875,
      "objective/train/avg_token_score": 0.014510654844343662,
      "objective/train/docs_used": 1354954,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5177143812179565,
      "objective/train/original_loss": 4.093996047973633,
      "objective/train/theoretical_loss": 3.378940488946228,
      "objective/train/tokens_used": 764292576,
      "objective/train/unlikelihood_loss": 0.049194738268852234,
      "theoretical_loss": 3.378940488946228,
      "tokens_seen": 2405433344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002758181818181818,
      "loss": 1.2485,
      "theoretical_loss": 3.378940488946228,
      "tokens_seen": 2405433344
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002757979797979798,
      "loss": 1.4525,
      "theoretical_loss": 3.3789330890158475,
      "tokens_seen": 2405498880
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002757777777777778,
      "loss": 1.5072,
      "theoretical_loss": 3.378925689343518,
      "tokens_seen": 2405564416
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027575757575757574,
      "loss": 1.4673,
      "theoretical_loss": 3.3789182899292225,
      "tokens_seen": 2405629952
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027573737373737375,
      "loss": 1.6121,
      "theoretical_loss": 3.3789108907729455,
      "tokens_seen": 2405695488
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002757171717171717,
      "loss": 1.4865,
      "theoretical_loss": 3.378903491874671,
      "tokens_seen": 2405761024
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756969696969697,
      "loss": 1.4452,
      "theoretical_loss": 3.3788960932343834,
      "tokens_seen": 2405826560
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756767676767677,
      "loss": 1.6236,
      "theoretical_loss": 3.378888694852066,
      "tokens_seen": 2405892096
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756565656565657,
      "loss": 1.3206,
      "theoretical_loss": 3.3788812967277027,
      "tokens_seen": 2405957632
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027563636363636364,
      "loss": 1.3018,
      "theoretical_loss": 3.3788738988612783,
      "tokens_seen": 2406023168
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002756161616161616,
      "loss": 1.3123,
      "theoretical_loss": 3.3788665012527757,
      "tokens_seen": 2406088704
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002755959595959596,
      "loss": 1.314,
      "theoretical_loss": 3.37885910390218,
      "tokens_seen": 2406154240
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027557575757575757,
      "loss": 1.374,
      "theoretical_loss": 3.378851706809474,
      "tokens_seen": 2406219776
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002755555555555556,
      "loss": 1.4314,
      "theoretical_loss": 3.3788443099746432,
      "tokens_seen": 2406285312
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027553535353535354,
      "loss": 1.3177,
      "theoretical_loss": 3.37883691339767,
      "tokens_seen": 2406350848
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002755151515151515,
      "loss": 1.2732,
      "theoretical_loss": 3.3788295170785396,
      "tokens_seen": 2406416384
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002754949494949495,
      "loss": 1.5919,
      "theoretical_loss": 3.378822121017235,
      "tokens_seen": 2406481920
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002754747474747475,
      "loss": 1.2327,
      "theoretical_loss": 3.378814725213741,
      "tokens_seen": 2406547456
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027545454545454547,
      "loss": 1.3749,
      "theoretical_loss": 3.378807329668041,
      "tokens_seen": 2406612992
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027543434343434343,
      "loss": 1.2338,
      "theoretical_loss": 3.3787999343801194,
      "tokens_seen": 2406678528
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002754141414141414,
      "loss": 1.6696,
      "theoretical_loss": 3.37879253934996,
      "tokens_seen": 2406744064
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753939393939394,
      "loss": 1.5424,
      "theoretical_loss": 3.378785144577547,
      "tokens_seen": 2406809600
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753737373737374,
      "loss": 1.4042,
      "theoretical_loss": 3.3787777500628646,
      "tokens_seen": 2406875136
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00027535353535353537,
      "loss": 1.4527,
      "theoretical_loss": 3.378770355805896,
      "tokens_seen": 2406940672
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002753333333333333,
      "loss": 1.72,
      "theoretical_loss": 3.3787629618066255,
      "tokens_seen": 2407006208
    },
    {
      "epoch": 0.45,
      "objective/train/aligned_token_count": 0.4144672453403473,
      "objective/train/avg_token_score": 0.003465415211394429,
      "objective/train/docs_used": 1356171,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4051523208618164,
      "objective/train/original_loss": 4.213868618011475,
      "objective/train/theoretical_loss": 3.3787555680650376,
      "objective/train/tokens_used": 765930976,
      "objective/train/unlikelihood_loss": 0.060969460755586624,
      "theoretical_loss": 3.3787555680650376,
      "tokens_seen": 2407071744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002753131313131313,
      "loss": 1.3332,
      "theoretical_loss": 3.3787555680650376,
      "tokens_seen": 2407071744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027529292929292934,
      "loss": 1.1632,
      "theoretical_loss": 3.3787481745811156,
      "tokens_seen": 2407137280
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002752727272727273,
      "loss": 1.1198,
      "theoretical_loss": 3.3787407813548445,
      "tokens_seen": 2407202816
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027525252525252526,
      "loss": 1.2137,
      "theoretical_loss": 3.378733388386207,
      "tokens_seen": 2407268352
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002752323232323232,
      "loss": 1.098,
      "theoretical_loss": 3.3787259956751883,
      "tokens_seen": 2407333888
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027521212121212117,
      "loss": 1.2879,
      "theoretical_loss": 3.3787186032217718,
      "tokens_seen": 2407399424
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027519191919191924,
      "loss": 1.2822,
      "theoretical_loss": 3.378711211025941,
      "tokens_seen": 2407464960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002751717171717172,
      "loss": 1.2128,
      "theoretical_loss": 3.3787038190876815,
      "tokens_seen": 2407530496
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027515151515151515,
      "loss": 1.4194,
      "theoretical_loss": 3.3786964274069757,
      "tokens_seen": 2407596032
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002751313131313131,
      "loss": 1.5592,
      "theoretical_loss": 3.3786890359838084,
      "tokens_seen": 2407661568
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027511111111111106,
      "loss": 1.2421,
      "theoretical_loss": 3.3786816448181636,
      "tokens_seen": 2407727104
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027509090909090913,
      "loss": 1.3789,
      "theoretical_loss": 3.378674253910025,
      "tokens_seen": 2407792640
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002750707070707071,
      "loss": 1.349,
      "theoretical_loss": 3.378666863259377,
      "tokens_seen": 2407858176
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027505050505050504,
      "loss": 1.3867,
      "theoretical_loss": 3.3786594728662034,
      "tokens_seen": 2407923712
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000275030303030303,
      "loss": 1.648,
      "theoretical_loss": 3.378652082730488,
      "tokens_seen": 2407989248
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000275010101010101,
      "loss": 1.3117,
      "theoretical_loss": 3.378644692852215,
      "tokens_seen": 2408054784
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274989898989899,
      "loss": 1.352,
      "theoretical_loss": 3.3786373032313692,
      "tokens_seen": 2408120320
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274969696969697,
      "loss": 1.5158,
      "theoretical_loss": 3.3786299138679334,
      "tokens_seen": 2408185856
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027494949494949494,
      "loss": 1.18,
      "theoretical_loss": 3.378622524761892,
      "tokens_seen": 2408251392
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002749292929292929,
      "loss": 1.5806,
      "theoretical_loss": 3.3786151359132295,
      "tokens_seen": 2408316928
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027490909090909096,
      "loss": 1.4872,
      "theoretical_loss": 3.37860774732193,
      "tokens_seen": 2408382464
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002748888888888889,
      "loss": 1.5229,
      "theoretical_loss": 3.3786003589879763,
      "tokens_seen": 2408448000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027486868686868687,
      "loss": 1.5913,
      "theoretical_loss": 3.378592970911354,
      "tokens_seen": 2408513536
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027484848484848483,
      "loss": 1.5848,
      "theoretical_loss": 3.378585583092046,
      "tokens_seen": 2408579072
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027482828282828284,
      "loss": 1.3659,
      "theoretical_loss": 3.3785781955300367,
      "tokens_seen": 2408644608
    },
    {
      "epoch": 0.46,
      "objective/train/aligned_token_count": 0.43866080045700073,
      "objective/train/avg_token_score": 0.021133270114660263,
      "objective/train/docs_used": 1356770,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4528706073760986,
      "objective/train/original_loss": 4.129415512084961,
      "objective/train/theoretical_loss": 3.3785708082253105,
      "objective/train/tokens_used": 767569376,
      "objective/train/unlikelihood_loss": 0.07518959790468216,
      "theoretical_loss": 3.3785708082253105,
      "tokens_seen": 2408710144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027480808080808085,
      "loss": 1.267,
      "theoretical_loss": 3.3785708082253105,
      "tokens_seen": 2408710144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002747878787878788,
      "loss": 1.4789,
      "theoretical_loss": 3.378563421177851,
      "tokens_seen": 2408775680
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027476767676767676,
      "loss": 1.3845,
      "theoretical_loss": 3.378556034387642,
      "tokens_seen": 2408841216
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002747474747474747,
      "loss": 1.3282,
      "theoretical_loss": 3.378548647854669,
      "tokens_seen": 2408906752
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027472727272727273,
      "loss": 1.2873,
      "theoretical_loss": 3.378541261578914,
      "tokens_seen": 2408972288
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027470707070707074,
      "loss": 1.3209,
      "theoretical_loss": 3.378533875560362,
      "tokens_seen": 2409037824
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002746868686868687,
      "loss": 1.3971,
      "theoretical_loss": 3.3785264897989977,
      "tokens_seen": 2409103360
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027466666666666666,
      "loss": 1.229,
      "theoretical_loss": 3.378519104294804,
      "tokens_seen": 2409168896
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027464646464646467,
      "loss": 1.3819,
      "theoretical_loss": 3.378511719047766,
      "tokens_seen": 2409234432
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002746262626262626,
      "loss": 1.3284,
      "theoretical_loss": 3.3785043340578667,
      "tokens_seen": 2409299968
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027460606060606064,
      "loss": 1.2549,
      "theoretical_loss": 3.3784969493250907,
      "tokens_seen": 2409365504
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002745858585858586,
      "loss": 1.3806,
      "theoretical_loss": 3.378489564849422,
      "tokens_seen": 2409431040
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027456565656565655,
      "loss": 1.3396,
      "theoretical_loss": 3.378482180630845,
      "tokens_seen": 2409496576
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027454545454545456,
      "loss": 1.3416,
      "theoretical_loss": 3.378474796669343,
      "tokens_seen": 2409562112
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002745252525252525,
      "loss": 1.5553,
      "theoretical_loss": 3.378467412964901,
      "tokens_seen": 2409627648
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027450505050505053,
      "loss": 1.3319,
      "theoretical_loss": 3.3784600295175022,
      "tokens_seen": 2409693184
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002744848484848485,
      "loss": 1.5643,
      "theoretical_loss": 3.3784526463271307,
      "tokens_seen": 2409758720
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002744646464646465,
      "loss": 1.2391,
      "theoretical_loss": 3.3784452633937714,
      "tokens_seen": 2409824256
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027444444444444445,
      "loss": 1.5292,
      "theoretical_loss": 3.378437880717408,
      "tokens_seen": 2409889792
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002744242424242424,
      "loss": 1.5734,
      "theoretical_loss": 3.378430498298024,
      "tokens_seen": 2409955328
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002744040404040404,
      "loss": 1.1948,
      "theoretical_loss": 3.378423116135604,
      "tokens_seen": 2410020864
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002743838383838384,
      "loss": 1.393,
      "theoretical_loss": 3.3784157342301318,
      "tokens_seen": 2410086400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002743636363636364,
      "loss": 1.2955,
      "theoretical_loss": 3.378408352581592,
      "tokens_seen": 2410151936
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027434343434343435,
      "loss": 1.4053,
      "theoretical_loss": 3.3784009711899676,
      "tokens_seen": 2410217472
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002743232323232323,
      "loss": 1.5155,
      "theoretical_loss": 3.3783935900552438,
      "tokens_seen": 2410283008
    },
    {
      "epoch": 0.46,
      "objective/train/aligned_token_count": 0.43169599771499634,
      "objective/train/avg_token_score": 0.011193030513823032,
      "objective/train/docs_used": 1358132,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4506155252456665,
      "objective/train/original_loss": 4.276076793670654,
      "objective/train/theoretical_loss": 3.378386209177404,
      "objective/train/tokens_used": 769207776,
      "objective/train/unlikelihood_loss": 0.0632748156785965,
      "theoretical_loss": 3.378386209177404,
      "tokens_seen": 2410348544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002743030303030303,
      "loss": 1.3454,
      "theoretical_loss": 3.378386209177404,
      "tokens_seen": 2410348544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742828282828283,
      "loss": 1.3211,
      "theoretical_loss": 3.378378828556433,
      "tokens_seen": 2410414080
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742626262626263,
      "loss": 1.2745,
      "theoretical_loss": 3.378371448192314,
      "tokens_seen": 2410479616
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027424242424242424,
      "loss": 1.5387,
      "theoretical_loss": 3.3783640680850313,
      "tokens_seen": 2410545152
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742222222222222,
      "loss": 1.4602,
      "theoretical_loss": 3.378356688234569,
      "tokens_seen": 2410610688
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002742020202020202,
      "loss": 1.3643,
      "theoretical_loss": 3.378349308640912,
      "tokens_seen": 2410676224
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002741818181818182,
      "loss": 1.5405,
      "theoretical_loss": 3.3783419293040433,
      "tokens_seen": 2410741760
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002741616161616162,
      "loss": 1.4663,
      "theoretical_loss": 3.378334550223947,
      "tokens_seen": 2410807296
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027414141414141413,
      "loss": 1.6071,
      "theoretical_loss": 3.378327171400608,
      "tokens_seen": 2410872832
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002741212121212121,
      "loss": 1.4572,
      "theoretical_loss": 3.3783197928340103,
      "tokens_seen": 2410938368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027410101010101015,
      "loss": 1.2605,
      "theoretical_loss": 3.378312414524137,
      "tokens_seen": 2411003904
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002740808080808081,
      "loss": 1.3273,
      "theoretical_loss": 3.378305036470973,
      "tokens_seen": 2411069440
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027406060606060607,
      "loss": 1.3542,
      "theoretical_loss": 3.378297658674503,
      "tokens_seen": 2411134976
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274040404040404,
      "loss": 1.2984,
      "theoretical_loss": 3.3782902811347095,
      "tokens_seen": 2411200512
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000274020202020202,
      "loss": 1.3833,
      "theoretical_loss": 3.3782829038515776,
      "tokens_seen": 2411266048
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027400000000000005,
      "loss": 1.2938,
      "theoretical_loss": 3.378275526825091,
      "tokens_seen": 2411331584
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000273979797979798,
      "loss": 1.2121,
      "theoretical_loss": 3.3782681500552343,
      "tokens_seen": 2411397120
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027395959595959596,
      "loss": 1.4176,
      "theoretical_loss": 3.378260773541991,
      "tokens_seen": 2411462656
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002739393939393939,
      "loss": 1.5123,
      "theoretical_loss": 3.378253397285346,
      "tokens_seen": 2411528192
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027391919191919193,
      "loss": 1.1287,
      "theoretical_loss": 3.3782460212852827,
      "tokens_seen": 2411593728
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027389898989898994,
      "loss": 1.4273,
      "theoretical_loss": 3.378238645541785,
      "tokens_seen": 2411659264
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002738787878787879,
      "loss": 1.1709,
      "theoretical_loss": 3.3782312700548376,
      "tokens_seen": 2411724800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027385858585858585,
      "loss": 1.399,
      "theoretical_loss": 3.3782238948244245,
      "tokens_seen": 2411790336
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002738383838383838,
      "loss": 1.5222,
      "theoretical_loss": 3.37821651985053,
      "tokens_seen": 2411855872
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002738181818181818,
      "loss": 1.4341,
      "theoretical_loss": 3.3782091451331375,
      "tokens_seen": 2411921408
    },
    {
      "epoch": 0.46,
      "objective/train/aligned_token_count": 0.39992669224739075,
      "objective/train/avg_token_score": 0.10333684831857681,
      "objective/train/docs_used": 1358876,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2629338502883911,
      "objective/train/original_loss": 4.434542655944824,
      "objective/train/theoretical_loss": 3.3782017706722316,
      "objective/train/tokens_used": 770846176,
      "objective/train/unlikelihood_loss": 0.06632868945598602,
      "theoretical_loss": 3.3782017706722316,
      "tokens_seen": 2411986944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027379797979797983,
      "loss": 1.2594,
      "theoretical_loss": 3.3782017706722316,
      "tokens_seen": 2411986944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737777777777778,
      "loss": 1.4312,
      "theoretical_loss": 3.3781943964677965,
      "tokens_seen": 2412052480
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027375757575757575,
      "loss": 1.5969,
      "theoretical_loss": 3.3781870225198163,
      "tokens_seen": 2412118016
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737373737373737,
      "loss": 1.4373,
      "theoretical_loss": 3.378179648828275,
      "tokens_seen": 2412183552
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002737171717171717,
      "loss": 1.3344,
      "theoretical_loss": 3.3781722753931565,
      "tokens_seen": 2412249088
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736969696969697,
      "loss": 1.5046,
      "theoretical_loss": 3.378164902214445,
      "tokens_seen": 2412314624
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736767676767677,
      "loss": 1.5869,
      "theoretical_loss": 3.3781575292921247,
      "tokens_seen": 2412380160
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027365656565656564,
      "loss": 1.2195,
      "theoretical_loss": 3.37815015662618,
      "tokens_seen": 2412445696
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027363636363636365,
      "loss": 1.4255,
      "theoretical_loss": 3.3781427842165948,
      "tokens_seen": 2412511232
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002736161616161616,
      "loss": 1.5395,
      "theoretical_loss": 3.3781354120633527,
      "tokens_seen": 2412576768
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002735959595959596,
      "loss": 1.3958,
      "theoretical_loss": 3.378128040166439,
      "tokens_seen": 2412642304
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002735757575757576,
      "loss": 1.412,
      "theoretical_loss": 3.3781206685258365,
      "tokens_seen": 2412707840
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027355555555555553,
      "loss": 1.5371,
      "theoretical_loss": 3.3781132971415304,
      "tokens_seen": 2412773376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027353535353535354,
      "loss": 1.4512,
      "theoretical_loss": 3.3781059260135042,
      "tokens_seen": 2412838912
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002735151515151515,
      "loss": 1.47,
      "theoretical_loss": 3.3780985551417424,
      "tokens_seen": 2412904448
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002734949494949495,
      "loss": 1.3362,
      "theoretical_loss": 3.3780911845262285,
      "tokens_seen": 2412969984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027347474747474747,
      "loss": 1.3795,
      "theoretical_loss": 3.378083814166948,
      "tokens_seen": 2413035520
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002734545454545455,
      "loss": 1.4535,
      "theoretical_loss": 3.378076444063883,
      "tokens_seen": 2413101056
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027343434343434343,
      "loss": 1.4555,
      "theoretical_loss": 3.378069074217019,
      "tokens_seen": 2413166592
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002734141414141414,
      "loss": 1.2891,
      "theoretical_loss": 3.3780617046263406,
      "tokens_seen": 2413232128
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002733939393939394,
      "loss": 1.6228,
      "theoretical_loss": 3.378054335291831,
      "tokens_seen": 2413297664
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027337373737373736,
      "loss": 1.2361,
      "theoretical_loss": 3.378046966213474,
      "tokens_seen": 2413363200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027335353535353537,
      "loss": 1.4203,
      "theoretical_loss": 3.378039597391255,
      "tokens_seen": 2413428736
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027333333333333333,
      "loss": 1.3833,
      "theoretical_loss": 3.378032228825157,
      "tokens_seen": 2413494272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002733131313131313,
      "loss": 1.3513,
      "theoretical_loss": 3.378024860515165,
      "tokens_seen": 2413559808
    },
    {
      "epoch": 0.46,
      "objective/train/aligned_token_count": 0.4924853444099426,
      "objective/train/avg_token_score": 0.010232564061880112,
      "objective/train/docs_used": 1360381,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4357194900512695,
      "objective/train/original_loss": 4.017053604125977,
      "objective/train/theoretical_loss": 3.3780174924612623,
      "objective/train/tokens_used": 772484576,
      "objective/train/unlikelihood_loss": 0.06597133725881577,
      "theoretical_loss": 3.3780174924612623,
      "tokens_seen": 2413625344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002732929292929293,
      "loss": 1.4329,
      "theoretical_loss": 3.3780174924612623,
      "tokens_seen": 2413625344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002732727272727273,
      "loss": 1.5522,
      "theoretical_loss": 3.3780101246634335,
      "tokens_seen": 2413690880
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027325252525252526,
      "loss": 1.3567,
      "theoretical_loss": 3.3780027571216626,
      "tokens_seen": 2413756416
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002732323232323232,
      "loss": 1.4132,
      "theoretical_loss": 3.3779953898359345,
      "tokens_seen": 2413821952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027321212121212123,
      "loss": 1.2629,
      "theoretical_loss": 3.3779880228062322,
      "tokens_seen": 2413887488
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002731919191919192,
      "loss": 1.3098,
      "theoretical_loss": 3.3779806560325407,
      "tokens_seen": 2413953024
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002731717171717172,
      "loss": 1.1355,
      "theoretical_loss": 3.3779732895148435,
      "tokens_seen": 2414018560
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027315151515151516,
      "loss": 1.5461,
      "theoretical_loss": 3.3779659232531256,
      "tokens_seen": 2414084096
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002731313131313131,
      "loss": 1.5903,
      "theoretical_loss": 3.3779585572473705,
      "tokens_seen": 2414149632
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002731111111111111,
      "loss": 1.4386,
      "theoretical_loss": 3.377951191497562,
      "tokens_seen": 2414215168
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027309090909090913,
      "loss": 1.4882,
      "theoretical_loss": 3.377943826003685,
      "tokens_seen": 2414280704
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002730707070707071,
      "loss": 1.3833,
      "theoretical_loss": 3.377936460765724,
      "tokens_seen": 2414346240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027305050505050505,
      "loss": 1.4857,
      "theoretical_loss": 3.377929095783662,
      "tokens_seen": 2414411776
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000273030303030303,
      "loss": 1.4072,
      "theoretical_loss": 3.3779217310574836,
      "tokens_seen": 2414477312
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000273010101010101,
      "loss": 1.4291,
      "theoretical_loss": 3.3779143665871736,
      "tokens_seen": 2414542848
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027298989898989903,
      "loss": 1.5379,
      "theoretical_loss": 3.3779070023727154,
      "tokens_seen": 2414608384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000272969696969697,
      "loss": 1.3389,
      "theoretical_loss": 3.377899638414094,
      "tokens_seen": 2414673920
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027294949494949494,
      "loss": 1.5942,
      "theoretical_loss": 3.377892274711292,
      "tokens_seen": 2414739456
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002729292929292929,
      "loss": 1.3806,
      "theoretical_loss": 3.3778849112642955,
      "tokens_seen": 2414804992
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027290909090909096,
      "loss": 1.3815,
      "theoretical_loss": 3.377877548073087,
      "tokens_seen": 2414870528
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002728888888888889,
      "loss": 1.6485,
      "theoretical_loss": 3.377870185137652,
      "tokens_seen": 2414936064
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002728686868686869,
      "loss": 1.4284,
      "theoretical_loss": 3.377862822457974,
      "tokens_seen": 2415001600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027284848484848483,
      "loss": 1.2452,
      "theoretical_loss": 3.3778554600340374,
      "tokens_seen": 2415067136
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002728282828282828,
      "loss": 1.3643,
      "theoretical_loss": 3.377848097865826,
      "tokens_seen": 2415132672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027280808080808086,
      "loss": 1.4281,
      "theoretical_loss": 3.3778407359533245,
      "tokens_seen": 2415198208
    },
    {
      "epoch": 0.46,
      "objective/train/aligned_token_count": 0.34909579157829285,
      "objective/train/avg_token_score": 0.018620479851961136,
      "objective/train/docs_used": 1361019,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9541930556297302,
      "objective/train/original_loss": 4.412327766418457,
      "objective/train/theoretical_loss": 3.3778333742965163,
      "objective/train/tokens_used": 774122976,
      "objective/train/unlikelihood_loss": 0.0790051519870758,
      "theoretical_loss": 3.3778333742965163,
      "tokens_seen": 2415263744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002727878787878788,
      "loss": 1.2981,
      "theoretical_loss": 3.3778333742965163,
      "tokens_seen": 2415263744
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027276767676767677,
      "loss": 1.6096,
      "theoretical_loss": 3.3778260128953868,
      "tokens_seen": 2415329280
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002727474747474747,
      "loss": 1.5292,
      "theoretical_loss": 3.377818651749919,
      "tokens_seen": 2415394816
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027272727272727274,
      "loss": 1.286,
      "theoretical_loss": 3.3778112908600977,
      "tokens_seen": 2415460352
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027270707070707075,
      "loss": 1.4034,
      "theoretical_loss": 3.377803930225907,
      "tokens_seen": 2415525888
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002726868686868687,
      "loss": 1.2416,
      "theoretical_loss": 3.377796569847331,
      "tokens_seen": 2415591424
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027266666666666666,
      "loss": 1.6903,
      "theoretical_loss": 3.377789209724354,
      "tokens_seen": 2415656960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002726464646464646,
      "loss": 1.2986,
      "theoretical_loss": 3.37778184985696,
      "tokens_seen": 2415722496
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027262626262626263,
      "loss": 1.4783,
      "theoretical_loss": 3.3777744902451334,
      "tokens_seen": 2415788032
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027260606060606064,
      "loss": 1.5071,
      "theoretical_loss": 3.3777671308888584,
      "tokens_seen": 2415853568
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002725858585858586,
      "loss": 1.2024,
      "theoretical_loss": 3.377759771788119,
      "tokens_seen": 2415919104
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027256565656565656,
      "loss": 1.6053,
      "theoretical_loss": 3.3777524129428995,
      "tokens_seen": 2415984640
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027254545454545457,
      "loss": 1.295,
      "theoretical_loss": 3.3777450543531837,
      "tokens_seen": 2416050176
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002725252525252525,
      "loss": 1.4427,
      "theoretical_loss": 3.377737696018957,
      "tokens_seen": 2416115712
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027250505050505053,
      "loss": 1.5971,
      "theoretical_loss": 3.377730337940202,
      "tokens_seen": 2416181248
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002724848484848485,
      "loss": 1.2938,
      "theoretical_loss": 3.377722980116904,
      "tokens_seen": 2416246784
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027246464646464645,
      "loss": 1.5034,
      "theoretical_loss": 3.3777156225490472,
      "tokens_seen": 2416312320
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027244444444444446,
      "loss": 1.5855,
      "theoretical_loss": 3.377708265236615,
      "tokens_seen": 2416377856
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002724242424242424,
      "loss": 1.4024,
      "theoretical_loss": 3.3777009081795923,
      "tokens_seen": 2416443392
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002724040404040404,
      "loss": 1.3726,
      "theoretical_loss": 3.377693551377963,
      "tokens_seen": 2416508928
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723838383838384,
      "loss": 1.5171,
      "theoretical_loss": 3.377686194831712,
      "tokens_seen": 2416574464
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723636363636364,
      "loss": 1.2153,
      "theoretical_loss": 3.377678838540822,
      "tokens_seen": 2416640000
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027234343434343435,
      "loss": 1.2217,
      "theoretical_loss": 3.3776714825052787,
      "tokens_seen": 2416705536
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723232323232323,
      "loss": 1.6014,
      "theoretical_loss": 3.3776641267250658,
      "tokens_seen": 2416771072
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002723030303030303,
      "loss": 1.2608,
      "theoretical_loss": 3.3776567712001673,
      "tokens_seen": 2416836608
    },
    {
      "epoch": 0.46,
      "objective/train/aligned_token_count": 0.40218719840049744,
      "objective/train/avg_token_score": 0.01878960430622101,
      "objective/train/docs_used": 1362090,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.262105107307434,
      "objective/train/original_loss": 4.663071632385254,
      "objective/train/theoretical_loss": 3.3776494159305677,
      "objective/train/tokens_used": 775761376,
      "objective/train/unlikelihood_loss": 0.04290124773979187,
      "theoretical_loss": 3.3776494159305677,
      "tokens_seen": 2416902144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002722828282828283,
      "loss": 1.3293,
      "theoretical_loss": 3.3776494159305677,
      "tokens_seen": 2416902144
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002722626262626263,
      "loss": 1.4147,
      "theoretical_loss": 3.377642060916251,
      "tokens_seen": 2416967680
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027224242424242424,
      "loss": 1.4418,
      "theoretical_loss": 3.3776347061572016,
      "tokens_seen": 2417033216
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002722222222222222,
      "loss": 1.4364,
      "theoretical_loss": 3.3776273516534037,
      "tokens_seen": 2417098752
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002722020202020202,
      "loss": 1.4055,
      "theoretical_loss": 3.377619997404841,
      "tokens_seen": 2417164288
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002721818181818182,
      "loss": 1.4055,
      "theoretical_loss": 3.377612643411499,
      "tokens_seen": 2417229824
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002721616161616162,
      "loss": 1.6457,
      "theoretical_loss": 3.3776052896733604,
      "tokens_seen": 2417295360
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027214141414141414,
      "loss": 1.4817,
      "theoretical_loss": 3.3775979361904103,
      "tokens_seen": 2417360896
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002721212121212121,
      "loss": 1.2333,
      "theoretical_loss": 3.377590582962633,
      "tokens_seen": 2417426432
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002721010101010101,
      "loss": 1.5276,
      "theoretical_loss": 3.3775832299900124,
      "tokens_seen": 2417491968
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002720808080808081,
      "loss": 1.4756,
      "theoretical_loss": 3.3775758772725326,
      "tokens_seen": 2417557504
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002720606060606061,
      "loss": 1.4306,
      "theoretical_loss": 3.377568524810178,
      "tokens_seen": 2417623040
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027204040404040403,
      "loss": 1.3386,
      "theoretical_loss": 3.377561172602933,
      "tokens_seen": 2417688576
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000272020202020202,
      "loss": 1.4023,
      "theoretical_loss": 3.377553820650782,
      "tokens_seen": 2417754112
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027200000000000005,
      "loss": 1.5993,
      "theoretical_loss": 3.3775464689537085,
      "tokens_seen": 2417819648
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000271979797979798,
      "loss": 1.4839,
      "theoretical_loss": 3.3775391175116973,
      "tokens_seen": 2417885184
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027195959595959597,
      "loss": 1.4307,
      "theoretical_loss": 3.3775317663247324,
      "tokens_seen": 2417950720
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002719393939393939,
      "loss": 1.3631,
      "theoretical_loss": 3.3775244153927986,
      "tokens_seen": 2418016256
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002719191919191919,
      "loss": 1.4699,
      "theoretical_loss": 3.3775170647158794,
      "tokens_seen": 2418081792
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027189898989898994,
      "loss": 1.4571,
      "theoretical_loss": 3.3775097142939594,
      "tokens_seen": 2418147328
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002718787878787879,
      "loss": 1.4639,
      "theoretical_loss": 3.3775023641270225,
      "tokens_seen": 2418212864
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027185858585858586,
      "loss": 1.5969,
      "theoretical_loss": 3.3774950142150537,
      "tokens_seen": 2418278400
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002718383838383838,
      "loss": 1.2887,
      "theoretical_loss": 3.3774876645580365,
      "tokens_seen": 2418343936
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027181818181818177,
      "loss": 1.5276,
      "theoretical_loss": 3.3774803151559554,
      "tokens_seen": 2418409472
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027179797979797984,
      "loss": 1.391,
      "theoretical_loss": 3.3774729660087948,
      "tokens_seen": 2418475008
    },
    {
      "epoch": 0.46,
      "objective/train/aligned_token_count": 0.34707966446876526,
      "objective/train/avg_token_score": 0.015130029059946537,
      "objective/train/docs_used": 1362527,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9828377366065979,
      "objective/train/original_loss": 4.681084632873535,
      "objective/train/theoretical_loss": 3.3774656171165387,
      "objective/train/tokens_used": 777399776,
      "objective/train/unlikelihood_loss": 0.053071580827236176,
      "theoretical_loss": 3.3774656171165387,
      "tokens_seen": 2418540544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002717777777777778,
      "loss": 1.2464,
      "theoretical_loss": 3.3774656171165387,
      "tokens_seen": 2418540544
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027175757575757575,
      "loss": 1.5075,
      "theoretical_loss": 3.3774582684791716,
      "tokens_seen": 2418606080
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002717373737373737,
      "loss": 1.3444,
      "theoretical_loss": 3.3774509200966776,
      "tokens_seen": 2418671616
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002717171717171717,
      "loss": 1.4334,
      "theoretical_loss": 3.377443571969041,
      "tokens_seen": 2418737152
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027169696969696973,
      "loss": 1.2845,
      "theoretical_loss": 3.377436224096246,
      "tokens_seen": 2418802688
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002716767676767677,
      "loss": 1.3205,
      "theoretical_loss": 3.3774288764782767,
      "tokens_seen": 2418868224
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027165656565656564,
      "loss": 1.4124,
      "theoretical_loss": 3.377421529115118,
      "tokens_seen": 2418933760
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002716363636363636,
      "loss": 1.4118,
      "theoretical_loss": 3.377414182006753,
      "tokens_seen": 2418999296
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002716161616161616,
      "loss": 1.3825,
      "theoretical_loss": 3.377406835153167,
      "tokens_seen": 2419064832
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715959595959596,
      "loss": 1.4019,
      "theoretical_loss": 3.3773994885543446,
      "tokens_seen": 2419130368
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002715757575757576,
      "loss": 1.4273,
      "theoretical_loss": 3.377392142210269,
      "tokens_seen": 2419195904
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027155555555555554,
      "loss": 1.4283,
      "theoretical_loss": 3.3773847961209245,
      "tokens_seen": 2419261440
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027153535353535355,
      "loss": 1.5825,
      "theoretical_loss": 3.377377450286296,
      "tokens_seen": 2419326976
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027151515151515156,
      "loss": 1.3173,
      "theoretical_loss": 3.377370104706367,
      "tokens_seen": 2419392512
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002714949494949495,
      "loss": 1.3996,
      "theoretical_loss": 3.377362759381123,
      "tokens_seen": 2419458048
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027147474747474747,
      "loss": 1.4768,
      "theoretical_loss": 3.3773554143105473,
      "tokens_seen": 2419523584
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027145454545454543,
      "loss": 1.4148,
      "theoretical_loss": 3.3773480694946243,
      "tokens_seen": 2419589120
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027143434343434344,
      "loss": 1.6371,
      "theoretical_loss": 3.3773407249333385,
      "tokens_seen": 2419654656
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027141414141414145,
      "loss": 1.4929,
      "theoretical_loss": 3.3773333806266743,
      "tokens_seen": 2419720192
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002713939393939394,
      "loss": 1.5436,
      "theoretical_loss": 3.3773260365746154,
      "tokens_seen": 2419785728
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027137373737373737,
      "loss": 1.1232,
      "theoretical_loss": 3.3773186927771466,
      "tokens_seen": 2419851264
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002713535353535354,
      "loss": 1.2031,
      "theoretical_loss": 3.377311349234252,
      "tokens_seen": 2419916800
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027133333333333333,
      "loss": 1.3913,
      "theoretical_loss": 3.3773040059459154,
      "tokens_seen": 2419982336
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027131313131313134,
      "loss": 1.4005,
      "theoretical_loss": 3.377296662912122,
      "tokens_seen": 2420047872
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002712929292929293,
      "loss": 1.274,
      "theoretical_loss": 3.3772893201328555,
      "tokens_seen": 2420113408
    },
    {
      "epoch": 0.46,
      "objective/train/aligned_token_count": 0.4501466155052185,
      "objective/train/avg_token_score": 0.003827194683253765,
      "objective/train/docs_used": 1363833,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4754382371902466,
      "objective/train/original_loss": 4.224735260009766,
      "objective/train/theoretical_loss": 3.3772819776081007,
      "objective/train/tokens_used": 779038176,
      "objective/train/unlikelihood_loss": 0.05591339245438576,
      "theoretical_loss": 3.3772819776081007,
      "tokens_seen": 2420178944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027127272727272726,
      "loss": 1.4997,
      "theoretical_loss": 3.3772819776081007,
      "tokens_seen": 2420178944
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027125252525252527,
      "loss": 1.2624,
      "theoretical_loss": 3.3772746353378413,
      "tokens_seen": 2420244480
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002712323232323232,
      "loss": 1.371,
      "theoretical_loss": 3.377267293322062,
      "tokens_seen": 2420310016
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027121212121212124,
      "loss": 1.5152,
      "theoretical_loss": 3.3772599515607467,
      "tokens_seen": 2420375552
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002711919191919192,
      "loss": 1.2702,
      "theoretical_loss": 3.3772526100538798,
      "tokens_seen": 2420441088
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002711717171717172,
      "loss": 1.4656,
      "theoretical_loss": 3.3772452688014454,
      "tokens_seen": 2420506624
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027115151515151516,
      "loss": 1.3483,
      "theoretical_loss": 3.377237927803429,
      "tokens_seen": 2420572160
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002711313131313131,
      "loss": 1.305,
      "theoretical_loss": 3.377230587059813,
      "tokens_seen": 2420637696
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027111111111111113,
      "loss": 1.3341,
      "theoretical_loss": 3.3772232465705834,
      "tokens_seen": 2420703232
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002710909090909091,
      "loss": 1.5441,
      "theoretical_loss": 3.3772159063357234,
      "tokens_seen": 2420768768
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002710707070707071,
      "loss": 1.5999,
      "theoretical_loss": 3.3772085663552174,
      "tokens_seen": 2420834304
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027105050505050505,
      "loss": 1.0797,
      "theoretical_loss": 3.37720122662905,
      "tokens_seen": 2420899840
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000271030303030303,
      "loss": 1.459,
      "theoretical_loss": 3.377193887157206,
      "tokens_seen": 2420965376
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000271010101010101,
      "loss": 1.3177,
      "theoretical_loss": 3.3771865479396688,
      "tokens_seen": 2421030912
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027098989898989903,
      "loss": 1.6586,
      "theoretical_loss": 3.3771792089764237,
      "tokens_seen": 2421096448
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.000270969696969697,
      "loss": 1.3761,
      "theoretical_loss": 3.3771718702674534,
      "tokens_seen": 2421161984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027094949494949495,
      "loss": 1.4629,
      "theoretical_loss": 3.3771645318127437,
      "tokens_seen": 2421227520
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002709292929292929,
      "loss": 1.4947,
      "theoretical_loss": 3.3771571936122786,
      "tokens_seen": 2421293056
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002709090909090909,
      "loss": 1.3086,
      "theoretical_loss": 3.3771498556660418,
      "tokens_seen": 2421358592
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002708888888888889,
      "loss": 1.4972,
      "theoretical_loss": 3.3771425179740184,
      "tokens_seen": 2421424128
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002708686868686869,
      "loss": 1.4944,
      "theoretical_loss": 3.3771351805361918,
      "tokens_seen": 2421489664
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027084848484848484,
      "loss": 1.3975,
      "theoretical_loss": 3.3771278433525476,
      "tokens_seen": 2421555200
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002708282828282828,
      "loss": 1.5136,
      "theoretical_loss": 3.3771205064230685,
      "tokens_seen": 2421620736
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027080808080808086,
      "loss": 1.404,
      "theoretical_loss": 3.3771131697477403,
      "tokens_seen": 2421686272
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002707878787878788,
      "loss": 1.6175,
      "theoretical_loss": 3.3771058333265467,
      "tokens_seen": 2421751808
    },
    {
      "epoch": 0.46,
      "objective/train/aligned_token_count": 0.374450147151947,
      "objective/train/avg_token_score": 0.017216596752405167,
      "objective/train/docs_used": 1364650,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2159229516983032,
      "objective/train/original_loss": 4.7002854347229,
      "objective/train/theoretical_loss": 3.3770984971594715,
      "objective/train/tokens_used": 780676576,
      "objective/train/unlikelihood_loss": 0.059372156858444214,
      "theoretical_loss": 3.3770984971594715,
      "tokens_seen": 2421817344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002707676767676768,
      "loss": 1.4491,
      "theoretical_loss": 3.3770984971594715,
      "tokens_seen": 2421817344
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027074747474747473,
      "loss": 1.6203,
      "theoretical_loss": 3.3770911612464998,
      "tokens_seen": 2421882880
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002707272727272727,
      "loss": 1.4145,
      "theoretical_loss": 3.377083825587616,
      "tokens_seen": 2421948416
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027070707070707075,
      "loss": 1.5303,
      "theoretical_loss": 3.377076490182804,
      "tokens_seen": 2422013952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002706868686868687,
      "loss": 1.4276,
      "theoretical_loss": 3.3770691550320477,
      "tokens_seen": 2422079488
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027066666666666667,
      "loss": 1.2244,
      "theoretical_loss": 3.3770618201353324,
      "tokens_seen": 2422145024
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002706464646464646,
      "loss": 1.5879,
      "theoretical_loss": 3.377054485492642,
      "tokens_seen": 2422210560
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027062626262626264,
      "loss": 1.5181,
      "theoretical_loss": 3.3770471511039606,
      "tokens_seen": 2422276096
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027060606060606065,
      "loss": 1.3388,
      "theoretical_loss": 3.3770398169692726,
      "tokens_seen": 2422341632
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002705858585858586,
      "loss": 1.4006,
      "theoretical_loss": 3.3770324830885627,
      "tokens_seen": 2422407168
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027056565656565656,
      "loss": 1.7885,
      "theoretical_loss": 3.377025149461815,
      "tokens_seen": 2422472704
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002705454545454545,
      "loss": 1.5197,
      "theoretical_loss": 3.3770178160890136,
      "tokens_seen": 2422538240
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027052525252525253,
      "loss": 1.4357,
      "theoretical_loss": 3.3770104829701437,
      "tokens_seen": 2422603776
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027050505050505054,
      "loss": 1.3793,
      "theoretical_loss": 3.3770031501051885,
      "tokens_seen": 2422669312
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704848484848485,
      "loss": 1.4988,
      "theoretical_loss": 3.376995817494133,
      "tokens_seen": 2422734848
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027046464646464645,
      "loss": 1.5143,
      "theoretical_loss": 3.376988485136961,
      "tokens_seen": 2422800384
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704444444444444,
      "loss": 1.3327,
      "theoretical_loss": 3.3769811530336575,
      "tokens_seen": 2422865920
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002704242424242424,
      "loss": 1.3438,
      "theoretical_loss": 3.376973821184207,
      "tokens_seen": 2422931456
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027040404040404043,
      "loss": 1.2838,
      "theoretical_loss": 3.3769664895885927,
      "tokens_seen": 2422996992
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002703838383838384,
      "loss": 1.5069,
      "theoretical_loss": 3.3769591582468004,
      "tokens_seen": 2423062528
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027036363636363635,
      "loss": 1.4779,
      "theoretical_loss": 3.376951827158813,
      "tokens_seen": 2423128064
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00027034343434343436,
      "loss": 1.5118,
      "theoretical_loss": 3.376944496324616,
      "tokens_seen": 2423193600
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002703232323232323,
      "loss": 1.5388,
      "theoretical_loss": 3.3769371657441933,
      "tokens_seen": 2423259136
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002703030303030303,
      "loss": 1.4908,
      "theoretical_loss": 3.376929835417529,
      "tokens_seen": 2423324672
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002702828282828283,
      "loss": 1.3317,
      "theoretical_loss": 3.376922505344608,
      "tokens_seen": 2423390208
    },
    {
      "epoch": 0.46,
      "objective/train/aligned_token_count": 0.5276759266853333,
      "objective/train/avg_token_score": 0.01432180404663086,
      "objective/train/docs_used": 1365365,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7265634536743164,
      "objective/train/original_loss": 4.5084404945373535,
      "objective/train/theoretical_loss": 3.376915175525414,
      "objective/train/tokens_used": 782314976,
      "objective/train/unlikelihood_loss": 0.04520232230424881,
      "theoretical_loss": 3.376915175525414,
      "tokens_seen": 2423455744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027026262626262624,
      "loss": 1.5789,
      "theoretical_loss": 3.376915175525414,
      "tokens_seen": 2423455744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027024242424242425,
      "loss": 1.3682,
      "theoretical_loss": 3.3769078459599324,
      "tokens_seen": 2423521280
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002702222222222222,
      "loss": 1.4071,
      "theoretical_loss": 3.3769005166481465,
      "tokens_seen": 2423586816
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002702020202020202,
      "loss": 1.4341,
      "theoretical_loss": 3.376893187590041,
      "tokens_seen": 2423652352
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701818181818182,
      "loss": 1.5279,
      "theoretical_loss": 3.3768858587856005,
      "tokens_seen": 2423717888
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701616161616162,
      "loss": 1.3947,
      "theoretical_loss": 3.376878530234809,
      "tokens_seen": 2423783424
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027014141414141414,
      "loss": 1.3634,
      "theoretical_loss": 3.376871201937651,
      "tokens_seen": 2423848960
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701212121212121,
      "loss": 1.3195,
      "theoretical_loss": 3.376863873894111,
      "tokens_seen": 2423914496
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701010101010101,
      "loss": 1.5182,
      "theoretical_loss": 3.3768565461041735,
      "tokens_seen": 2423980032
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002701010101010101,
      "loss": 1.4401,
      "theoretical_loss": 3.376849218567822,
      "tokens_seen": 2424045568
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027008080808080807,
      "loss": 1.5109,
      "theoretical_loss": 3.376841891285042,
      "tokens_seen": 2424111104
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002700606060606061,
      "loss": 1.4126,
      "theoretical_loss": 3.3768345642558173,
      "tokens_seen": 2424176640
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027004040404040404,
      "loss": 1.5159,
      "theoretical_loss": 3.3768272374801325,
      "tokens_seen": 2424242176
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000270020202020202,
      "loss": 1.3438,
      "theoretical_loss": 3.3768199109579715,
      "tokens_seen": 2424307712
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00027,
      "loss": 1.196,
      "theoretical_loss": 3.376812584689319,
      "tokens_seen": 2424373248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000269979797979798,
      "loss": 1.4208,
      "theoretical_loss": 3.3768052586741595,
      "tokens_seen": 2424438784
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026995959595959597,
      "loss": 1.4078,
      "theoretical_loss": 3.3767979329124773,
      "tokens_seen": 2424504320
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026993939393939393,
      "loss": 1.3336,
      "theoretical_loss": 3.3767906074042564,
      "tokens_seen": 2424569856
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002699191919191919,
      "loss": 1.3594,
      "theoretical_loss": 3.376783282149482,
      "tokens_seen": 2424635392
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002698989898989899,
      "loss": 1.3883,
      "theoretical_loss": 3.376775957148138,
      "tokens_seen": 2424700928
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002698787878787879,
      "loss": 1.4308,
      "theoretical_loss": 3.376768632400208,
      "tokens_seen": 2424766464
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026985858585858586,
      "loss": 1.4464,
      "theoretical_loss": 3.376761307905678,
      "tokens_seen": 2424832000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002698383838383838,
      "loss": 1.6151,
      "theoretical_loss": 3.376753983664531,
      "tokens_seen": 2424897536
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026981818181818183,
      "loss": 1.2595,
      "theoretical_loss": 3.376746659676752,
      "tokens_seen": 2424963072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026979797979797984,
      "loss": 1.3309,
      "theoretical_loss": 3.3767393359423257,
      "tokens_seen": 2425028608
    },
    {
      "epoch": 0.47,
      "objective/train/aligned_token_count": 0.3789711594581604,
      "objective/train/avg_token_score": 0.03819414973258972,
      "objective/train/docs_used": 1366807,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2040443420410156,
      "objective/train/original_loss": 4.851312637329102,
      "objective/train/theoretical_loss": 3.3767320124612357,
      "objective/train/tokens_used": 783953376,
      "objective/train/unlikelihood_loss": 0.055656250566244125,
      "theoretical_loss": 3.3767320124612357,
      "tokens_seen": 2425094144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697777777777778,
      "loss": 1.4257,
      "theoretical_loss": 3.3767320124612357,
      "tokens_seen": 2425094144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026975757575757576,
      "loss": 1.4545,
      "theoretical_loss": 3.376724689233467,
      "tokens_seen": 2425159680
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697373737373737,
      "loss": 1.4226,
      "theoretical_loss": 3.3767173662590038,
      "tokens_seen": 2425225216
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002697171717171717,
      "loss": 1.6581,
      "theoretical_loss": 3.3767100435378303,
      "tokens_seen": 2425290752
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026969696969696974,
      "loss": 1.3555,
      "theoretical_loss": 3.3767027210699307,
      "tokens_seen": 2425356288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002696767676767677,
      "loss": 1.1758,
      "theoretical_loss": 3.3766953988552904,
      "tokens_seen": 2425421824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026965656565656565,
      "loss": 1.4061,
      "theoretical_loss": 3.376688076893893,
      "tokens_seen": 2425487360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002696363636363636,
      "loss": 1.5817,
      "theoretical_loss": 3.376680755185723,
      "tokens_seen": 2425552896
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026961616161616167,
      "loss": 1.3592,
      "theoretical_loss": 3.376673433730765,
      "tokens_seen": 2425618432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026959595959595963,
      "loss": 1.5262,
      "theoretical_loss": 3.376666112529003,
      "tokens_seen": 2425683968
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002695757575757576,
      "loss": 1.3167,
      "theoretical_loss": 3.376658791580422,
      "tokens_seen": 2425749504
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026955555555555554,
      "loss": 1.2868,
      "theoretical_loss": 3.376651470885006,
      "tokens_seen": 2425815040
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002695353535353535,
      "loss": 1.4404,
      "theoretical_loss": 3.3766441504427394,
      "tokens_seen": 2425880576
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026951515151515156,
      "loss": 1.5302,
      "theoretical_loss": 3.3766368302536067,
      "tokens_seen": 2425946112
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002694949494949495,
      "loss": 1.1938,
      "theoretical_loss": 3.376629510317592,
      "tokens_seen": 2426011648
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002694747474747475,
      "loss": 1.3995,
      "theoretical_loss": 3.3766221906346807,
      "tokens_seen": 2426077184
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026945454545454543,
      "loss": 1.4581,
      "theoretical_loss": 3.3766148712048563,
      "tokens_seen": 2426142720
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026943434343434345,
      "loss": 1.5112,
      "theoretical_loss": 3.376607552028103,
      "tokens_seen": 2426208256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026941414141414146,
      "loss": 1.3009,
      "theoretical_loss": 3.376600233104406,
      "tokens_seen": 2426273792
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002693939393939394,
      "loss": 1.3945,
      "theoretical_loss": 3.376592914433749,
      "tokens_seen": 2426339328
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026937373737373737,
      "loss": 1.5286,
      "theoretical_loss": 3.3765855960161173,
      "tokens_seen": 2426404864
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026935353535353533,
      "loss": 1.8625,
      "theoretical_loss": 3.3765782778514946,
      "tokens_seen": 2426470400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026933333333333334,
      "loss": 1.2729,
      "theoretical_loss": 3.3765709599398654,
      "tokens_seen": 2426535936
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026931313131313135,
      "loss": 1.3806,
      "theoretical_loss": 3.3765636422812144,
      "tokens_seen": 2426601472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002692929292929293,
      "loss": 1.2773,
      "theoretical_loss": 3.376556324875526,
      "tokens_seen": 2426667008
    },
    {
      "epoch": 0.47,
      "objective/train/aligned_token_count": 0.4211265742778778,
      "objective/train/avg_token_score": 0.0058179278858006,
      "objective/train/docs_used": 1367391,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.287906289100647,
      "objective/train/original_loss": 4.212308883666992,
      "objective/train/theoretical_loss": 3.376549007722784,
      "objective/train/tokens_used": 785591776,
      "objective/train/unlikelihood_loss": 0.05801492929458618,
      "theoretical_loss": 3.376549007722784,
      "tokens_seen": 2426732544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026927272727272726,
      "loss": 1.351,
      "theoretical_loss": 3.376549007722784,
      "tokens_seen": 2426732544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002692525252525253,
      "loss": 1.3515,
      "theoretical_loss": 3.3765416908229735,
      "tokens_seen": 2426798080
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026923232323232323,
      "loss": 1.6064,
      "theoretical_loss": 3.376534374176079,
      "tokens_seen": 2426863616
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026921212121212124,
      "loss": 1.5312,
      "theoretical_loss": 3.3765270577820847,
      "tokens_seen": 2426929152
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002691919191919192,
      "loss": 1.4,
      "theoretical_loss": 3.3765197416409745,
      "tokens_seen": 2426994688
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026917171717171716,
      "loss": 1.4882,
      "theoretical_loss": 3.376512425752734,
      "tokens_seen": 2427060224
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026915151515151517,
      "loss": 1.4684,
      "theoretical_loss": 3.376505110117346,
      "tokens_seen": 2427125760
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002691313131313131,
      "loss": 1.3819,
      "theoretical_loss": 3.3764977947347967,
      "tokens_seen": 2427191296
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026911111111111114,
      "loss": 1.419,
      "theoretical_loss": 3.37649047960507,
      "tokens_seen": 2427256832
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002690909090909091,
      "loss": 1.4821,
      "theoretical_loss": 3.3764831647281492,
      "tokens_seen": 2427322368
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002690707070707071,
      "loss": 1.3745,
      "theoretical_loss": 3.37647585010402,
      "tokens_seen": 2427387904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026905050505050506,
      "loss": 1.185,
      "theoretical_loss": 3.3764685357326667,
      "tokens_seen": 2427453440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000269030303030303,
      "loss": 1.5601,
      "theoretical_loss": 3.3764612216140732,
      "tokens_seen": 2427518976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026901010101010103,
      "loss": 1.3534,
      "theoretical_loss": 3.376453907748224,
      "tokens_seen": 2427584512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000268989898989899,
      "loss": 1.4042,
      "theoretical_loss": 3.3764465941351043,
      "tokens_seen": 2427650048
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000268969696969697,
      "loss": 1.5393,
      "theoretical_loss": 3.3764392807746977,
      "tokens_seen": 2427715584
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026894949494949495,
      "loss": 1.4397,
      "theoretical_loss": 3.376431967666989,
      "tokens_seen": 2427781120
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002689292929292929,
      "loss": 1.4097,
      "theoretical_loss": 3.3764246548119625,
      "tokens_seen": 2427846656
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002689090909090909,
      "loss": 1.503,
      "theoretical_loss": 3.376417342209603,
      "tokens_seen": 2427912192
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026888888888888893,
      "loss": 1.2875,
      "theoretical_loss": 3.3764100298598945,
      "tokens_seen": 2427977728
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002688686868686869,
      "loss": 1.317,
      "theoretical_loss": 3.3764027177628217,
      "tokens_seen": 2428043264
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026884848484848485,
      "loss": 1.3102,
      "theoretical_loss": 3.376395405918369,
      "tokens_seen": 2428108800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002688282828282828,
      "loss": 1.4039,
      "theoretical_loss": 3.376388094326521,
      "tokens_seen": 2428174336
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002688080808080808,
      "loss": 1.2587,
      "theoretical_loss": 3.376380782987262,
      "tokens_seen": 2428239872
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687878787878788,
      "loss": 1.4214,
      "theoretical_loss": 3.3763734719005765,
      "tokens_seen": 2428305408
    },
    {
      "epoch": 0.47,
      "objective/train/aligned_token_count": 0.5036656856536865,
      "objective/train/avg_token_score": 0.011943654157221317,
      "objective/train/docs_used": 1368496,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6017712354660034,
      "objective/train/original_loss": 4.273252487182617,
      "objective/train/theoretical_loss": 3.376366161066449,
      "objective/train/tokens_used": 787230176,
      "objective/train/unlikelihood_loss": 0.05992387607693672,
      "theoretical_loss": 3.376366161066449,
      "tokens_seen": 2428370944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687676767676768,
      "loss": 1.5512,
      "theoretical_loss": 3.376366161066449,
      "tokens_seen": 2428370944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026874747474747474,
      "loss": 1.6345,
      "theoretical_loss": 3.3763588504848636,
      "tokens_seen": 2428436480
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002687272727272727,
      "loss": 1.5013,
      "theoretical_loss": 3.376351540155805,
      "tokens_seen": 2428502016
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026870707070707076,
      "loss": 1.244,
      "theoretical_loss": 3.3763442300792583,
      "tokens_seen": 2428567552
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686868686868687,
      "loss": 1.5931,
      "theoretical_loss": 3.376336920255207,
      "tokens_seen": 2428633088
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686666666666667,
      "loss": 1.308,
      "theoretical_loss": 3.376329610683636,
      "tokens_seen": 2428698624
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026864646464646463,
      "loss": 1.2824,
      "theoretical_loss": 3.37632230136453,
      "tokens_seen": 2428764160
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002686262626262626,
      "loss": 1.2302,
      "theoretical_loss": 3.3763149922978726,
      "tokens_seen": 2428829696
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026860606060606065,
      "loss": 1.4417,
      "theoretical_loss": 3.376307683483649,
      "tokens_seen": 2428895232
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002685858585858586,
      "loss": 1.5679,
      "theoretical_loss": 3.376300374921844,
      "tokens_seen": 2428960768
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026856565656565657,
      "loss": 1.4477,
      "theoretical_loss": 3.376293066612441,
      "tokens_seen": 2429026304
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002685454545454545,
      "loss": 1.4836,
      "theoretical_loss": 3.3762857585554253,
      "tokens_seen": 2429091840
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002685252525252525,
      "loss": 1.4442,
      "theoretical_loss": 3.3762784507507813,
      "tokens_seen": 2429157376
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026850505050505055,
      "loss": 1.4629,
      "theoretical_loss": 3.376271143198493,
      "tokens_seen": 2429222912
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002684848484848485,
      "loss": 1.4456,
      "theoretical_loss": 3.3762638358985457,
      "tokens_seen": 2429288448
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026846464646464646,
      "loss": 1.294,
      "theoretical_loss": 3.376256528850923,
      "tokens_seen": 2429353984
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002684444444444444,
      "loss": 1.6633,
      "theoretical_loss": 3.37624922205561,
      "tokens_seen": 2429419520
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026842424242424243,
      "loss": 1.3302,
      "theoretical_loss": 3.3762419155125905,
      "tokens_seen": 2429485056
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026840404040404044,
      "loss": 1.3347,
      "theoretical_loss": 3.37623460922185,
      "tokens_seen": 2429550592
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002683838383838384,
      "loss": 1.4155,
      "theoretical_loss": 3.3762273031833723,
      "tokens_seen": 2429616128
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026836363636363635,
      "loss": 1.5818,
      "theoretical_loss": 3.3762199973971416,
      "tokens_seen": 2429681664
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002683434343434343,
      "loss": 1.2772,
      "theoretical_loss": 3.3762126918631434,
      "tokens_seen": 2429747200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002683232323232323,
      "loss": 1.6304,
      "theoretical_loss": 3.376205386581361,
      "tokens_seen": 2429812736
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026830303030303033,
      "loss": 1.3993,
      "theoretical_loss": 3.3761980815517796,
      "tokens_seen": 2429878272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682828282828283,
      "loss": 1.4549,
      "theoretical_loss": 3.3761907767743837,
      "tokens_seen": 2429943808
    },
    {
      "epoch": 0.47,
      "objective/train/aligned_token_count": 0.4417766332626343,
      "objective/train/avg_token_score": 0.012998738326132298,
      "objective/train/docs_used": 1368772,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5431228876113892,
      "objective/train/original_loss": 4.493236541748047,
      "objective/train/theoretical_loss": 3.3761834722491577,
      "objective/train/tokens_used": 788868576,
      "objective/train/unlikelihood_loss": 0.04973199963569641,
      "theoretical_loss": 3.3761834722491577,
      "tokens_seen": 2430009344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026826262626262624,
      "loss": 1.4142,
      "theoretical_loss": 3.3761834722491577,
      "tokens_seen": 2430009344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026824242424242426,
      "loss": 1.6739,
      "theoretical_loss": 3.376176167976086,
      "tokens_seen": 2430074880
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682222222222222,
      "loss": 1.4937,
      "theoretical_loss": 3.376168863955153,
      "tokens_seen": 2430140416
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002682020202020202,
      "loss": 1.599,
      "theoretical_loss": 3.3761615601863433,
      "tokens_seen": 2430205952
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681818181818182,
      "loss": 1.3969,
      "theoretical_loss": 3.3761542566696416,
      "tokens_seen": 2430271488
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026816161616161614,
      "loss": 1.499,
      "theoretical_loss": 3.376146953405032,
      "tokens_seen": 2430337024
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026814141414141415,
      "loss": 1.1976,
      "theoretical_loss": 3.3761396503924996,
      "tokens_seen": 2430402560
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681212121212121,
      "loss": 1.2928,
      "theoretical_loss": 3.376132347632028,
      "tokens_seen": 2430468096
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002681010101010101,
      "loss": 1.5085,
      "theoretical_loss": 3.3761250451236027,
      "tokens_seen": 2430533632
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002680808080808081,
      "loss": 1.3558,
      "theoretical_loss": 3.376117742867208,
      "tokens_seen": 2430599168
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002680606060606061,
      "loss": 1.5296,
      "theoretical_loss": 3.3761104408628277,
      "tokens_seen": 2430664704
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026804040404040404,
      "loss": 1.3762,
      "theoretical_loss": 3.376103139110447,
      "tokens_seen": 2430730240
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026802020202020205,
      "loss": 1.4348,
      "theoretical_loss": 3.37609583761005,
      "tokens_seen": 2430795776
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000268,
      "loss": 1.3708,
      "theoretical_loss": 3.376088536361621,
      "tokens_seen": 2430861312
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026797979797979797,
      "loss": 1.564,
      "theoretical_loss": 3.3760812353651457,
      "tokens_seen": 2430926848
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000267959595959596,
      "loss": 1.4627,
      "theoretical_loss": 3.3760739346206075,
      "tokens_seen": 2430992384
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026793939393939393,
      "loss": 1.3129,
      "theoretical_loss": 3.376066634127991,
      "tokens_seen": 2431057920
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026791919191919194,
      "loss": 1.564,
      "theoretical_loss": 3.376059333887281,
      "tokens_seen": 2431123456
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002678989898989899,
      "loss": 1.5397,
      "theoretical_loss": 3.3760520338984623,
      "tokens_seen": 2431188992
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002678787878787879,
      "loss": 1.6729,
      "theoretical_loss": 3.3760447341615185,
      "tokens_seen": 2431254528
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026785858585858587,
      "loss": 1.2465,
      "theoretical_loss": 3.376037434676435,
      "tokens_seen": 2431320064
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002678383838383838,
      "loss": 1.2975,
      "theoretical_loss": 3.3760301354431963,
      "tokens_seen": 2431385600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026781818181818184,
      "loss": 1.447,
      "theoretical_loss": 3.376022836461786,
      "tokens_seen": 2431451136
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002677979797979798,
      "loss": 1.3259,
      "theoretical_loss": 3.37601553773219,
      "tokens_seen": 2431516672
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002677777777777778,
      "loss": 1.3794,
      "theoretical_loss": 3.3760082392543915,
      "tokens_seen": 2431582208
    },
    {
      "epoch": 0.47,
      "objective/train/aligned_token_count": 0.413184255361557,
      "objective/train/avg_token_score": 0.007847166620194912,
      "objective/train/docs_used": 1370076,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3364901542663574,
      "objective/train/original_loss": 4.432122230529785,
      "objective/train/theoretical_loss": 3.3760009410283756,
      "objective/train/tokens_used": 790506976,
      "objective/train/unlikelihood_loss": 0.0553634874522686,
      "theoretical_loss": 3.3760009410283756,
      "tokens_seen": 2431647744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026775757575757576,
      "loss": 1.1973,
      "theoretical_loss": 3.3760009410283756,
      "tokens_seen": 2431647744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002677373737373737,
      "loss": 1.3862,
      "theoretical_loss": 3.3759936430541275,
      "tokens_seen": 2431713280
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026771717171717173,
      "loss": 1.237,
      "theoretical_loss": 3.3759863453316306,
      "tokens_seen": 2431778816
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026769696969696974,
      "loss": 1.4097,
      "theoretical_loss": 3.37597904786087,
      "tokens_seen": 2431844352
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002676767676767677,
      "loss": 1.4091,
      "theoretical_loss": 3.37597175064183,
      "tokens_seen": 2431909888
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026765656565656566,
      "loss": 1.5678,
      "theoretical_loss": 3.375964453674495,
      "tokens_seen": 2431975424
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002676363636363636,
      "loss": 1.47,
      "theoretical_loss": 3.3759571569588505,
      "tokens_seen": 2432040960
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002676161616161616,
      "loss": 1.5157,
      "theoretical_loss": 3.37594986049488,
      "tokens_seen": 2432106496
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026759595959595963,
      "loss": 1.4107,
      "theoretical_loss": 3.3759425642825684,
      "tokens_seen": 2432172032
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002675757575757576,
      "loss": 1.4374,
      "theoretical_loss": 3.3759352683219004,
      "tokens_seen": 2432237568
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026755555555555555,
      "loss": 1.3261,
      "theoretical_loss": 3.37592797261286,
      "tokens_seen": 2432303104
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002675353535353535,
      "loss": 1.6439,
      "theoretical_loss": 3.3759206771554324,
      "tokens_seen": 2432368640
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026751515151515157,
      "loss": 1.6005,
      "theoretical_loss": 3.3759133819496014,
      "tokens_seen": 2432434176
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002674949494949495,
      "loss": 1.4752,
      "theoretical_loss": 3.3759060869953528,
      "tokens_seen": 2432499712
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002674747474747475,
      "loss": 1.3351,
      "theoretical_loss": 3.3758987922926695,
      "tokens_seen": 2432565248
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026745454545454544,
      "loss": 1.4102,
      "theoretical_loss": 3.3758914978415375,
      "tokens_seen": 2432630784
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002674343434343434,
      "loss": 1.5238,
      "theoretical_loss": 3.3758842036419408,
      "tokens_seen": 2432696320
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026741414141414146,
      "loss": 1.4351,
      "theoretical_loss": 3.3758769096938632,
      "tokens_seen": 2432761856
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002673939393939394,
      "loss": 1.3411,
      "theoretical_loss": 3.3758696159972903,
      "tokens_seen": 2432827392
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002673737373737374,
      "loss": 1.3929,
      "theoretical_loss": 3.3758623225522064,
      "tokens_seen": 2432892928
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026735353535353533,
      "loss": 1.3573,
      "theoretical_loss": 3.375855029358596,
      "tokens_seen": 2432958464
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026733333333333334,
      "loss": 1.2822,
      "theoretical_loss": 3.3758477364164436,
      "tokens_seen": 2433024000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026731313131313136,
      "loss": 1.2835,
      "theoretical_loss": 3.3758404437257337,
      "tokens_seen": 2433089536
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672929292929293,
      "loss": 1.3646,
      "theoretical_loss": 3.3758331512864506,
      "tokens_seen": 2433155072
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026727272727272727,
      "loss": 1.1019,
      "theoretical_loss": 3.3758258590985797,
      "tokens_seen": 2433220608
    },
    {
      "epoch": 0.47,
      "objective/train/aligned_token_count": 0.5023826956748962,
      "objective/train/avg_token_score": 0.0052525377832353115,
      "objective/train/docs_used": 1370835,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6614478826522827,
      "objective/train/original_loss": 4.050638675689697,
      "objective/train/theoretical_loss": 3.3758185671621046,
      "objective/train/tokens_used": 792145376,
      "objective/train/unlikelihood_loss": 0.0546875,
      "theoretical_loss": 3.3758185671621046,
      "tokens_seen": 2433286144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002672525252525252,
      "loss": 1.4266,
      "theoretical_loss": 3.3758185671621046,
      "tokens_seen": 2433286144
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026723232323232324,
      "loss": 1.467,
      "theoretical_loss": 3.3758112754770107,
      "tokens_seen": 2433351680
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026721212121212125,
      "loss": 1.4227,
      "theoretical_loss": 3.375803984043282,
      "tokens_seen": 2433417216
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002671919191919192,
      "loss": 1.4943,
      "theoretical_loss": 3.375796692860903,
      "tokens_seen": 2433482752
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026717171717171716,
      "loss": 1.3065,
      "theoretical_loss": 3.3757894019298584,
      "tokens_seen": 2433548288
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002671515151515151,
      "loss": 1.7132,
      "theoretical_loss": 3.375782111250133,
      "tokens_seen": 2433613824
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026713131313131313,
      "loss": 1.312,
      "theoretical_loss": 3.3757748208217118,
      "tokens_seen": 2433679360
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026711111111111114,
      "loss": 1.453,
      "theoretical_loss": 3.3757675306445782,
      "tokens_seen": 2433744896
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002670909090909091,
      "loss": 1.6138,
      "theoretical_loss": 3.3757602407187175,
      "tokens_seen": 2433810432
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026707070707070705,
      "loss": 1.465,
      "theoretical_loss": 3.3757529510441144,
      "tokens_seen": 2433875968
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026705050505050507,
      "loss": 1.3935,
      "theoretical_loss": 3.375745661620753,
      "tokens_seen": 2433941504
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000267030303030303,
      "loss": 1.3013,
      "theoretical_loss": 3.375738372448618,
      "tokens_seen": 2434007040
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026701010101010103,
      "loss": 1.5825,
      "theoretical_loss": 3.375731083527694,
      "tokens_seen": 2434072576
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000266989898989899,
      "loss": 1.4637,
      "theoretical_loss": 3.375723794857966,
      "tokens_seen": 2434138112
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026696969696969695,
      "loss": 1.4819,
      "theoretical_loss": 3.375716506439418,
      "tokens_seen": 2434203648
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026694949494949496,
      "loss": 1.4523,
      "theoretical_loss": 3.375709218272035,
      "tokens_seen": 2434269184
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002669292929292929,
      "loss": 1.3722,
      "theoretical_loss": 3.3757019303558016,
      "tokens_seen": 2434334720
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002669090909090909,
      "loss": 1.5882,
      "theoretical_loss": 3.3756946426907017,
      "tokens_seen": 2434400256
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002668888888888889,
      "loss": 1.2121,
      "theoretical_loss": 3.375687355276721,
      "tokens_seen": 2434465792
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002668686868686869,
      "loss": 1.3749,
      "theoretical_loss": 3.375680068113843,
      "tokens_seen": 2434531328
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026684848484848485,
      "loss": 1.4169,
      "theoretical_loss": 3.375672781202053,
      "tokens_seen": 2434596864
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002668282828282828,
      "loss": 1.3154,
      "theoretical_loss": 3.375665494541335,
      "tokens_seen": 2434662400
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002668080808080808,
      "loss": 1.3272,
      "theoretical_loss": 3.375658208131674,
      "tokens_seen": 2434727936
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667878787878788,
      "loss": 1.3647,
      "theoretical_loss": 3.3756509219730546,
      "tokens_seen": 2434793472
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667676767676768,
      "loss": 1.2965,
      "theoretical_loss": 3.3756436360654614,
      "tokens_seen": 2434859008
    },
    {
      "epoch": 0.47,
      "objective/train/aligned_token_count": 0.44189882278442383,
      "objective/train/avg_token_score": 0.008371703326702118,
      "objective/train/docs_used": 1372111,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4221819639205933,
      "objective/train/original_loss": 4.99832820892334,
      "objective/train/theoretical_loss": 3.375636350408879,
      "objective/train/tokens_used": 793783776,
      "objective/train/unlikelihood_loss": 0.05759654939174652,
      "theoretical_loss": 3.375636350408879,
      "tokens_seen": 2434924544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026674747474747474,
      "loss": 1.5208,
      "theoretical_loss": 3.375636350408879,
      "tokens_seen": 2434924544
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667272727272727,
      "loss": 1.5526,
      "theoretical_loss": 3.3756290650032916,
      "tokens_seen": 2434990080
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002667070707070707,
      "loss": 1.4955,
      "theoretical_loss": 3.3756217798486845,
      "tokens_seen": 2435055616
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666868686868687,
      "loss": 1.2515,
      "theoretical_loss": 3.3756144949450415,
      "tokens_seen": 2435121152
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666666666666667,
      "loss": 1.4224,
      "theoretical_loss": 3.375607210292348,
      "tokens_seen": 2435186688
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026664646464646464,
      "loss": 1.4783,
      "theoretical_loss": 3.375599925890588,
      "tokens_seen": 2435252224
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666262626262626,
      "loss": 1.3437,
      "theoretical_loss": 3.3755926417397464,
      "tokens_seen": 2435317760
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002666060606060606,
      "loss": 1.1874,
      "theoretical_loss": 3.3755853578398076,
      "tokens_seen": 2435383296
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002665858585858586,
      "loss": 1.3096,
      "theoretical_loss": 3.375578074190756,
      "tokens_seen": 2435448832
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026656565656565657,
      "loss": 1.3301,
      "theoretical_loss": 3.375570790792577,
      "tokens_seen": 2435514368
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026654545454545453,
      "loss": 1.488,
      "theoretical_loss": 3.375563507645255,
      "tokens_seen": 2435579904
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002665252525252525,
      "loss": 1.4217,
      "theoretical_loss": 3.375556224748774,
      "tokens_seen": 2435645440
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026650505050505055,
      "loss": 1.4493,
      "theoretical_loss": 3.375548942103119,
      "tokens_seen": 2435710976
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002664848484848485,
      "loss": 1.2493,
      "theoretical_loss": 3.3755416597082744,
      "tokens_seen": 2435776512
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026646464646464646,
      "loss": 1.3371,
      "theoretical_loss": 3.3755343775642253,
      "tokens_seen": 2435842048
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002664444444444444,
      "loss": 1.4032,
      "theoretical_loss": 3.3755270956709555,
      "tokens_seen": 2435907584
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002664242424242424,
      "loss": 1.2487,
      "theoretical_loss": 3.3755198140284506,
      "tokens_seen": 2435973120
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026640404040404044,
      "loss": 1.2914,
      "theoretical_loss": 3.3755125326366944,
      "tokens_seen": 2436038656
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002663838383838384,
      "loss": 1.5862,
      "theoretical_loss": 3.3755052514956723,
      "tokens_seen": 2436104192
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026636363636363636,
      "loss": 1.5442,
      "theoretical_loss": 3.3754979706053683,
      "tokens_seen": 2436169728
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002663434343434343,
      "loss": 1.4923,
      "theoretical_loss": 3.3754906899657673,
      "tokens_seen": 2436235264
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002663232323232324,
      "loss": 1.577,
      "theoretical_loss": 3.3754834095768533,
      "tokens_seen": 2436300800
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026630303030303034,
      "loss": 1.4873,
      "theoretical_loss": 3.375476129438612,
      "tokens_seen": 2436366336
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002662828282828283,
      "loss": 1.5438,
      "theoretical_loss": 3.3754688495510274,
      "tokens_seen": 2436431872
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026626262626262625,
      "loss": 1.4676,
      "theoretical_loss": 3.3754615699140844,
      "tokens_seen": 2436497408
    },
    {
      "epoch": 0.47,
      "objective/train/aligned_token_count": 0.49767839908599854,
      "objective/train/avg_token_score": 0.004969476256519556,
      "objective/train/docs_used": 1372668,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6521486043930054,
      "objective/train/original_loss": 4.205765724182129,
      "objective/train/theoretical_loss": 3.3754542905277667,
      "objective/train/tokens_used": 795422176,
      "objective/train/unlikelihood_loss": 0.047138575464487076,
      "theoretical_loss": 3.3754542905277667,
      "tokens_seen": 2436562944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002662424242424242,
      "loss": 1.6008,
      "theoretical_loss": 3.3754542905277667,
      "tokens_seen": 2436562944
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026622222222222227,
      "loss": 1.2792,
      "theoretical_loss": 3.3754470113920605,
      "tokens_seen": 2436628480
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026620202020202023,
      "loss": 1.4652,
      "theoretical_loss": 3.3754397325069494,
      "tokens_seen": 2436694016
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002661818181818182,
      "loss": 1.4571,
      "theoretical_loss": 3.375432453872418,
      "tokens_seen": 2436759552
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026616161616161614,
      "loss": 1.4756,
      "theoretical_loss": 3.3754251754884512,
      "tokens_seen": 2436825088
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026614141414141415,
      "loss": 1.3139,
      "theoretical_loss": 3.3754178973550335,
      "tokens_seen": 2436890624
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026612121212121217,
      "loss": 1.2889,
      "theoretical_loss": 3.37541061947215,
      "tokens_seen": 2436956160
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002661010101010101,
      "loss": 1.3607,
      "theoretical_loss": 3.375403341839785,
      "tokens_seen": 2437021696
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002660808080808081,
      "loss": 1.2331,
      "theoretical_loss": 3.375396064457923,
      "tokens_seen": 2437087232
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026606060606060604,
      "loss": 1.5838,
      "theoretical_loss": 3.3753887873265485,
      "tokens_seen": 2437152768
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026604040404040405,
      "loss": 1.4708,
      "theoretical_loss": 3.3753815104456466,
      "tokens_seen": 2437218304
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026602020202020206,
      "loss": 1.2511,
      "theoretical_loss": 3.3753742338152017,
      "tokens_seen": 2437283840
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000266,
      "loss": 1.5743,
      "theoretical_loss": 3.3753669574351983,
      "tokens_seen": 2437349376
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026597979797979797,
      "loss": 1.4451,
      "theoretical_loss": 3.3753596813056213,
      "tokens_seen": 2437414912
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.000265959595959596,
      "loss": 1.274,
      "theoretical_loss": 3.375352405426456,
      "tokens_seen": 2437480448
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026593939393939394,
      "loss": 1.3445,
      "theoretical_loss": 3.3753451297976858,
      "tokens_seen": 2437545984
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026591919191919195,
      "loss": 1.4743,
      "theoretical_loss": 3.375337854419296,
      "tokens_seen": 2437611520
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002658989898989899,
      "loss": 1.3312,
      "theoretical_loss": 3.375330579291271,
      "tokens_seen": 2437677056
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026587878787878786,
      "loss": 1.287,
      "theoretical_loss": 3.3753233044135955,
      "tokens_seen": 2437742592
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002658585858585859,
      "loss": 1.3386,
      "theoretical_loss": 3.375316029786254,
      "tokens_seen": 2437808128
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026583838383838383,
      "loss": 1.455,
      "theoretical_loss": 3.375308755409232,
      "tokens_seen": 2437873664
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026581818181818184,
      "loss": 1.412,
      "theoretical_loss": 3.3753014812825133,
      "tokens_seen": 2437939200
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002657979797979798,
      "loss": 1.4419,
      "theoretical_loss": 3.3752942074060828,
      "tokens_seen": 2438004736
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002657777777777778,
      "loss": 1.2685,
      "theoretical_loss": 3.3752869337799254,
      "tokens_seen": 2438070272
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026575757575757577,
      "loss": 1.4495,
      "theoretical_loss": 3.3752796604040256,
      "tokens_seen": 2438135808
    },
    {
      "epoch": 0.47,
      "objective/train/aligned_token_count": 0.45625612139701843,
      "objective/train/avg_token_score": 0.016048340126872063,
      "objective/train/docs_used": 1374127,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4453073740005493,
      "objective/train/original_loss": 4.2503132820129395,
      "objective/train/theoretical_loss": 3.375272387278368,
      "objective/train/tokens_used": 797060576,
      "objective/train/unlikelihood_loss": 0.07508569210767746,
      "theoretical_loss": 3.375272387278368,
      "tokens_seen": 2438201344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002657373737373737,
      "loss": 1.2916,
      "theoretical_loss": 3.375272387278368,
      "tokens_seen": 2438201344
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026571717171717174,
      "loss": 1.2794,
      "theoretical_loss": 3.3752651144029366,
      "tokens_seen": 2438266880
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002656969696969697,
      "loss": 1.39,
      "theoretical_loss": 3.3752578417777177,
      "tokens_seen": 2438332416
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002656767676767677,
      "loss": 1.2217,
      "theoretical_loss": 3.375250569402694,
      "tokens_seen": 2438397952
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026565656565656566,
      "loss": 1.5209,
      "theoretical_loss": 3.3752432972778523,
      "tokens_seen": 2438463488
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002656363636363636,
      "loss": 1.3985,
      "theoretical_loss": 3.3752360254031757,
      "tokens_seen": 2438529024
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026561616161616163,
      "loss": 1.5183,
      "theoretical_loss": 3.375228753778649,
      "tokens_seen": 2438594560
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026559595959595964,
      "loss": 1.3592,
      "theoretical_loss": 3.3752214824042572,
      "tokens_seen": 2438660096
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002655757575757576,
      "loss": 1.6963,
      "theoretical_loss": 3.375214211279985,
      "tokens_seen": 2438725632
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026555555555555555,
      "loss": 1.3304,
      "theoretical_loss": 3.3752069404058176,
      "tokens_seen": 2438791168
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002655353535353535,
      "loss": 1.5981,
      "theoretical_loss": 3.375199669781739,
      "tokens_seen": 2438856704
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002655151515151515,
      "loss": 1.4888,
      "theoretical_loss": 3.3751923994077337,
      "tokens_seen": 2438922240
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026549494949494953,
      "loss": 1.493,
      "theoretical_loss": 3.3751851292837864,
      "tokens_seen": 2438987776
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002654747474747475,
      "loss": 1.2447,
      "theoretical_loss": 3.3751778594098822,
      "tokens_seen": 2439053312
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026545454545454545,
      "loss": 1.6882,
      "theoretical_loss": 3.375170589786006,
      "tokens_seen": 2439118848
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002654343434343434,
      "loss": 1.5198,
      "theoretical_loss": 3.375163320412142,
      "tokens_seen": 2439184384
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026541414141414147,
      "loss": 1.3531,
      "theoretical_loss": 3.3751560512882746,
      "tokens_seen": 2439249920
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002653939393939394,
      "loss": 1.3394,
      "theoretical_loss": 3.3751487824143895,
      "tokens_seen": 2439315456
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002653737373737374,
      "loss": 1.4677,
      "theoretical_loss": 3.3751415137904703,
      "tokens_seen": 2439380992
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026535353535353534,
      "loss": 1.3079,
      "theoretical_loss": 3.3751342454165023,
      "tokens_seen": 2439446528
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002653333333333333,
      "loss": 1.3172,
      "theoretical_loss": 3.3751269772924704,
      "tokens_seen": 2439512064
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026531313131313136,
      "loss": 1.324,
      "theoretical_loss": 3.3751197094183585,
      "tokens_seen": 2439577600
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002652929292929293,
      "loss": 1.4946,
      "theoretical_loss": 3.375112441794152,
      "tokens_seen": 2439643136
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002652727272727273,
      "loss": 1.5687,
      "theoretical_loss": 3.3751051744198355,
      "tokens_seen": 2439708672
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00026525252525252523,
      "loss": 1.5891,
      "theoretical_loss": 3.375097907295393,
      "tokens_seen": 2439774208
    },
    {
      "epoch": 0.47,
      "objective/train/aligned_token_count": 0.23955278098583221,
      "objective/train/avg_token_score": 0.044977594166994095,
      "objective/train/docs_used": 1374810,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8685998320579529,
      "objective/train/original_loss": 5.431096076965332,
      "objective/train/theoretical_loss": 3.37509064042081,
      "objective/train/tokens_used": 798698976,
      "objective/train/unlikelihood_loss": 0.03156616911292076,
      "theoretical_loss": 3.37509064042081,
      "tokens_seen": 2439839744
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002652323232323232,
      "loss": 1.2846,
      "theoretical_loss": 3.37509064042081,
      "tokens_seen": 2439839744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026521212121212125,
      "loss": 1.3554,
      "theoretical_loss": 3.375083373796071,
      "tokens_seen": 2439905280
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002651919191919192,
      "loss": 1.5413,
      "theoretical_loss": 3.3750761074211604,
      "tokens_seen": 2439970816
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026517171717171717,
      "loss": 1.3501,
      "theoretical_loss": 3.3750688412960637,
      "tokens_seen": 2440036352
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002651515151515151,
      "loss": 1.388,
      "theoretical_loss": 3.3750615754207645,
      "tokens_seen": 2440101888
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026513131313131314,
      "loss": 1.5143,
      "theoretical_loss": 3.375054309795248,
      "tokens_seen": 2440167424
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026511111111111115,
      "loss": 1.4478,
      "theoretical_loss": 3.375047044419499,
      "tokens_seen": 2440232960
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002650909090909091,
      "loss": 1.3942,
      "theoretical_loss": 3.3750397792935023,
      "tokens_seen": 2440298496
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026507070707070706,
      "loss": 1.3847,
      "theoretical_loss": 3.375032514417242,
      "tokens_seen": 2440364032
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000265050505050505,
      "loss": 1.4802,
      "theoretical_loss": 3.375025249790704,
      "tokens_seen": 2440429568
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026503030303030303,
      "loss": 1.3851,
      "theoretical_loss": 3.375017985413872,
      "tokens_seen": 2440495104
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026501010101010104,
      "loss": 1.2815,
      "theoretical_loss": 3.375010721286731,
      "tokens_seen": 2440560640
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000264989898989899,
      "loss": 1.3419,
      "theoretical_loss": 3.3750034574092656,
      "tokens_seen": 2440626176
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026496969696969695,
      "loss": 1.2877,
      "theoretical_loss": 3.3749961937814605,
      "tokens_seen": 2440691712
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026494949494949496,
      "loss": 1.2352,
      "theoretical_loss": 3.3749889304033007,
      "tokens_seen": 2440757248
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002649292929292929,
      "loss": 1.497,
      "theoretical_loss": 3.3749816672747706,
      "tokens_seen": 2440822784
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026490909090909093,
      "loss": 1.4714,
      "theoretical_loss": 3.3749744043958554,
      "tokens_seen": 2440888320
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002648888888888889,
      "loss": 1.5251,
      "theoretical_loss": 3.374967141766539,
      "tokens_seen": 2440953856
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026486868686868685,
      "loss": 1.4611,
      "theoretical_loss": 3.374959879386807,
      "tokens_seen": 2441019392
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026484848484848486,
      "loss": 1.3546,
      "theoretical_loss": 3.3749526172566435,
      "tokens_seen": 2441084928
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002648282828282828,
      "loss": 1.2905,
      "theoretical_loss": 3.3749453553760334,
      "tokens_seen": 2441150464
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002648080808080808,
      "loss": 1.4797,
      "theoretical_loss": 3.374938093744962,
      "tokens_seen": 2441216000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002647878787878788,
      "loss": 1.5619,
      "theoretical_loss": 3.374930832363413,
      "tokens_seen": 2441281536
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002647676767676768,
      "loss": 1.3283,
      "theoretical_loss": 3.3749235712313714,
      "tokens_seen": 2441347072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026474747474747475,
      "loss": 1.441,
      "theoretical_loss": 3.3749163103488224,
      "tokens_seen": 2441412608
    },
    {
      "epoch": 0.48,
      "objective/train/aligned_token_count": 0.3643695116043091,
      "objective/train/avg_token_score": 0.00988643616437912,
      "objective/train/docs_used": 1375482,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1439762115478516,
      "objective/train/original_loss": 4.684963703155518,
      "objective/train/theoretical_loss": 3.374909049715751,
      "objective/train/tokens_used": 800337376,
      "objective/train/unlikelihood_loss": 0.04493914917111397,
      "theoretical_loss": 3.374909049715751,
      "tokens_seen": 2441478144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002647272727272727,
      "loss": 1.4979,
      "theoretical_loss": 3.374909049715751,
      "tokens_seen": 2441478144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002647070707070707,
      "loss": 1.2366,
      "theoretical_loss": 3.3749017893321405,
      "tokens_seen": 2441543680
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002646868686868687,
      "loss": 1.4891,
      "theoretical_loss": 3.3748945291979773,
      "tokens_seen": 2441609216
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002646666666666667,
      "loss": 1.328,
      "theoretical_loss": 3.374887269313245,
      "tokens_seen": 2441674752
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026464646464646464,
      "loss": 1.3269,
      "theoretical_loss": 3.374880009677929,
      "tokens_seen": 2441740288
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026462626262626265,
      "loss": 1.3155,
      "theoretical_loss": 3.3748727502920133,
      "tokens_seen": 2441805824
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002646060606060606,
      "loss": 1.4488,
      "theoretical_loss": 3.3748654911554836,
      "tokens_seen": 2441871360
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645858585858586,
      "loss": 1.4348,
      "theoretical_loss": 3.374858232268324,
      "tokens_seen": 2441936896
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645656565656566,
      "loss": 1.4197,
      "theoretical_loss": 3.3748509736305192,
      "tokens_seen": 2442002432
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026454545454545453,
      "loss": 1.2735,
      "theoretical_loss": 3.3748437152420543,
      "tokens_seen": 2442067968
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026452525252525255,
      "loss": 1.3401,
      "theoretical_loss": 3.374836457102914,
      "tokens_seen": 2442133504
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002645050505050505,
      "loss": 1.6292,
      "theoretical_loss": 3.3748291992130826,
      "tokens_seen": 2442199040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002644848484848485,
      "loss": 1.5079,
      "theoretical_loss": 3.374821941572545,
      "tokens_seen": 2442264576
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026446464646464647,
      "loss": 1.5798,
      "theoretical_loss": 3.3748146841812865,
      "tokens_seen": 2442330112
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026444444444444443,
      "loss": 1.6818,
      "theoretical_loss": 3.3748074270392916,
      "tokens_seen": 2442395648
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026442424242424244,
      "loss": 1.5543,
      "theoretical_loss": 3.3748001701465444,
      "tokens_seen": 2442461184
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026440404040404045,
      "loss": 1.5054,
      "theoretical_loss": 3.374792913503031,
      "tokens_seen": 2442526720
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002643838383838384,
      "loss": 1.4145,
      "theoretical_loss": 3.3747856571087342,
      "tokens_seen": 2442592256
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026436363636363636,
      "loss": 1.2058,
      "theoretical_loss": 3.3747784009636406,
      "tokens_seen": 2442657792
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002643434343434343,
      "loss": 1.5002,
      "theoretical_loss": 3.3747711450677342,
      "tokens_seen": 2442723328
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026432323232323233,
      "loss": 1.5474,
      "theoretical_loss": 3.3747638894209997,
      "tokens_seen": 2442788864
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026430303030303034,
      "loss": 1.4447,
      "theoretical_loss": 3.374756634023422,
      "tokens_seen": 2442854400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002642828282828283,
      "loss": 1.3354,
      "theoretical_loss": 3.374749378874985,
      "tokens_seen": 2442919936
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026426262626262626,
      "loss": 1.5456,
      "theoretical_loss": 3.3747421239756754,
      "tokens_seen": 2442985472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002642424242424242,
      "loss": 1.3787,
      "theoretical_loss": 3.374734869325476,
      "tokens_seen": 2443051008
    },
    {
      "epoch": 0.48,
      "objective/train/aligned_token_count": 0.3561827838420868,
      "objective/train/avg_token_score": 0.012605557218194008,
      "objective/train/docs_used": 1376516,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.129889965057373,
      "objective/train/original_loss": 4.680431365966797,
      "objective/train/theoretical_loss": 3.374727614924373,
      "objective/train/tokens_used": 801975776,
      "objective/train/unlikelihood_loss": 0.0587725006043911,
      "theoretical_loss": 3.374727614924373,
      "tokens_seen": 2443116544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002642222222222223,
      "loss": 1.3758,
      "theoretical_loss": 3.374727614924373,
      "tokens_seen": 2443116544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026420202020202023,
      "loss": 1.5178,
      "theoretical_loss": 3.3747203607723506,
      "tokens_seen": 2443182080
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002641818181818182,
      "loss": 1.276,
      "theoretical_loss": 3.374713106869393,
      "tokens_seen": 2443247616
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026416161616161615,
      "loss": 1.4022,
      "theoretical_loss": 3.3747058532154854,
      "tokens_seen": 2443313152
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002641414141414141,
      "loss": 1.2668,
      "theoretical_loss": 3.374698599810613,
      "tokens_seen": 2443378688
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026412121212121217,
      "loss": 1.4591,
      "theoretical_loss": 3.3746913466547603,
      "tokens_seen": 2443444224
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026410101010101013,
      "loss": 1.522,
      "theoretical_loss": 3.374684093747912,
      "tokens_seen": 2443509760
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002640808080808081,
      "loss": 1.4481,
      "theoretical_loss": 3.3746768410900527,
      "tokens_seen": 2443575296
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026406060606060604,
      "loss": 1.4744,
      "theoretical_loss": 3.3746695886811677,
      "tokens_seen": 2443640832
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026404040404040405,
      "loss": 1.2837,
      "theoretical_loss": 3.374662336521241,
      "tokens_seen": 2443706368
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026402020202020206,
      "loss": 1.3005,
      "theoretical_loss": 3.374655084610258,
      "tokens_seen": 2443771904
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000264,
      "loss": 1.4709,
      "theoretical_loss": 3.3746478329482033,
      "tokens_seen": 2443837440
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000263979797979798,
      "loss": 1.4059,
      "theoretical_loss": 3.3746405815350613,
      "tokens_seen": 2443902976
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026395959595959593,
      "loss": 1.3564,
      "theoretical_loss": 3.3746333303708176,
      "tokens_seen": 2443968512
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026393939393939395,
      "loss": 1.5956,
      "theoretical_loss": 3.3746260794554566,
      "tokens_seen": 2444034048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026391919191919196,
      "loss": 1.4369,
      "theoretical_loss": 3.3746188287889627,
      "tokens_seen": 2444099584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002638989898989899,
      "loss": 1.6109,
      "theoretical_loss": 3.3746115783713213,
      "tokens_seen": 2444165120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026387878787878787,
      "loss": 1.2371,
      "theoretical_loss": 3.3746043282025164,
      "tokens_seen": 2444230656
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002638585858585858,
      "loss": 1.5408,
      "theoretical_loss": 3.3745970782825334,
      "tokens_seen": 2444296192
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026383838383838384,
      "loss": 1.5291,
      "theoretical_loss": 3.3745898286113576,
      "tokens_seen": 2444361728
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026381818181818185,
      "loss": 1.4975,
      "theoretical_loss": 3.3745825791889725,
      "tokens_seen": 2444427264
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002637979797979798,
      "loss": 1.3777,
      "theoretical_loss": 3.3745753300153636,
      "tokens_seen": 2444492800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026377777777777776,
      "loss": 1.1608,
      "theoretical_loss": 3.3745680810905156,
      "tokens_seen": 2444558336
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002637575757575758,
      "loss": 1.2237,
      "theoretical_loss": 3.3745608324144136,
      "tokens_seen": 2444623872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026373737373737373,
      "loss": 1.395,
      "theoretical_loss": 3.374553583987042,
      "tokens_seen": 2444689408
    },
    {
      "epoch": 0.48,
      "objective/train/aligned_token_count": 0.4204545319080353,
      "objective/train/avg_token_score": 0.0038383633363991976,
      "objective/train/docs_used": 1376516,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3151545524597168,
      "objective/train/original_loss": 3.912013530731201,
      "objective/train/theoretical_loss": 3.374546335808386,
      "objective/train/tokens_used": 803614176,
      "objective/train/unlikelihood_loss": 0.07800620049238205,
      "theoretical_loss": 3.374546335808386,
      "tokens_seen": 2444754944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026371717171717174,
      "loss": 1.3375,
      "theoretical_loss": 3.374546335808386,
      "tokens_seen": 2444754944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002636969696969697,
      "loss": 1.5382,
      "theoretical_loss": 3.3745390878784294,
      "tokens_seen": 2444820480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026367676767676766,
      "loss": 1.5161,
      "theoretical_loss": 3.3745318401971582,
      "tokens_seen": 2444886016
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026365656565656567,
      "loss": 1.4187,
      "theoretical_loss": 3.3745245927645566,
      "tokens_seen": 2444951552
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002636363636363636,
      "loss": 1.2657,
      "theoretical_loss": 3.37451734558061,
      "tokens_seen": 2445017088
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026361616161616163,
      "loss": 1.5759,
      "theoretical_loss": 3.3745100986453025,
      "tokens_seen": 2445082624
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635959595959596,
      "loss": 1.3674,
      "theoretical_loss": 3.374502851958619,
      "tokens_seen": 2445148160
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635757575757576,
      "loss": 1.5289,
      "theoretical_loss": 3.3744956055205444,
      "tokens_seen": 2445213696
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026355555555555556,
      "loss": 1.5036,
      "theoretical_loss": 3.3744883593310635,
      "tokens_seen": 2445279232
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635353535353535,
      "loss": 1.202,
      "theoretical_loss": 3.3744811133901615,
      "tokens_seen": 2445344768
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002635151515151515,
      "loss": 1.2514,
      "theoretical_loss": 3.3744738676978225,
      "tokens_seen": 2445410304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634949494949495,
      "loss": 1.3252,
      "theoretical_loss": 3.374466622254032,
      "tokens_seen": 2445475840
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634747474747475,
      "loss": 1.394,
      "theoretical_loss": 3.3744593770587743,
      "tokens_seen": 2445541376
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026345454545454545,
      "loss": 1.2229,
      "theoretical_loss": 3.3744521321120344,
      "tokens_seen": 2445606912
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634343434343434,
      "loss": 1.4442,
      "theoretical_loss": 3.374444887413797,
      "tokens_seen": 2445672448
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002634141414141414,
      "loss": 1.4369,
      "theoretical_loss": 3.3744376429640477,
      "tokens_seen": 2445737984
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026339393939393943,
      "loss": 1.4765,
      "theoretical_loss": 3.37443039876277,
      "tokens_seen": 2445803520
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002633737373737374,
      "loss": 1.675,
      "theoretical_loss": 3.3744231548099495,
      "tokens_seen": 2445869056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026335353535353534,
      "loss": 1.4516,
      "theoretical_loss": 3.3744159111055714,
      "tokens_seen": 2445934592
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002633333333333333,
      "loss": 1.5569,
      "theoretical_loss": 3.3744086676496194,
      "tokens_seen": 2446000128
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002633131313131313,
      "loss": 1.4293,
      "theoretical_loss": 3.3744014244420795,
      "tokens_seen": 2446065664
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002632929292929293,
      "loss": 1.3152,
      "theoretical_loss": 3.3743941814829355,
      "tokens_seen": 2446131200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002632727272727273,
      "loss": 1.3338,
      "theoretical_loss": 3.374386938772173,
      "tokens_seen": 2446196736
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026325252525252524,
      "loss": 1.6743,
      "theoretical_loss": 3.3743796963097763,
      "tokens_seen": 2446262272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002632323232323232,
      "loss": 1.3675,
      "theoretical_loss": 3.374372454095731,
      "tokens_seen": 2446327808
    },
    {
      "epoch": 0.48,
      "objective/train/aligned_token_count": 0.3816593289375305,
      "objective/train/avg_token_score": 0.07048311084508896,
      "objective/train/docs_used": 1377582,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1800470352172852,
      "objective/train/original_loss": 4.771828651428223,
      "objective/train/theoretical_loss": 3.3743652121300207,
      "objective/train/tokens_used": 805252576,
      "objective/train/unlikelihood_loss": 0.05653870478272438,
      "theoretical_loss": 3.3743652121300207,
      "tokens_seen": 2446393344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026321212121212126,
      "loss": 1.1173,
      "theoretical_loss": 3.3743652121300207,
      "tokens_seen": 2446393344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631919191919192,
      "loss": 1.5351,
      "theoretical_loss": 3.3743579704126314,
      "tokens_seen": 2446458880
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631717171717172,
      "loss": 1.3961,
      "theoretical_loss": 3.374350728943547,
      "tokens_seen": 2446524416
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026315151515151513,
      "loss": 1.2661,
      "theoretical_loss": 3.3743434877227534,
      "tokens_seen": 2446589952
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002631313131313131,
      "loss": 1.4444,
      "theoretical_loss": 3.3743362467502345,
      "tokens_seen": 2446655488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026311111111111115,
      "loss": 1.3053,
      "theoretical_loss": 3.3743290060259756,
      "tokens_seen": 2446721024
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002630909090909091,
      "loss": 1.4658,
      "theoretical_loss": 3.3743217655499613,
      "tokens_seen": 2446786560
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026307070707070707,
      "loss": 1.5152,
      "theoretical_loss": 3.374314525322177,
      "tokens_seen": 2446852096
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000263050505050505,
      "loss": 1.607,
      "theoretical_loss": 3.3743072853426064,
      "tokens_seen": 2446917632
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026303030303030303,
      "loss": 1.5177,
      "theoretical_loss": 3.3743000456112355,
      "tokens_seen": 2446983168
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026301010101010104,
      "loss": 1.6303,
      "theoretical_loss": 3.3742928061280484,
      "tokens_seen": 2447048704
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262989898989899,
      "loss": 1.6065,
      "theoretical_loss": 3.3742855668930307,
      "tokens_seen": 2447114240
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026296969696969696,
      "loss": 1.4341,
      "theoretical_loss": 3.374278327906166,
      "tokens_seen": 2447179776
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002629494949494949,
      "loss": 1.5214,
      "theoretical_loss": 3.374271089167441,
      "tokens_seen": 2447245312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262929292929293,
      "loss": 1.3223,
      "theoretical_loss": 3.3742638506768383,
      "tokens_seen": 2447310848
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026290909090909094,
      "loss": 1.5969,
      "theoretical_loss": 3.3742566124343445,
      "tokens_seen": 2447376384
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002628888888888889,
      "loss": 1.5297,
      "theoretical_loss": 3.3742493744399438,
      "tokens_seen": 2447441920
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026286868686868685,
      "loss": 1.4838,
      "theoretical_loss": 3.3742421366936215,
      "tokens_seen": 2447507456
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026284848484848486,
      "loss": 1.4565,
      "theoretical_loss": 3.3742348991953617,
      "tokens_seen": 2447572992
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002628282828282829,
      "loss": 1.4441,
      "theoretical_loss": 3.3742276619451497,
      "tokens_seen": 2447638528
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026280808080808083,
      "loss": 1.6128,
      "theoretical_loss": 3.3742204249429704,
      "tokens_seen": 2447704064
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002627878787878788,
      "loss": 1.4671,
      "theoretical_loss": 3.374213188188808,
      "tokens_seen": 2447769600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026276767676767674,
      "loss": 1.3078,
      "theoretical_loss": 3.374205951682649,
      "tokens_seen": 2447835136
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026274747474747475,
      "loss": 1.5491,
      "theoretical_loss": 3.3741987154244764,
      "tokens_seen": 2447900672
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026272727272727277,
      "loss": 1.3998,
      "theoretical_loss": 3.374191479414276,
      "tokens_seen": 2447966208
    },
    {
      "epoch": 0.48,
      "objective/train/aligned_token_count": 0.3072458505630493,
      "objective/train/avg_token_score": 0.006326456554234028,
      "objective/train/docs_used": 1378304,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9334326386451721,
      "objective/train/original_loss": 4.613555431365967,
      "objective/train/theoretical_loss": 3.3741842436520324,
      "objective/train/tokens_used": 806890976,
      "objective/train/unlikelihood_loss": 0.06921292841434479,
      "theoretical_loss": 3.3741842436520324,
      "tokens_seen": 2448031744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002627070707070707,
      "loss": 1.4832,
      "theoretical_loss": 3.3741842436520324,
      "tokens_seen": 2448031744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002626868686868687,
      "loss": 1.292,
      "theoretical_loss": 3.374177008137731,
      "tokens_seen": 2448097280
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002626666666666667,
      "loss": 1.4735,
      "theoretical_loss": 3.3741697728713556,
      "tokens_seen": 2448162816
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026264646464646465,
      "loss": 1.1619,
      "theoretical_loss": 3.374162537852892,
      "tokens_seen": 2448228352
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026262626262626266,
      "loss": 1.3301,
      "theoretical_loss": 3.374155303082325,
      "tokens_seen": 2448293888
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002626060606060606,
      "loss": 1.3671,
      "theoretical_loss": 3.374148068559639,
      "tokens_seen": 2448359424
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026258585858585857,
      "loss": 1.3472,
      "theoretical_loss": 3.374140834284819,
      "tokens_seen": 2448424960
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002625656565656566,
      "loss": 1.291,
      "theoretical_loss": 3.37413360025785,
      "tokens_seen": 2448490496
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026254545454545454,
      "loss": 1.3715,
      "theoretical_loss": 3.3741263664787176,
      "tokens_seen": 2448556032
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026252525252525255,
      "loss": 1.388,
      "theoretical_loss": 3.3741191329474054,
      "tokens_seen": 2448621568
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002625050505050505,
      "loss": 1.4565,
      "theoretical_loss": 3.3741118996638986,
      "tokens_seen": 2448687104
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002624848484848485,
      "loss": 1.2732,
      "theoretical_loss": 3.3741046666281824,
      "tokens_seen": 2448752640
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002624646464646465,
      "loss": 1.4121,
      "theoretical_loss": 3.374097433840242,
      "tokens_seen": 2448818176
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026244444444444443,
      "loss": 1.4158,
      "theoretical_loss": 3.3740902013000613,
      "tokens_seen": 2448883712
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026242424242424244,
      "loss": 1.2275,
      "theoretical_loss": 3.374082969007626,
      "tokens_seen": 2448949248
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002624040404040404,
      "loss": 1.5712,
      "theoretical_loss": 3.374075736962921,
      "tokens_seen": 2449014784
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002623838383838384,
      "loss": 1.4462,
      "theoretical_loss": 3.3740685051659307,
      "tokens_seen": 2449080320
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026236363636363637,
      "loss": 1.5346,
      "theoretical_loss": 3.3740612736166398,
      "tokens_seen": 2449145856
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002623434343434343,
      "loss": 1.5302,
      "theoretical_loss": 3.3740540423150343,
      "tokens_seen": 2449211392
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026232323232323234,
      "loss": 1.1535,
      "theoretical_loss": 3.3740468112610977,
      "tokens_seen": 2449276928
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026230303030303035,
      "loss": 1.2492,
      "theoretical_loss": 3.374039580454816,
      "tokens_seen": 2449342464
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002622828282828283,
      "loss": 1.4082,
      "theoretical_loss": 3.374032349896174,
      "tokens_seen": 2449408000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026226262626262626,
      "loss": 1.5108,
      "theoretical_loss": 3.3740251195851556,
      "tokens_seen": 2449473536
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002622424242424242,
      "loss": 1.5555,
      "theoretical_loss": 3.3740178895217463,
      "tokens_seen": 2449539072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026222222222222223,
      "loss": 1.471,
      "theoretical_loss": 3.3740106597059314,
      "tokens_seen": 2449604608
    },
    {
      "epoch": 0.48,
      "objective/train/aligned_token_count": 0.48332110047340393,
      "objective/train/avg_token_score": 0.011905195191502571,
      "objective/train/docs_used": 1378955,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.505688190460205,
      "objective/train/original_loss": 3.748351573944092,
      "objective/train/theoretical_loss": 3.3740034301376953,
      "objective/train/tokens_used": 808529376,
      "objective/train/unlikelihood_loss": 0.07834320515394211,
      "theoretical_loss": 3.3740034301376953,
      "tokens_seen": 2449670144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026220202020202024,
      "loss": 1.43,
      "theoretical_loss": 3.3740034301376953,
      "tokens_seen": 2449670144
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002621818181818182,
      "loss": 1.379,
      "theoretical_loss": 3.3739962008170234,
      "tokens_seen": 2449735680
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026216161616161615,
      "loss": 1.5039,
      "theoretical_loss": 3.3739889717439,
      "tokens_seen": 2449801216
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002621414141414141,
      "loss": 1.4276,
      "theoretical_loss": 3.37398174291831,
      "tokens_seen": 2449866752
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002621212121212122,
      "loss": 1.346,
      "theoretical_loss": 3.3739745143402384,
      "tokens_seen": 2449932288
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026210101010101013,
      "loss": 1.5119,
      "theoretical_loss": 3.373967286009671,
      "tokens_seen": 2449997824
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002620808080808081,
      "loss": 1.4926,
      "theoretical_loss": 3.3739600579265914,
      "tokens_seen": 2450063360
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026206060606060605,
      "loss": 1.23,
      "theoretical_loss": 3.3739528300909853,
      "tokens_seen": 2450128896
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262040404040404,
      "loss": 1.452,
      "theoretical_loss": 3.3739456025028374,
      "tokens_seen": 2450194432
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026202020202020207,
      "loss": 1.3912,
      "theoretical_loss": 3.3739383751621324,
      "tokens_seen": 2450259968
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000262,
      "loss": 1.372,
      "theoretical_loss": 3.3739311480688556,
      "tokens_seen": 2450325504
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000261979797979798,
      "loss": 1.1734,
      "theoretical_loss": 3.3739239212229912,
      "tokens_seen": 2450391040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026195959595959594,
      "loss": 1.3622,
      "theoretical_loss": 3.373916694624525,
      "tokens_seen": 2450456576
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002619393939393939,
      "loss": 1.2582,
      "theoretical_loss": 3.3739094682734416,
      "tokens_seen": 2450522112
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026191919191919196,
      "loss": 1.4113,
      "theoretical_loss": 3.373902242169726,
      "tokens_seen": 2450587648
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618989898989899,
      "loss": 1.2586,
      "theoretical_loss": 3.3738950163133623,
      "tokens_seen": 2450653184
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002618787878787879,
      "loss": 1.472,
      "theoretical_loss": 3.3738877907043365,
      "tokens_seen": 2450718720
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026185858585858583,
      "loss": 1.3455,
      "theoretical_loss": 3.3738805653426334,
      "tokens_seen": 2450784256
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026183838383838384,
      "loss": 1.3272,
      "theoretical_loss": 3.3738733402282373,
      "tokens_seen": 2450849792
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026181818181818185,
      "loss": 1.4178,
      "theoretical_loss": 3.3738661153611336,
      "tokens_seen": 2450915328
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002617979797979798,
      "loss": 1.39,
      "theoretical_loss": 3.3738588907413067,
      "tokens_seen": 2450980864
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026177777777777777,
      "loss": 1.3406,
      "theoretical_loss": 3.373851666368742,
      "tokens_seen": 2451046400
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002617575757575757,
      "loss": 1.5751,
      "theoretical_loss": 3.373844442243425,
      "tokens_seen": 2451111936
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026173737373737374,
      "loss": 1.4922,
      "theoretical_loss": 3.373837218365339,
      "tokens_seen": 2451177472
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026171717171717175,
      "loss": 1.3852,
      "theoretical_loss": 3.3738299947344705,
      "tokens_seen": 2451243008
    },
    {
      "epoch": 0.48,
      "objective/train/aligned_token_count": 0.3926564157009125,
      "objective/train/avg_token_score": 0.007126107346266508,
      "objective/train/docs_used": 1380357,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2451292276382446,
      "objective/train/original_loss": 4.954854488372803,
      "objective/train/theoretical_loss": 3.3738227713508038,
      "objective/train/tokens_used": 810167776,
      "objective/train/unlikelihood_loss": 0.04100557789206505,
      "theoretical_loss": 3.3738227713508038,
      "tokens_seen": 2451308544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002616969696969697,
      "loss": 1.2577,
      "theoretical_loss": 3.3738227713508038,
      "tokens_seen": 2451308544
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026167676767676766,
      "loss": 1.3339,
      "theoretical_loss": 3.3738155482143233,
      "tokens_seen": 2451374080
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026165656565656567,
      "loss": 1.6091,
      "theoretical_loss": 3.373808325325015,
      "tokens_seen": 2451439616
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026163636363636363,
      "loss": 1.285,
      "theoretical_loss": 3.3738011026828634,
      "tokens_seen": 2451505152
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026161616161616164,
      "loss": 1.3214,
      "theoretical_loss": 3.373793880287853,
      "tokens_seen": 2451570688
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002615959595959596,
      "loss": 1.3332,
      "theoretical_loss": 3.373786658139969,
      "tokens_seen": 2451636224
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026157575757575755,
      "loss": 1.5666,
      "theoretical_loss": 3.3737794362391966,
      "tokens_seen": 2451701760
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026155555555555556,
      "loss": 1.5484,
      "theoretical_loss": 3.3737722145855207,
      "tokens_seen": 2451767296
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002615353535353535,
      "loss": 1.2139,
      "theoretical_loss": 3.3737649931789258,
      "tokens_seen": 2451832832
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026151515151515153,
      "loss": 1.5029,
      "theoretical_loss": 3.373757772019397,
      "tokens_seen": 2451898368
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002614949494949495,
      "loss": 1.463,
      "theoretical_loss": 3.37375055110692,
      "tokens_seen": 2451963904
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002614747474747475,
      "loss": 1.3156,
      "theoretical_loss": 3.373743330441479,
      "tokens_seen": 2452029440
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026145454545454546,
      "loss": 1.5242,
      "theoretical_loss": 3.3737361100230587,
      "tokens_seen": 2452094976
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002614343434343434,
      "loss": 1.3212,
      "theoretical_loss": 3.3737288898516447,
      "tokens_seen": 2452160512
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002614141414141414,
      "loss": 1.7635,
      "theoretical_loss": 3.3737216699272214,
      "tokens_seen": 2452226048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613939393939394,
      "loss": 1.2969,
      "theoretical_loss": 3.3737144502497745,
      "tokens_seen": 2452291584
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613737373737374,
      "loss": 1.4482,
      "theoretical_loss": 3.373707230819288,
      "tokens_seen": 2452357120
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026135353535353535,
      "loss": 1.2466,
      "theoretical_loss": 3.3737000116357474,
      "tokens_seen": 2452422656
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613333333333333,
      "loss": 1.671,
      "theoretical_loss": 3.3736927926991376,
      "tokens_seen": 2452488192
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002613131313131313,
      "loss": 1.2645,
      "theoretical_loss": 3.373685574009444,
      "tokens_seen": 2452553728
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026129292929292933,
      "loss": 1.5853,
      "theoretical_loss": 3.3736783555666507,
      "tokens_seen": 2452619264
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002612727272727273,
      "loss": 1.1897,
      "theoretical_loss": 3.373671137370743,
      "tokens_seen": 2452684800
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026125252525252524,
      "loss": 1.4363,
      "theoretical_loss": 3.373663919421706,
      "tokens_seen": 2452750336
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002612323232323232,
      "loss": 1.4949,
      "theoretical_loss": 3.3736567017195247,
      "tokens_seen": 2452815872
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002612121212121212,
      "loss": 1.32,
      "theoretical_loss": 3.3736494842641838,
      "tokens_seen": 2452881408
    },
    {
      "epoch": 0.48,
      "objective/train/aligned_token_count": 0.48625367879867554,
      "objective/train/avg_token_score": 0.008265034295618534,
      "objective/train/docs_used": 1380887,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3857629299163818,
      "objective/train/original_loss": 4.038858413696289,
      "objective/train/theoretical_loss": 3.3736422670556685,
      "objective/train/tokens_used": 811806176,
      "objective/train/unlikelihood_loss": 0.06039511412382126,
      "theoretical_loss": 3.3736422670556685,
      "tokens_seen": 2452946944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002611919191919192,
      "loss": 1.3547,
      "theoretical_loss": 3.3736422670556685,
      "tokens_seen": 2452946944
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002611717171717172,
      "loss": 1.3348,
      "theoretical_loss": 3.373635050093964,
      "tokens_seen": 2453012480
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026115151515151514,
      "loss": 1.2655,
      "theoretical_loss": 3.373627833379054,
      "tokens_seen": 2453078016
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026113131313131315,
      "loss": 1.4173,
      "theoretical_loss": 3.373620616910925,
      "tokens_seen": 2453143552
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026111111111111116,
      "loss": 1.4813,
      "theoretical_loss": 3.3736134006895613,
      "tokens_seen": 2453209088
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002610909090909091,
      "loss": 1.4117,
      "theoretical_loss": 3.373606184714948,
      "tokens_seen": 2453274624
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026107070707070707,
      "loss": 1.4735,
      "theoretical_loss": 3.37359896898707,
      "tokens_seen": 2453340160
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026105050505050503,
      "loss": 1.4336,
      "theoretical_loss": 3.373591753505912,
      "tokens_seen": 2453405696
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026103030303030304,
      "loss": 1.5899,
      "theoretical_loss": 3.3735845382714595,
      "tokens_seen": 2453471232
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026101010101010105,
      "loss": 1.4258,
      "theoretical_loss": 3.373577323283697,
      "tokens_seen": 2453536768
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000260989898989899,
      "loss": 1.4302,
      "theoretical_loss": 3.3735701085426104,
      "tokens_seen": 2453602304
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026096969696969696,
      "loss": 1.3141,
      "theoretical_loss": 3.3735628940481837,
      "tokens_seen": 2453667840
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002609494949494949,
      "loss": 1.5696,
      "theoretical_loss": 3.3735556798004014,
      "tokens_seen": 2453733376
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.000260929292929293,
      "loss": 1.0154,
      "theoretical_loss": 3.3735484657992503,
      "tokens_seen": 2453798912
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026090909090909094,
      "loss": 1.6383,
      "theoretical_loss": 3.3735412520447134,
      "tokens_seen": 2453864448
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608888888888889,
      "loss": 1.4743,
      "theoretical_loss": 3.373534038536777,
      "tokens_seen": 2453929984
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026086868686868686,
      "loss": 1.3615,
      "theoretical_loss": 3.3735268252754262,
      "tokens_seen": 2453995520
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608484848484848,
      "loss": 1.5988,
      "theoretical_loss": 3.373519612260645,
      "tokens_seen": 2454061056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002608282828282829,
      "loss": 1.3725,
      "theoretical_loss": 3.373512399492419,
      "tokens_seen": 2454126592
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026080808080808084,
      "loss": 1.4133,
      "theoretical_loss": 3.373505186970733,
      "tokens_seen": 2454192128
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002607878787878788,
      "loss": 1.523,
      "theoretical_loss": 3.3734979746955718,
      "tokens_seen": 2454257664
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026076767676767675,
      "loss": 1.4383,
      "theoretical_loss": 3.373490762666921,
      "tokens_seen": 2454323200
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026074747474747476,
      "loss": 1.5212,
      "theoretical_loss": 3.373483550884765,
      "tokens_seen": 2454388736
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026072727272727277,
      "loss": 1.3598,
      "theoretical_loss": 3.3734763393490894,
      "tokens_seen": 2454454272
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026070707070707073,
      "loss": 1.4001,
      "theoretical_loss": 3.3734691280598783,
      "tokens_seen": 2454519808
    },
    {
      "epoch": 0.48,
      "objective/train/aligned_token_count": 0.32032012939453125,
      "objective/train/avg_token_score": 0.012385210022330284,
      "objective/train/docs_used": 1382209,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0630464553833008,
      "objective/train/original_loss": 4.590609550476074,
      "objective/train/theoretical_loss": 3.3734619170171176,
      "objective/train/tokens_used": 813444576,
      "objective/train/unlikelihood_loss": 0.07037441432476044,
      "theoretical_loss": 3.3734619170171176,
      "tokens_seen": 2454585344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002606868686868687,
      "loss": 1.5501,
      "theoretical_loss": 3.3734619170171176,
      "tokens_seen": 2454585344
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026066666666666664,
      "loss": 1.365,
      "theoretical_loss": 3.373454706220792,
      "tokens_seen": 2454650880
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026064646464646465,
      "loss": 1.3438,
      "theoretical_loss": 3.3734474956708858,
      "tokens_seen": 2454716416
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026062626262626266,
      "loss": 1.2973,
      "theoretical_loss": 3.373440285367385,
      "tokens_seen": 2454781952
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002606060606060606,
      "loss": 1.448,
      "theoretical_loss": 3.3734330753102744,
      "tokens_seen": 2454847488
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002605858585858586,
      "loss": 1.4042,
      "theoretical_loss": 3.3734258654995384,
      "tokens_seen": 2454913024
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026056565656565653,
      "loss": 1.3402,
      "theoretical_loss": 3.3734186559351627,
      "tokens_seen": 2454978560
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026054545454545455,
      "loss": 1.523,
      "theoretical_loss": 3.373411446617132,
      "tokens_seen": 2455044096
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026052525252525256,
      "loss": 1.2798,
      "theoretical_loss": 3.373404237545431,
      "tokens_seen": 2455109632
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002605050505050505,
      "loss": 1.5493,
      "theoretical_loss": 3.3733970287200457,
      "tokens_seen": 2455175168
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026048484848484847,
      "loss": 1.6334,
      "theoretical_loss": 3.37338982014096,
      "tokens_seen": 2455240704
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002604646464646465,
      "loss": 1.4967,
      "theoretical_loss": 3.3733826118081596,
      "tokens_seen": 2455306240
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026044444444444444,
      "loss": 1.2798,
      "theoretical_loss": 3.3733754037216293,
      "tokens_seen": 2455371776
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026042424242424245,
      "loss": 1.3696,
      "theoretical_loss": 3.373368195881354,
      "tokens_seen": 2455437312
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002604040404040404,
      "loss": 1.3142,
      "theoretical_loss": 3.3733609882873186,
      "tokens_seen": 2455502848
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026038383838383836,
      "loss": 1.3637,
      "theoretical_loss": 3.3733537809395084,
      "tokens_seen": 2455568384
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002603636363636364,
      "loss": 1.2552,
      "theoretical_loss": 3.3733465738379085,
      "tokens_seen": 2455633920
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026034343434343433,
      "loss": 1.4304,
      "theoretical_loss": 3.3733393669825036,
      "tokens_seen": 2455699456
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026032323232323234,
      "loss": 1.4965,
      "theoretical_loss": 3.3733321603732787,
      "tokens_seen": 2455764992
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002603030303030303,
      "loss": 1.3731,
      "theoretical_loss": 3.3733249540102195,
      "tokens_seen": 2455830528
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602828282828283,
      "loss": 1.4535,
      "theoretical_loss": 3.37331774789331,
      "tokens_seen": 2455896064
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026026262626262627,
      "loss": 1.4218,
      "theoretical_loss": 3.373310542022536,
      "tokens_seen": 2455961600
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602424242424242,
      "loss": 1.4328,
      "theoretical_loss": 3.3733033363978824,
      "tokens_seen": 2456027136
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00026022222222222223,
      "loss": 1.281,
      "theoretical_loss": 3.373296131019334,
      "tokens_seen": 2456092672
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002602020202020202,
      "loss": 1.2733,
      "theoretical_loss": 3.3732889258868757,
      "tokens_seen": 2456158208
    },
    {
      "epoch": 0.48,
      "objective/train/aligned_token_count": 0.4045698940753937,
      "objective/train/avg_token_score": 0.023738790303468704,
      "objective/train/docs_used": 1382688,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5022037029266357,
      "objective/train/original_loss": 4.592807292938232,
      "objective/train/theoretical_loss": 3.373281721000493,
      "objective/train/tokens_used": 815082976,
      "objective/train/unlikelihood_loss": 0.0653972402215004,
      "theoretical_loss": 3.373281721000493,
      "tokens_seen": 2456223744
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002601818181818182,
      "loss": 1.4543,
      "theoretical_loss": 3.373281721000493,
      "tokens_seen": 2456223744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026016161616161616,
      "loss": 1.3891,
      "theoretical_loss": 3.3732745163601705,
      "tokens_seen": 2456289280
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002601414141414141,
      "loss": 1.5424,
      "theoretical_loss": 3.373267311965894,
      "tokens_seen": 2456354816
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026012121212121213,
      "loss": 1.3051,
      "theoretical_loss": 3.373260107817647,
      "tokens_seen": 2456420352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026010101010101014,
      "loss": 1.3983,
      "theoretical_loss": 3.373252903915416,
      "tokens_seen": 2456485888
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002600808080808081,
      "loss": 1.5013,
      "theoretical_loss": 3.3732457002591856,
      "tokens_seen": 2456551424
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026006060606060605,
      "loss": 1.3486,
      "theoretical_loss": 3.3732384968489404,
      "tokens_seen": 2456616960
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000260040404040404,
      "loss": 1.3055,
      "theoretical_loss": 3.3732312936846665,
      "tokens_seen": 2456682496
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000260020202020202,
      "loss": 1.6563,
      "theoretical_loss": 3.3732240907663473,
      "tokens_seen": 2456748032
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00026000000000000003,
      "loss": 1.3382,
      "theoretical_loss": 3.3732168880939692,
      "tokens_seen": 2456813568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000259979797979798,
      "loss": 1.3816,
      "theoretical_loss": 3.373209685667517,
      "tokens_seen": 2456879104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025995959595959595,
      "loss": 1.3141,
      "theoretical_loss": 3.3732024834869754,
      "tokens_seen": 2456944640
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002599393939393939,
      "loss": 1.3117,
      "theoretical_loss": 3.37319528155233,
      "tokens_seen": 2457010176
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025991919191919197,
      "loss": 1.4567,
      "theoretical_loss": 3.3731880798635645,
      "tokens_seen": 2457075712
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002598989898989899,
      "loss": 1.4919,
      "theoretical_loss": 3.3731808784206656,
      "tokens_seen": 2457141248
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002598787878787879,
      "loss": 1.5879,
      "theoretical_loss": 3.3731736772236176,
      "tokens_seen": 2457206784
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025985858585858584,
      "loss": 1.4237,
      "theoretical_loss": 3.3731664762724054,
      "tokens_seen": 2457272320
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002598383838383838,
      "loss": 1.5234,
      "theoretical_loss": 3.3731592755670143,
      "tokens_seen": 2457337856
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025981818181818186,
      "loss": 1.558,
      "theoretical_loss": 3.3731520751074293,
      "tokens_seen": 2457403392
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002597979797979798,
      "loss": 1.1901,
      "theoretical_loss": 3.3731448748936357,
      "tokens_seen": 2457468928
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002597777777777778,
      "loss": 1.4367,
      "theoretical_loss": 3.3731376749256183,
      "tokens_seen": 2457534464
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025975757575757573,
      "loss": 1.2589,
      "theoretical_loss": 3.3731304752033617,
      "tokens_seen": 2457600000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025973737373737374,
      "loss": 1.293,
      "theoretical_loss": 3.3731232757268517,
      "tokens_seen": 2457665536
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025971717171717175,
      "loss": 1.3418,
      "theoretical_loss": 3.373116076496073,
      "tokens_seen": 2457731072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002596969696969697,
      "loss": 1.5131,
      "theoretical_loss": 3.373108877511011,
      "tokens_seen": 2457796608
    },
    {
      "epoch": 0.49,
      "objective/train/aligned_token_count": 0.46609237790107727,
      "objective/train/avg_token_score": 0.019508633762598038,
      "objective/train/docs_used": 1383732,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3144094944000244,
      "objective/train/original_loss": 3.8763251304626465,
      "objective/train/theoretical_loss": 3.3731016787716506,
      "objective/train/tokens_used": 816721376,
      "objective/train/unlikelihood_loss": 0.08185066282749176,
      "theoretical_loss": 3.3731016787716506,
      "tokens_seen": 2457862144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025967676767676767,
      "loss": 1.49,
      "theoretical_loss": 3.3731016787716506,
      "tokens_seen": 2457862144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002596565656565656,
      "loss": 1.5375,
      "theoretical_loss": 3.3730944802779765,
      "tokens_seen": 2457927680
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025963636363636363,
      "loss": 1.4606,
      "theoretical_loss": 3.373087282029974,
      "tokens_seen": 2457993216
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025961616161616165,
      "loss": 1.2722,
      "theoretical_loss": 3.3730800840276283,
      "tokens_seen": 2458058752
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002595959595959596,
      "loss": 1.3971,
      "theoretical_loss": 3.3730728862709243,
      "tokens_seen": 2458124288
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025957575757575756,
      "loss": 1.4578,
      "theoretical_loss": 3.373065688759847,
      "tokens_seen": 2458189824
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025955555555555557,
      "loss": 1.3365,
      "theoretical_loss": 3.373058491494382,
      "tokens_seen": 2458255360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002595353535353535,
      "loss": 1.2273,
      "theoretical_loss": 3.3730512944745135,
      "tokens_seen": 2458320896
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025951515151515154,
      "loss": 1.3944,
      "theoretical_loss": 3.3730440977002276,
      "tokens_seen": 2458386432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002594949494949495,
      "loss": 1.4384,
      "theoretical_loss": 3.3730369011715085,
      "tokens_seen": 2458451968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025947474747474745,
      "loss": 1.2571,
      "theoretical_loss": 3.3730297048883413,
      "tokens_seen": 2458517504
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025945454545454546,
      "loss": 1.4183,
      "theoretical_loss": 3.373022508850712,
      "tokens_seen": 2458583040
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002594343434343435,
      "loss": 1.3415,
      "theoretical_loss": 3.3730153130586045,
      "tokens_seen": 2458648576
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025941414141414143,
      "loss": 1.3365,
      "theoretical_loss": 3.373008117512005,
      "tokens_seen": 2458714112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002593939393939394,
      "loss": 1.5258,
      "theoretical_loss": 3.373000922210897,
      "tokens_seen": 2458779648
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002593737373737374,
      "loss": 1.4641,
      "theoretical_loss": 3.3729937271552677,
      "tokens_seen": 2458845184
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025935353535353536,
      "loss": 1.2659,
      "theoretical_loss": 3.3729865323451005,
      "tokens_seen": 2458910720
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025933333333333337,
      "loss": 1.5396,
      "theoretical_loss": 3.3729793377803814,
      "tokens_seen": 2458976256
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002593131313131313,
      "loss": 1.3241,
      "theoretical_loss": 3.372972143461095,
      "tokens_seen": 2459041792
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002592929292929293,
      "loss": 1.1599,
      "theoretical_loss": 3.3729649493872262,
      "tokens_seen": 2459107328
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002592727272727273,
      "loss": 1.2394,
      "theoretical_loss": 3.372957755558761,
      "tokens_seen": 2459172864
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025925252525252525,
      "loss": 1.4391,
      "theoretical_loss": 3.3729505619756837,
      "tokens_seen": 2459238400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025923232323232326,
      "loss": 1.5163,
      "theoretical_loss": 3.372943368637979,
      "tokens_seen": 2459303936
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002592121212121212,
      "loss": 1.3852,
      "theoretical_loss": 3.3729361755456333,
      "tokens_seen": 2459369472
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025919191919191923,
      "loss": 1.3292,
      "theoretical_loss": 3.372928982698631,
      "tokens_seen": 2459435008
    },
    {
      "epoch": 0.49,
      "objective/train/aligned_token_count": 0.3010141849517822,
      "objective/train/avg_token_score": 0.02664129249751568,
      "objective/train/docs_used": 1384412,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0720702409744263,
      "objective/train/original_loss": 4.936168193817139,
      "objective/train/theoretical_loss": 3.372921790096957,
      "objective/train/tokens_used": 818359776,
      "objective/train/unlikelihood_loss": 0.042066656053066254,
      "theoretical_loss": 3.372921790096957,
      "tokens_seen": 2459500544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002591717171717172,
      "loss": 1.3657,
      "theoretical_loss": 3.372921790096957,
      "tokens_seen": 2459500544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025915151515151514,
      "loss": 1.3255,
      "theoretical_loss": 3.3729145977405963,
      "tokens_seen": 2459566080
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025913131313131315,
      "loss": 1.4326,
      "theoretical_loss": 3.372907405629535,
      "tokens_seen": 2459631616
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002591111111111111,
      "loss": 1.5671,
      "theoretical_loss": 3.3729002137637565,
      "tokens_seen": 2459697152
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002590909090909091,
      "loss": 1.315,
      "theoretical_loss": 3.3728930221432476,
      "tokens_seen": 2459762688
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002590707070707071,
      "loss": 1.4237,
      "theoretical_loss": 3.3728858307679923,
      "tokens_seen": 2459828224
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025905050505050503,
      "loss": 1.4501,
      "theoretical_loss": 3.3728786396379764,
      "tokens_seen": 2459893760
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025903030303030304,
      "loss": 1.4712,
      "theoretical_loss": 3.3728714487531843,
      "tokens_seen": 2459959296
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025901010101010106,
      "loss": 1.6495,
      "theoretical_loss": 3.372864258113602,
      "tokens_seen": 2460024832
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000258989898989899,
      "loss": 1.3786,
      "theoretical_loss": 3.3728570677192136,
      "tokens_seen": 2460090368
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025896969696969697,
      "loss": 1.2898,
      "theoretical_loss": 3.3728498775700047,
      "tokens_seen": 2460155904
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002589494949494949,
      "loss": 1.6559,
      "theoretical_loss": 3.3728426876659605,
      "tokens_seen": 2460221440
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025892929292929294,
      "loss": 1.5806,
      "theoretical_loss": 3.372835498007066,
      "tokens_seen": 2460286976
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025890909090909095,
      "loss": 1.1456,
      "theoretical_loss": 3.3728283085933066,
      "tokens_seen": 2460352512
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002588888888888889,
      "loss": 1.2652,
      "theoretical_loss": 3.3728211194246667,
      "tokens_seen": 2460418048
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025886868686868686,
      "loss": 1.2788,
      "theoretical_loss": 3.3728139305011324,
      "tokens_seen": 2460483584
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002588484848484848,
      "loss": 1.4074,
      "theoretical_loss": 3.3728067418226875,
      "tokens_seen": 2460549120
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025882828282828283,
      "loss": 1.3893,
      "theoretical_loss": 3.372799553389319,
      "tokens_seen": 2460614656
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025880808080808084,
      "loss": 1.3848,
      "theoretical_loss": 3.37279236520101,
      "tokens_seen": 2460680192
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587878787878788,
      "loss": 1.2345,
      "theoretical_loss": 3.3727851772577466,
      "tokens_seen": 2460745728
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025876767676767675,
      "loss": 1.2816,
      "theoretical_loss": 3.372777989559514,
      "tokens_seen": 2460811264
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587474747474747,
      "loss": 1.4555,
      "theoretical_loss": 3.372770802106297,
      "tokens_seen": 2460876800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002587272727272728,
      "loss": 1.3982,
      "theoretical_loss": 3.3727636148980813,
      "tokens_seen": 2460942336
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025870707070707073,
      "loss": 1.3423,
      "theoretical_loss": 3.3727564279348514,
      "tokens_seen": 2461007872
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002586868686868687,
      "loss": 1.3613,
      "theoretical_loss": 3.3727492412165927,
      "tokens_seen": 2461073408
    },
    {
      "epoch": 0.49,
      "objective/train/aligned_token_count": 0.42271506786346436,
      "objective/train/avg_token_score": 0.03223308175802231,
      "objective/train/docs_used": 1385122,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3977241516113281,
      "objective/train/original_loss": 4.394418239593506,
      "objective/train/theoretical_loss": 3.37274205474329,
      "objective/train/tokens_used": 819998176,
      "objective/train/unlikelihood_loss": 0.0575607530772686,
      "theoretical_loss": 3.37274205474329,
      "tokens_seen": 2461138944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025866666666666665,
      "loss": 1.3914,
      "theoretical_loss": 3.37274205474329,
      "tokens_seen": 2461138944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002586464646464646,
      "loss": 1.3943,
      "theoretical_loss": 3.3727348685149288,
      "tokens_seen": 2461204480
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025862626262626267,
      "loss": 1.2706,
      "theoretical_loss": 3.3727276825314942,
      "tokens_seen": 2461270016
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002586060606060606,
      "loss": 1.3098,
      "theoretical_loss": 3.3727204967929714,
      "tokens_seen": 2461335552
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002585858585858586,
      "loss": 1.2619,
      "theoretical_loss": 3.372713311299345,
      "tokens_seen": 2461401088
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025856565656565654,
      "loss": 1.4226,
      "theoretical_loss": 3.3727061260506006,
      "tokens_seen": 2461466624
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025854545454545455,
      "loss": 1.2879,
      "theoretical_loss": 3.3726989410467234,
      "tokens_seen": 2461532160
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025852525252525256,
      "loss": 1.6175,
      "theoretical_loss": 3.372691756287699,
      "tokens_seen": 2461597696
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002585050505050505,
      "loss": 1.2485,
      "theoretical_loss": 3.3726845717735108,
      "tokens_seen": 2461663232
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002584848484848485,
      "loss": 1.4126,
      "theoretical_loss": 3.3726773875041456,
      "tokens_seen": 2461728768
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025846464646464643,
      "loss": 1.5699,
      "theoretical_loss": 3.372670203479588,
      "tokens_seen": 2461794304
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025844444444444444,
      "loss": 1.5026,
      "theoretical_loss": 3.3726630196998233,
      "tokens_seen": 2461859840
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025842424242424246,
      "loss": 1.4047,
      "theoretical_loss": 3.372655836164836,
      "tokens_seen": 2461925376
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002584040404040404,
      "loss": 1.2775,
      "theoretical_loss": 3.372648652874612,
      "tokens_seen": 2461990912
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025838383838383837,
      "loss": 1.3722,
      "theoretical_loss": 3.3726414698291363,
      "tokens_seen": 2462056448
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002583636363636364,
      "loss": 1.3366,
      "theoretical_loss": 3.372634287028394,
      "tokens_seen": 2462121984
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025834343434343434,
      "loss": 1.2525,
      "theoretical_loss": 3.37262710447237,
      "tokens_seen": 2462187520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025832323232323235,
      "loss": 1.2427,
      "theoretical_loss": 3.3726199221610496,
      "tokens_seen": 2462253056
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002583030303030303,
      "loss": 1.3191,
      "theoretical_loss": 3.3726127400944184,
      "tokens_seen": 2462318592
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025828282828282826,
      "loss": 1.1799,
      "theoretical_loss": 3.3726055582724603,
      "tokens_seen": 2462384128
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025826262626262627,
      "loss": 1.4801,
      "theoretical_loss": 3.372598376695162,
      "tokens_seen": 2462449664
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025824242424242423,
      "loss": 1.4991,
      "theoretical_loss": 3.3725911953625074,
      "tokens_seen": 2462515200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025822222222222224,
      "loss": 1.2817,
      "theoretical_loss": 3.3725840142744827,
      "tokens_seen": 2462580736
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002582020202020202,
      "loss": 1.3997,
      "theoretical_loss": 3.372576833431072,
      "tokens_seen": 2462646272
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002581818181818182,
      "loss": 1.365,
      "theoretical_loss": 3.3725696528322615,
      "tokens_seen": 2462711808
    },
    {
      "epoch": 0.49,
      "objective/train/aligned_token_count": 0.5036045908927917,
      "objective/train/avg_token_score": 0.012585188262164593,
      "objective/train/docs_used": 1386383,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5493502616882324,
      "objective/train/original_loss": 3.6951632499694824,
      "objective/train/theoretical_loss": 3.3725624724780356,
      "objective/train/tokens_used": 821636576,
      "objective/train/unlikelihood_loss": 0.07116042822599411,
      "theoretical_loss": 3.3725624724780356,
      "tokens_seen": 2462777344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025816161616161617,
      "loss": 1.5629,
      "theoretical_loss": 3.3725624724780356,
      "tokens_seen": 2462777344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002581414141414141,
      "loss": 1.4754,
      "theoretical_loss": 3.37255529236838,
      "tokens_seen": 2462842880
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025812121212121213,
      "loss": 1.308,
      "theoretical_loss": 3.3725481125032792,
      "tokens_seen": 2462908416
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002581010101010101,
      "loss": 1.2993,
      "theoretical_loss": 3.372540932882719,
      "tokens_seen": 2462973952
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002580808080808081,
      "loss": 1.5228,
      "theoretical_loss": 3.3725337535066844,
      "tokens_seen": 2463039488
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025806060606060606,
      "loss": 1.3431,
      "theoretical_loss": 3.37252657437516,
      "tokens_seen": 2463105024
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000258040404040404,
      "loss": 1.5019,
      "theoretical_loss": 3.3725193954881316,
      "tokens_seen": 2463170560
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000258020202020202,
      "loss": 1.4198,
      "theoretical_loss": 3.3725122168455846,
      "tokens_seen": 2463236096
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025800000000000004,
      "loss": 1.4353,
      "theoretical_loss": 3.3725050384475033,
      "tokens_seen": 2463301632
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000257979797979798,
      "loss": 1.6844,
      "theoretical_loss": 3.372497860293874,
      "tokens_seen": 2463367168
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025795959595959595,
      "loss": 1.3283,
      "theoretical_loss": 3.3724906823846807,
      "tokens_seen": 2463432704
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002579393939393939,
      "loss": 1.4559,
      "theoretical_loss": 3.3724835047199093,
      "tokens_seen": 2463498240
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002579191919191919,
      "loss": 1.3375,
      "theoretical_loss": 3.3724763272995446,
      "tokens_seen": 2463563776
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025789898989898993,
      "loss": 1.1249,
      "theoretical_loss": 3.372469150123572,
      "tokens_seen": 2463629312
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002578787878787879,
      "loss": 1.2817,
      "theoretical_loss": 3.3724619731919767,
      "tokens_seen": 2463694848
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025785858585858584,
      "loss": 1.265,
      "theoretical_loss": 3.372454796504744,
      "tokens_seen": 2463760384
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002578383838383838,
      "loss": 1.4694,
      "theoretical_loss": 3.3724476200618585,
      "tokens_seen": 2463825920
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025781818181818187,
      "loss": 1.4269,
      "theoretical_loss": 3.372440443863306,
      "tokens_seen": 2463891456
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002577979797979798,
      "loss": 1.1265,
      "theoretical_loss": 3.3724332679090714,
      "tokens_seen": 2463956992
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002577777777777778,
      "loss": 1.4013,
      "theoretical_loss": 3.37242609219914,
      "tokens_seen": 2464022528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025775757575757574,
      "loss": 1.7429,
      "theoretical_loss": 3.372418916733497,
      "tokens_seen": 2464088064
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025773737373737375,
      "loss": 1.288,
      "theoretical_loss": 3.3724117415121273,
      "tokens_seen": 2464153600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025771717171717176,
      "loss": 1.5554,
      "theoretical_loss": 3.372404566535016,
      "tokens_seen": 2464219136
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576969696969697,
      "loss": 1.4095,
      "theoretical_loss": 3.3723973918021493,
      "tokens_seen": 2464284672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025767676767676767,
      "loss": 1.3596,
      "theoretical_loss": 3.3723902173135114,
      "tokens_seen": 2464350208
    },
    {
      "epoch": 0.49,
      "objective/train/aligned_token_count": 0.3853861093521118,
      "objective/train/avg_token_score": 0.009328269399702549,
      "objective/train/docs_used": 1387153,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4622187614440918,
      "objective/train/original_loss": 4.738178730010986,
      "objective/train/theoretical_loss": 3.3723830430690875,
      "objective/train/tokens_used": 823274976,
      "objective/train/unlikelihood_loss": 0.04931565746665001,
      "theoretical_loss": 3.3723830430690875,
      "tokens_seen": 2464415744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025765656565656563,
      "loss": 1.4294,
      "theoretical_loss": 3.3723830430690875,
      "tokens_seen": 2464415744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002576363636363637,
      "loss": 1.3113,
      "theoretical_loss": 3.3723758690688634,
      "tokens_seen": 2464481280
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025761616161616165,
      "loss": 1.4991,
      "theoretical_loss": 3.372368695312824,
      "tokens_seen": 2464546816
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002575959595959596,
      "loss": 1.3435,
      "theoretical_loss": 3.372361521800954,
      "tokens_seen": 2464612352
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025757575757575756,
      "loss": 1.3103,
      "theoretical_loss": 3.3723543485332392,
      "tokens_seen": 2464677888
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002575555555555555,
      "loss": 1.3804,
      "theoretical_loss": 3.372347175509665,
      "tokens_seen": 2464743424
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002575353535353536,
      "loss": 1.4953,
      "theoretical_loss": 3.3723400027302164,
      "tokens_seen": 2464808960
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025751515151515154,
      "loss": 1.439,
      "theoretical_loss": 3.372332830194878,
      "tokens_seen": 2464874496
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002574949494949495,
      "loss": 1.3898,
      "theoretical_loss": 3.3723256579036356,
      "tokens_seen": 2464940032
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025747474747474746,
      "loss": 1.3745,
      "theoretical_loss": 3.372318485856474,
      "tokens_seen": 2465005568
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002574545454545454,
      "loss": 1.4681,
      "theoretical_loss": 3.3723113140533787,
      "tokens_seen": 2465071104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002574343434343435,
      "loss": 1.4551,
      "theoretical_loss": 3.372304142494335,
      "tokens_seen": 2465136640
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025741414141414144,
      "loss": 1.2745,
      "theoretical_loss": 3.372296971179328,
      "tokens_seen": 2465202176
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002573939393939394,
      "loss": 1.3047,
      "theoretical_loss": 3.372289800108343,
      "tokens_seen": 2465267712
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025737373737373735,
      "loss": 1.6231,
      "theoretical_loss": 3.372282629281365,
      "tokens_seen": 2465333248
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025735353535353536,
      "loss": 1.4382,
      "theoretical_loss": 3.372275458698379,
      "tokens_seen": 2465398784
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025733333333333337,
      "loss": 1.312,
      "theoretical_loss": 3.372268288359371,
      "tokens_seen": 2465464320
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025731313131313133,
      "loss": 1.3217,
      "theoretical_loss": 3.3722611182643254,
      "tokens_seen": 2465529856
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002572929292929293,
      "loss": 1.4753,
      "theoretical_loss": 3.3722539484132277,
      "tokens_seen": 2465595392
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025727272727272724,
      "loss": 1.5469,
      "theoretical_loss": 3.3722467788060633,
      "tokens_seen": 2465660928
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025725252525252525,
      "loss": 1.4707,
      "theoretical_loss": 3.3722396094428175,
      "tokens_seen": 2465726464
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025723232323232326,
      "loss": 1.5088,
      "theoretical_loss": 3.372232440323475,
      "tokens_seen": 2465792000
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002572121212121212,
      "loss": 1.2793,
      "theoretical_loss": 3.372225271448021,
      "tokens_seen": 2465857536
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571919191919192,
      "loss": 1.3416,
      "theoretical_loss": 3.3722181028164417,
      "tokens_seen": 2465923072
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571717171717172,
      "loss": 1.5166,
      "theoretical_loss": 3.3722109344287214,
      "tokens_seen": 2465988608
    },
    {
      "epoch": 0.49,
      "objective/train/aligned_token_count": 0.3866080045700073,
      "objective/train/avg_token_score": 0.020604686811566353,
      "objective/train/docs_used": 1388132,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.195600152015686,
      "objective/train/original_loss": 4.573416233062744,
      "objective/train/theoretical_loss": 3.3722037662848456,
      "objective/train/tokens_used": 824913376,
      "objective/train/unlikelihood_loss": 0.05502839758992195,
      "theoretical_loss": 3.3722037662848456,
      "tokens_seen": 2466054144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025715151515151515,
      "loss": 1.4404,
      "theoretical_loss": 3.3722037662848456,
      "tokens_seen": 2466054144
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025713131313131316,
      "loss": 1.5096,
      "theoretical_loss": 3.3721965983847992,
      "tokens_seen": 2466119680
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002571111111111111,
      "loss": 1.5217,
      "theoretical_loss": 3.372189430728568,
      "tokens_seen": 2466185216
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025709090909090907,
      "loss": 1.4246,
      "theoretical_loss": 3.3721822633161374,
      "tokens_seen": 2466250752
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002570707070707071,
      "loss": 1.4014,
      "theoretical_loss": 3.3721750961474917,
      "tokens_seen": 2466316288
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025705050505050504,
      "loss": 1.2886,
      "theoretical_loss": 3.3721679292226163,
      "tokens_seen": 2466381824
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025703030303030305,
      "loss": 1.4598,
      "theoretical_loss": 3.3721607625414975,
      "tokens_seen": 2466447360
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000257010101010101,
      "loss": 1.3768,
      "theoretical_loss": 3.3721535961041194,
      "tokens_seen": 2466512896
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256989898989899,
      "loss": 1.4212,
      "theoretical_loss": 3.3721464299104675,
      "tokens_seen": 2466578432
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256969696969697,
      "loss": 1.2841,
      "theoretical_loss": 3.3721392639605274,
      "tokens_seen": 2466643968
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025694949494949493,
      "loss": 1.3001,
      "theoretical_loss": 3.372132098254284,
      "tokens_seen": 2466709504
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025692929292929294,
      "loss": 1.3686,
      "theoretical_loss": 3.3721249327917224,
      "tokens_seen": 2466775040
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002569090909090909,
      "loss": 1.4561,
      "theoretical_loss": 3.3721177675728287,
      "tokens_seen": 2466840576
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002568888888888889,
      "loss": 1.1503,
      "theoretical_loss": 3.372110602597587,
      "tokens_seen": 2466906112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025686868686868687,
      "loss": 1.2782,
      "theoretical_loss": 3.372103437865983,
      "tokens_seen": 2466971648
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002568484848484848,
      "loss": 1.2826,
      "theoretical_loss": 3.372096273378002,
      "tokens_seen": 2467037184
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025682828282828284,
      "loss": 1.5478,
      "theoretical_loss": 3.3720891091336296,
      "tokens_seen": 2467102720
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025680808080808085,
      "loss": 1.1129,
      "theoretical_loss": 3.3720819451328508,
      "tokens_seen": 2467168256
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002567878787878788,
      "loss": 1.4652,
      "theoretical_loss": 3.37207478137565,
      "tokens_seen": 2467233792
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025676767676767676,
      "loss": 1.6238,
      "theoretical_loss": 3.372067617862014,
      "tokens_seen": 2467299328
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002567474747474747,
      "loss": 1.5662,
      "theoretical_loss": 3.3720604545919266,
      "tokens_seen": 2467364864
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025672727272727273,
      "loss": 1.3172,
      "theoretical_loss": 3.372053291565374,
      "tokens_seen": 2467430400
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025670707070707074,
      "loss": 1.2964,
      "theoretical_loss": 3.3720461287823413,
      "tokens_seen": 2467495936
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566868686868687,
      "loss": 1.2571,
      "theoretical_loss": 3.372038966242813,
      "tokens_seen": 2467561472
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025666666666666665,
      "loss": 1.2736,
      "theoretical_loss": 3.3720318039467756,
      "tokens_seen": 2467627008
    },
    {
      "epoch": 0.49,
      "objective/train/aligned_token_count": 0.3740224838256836,
      "objective/train/avg_token_score": 0.010944698937237263,
      "objective/train/docs_used": 1388592,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3328453302383423,
      "objective/train/original_loss": 4.4660563468933105,
      "objective/train/theoretical_loss": 3.372024641894213,
      "objective/train/tokens_used": 826551776,
      "objective/train/unlikelihood_loss": 0.06689217686653137,
      "theoretical_loss": 3.372024641894213,
      "tokens_seen": 2467692544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566464646464646,
      "loss": 1.3832,
      "theoretical_loss": 3.372024641894213,
      "tokens_seen": 2467692544
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002566262626262627,
      "loss": 1.2613,
      "theoretical_loss": 3.3720174800851117,
      "tokens_seen": 2467758080
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025660606060606063,
      "loss": 1.3515,
      "theoretical_loss": 3.3720103185194565,
      "tokens_seen": 2467823616
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565858585858586,
      "loss": 1.5459,
      "theoretical_loss": 3.3720031571972324,
      "tokens_seen": 2467889152
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025656565656565655,
      "loss": 1.3797,
      "theoretical_loss": 3.3719959961184247,
      "tokens_seen": 2467954688
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565454545454545,
      "loss": 1.4121,
      "theoretical_loss": 3.371988835283019,
      "tokens_seen": 2468020224
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025652525252525257,
      "loss": 1.4178,
      "theoretical_loss": 3.3719816746910003,
      "tokens_seen": 2468085760
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002565050505050505,
      "loss": 1.3048,
      "theoretical_loss": 3.371974514342354,
      "tokens_seen": 2468151296
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002564848484848485,
      "loss": 1.4083,
      "theoretical_loss": 3.371967354237065,
      "tokens_seen": 2468216832
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025646464646464644,
      "loss": 1.6041,
      "theoretical_loss": 3.371960194375119,
      "tokens_seen": 2468282368
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025644444444444445,
      "loss": 1.4215,
      "theoretical_loss": 3.371953034756501,
      "tokens_seen": 2468347904
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025642424242424246,
      "loss": 1.4265,
      "theoretical_loss": 3.371945875381197,
      "tokens_seen": 2468413440
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002564040404040404,
      "loss": 1.4365,
      "theoretical_loss": 3.371938716249191,
      "tokens_seen": 2468478976
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002563838383838384,
      "loss": 1.3443,
      "theoretical_loss": 3.37193155736047,
      "tokens_seen": 2468544512
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025636363636363633,
      "loss": 1.4689,
      "theoretical_loss": 3.3719243987150174,
      "tokens_seen": 2468610048
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025634343434343434,
      "loss": 1.4718,
      "theoretical_loss": 3.3719172403128193,
      "tokens_seen": 2468675584
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025632323232323235,
      "loss": 1.3394,
      "theoretical_loss": 3.371910082153861,
      "tokens_seen": 2468741120
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002563030303030303,
      "loss": 1.4569,
      "theoretical_loss": 3.3719029242381278,
      "tokens_seen": 2468806656
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025628282828282827,
      "loss": 1.4174,
      "theoretical_loss": 3.3718957665656046,
      "tokens_seen": 2468872192
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002562626262626263,
      "loss": 1.3537,
      "theoretical_loss": 3.3718886091362776,
      "tokens_seen": 2468937728
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025624242424242424,
      "loss": 1.5565,
      "theoretical_loss": 3.3718814519501312,
      "tokens_seen": 2469003264
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025622222222222225,
      "loss": 1.3856,
      "theoretical_loss": 3.371874295007151,
      "tokens_seen": 2469068800
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002562020202020202,
      "loss": 1.2487,
      "theoretical_loss": 3.371867138307322,
      "tokens_seen": 2469134336
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025618181818181816,
      "loss": 1.4255,
      "theoretical_loss": 3.3718599818506303,
      "tokens_seen": 2469199872
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025616161616161617,
      "loss": 1.3329,
      "theoretical_loss": 3.3718528256370606,
      "tokens_seen": 2469265408
    },
    {
      "epoch": 0.49,
      "objective/train/aligned_token_count": 0.23625366389751434,
      "objective/train/avg_token_score": 0.016802476719021797,
      "objective/train/docs_used": 1390136,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8069737553596497,
      "objective/train/original_loss": 4.469490051269531,
      "objective/train/theoretical_loss": 3.3718456696665977,
      "objective/train/tokens_used": 828190176,
      "objective/train/unlikelihood_loss": 0.06268247961997986,
      "theoretical_loss": 3.3718456696665977,
      "tokens_seen": 2469330944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025614141414141413,
      "loss": 1.2137,
      "theoretical_loss": 3.3718456696665977,
      "tokens_seen": 2469330944
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025612121212121214,
      "loss": 1.5413,
      "theoretical_loss": 3.371838513939228,
      "tokens_seen": 2469396480
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002561010101010101,
      "loss": 1.315,
      "theoretical_loss": 3.3718313584549358,
      "tokens_seen": 2469462016
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002560808080808081,
      "loss": 1.6637,
      "theoretical_loss": 3.371824203213707,
      "tokens_seen": 2469527552
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025606060606060606,
      "loss": 1.2927,
      "theoretical_loss": 3.3718170482155267,
      "tokens_seen": 2469593088
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256040404040404,
      "loss": 1.4857,
      "theoretical_loss": 3.3718098934603797,
      "tokens_seen": 2469658624
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025602020202020203,
      "loss": 1.3614,
      "theoretical_loss": 3.3718027389482526,
      "tokens_seen": 2469724160
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000256,
      "loss": 1.5806,
      "theoretical_loss": 3.371795584679129,
      "tokens_seen": 2469789696
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.000255979797979798,
      "loss": 1.4565,
      "theoretical_loss": 3.3717884306529955,
      "tokens_seen": 2469855232
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025595959595959596,
      "loss": 1.2254,
      "theoretical_loss": 3.371781276869837,
      "tokens_seen": 2469920768
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025593939393939397,
      "loss": 1.3346,
      "theoretical_loss": 3.371774123329639,
      "tokens_seen": 2469986304
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002559191919191919,
      "loss": 1.4818,
      "theoretical_loss": 3.3717669700323865,
      "tokens_seen": 2470051840
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025589898989898994,
      "loss": 1.6753,
      "theoretical_loss": 3.3717598169780647,
      "tokens_seen": 2470117376
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002558787878787879,
      "loss": 1.489,
      "theoretical_loss": 3.3717526641666593,
      "tokens_seen": 2470182912
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025585858585858585,
      "loss": 1.4065,
      "theoretical_loss": 3.371745511598155,
      "tokens_seen": 2470248448
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025583838383838386,
      "loss": 1.6745,
      "theoretical_loss": 3.3717383592725376,
      "tokens_seen": 2470313984
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002558181818181818,
      "loss": 1.478,
      "theoretical_loss": 3.371731207189793,
      "tokens_seen": 2470379520
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025579797979797983,
      "loss": 1.7185,
      "theoretical_loss": 3.371724055349905,
      "tokens_seen": 2470445056
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002557777777777778,
      "loss": 1.5756,
      "theoretical_loss": 3.3717169037528603,
      "tokens_seen": 2470510592
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025575757575757574,
      "loss": 1.3026,
      "theoretical_loss": 3.371709752398643,
      "tokens_seen": 2470576128
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025573737373737375,
      "loss": 1.3675,
      "theoretical_loss": 3.3717026012872395,
      "tokens_seen": 2470641664
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025571717171717176,
      "loss": 1.3296,
      "theoretical_loss": 3.3716954504186347,
      "tokens_seen": 2470707200
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002556969696969697,
      "loss": 1.5027,
      "theoretical_loss": 3.3716882997928135,
      "tokens_seen": 2470772736
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002556767676767677,
      "loss": 1.4023,
      "theoretical_loss": 3.371681149409762,
      "tokens_seen": 2470838272
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025565656565656563,
      "loss": 1.3631,
      "theoretical_loss": 3.371673999269465,
      "tokens_seen": 2470903808
    },
    {
      "epoch": 0.49,
      "objective/train/aligned_token_count": 0.3459188640117645,
      "objective/train/avg_token_score": 0.006191360764205456,
      "objective/train/docs_used": 1390574,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1513702869415283,
      "objective/train/original_loss": 4.329716682434082,
      "objective/train/theoretical_loss": 3.371666849371908,
      "objective/train/tokens_used": 829828576,
      "objective/train/unlikelihood_loss": 0.06842304766178131,
      "theoretical_loss": 3.371666849371908,
      "tokens_seen": 2470969344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025563636363636365,
      "loss": 1.4575,
      "theoretical_loss": 3.371666849371908,
      "tokens_seen": 2470969344
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025561616161616166,
      "loss": 1.3676,
      "theoretical_loss": 3.3716596997170765,
      "tokens_seen": 2471034880
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002555959595959596,
      "loss": 1.5556,
      "theoretical_loss": 3.371652550304955,
      "tokens_seen": 2471100416
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025557575757575757,
      "loss": 1.2096,
      "theoretical_loss": 3.3716454011355297,
      "tokens_seen": 2471165952
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025555555555555553,
      "loss": 1.5397,
      "theoretical_loss": 3.3716382522087858,
      "tokens_seen": 2471231488
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025553535353535354,
      "loss": 1.3759,
      "theoretical_loss": 3.371631103524708,
      "tokens_seen": 2471297024
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025551515151515155,
      "loss": 1.3587,
      "theoretical_loss": 3.3716239550832827,
      "tokens_seen": 2471362560
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554949494949495,
      "loss": 1.3298,
      "theoretical_loss": 3.3716168068844943,
      "tokens_seen": 2471428096
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025547474747474746,
      "loss": 1.4289,
      "theoretical_loss": 3.3716096589283286,
      "tokens_seen": 2471493632
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554545454545454,
      "loss": 1.5166,
      "theoretical_loss": 3.3716025112147707,
      "tokens_seen": 2471559168
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002554343434343435,
      "loss": 1.3191,
      "theoretical_loss": 3.371595363743806,
      "tokens_seen": 2471624704
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025541414141414144,
      "loss": 1.2512,
      "theoretical_loss": 3.37158821651542,
      "tokens_seen": 2471690240
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553939393939394,
      "loss": 1.459,
      "theoretical_loss": 3.3715810695295976,
      "tokens_seen": 2471755776
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025537373737373736,
      "loss": 1.7564,
      "theoretical_loss": 3.3715739227863244,
      "tokens_seen": 2471821312
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553535353535353,
      "loss": 1.3125,
      "theoretical_loss": 3.371566776285586,
      "tokens_seen": 2471886848
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002553333333333334,
      "loss": 1.5802,
      "theoretical_loss": 3.3715596300273676,
      "tokens_seen": 2471952384
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025531313131313133,
      "loss": 1.3269,
      "theoretical_loss": 3.371552484011654,
      "tokens_seen": 2472017920
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002552929292929293,
      "loss": 1.5289,
      "theoretical_loss": 3.3715453382384317,
      "tokens_seen": 2472083456
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025527272727272725,
      "loss": 1.3869,
      "theoretical_loss": 3.3715381927076846,
      "tokens_seen": 2472148992
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025525252525252526,
      "loss": 1.5582,
      "theoretical_loss": 3.371531047419399,
      "tokens_seen": 2472214528
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025523232323232327,
      "loss": 1.6277,
      "theoretical_loss": 3.3715239023735606,
      "tokens_seen": 2472280064
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025521212121212123,
      "loss": 1.607,
      "theoretical_loss": 3.3715167575701535,
      "tokens_seen": 2472345600
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002551919191919192,
      "loss": 1.2711,
      "theoretical_loss": 3.371509613009164,
      "tokens_seen": 2472411136
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025517171717171714,
      "loss": 1.5563,
      "theoretical_loss": 3.371502468690577,
      "tokens_seen": 2472476672
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025515151515151515,
      "loss": 1.4318,
      "theoretical_loss": 3.371495324614378,
      "tokens_seen": 2472542208
    },
    {
      "epoch": 0.49,
      "objective/train/aligned_token_count": 0.4980449676513672,
      "objective/train/avg_token_score": 0.014122984372079372,
      "objective/train/docs_used": 1391138,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.702610969543457,
      "objective/train/original_loss": 4.544701099395752,
      "objective/train/theoretical_loss": 3.3714881807805526,
      "objective/train/tokens_used": 831466976,
      "objective/train/unlikelihood_loss": 0.05094929039478302,
      "theoretical_loss": 3.3714881807805526,
      "tokens_seen": 2472607744
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025513131313131316,
      "loss": 1.4425,
      "theoretical_loss": 3.3714881807805526,
      "tokens_seen": 2472607744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002551111111111111,
      "loss": 1.3882,
      "theoretical_loss": 3.3714810371890858,
      "tokens_seen": 2472673280
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002550909090909091,
      "loss": 1.3394,
      "theoretical_loss": 3.371473893839963,
      "tokens_seen": 2472738816
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002550707070707071,
      "loss": 1.3514,
      "theoretical_loss": 3.3714667507331697,
      "tokens_seen": 2472804352
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025505050505050504,
      "loss": 1.5794,
      "theoretical_loss": 3.371459607868691,
      "tokens_seen": 2472869888
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025503030303030306,
      "loss": 1.6147,
      "theoretical_loss": 3.3714524652465125,
      "tokens_seen": 2472935424
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000255010101010101,
      "loss": 1.4091,
      "theoretical_loss": 3.3714453228666197,
      "tokens_seen": 2473000960
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025498989898989897,
      "loss": 1.3324,
      "theoretical_loss": 3.3714381807289975,
      "tokens_seen": 2473066496
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000254969696969697,
      "loss": 1.4144,
      "theoretical_loss": 3.3714310388336317,
      "tokens_seen": 2473132032
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025494949494949494,
      "loss": 1.446,
      "theoretical_loss": 3.3714238971805077,
      "tokens_seen": 2473197568
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025492929292929295,
      "loss": 1.4542,
      "theoretical_loss": 3.3714167557696104,
      "tokens_seen": 2473263104
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002549090909090909,
      "loss": 1.4365,
      "theoretical_loss": 3.3714096146009256,
      "tokens_seen": 2473328640
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002548888888888889,
      "loss": 1.5461,
      "theoretical_loss": 3.371402473674438,
      "tokens_seen": 2473394176
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002548686868686869,
      "loss": 1.5533,
      "theoretical_loss": 3.3713953329901343,
      "tokens_seen": 2473459712
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025484848484848483,
      "loss": 1.3742,
      "theoretical_loss": 3.3713881925479985,
      "tokens_seen": 2473525248
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025482828282828284,
      "loss": 1.6762,
      "theoretical_loss": 3.3713810523480165,
      "tokens_seen": 2473590784
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002548080808080808,
      "loss": 1.2392,
      "theoretical_loss": 3.371373912390174,
      "tokens_seen": 2473656320
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002547878787878788,
      "loss": 1.7106,
      "theoretical_loss": 3.3713667726744556,
      "tokens_seen": 2473721856
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025476767676767677,
      "loss": 1.2457,
      "theoretical_loss": 3.3713596332008473,
      "tokens_seen": 2473787392
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002547474747474747,
      "loss": 1.301,
      "theoretical_loss": 3.3713524939693347,
      "tokens_seen": 2473852928
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025472727272727273,
      "loss": 1.3288,
      "theoretical_loss": 3.371345354979902,
      "tokens_seen": 2473918464
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025470707070707075,
      "loss": 1.5108,
      "theoretical_loss": 3.371338216232536,
      "tokens_seen": 2473984000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002546868686868687,
      "loss": 1.507,
      "theoretical_loss": 3.371331077727221,
      "tokens_seen": 2474049536
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025466666666666666,
      "loss": 1.523,
      "theoretical_loss": 3.371323939463943,
      "tokens_seen": 2474115072
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002546464646464646,
      "loss": 1.3846,
      "theoretical_loss": 3.371316801442687,
      "tokens_seen": 2474180608
    },
    {
      "epoch": 0.5,
      "objective/train/aligned_token_count": 0.44837486743927,
      "objective/train/avg_token_score": 0.013653338886797428,
      "objective/train/docs_used": 1392460,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5986764430999756,
      "objective/train/original_loss": 4.5253400802612305,
      "objective/train/theoretical_loss": 3.371309663663439,
      "objective/train/tokens_used": 833105376,
      "objective/train/unlikelihood_loss": 0.049720846116542816,
      "theoretical_loss": 3.371309663663439,
      "tokens_seen": 2474246144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002546262626262626,
      "loss": 1.4931,
      "theoretical_loss": 3.371309663663439,
      "tokens_seen": 2474246144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025460606060606064,
      "loss": 1.5522,
      "theoretical_loss": 3.371302526126184,
      "tokens_seen": 2474311680
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545858585858586,
      "loss": 1.5066,
      "theoretical_loss": 3.371295388830907,
      "tokens_seen": 2474377216
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025456565656565655,
      "loss": 1.4194,
      "theoretical_loss": 3.3712882517775937,
      "tokens_seen": 2474442752
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545454545454545,
      "loss": 1.3779,
      "theoretical_loss": 3.3712811149662296,
      "tokens_seen": 2474508288
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002545252525252526,
      "loss": 1.2691,
      "theoretical_loss": 3.3712739783968004,
      "tokens_seen": 2474573824
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025450505050505053,
      "loss": 1.3915,
      "theoretical_loss": 3.371266842069291,
      "tokens_seen": 2474639360
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544848484848485,
      "loss": 1.5543,
      "theoretical_loss": 3.3712597059836864,
      "tokens_seen": 2474704896
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025446464646464644,
      "loss": 1.2629,
      "theoretical_loss": 3.371252570139973,
      "tokens_seen": 2474770432
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544444444444444,
      "loss": 1.2519,
      "theoretical_loss": 3.3712454345381353,
      "tokens_seen": 2474835968
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025442424242424247,
      "loss": 1.4973,
      "theoretical_loss": 3.3712382991781595,
      "tokens_seen": 2474901504
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002544040404040404,
      "loss": 1.5886,
      "theoretical_loss": 3.3712311640600303,
      "tokens_seen": 2474967040
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002543838383838384,
      "loss": 1.4052,
      "theoretical_loss": 3.3712240291837334,
      "tokens_seen": 2475032576
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025436363636363634,
      "loss": 1.3133,
      "theoretical_loss": 3.3712168945492547,
      "tokens_seen": 2475098112
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025434343434343435,
      "loss": 1.4247,
      "theoretical_loss": 3.3712097601565785,
      "tokens_seen": 2475163648
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025432323232323236,
      "loss": 1.6087,
      "theoretical_loss": 3.371202626005691,
      "tokens_seen": 2475229184
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002543030303030303,
      "loss": 1.4695,
      "theoretical_loss": 3.3711954920965774,
      "tokens_seen": 2475294720
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025428282828282827,
      "loss": 1.4895,
      "theoretical_loss": 3.3711883584292233,
      "tokens_seen": 2475360256
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025426262626262623,
      "loss": 1.3937,
      "theoretical_loss": 3.3711812250036135,
      "tokens_seen": 2475425792
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002542424242424243,
      "loss": 1.2993,
      "theoretical_loss": 3.371174091819734,
      "tokens_seen": 2475491328
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025422222222222225,
      "loss": 1.4834,
      "theoretical_loss": 3.37116695887757,
      "tokens_seen": 2475556864
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002542020202020202,
      "loss": 1.2769,
      "theoretical_loss": 3.371159826177107,
      "tokens_seen": 2475622400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025418181818181817,
      "loss": 1.478,
      "theoretical_loss": 3.3711526937183303,
      "tokens_seen": 2475687936
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541616161616161,
      "loss": 1.232,
      "theoretical_loss": 3.371145561501226,
      "tokens_seen": 2475753472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541414141414142,
      "loss": 1.2024,
      "theoretical_loss": 3.371138429525778,
      "tokens_seen": 2475819008
    },
    {
      "epoch": 0.5,
      "objective/train/aligned_token_count": 0.43493402004241943,
      "objective/train/avg_token_score": 0.007233086507767439,
      "objective/train/docs_used": 1393060,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.363207459449768,
      "objective/train/original_loss": 4.51606559753418,
      "objective/train/theoretical_loss": 3.371131297791973,
      "objective/train/tokens_used": 834743776,
      "objective/train/unlikelihood_loss": 0.040788426995277405,
      "theoretical_loss": 3.371131297791973,
      "tokens_seen": 2475884544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025412121212121214,
      "loss": 1.5316,
      "theoretical_loss": 3.371131297791973,
      "tokens_seen": 2475884544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002541010101010101,
      "loss": 1.201,
      "theoretical_loss": 3.3711241662997957,
      "tokens_seen": 2475950080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025408080808080806,
      "loss": 1.2957,
      "theoretical_loss": 3.371117035049232,
      "tokens_seen": 2476015616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025406060606060607,
      "loss": 1.4979,
      "theoretical_loss": 3.3711099040402672,
      "tokens_seen": 2476081152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002540404040404041,
      "loss": 1.3879,
      "theoretical_loss": 3.3711027732728867,
      "tokens_seen": 2476146688
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025402020202020204,
      "loss": 1.2955,
      "theoretical_loss": 3.3710956427470755,
      "tokens_seen": 2476212224
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000254,
      "loss": 1.5327,
      "theoretical_loss": 3.37108851246282,
      "tokens_seen": 2476277760
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025397979797979795,
      "loss": 1.5037,
      "theoretical_loss": 3.3710813824201047,
      "tokens_seen": 2476343296
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025395959595959596,
      "loss": 1.6313,
      "theoretical_loss": 3.3710742526189152,
      "tokens_seen": 2476408832
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000253939393939394,
      "loss": 1.5572,
      "theoretical_loss": 3.3710671230592375,
      "tokens_seen": 2476474368
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025391919191919193,
      "loss": 1.207,
      "theoretical_loss": 3.3710599937410564,
      "tokens_seen": 2476539904
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538989898989899,
      "loss": 1.4333,
      "theoretical_loss": 3.3710528646643576,
      "tokens_seen": 2476605440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538787878787879,
      "loss": 1.5483,
      "theoretical_loss": 3.3710457358291266,
      "tokens_seen": 2476670976
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025385858585858585,
      "loss": 1.2988,
      "theoretical_loss": 3.3710386072353486,
      "tokens_seen": 2476736512
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025383838383838387,
      "loss": 1.5437,
      "theoretical_loss": 3.371031478883009,
      "tokens_seen": 2476802048
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002538181818181818,
      "loss": 1.6016,
      "theoretical_loss": 3.371024350772094,
      "tokens_seen": 2476867584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002537979797979798,
      "loss": 1.485,
      "theoretical_loss": 3.3710172229025877,
      "tokens_seen": 2476933120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002537777777777778,
      "loss": 1.537,
      "theoretical_loss": 3.3710100952744764,
      "tokens_seen": 2476998656
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025375757575757575,
      "loss": 1.3531,
      "theoretical_loss": 3.371002967887746,
      "tokens_seen": 2477064192
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025373737373737376,
      "loss": 1.3003,
      "theoretical_loss": 3.3709958407423803,
      "tokens_seen": 2477129728
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002537171717171717,
      "loss": 1.4395,
      "theoretical_loss": 3.3709887138383663,
      "tokens_seen": 2477195264
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002536969696969697,
      "loss": 1.3846,
      "theoretical_loss": 3.370981587175689,
      "tokens_seen": 2477260800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002536767676767677,
      "loss": 1.4479,
      "theoretical_loss": 3.3709744607543337,
      "tokens_seen": 2477326336
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025365656565656564,
      "loss": 1.6053,
      "theoretical_loss": 3.370967334574286,
      "tokens_seen": 2477391872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025363636363636365,
      "loss": 1.5892,
      "theoretical_loss": 3.3709602086355313,
      "tokens_seen": 2477457408
    },
    {
      "epoch": 0.5,
      "objective/train/aligned_token_count": 0.5090420246124268,
      "objective/train/avg_token_score": 0.014255356043577194,
      "objective/train/docs_used": 1394031,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5706076622009277,
      "objective/train/original_loss": 3.748664140701294,
      "objective/train/theoretical_loss": 3.3709530829380543,
      "objective/train/tokens_used": 836382176,
      "objective/train/unlikelihood_loss": 0.06595630198717117,
      "theoretical_loss": 3.3709530829380543,
      "tokens_seen": 2477522944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002536161616161616,
      "loss": 1.5841,
      "theoretical_loss": 3.3709530829380543,
      "tokens_seen": 2477522944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002535959595959596,
      "loss": 1.5205,
      "theoretical_loss": 3.370945957481842,
      "tokens_seen": 2477588480
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002535757575757576,
      "loss": 1.4759,
      "theoretical_loss": 3.3709388322668783,
      "tokens_seen": 2477654016
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025355555555555553,
      "loss": 1.3991,
      "theoretical_loss": 3.37093170729315,
      "tokens_seen": 2477719552
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025353535353535354,
      "loss": 1.4871,
      "theoretical_loss": 3.3709245825606415,
      "tokens_seen": 2477785088
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025351515151515155,
      "loss": 1.5017,
      "theoretical_loss": 3.370917458069339,
      "tokens_seen": 2477850624
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002534949494949495,
      "loss": 1.3742,
      "theoretical_loss": 3.370910333819227,
      "tokens_seen": 2477916160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025347474747474747,
      "loss": 1.5152,
      "theoretical_loss": 3.370903209810292,
      "tokens_seen": 2477981696
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002534545454545454,
      "loss": 1.347,
      "theoretical_loss": 3.370896086042519,
      "tokens_seen": 2478047232
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025343434343434344,
      "loss": 1.2469,
      "theoretical_loss": 3.3708889625158935,
      "tokens_seen": 2478112768
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025341414141414145,
      "loss": 1.4436,
      "theoretical_loss": 3.3708818392304005,
      "tokens_seen": 2478178304
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533939393939394,
      "loss": 1.6625,
      "theoretical_loss": 3.3708747161860266,
      "tokens_seen": 2478243840
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025337373737373736,
      "loss": 1.2588,
      "theoretical_loss": 3.370867593382756,
      "tokens_seen": 2478309376
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533535353535353,
      "loss": 1.2211,
      "theoretical_loss": 3.370860470820575,
      "tokens_seen": 2478374912
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002533333333333334,
      "loss": 1.4201,
      "theoretical_loss": 3.3708533484994687,
      "tokens_seen": 2478440448
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025331313131313134,
      "loss": 1.439,
      "theoretical_loss": 3.3708462264194226,
      "tokens_seen": 2478505984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002532929292929293,
      "loss": 1.3853,
      "theoretical_loss": 3.3708391045804222,
      "tokens_seen": 2478571520
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025327272727272725,
      "loss": 1.5972,
      "theoretical_loss": 3.370831982982453,
      "tokens_seen": 2478637056
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002532525252525252,
      "loss": 1.3811,
      "theoretical_loss": 3.3708248616255005,
      "tokens_seen": 2478702592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002532323232323233,
      "loss": 1.4634,
      "theoretical_loss": 3.3708177405095503,
      "tokens_seen": 2478768128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025321212121212123,
      "loss": 1.5021,
      "theoretical_loss": 3.3708106196345877,
      "tokens_seen": 2478833664
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002531919191919192,
      "loss": 1.4107,
      "theoretical_loss": 3.370803499000598,
      "tokens_seen": 2478899200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025317171717171715,
      "loss": 1.3074,
      "theoretical_loss": 3.370796378607567,
      "tokens_seen": 2478964736
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025315151515151516,
      "loss": 1.4847,
      "theoretical_loss": 3.37078925845548,
      "tokens_seen": 2479030272
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025313131313131317,
      "loss": 1.6325,
      "theoretical_loss": 3.3707821385443224,
      "tokens_seen": 2479095808
    },
    {
      "epoch": 0.5,
      "objective/train/aligned_token_count": 0.5205278396606445,
      "objective/train/avg_token_score": 0.004616351332515478,
      "objective/train/docs_used": 1394307,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7414612770080566,
      "objective/train/original_loss": 3.779306411743164,
      "objective/train/theoretical_loss": 3.3707750188740797,
      "objective/train/tokens_used": 838020576,
      "objective/train/unlikelihood_loss": 0.07309654355049133,
      "theoretical_loss": 3.3707750188740797,
      "tokens_seen": 2479161344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002531111111111111,
      "loss": 1.7094,
      "theoretical_loss": 3.3707750188740797,
      "tokens_seen": 2479161344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002530909090909091,
      "loss": 1.5174,
      "theoretical_loss": 3.3707678994447376,
      "tokens_seen": 2479226880
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025307070707070704,
      "loss": 1.474,
      "theoretical_loss": 3.3707607802562816,
      "tokens_seen": 2479292416
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025305050505050505,
      "loss": 1.221,
      "theoretical_loss": 3.370753661308697,
      "tokens_seen": 2479357952
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025303030303030306,
      "loss": 1.3608,
      "theoretical_loss": 3.370746542601969,
      "tokens_seen": 2479423488
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000253010101010101,
      "loss": 1.3558,
      "theoretical_loss": 3.3707394241360835,
      "tokens_seen": 2479489024
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000252989898989899,
      "loss": 1.4984,
      "theoretical_loss": 3.370732305911026,
      "tokens_seen": 2479554560
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000252969696969697,
      "loss": 1.3019,
      "theoretical_loss": 3.370725187926782,
      "tokens_seen": 2479620096
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025294949494949494,
      "loss": 1.5722,
      "theoretical_loss": 3.3707180701833366,
      "tokens_seen": 2479685632
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025292929292929295,
      "loss": 1.3778,
      "theoretical_loss": 3.3707109526806756,
      "tokens_seen": 2479751168
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002529090909090909,
      "loss": 1.4837,
      "theoretical_loss": 3.3707038354187846,
      "tokens_seen": 2479816704
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025288888888888887,
      "loss": 1.5874,
      "theoretical_loss": 3.3706967183976486,
      "tokens_seen": 2479882240
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528686868686869,
      "loss": 1.4982,
      "theoretical_loss": 3.3706896016172543,
      "tokens_seen": 2479947776
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025284848484848484,
      "loss": 1.3269,
      "theoretical_loss": 3.370682485077585,
      "tokens_seen": 2480013312
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025282828282828285,
      "loss": 1.5219,
      "theoretical_loss": 3.3706753687786284,
      "tokens_seen": 2480078848
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002528080808080808,
      "loss": 1.454,
      "theoretical_loss": 3.3706682527203693,
      "tokens_seen": 2480144384
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002527878787878788,
      "loss": 1.3917,
      "theoretical_loss": 3.3706611369027923,
      "tokens_seen": 2480209920
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025276767676767677,
      "loss": 1.6233,
      "theoretical_loss": 3.370654021325884,
      "tokens_seen": 2480275456
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025274747474747473,
      "loss": 1.5398,
      "theoretical_loss": 3.37064690598963,
      "tokens_seen": 2480340992
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025272727272727274,
      "loss": 1.5386,
      "theoretical_loss": 3.3706397908940144,
      "tokens_seen": 2480406528
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002527070707070707,
      "loss": 1.5565,
      "theoretical_loss": 3.3706326760390244,
      "tokens_seen": 2480472064
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002526868686868687,
      "loss": 1.6136,
      "theoretical_loss": 3.370625561424644,
      "tokens_seen": 2480537600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025266666666666666,
      "loss": 1.3835,
      "theoretical_loss": 3.37061844705086,
      "tokens_seen": 2480603136
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002526464646464646,
      "loss": 1.2289,
      "theoretical_loss": 3.3706113329176572,
      "tokens_seen": 2480668672
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025262626262626263,
      "loss": 1.5684,
      "theoretical_loss": 3.370604219025021,
      "tokens_seen": 2480734208
    },
    {
      "debugging/Self-BLEU-5": 0.29522518101396983,
      "debugging/distinct-1-grams": 0.7404446023849008,
      "debugging/distinct-2-grams": 0.9621106381669762,
      "debugging/entropy-1-grams": 4.826739643784389,
      "debugging/entropy-2-grams": 5.449830011825935,
      "debugging/length": 541.0,
      "debugging/num_segments": 3,
      "epoch": 0.5,
      "objective/train/aligned_token_count": 0.6083211302757263,
      "objective/train/avg_token_score": 0.019920673221349716,
      "objective/train/docs_used": 1395647,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.797440528869629,
      "objective/train/original_loss": 3.1614341735839844,
      "objective/train/theoretical_loss": 3.3705971053729376,
      "objective/train/tokens_used": 839658976,
      "objective/train/unlikelihood_loss": 0.14986887574195862,
      "theoretical_loss": 3.3705971053729376,
      "tokens_seen": 2480799744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025260606060606064,
      "loss": 1.6603,
      "theoretical_loss": 3.3705971053729376,
      "tokens_seen": 2480799744
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002525858585858586,
      "loss": 1.4558,
      "theoretical_loss": 3.3705899919613915,
      "tokens_seen": 2480865280
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025256565656565656,
      "loss": 1.2445,
      "theoretical_loss": 3.370582878790369,
      "tokens_seen": 2480930816
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025254545454545457,
      "loss": 1.4313,
      "theoretical_loss": 3.370575765859856,
      "tokens_seen": 2480996352
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002525252525252525,
      "loss": 1.5145,
      "theoretical_loss": 3.370568653169837,
      "tokens_seen": 2481061888
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025250505050505054,
      "loss": 1.7166,
      "theoretical_loss": 3.3705615407202973,
      "tokens_seen": 2481127424
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002524848484848485,
      "loss": 1.447,
      "theoretical_loss": 3.370554428511224,
      "tokens_seen": 2481192960
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025246464646464645,
      "loss": 1.4322,
      "theoretical_loss": 3.370547316542601,
      "tokens_seen": 2481258496
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025244444444444446,
      "loss": 1.4969,
      "theoretical_loss": 3.3705402048144153,
      "tokens_seen": 2481324032
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025242424242424247,
      "loss": 1.4399,
      "theoretical_loss": 3.370533093326651,
      "tokens_seen": 2481389568
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025240404040404043,
      "loss": 1.3008,
      "theoretical_loss": 3.370525982079294,
      "tokens_seen": 2481455104
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002523838383838384,
      "loss": 1.393,
      "theoretical_loss": 3.3705188710723304,
      "tokens_seen": 2481520640
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025236363636363634,
      "loss": 1.4269,
      "theoretical_loss": 3.3705117603057455,
      "tokens_seen": 2481586176
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025234343434343435,
      "loss": 1.5876,
      "theoretical_loss": 3.370504649779525,
      "tokens_seen": 2481651712
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025232323232323236,
      "loss": 1.7248,
      "theoretical_loss": 3.3704975394936536,
      "tokens_seen": 2481717248
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002523030303030303,
      "loss": 1.2866,
      "theoretical_loss": 3.3704904294481173,
      "tokens_seen": 2481782784
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002522828282828283,
      "loss": 1.4932,
      "theoretical_loss": 3.3704833196429016,
      "tokens_seen": 2481848320
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025226262626262624,
      "loss": 1.3145,
      "theoretical_loss": 3.370476210077993,
      "tokens_seen": 2481913856
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025224242424242425,
      "loss": 1.2655,
      "theoretical_loss": 3.3704691007533754,
      "tokens_seen": 2481979392
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025222222222222226,
      "loss": 1.417,
      "theoretical_loss": 3.370461991669035,
      "tokens_seen": 2482044928
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002522020202020202,
      "loss": 1.5247,
      "theoretical_loss": 3.370454882824958,
      "tokens_seen": 2482110464
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025218181818181817,
      "loss": 1.2555,
      "theoretical_loss": 3.370447774221129,
      "tokens_seen": 2482176000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025216161616161613,
      "loss": 1.2501,
      "theoretical_loss": 3.370440665857534,
      "tokens_seen": 2482241536
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002521414141414142,
      "loss": 1.4846,
      "theoretical_loss": 3.3704335577341586,
      "tokens_seen": 2482307072
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025212121212121215,
      "loss": 1.3635,
      "theoretical_loss": 3.3704264498509877,
      "tokens_seen": 2482372608
    },
    {
      "epoch": 0.5,
      "objective/train/aligned_token_count": 0.3412145674228668,
      "objective/train/avg_token_score": 0.02692093700170517,
      "objective/train/docs_used": 1396325,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1937379837036133,
      "objective/train/original_loss": 5.343158721923828,
      "objective/train/theoretical_loss": 3.3704193422080078,
      "objective/train/tokens_used": 841297376,
      "objective/train/unlikelihood_loss": 0.03865502402186394,
      "theoretical_loss": 3.3704193422080078,
      "tokens_seen": 2482438144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002521010101010101,
      "loss": 1.2697,
      "theoretical_loss": 3.3704193422080078,
      "tokens_seen": 2482438144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025208080808080806,
      "loss": 1.234,
      "theoretical_loss": 3.3704122348052037,
      "tokens_seen": 2482503680
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000252060606060606,
      "loss": 1.5142,
      "theoretical_loss": 3.3704051276425613,
      "tokens_seen": 2482569216
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002520404040404041,
      "loss": 1.396,
      "theoretical_loss": 3.3703980207200663,
      "tokens_seen": 2482634752
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025202020202020204,
      "loss": 1.3311,
      "theoretical_loss": 3.3703909140377037,
      "tokens_seen": 2482700288
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000252,
      "loss": 1.2108,
      "theoretical_loss": 3.3703838075954593,
      "tokens_seen": 2482765824
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025197979797979796,
      "loss": 1.3661,
      "theoretical_loss": 3.370376701393319,
      "tokens_seen": 2482831360
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025195959595959597,
      "loss": 1.519,
      "theoretical_loss": 3.3703695954312676,
      "tokens_seen": 2482896896
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000251939393939394,
      "loss": 1.383,
      "theoretical_loss": 3.3703624897092914,
      "tokens_seen": 2482962432
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025191919191919194,
      "loss": 1.445,
      "theoretical_loss": 3.3703553842273757,
      "tokens_seen": 2483027968
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002518989898989899,
      "loss": 1.4253,
      "theoretical_loss": 3.370348278985506,
      "tokens_seen": 2483093504
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025187878787878785,
      "loss": 1.4266,
      "theoretical_loss": 3.3703411739836677,
      "tokens_seen": 2483159040
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025185858585858586,
      "loss": 1.353,
      "theoretical_loss": 3.3703340692218466,
      "tokens_seen": 2483224576
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025183838383838387,
      "loss": 1.5075,
      "theoretical_loss": 3.3703269647000282,
      "tokens_seen": 2483290112
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025181818181818183,
      "loss": 1.3903,
      "theoretical_loss": 3.3703198604181983,
      "tokens_seen": 2483355648
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002517979797979798,
      "loss": 1.5156,
      "theoretical_loss": 3.3703127563763418,
      "tokens_seen": 2483421184
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002517777777777778,
      "loss": 1.3331,
      "theoretical_loss": 3.370305652574445,
      "tokens_seen": 2483486720
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025175757575757575,
      "loss": 1.4398,
      "theoretical_loss": 3.370298549012493,
      "tokens_seen": 2483552256
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025173737373737376,
      "loss": 1.5963,
      "theoretical_loss": 3.370291445690471,
      "tokens_seen": 2483617792
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002517171717171717,
      "loss": 1.3035,
      "theoretical_loss": 3.370284342608366,
      "tokens_seen": 2483683328
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002516969696969697,
      "loss": 1.3825,
      "theoretical_loss": 3.370277239766162,
      "tokens_seen": 2483748864
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002516767676767677,
      "loss": 1.4681,
      "theoretical_loss": 3.3702701371638453,
      "tokens_seen": 2483814400
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025165656565656565,
      "loss": 1.3965,
      "theoretical_loss": 3.3702630348014013,
      "tokens_seen": 2483879936
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025163636363636366,
      "loss": 1.3121,
      "theoretical_loss": 3.3702559326788157,
      "tokens_seen": 2483945472
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002516161616161616,
      "loss": 1.259,
      "theoretical_loss": 3.370248830796074,
      "tokens_seen": 2484011008
    },
    {
      "epoch": 0.5,
      "objective/train/aligned_token_count": 0.33119502663612366,
      "objective/train/avg_token_score": 0.015470614656805992,
      "objective/train/docs_used": 1397604,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.093152642250061,
      "objective/train/original_loss": 4.3071722984313965,
      "objective/train/theoretical_loss": 3.3702417291531614,
      "objective/train/tokens_used": 842935776,
      "objective/train/unlikelihood_loss": 0.07489950954914093,
      "theoretical_loss": 3.3702417291531614,
      "tokens_seen": 2484076544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002515959595959596,
      "loss": 1.3493,
      "theoretical_loss": 3.3702417291531614,
      "tokens_seen": 2484076544
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002515757575757576,
      "loss": 1.6412,
      "theoretical_loss": 3.3702346277500643,
      "tokens_seen": 2484142080
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025155555555555554,
      "loss": 1.6882,
      "theoretical_loss": 3.370227526586768,
      "tokens_seen": 2484207616
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025153535353535355,
      "loss": 1.4557,
      "theoretical_loss": 3.3702204256632573,
      "tokens_seen": 2484273152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002515151515151515,
      "loss": 1.4239,
      "theoretical_loss": 3.3702133249795185,
      "tokens_seen": 2484338688
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002514949494949495,
      "loss": 1.4866,
      "theoretical_loss": 3.3702062245355373,
      "tokens_seen": 2484404224
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002514747474747475,
      "loss": 1.4809,
      "theoretical_loss": 3.370199124331299,
      "tokens_seen": 2484469760
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025145454545454543,
      "loss": 1.5036,
      "theoretical_loss": 3.3701920243667893,
      "tokens_seen": 2484535296
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025143434343434344,
      "loss": 1.3675,
      "theoretical_loss": 3.3701849246419937,
      "tokens_seen": 2484600832
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025141414141414145,
      "loss": 1.3449,
      "theoretical_loss": 3.3701778251568975,
      "tokens_seen": 2484666368
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002513939393939394,
      "loss": 1.2857,
      "theoretical_loss": 3.370170725911487,
      "tokens_seen": 2484731904
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025137373737373737,
      "loss": 1.4917,
      "theoretical_loss": 3.3701636269057467,
      "tokens_seen": 2484797440
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002513535353535353,
      "loss": 1.4037,
      "theoretical_loss": 3.3701565281396633,
      "tokens_seen": 2484862976
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025133333333333333,
      "loss": 1.3405,
      "theoretical_loss": 3.370149429613222,
      "tokens_seen": 2484928512
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025131313131313135,
      "loss": 1.5142,
      "theoretical_loss": 3.370142331326408,
      "tokens_seen": 2484994048
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002512929292929293,
      "loss": 1.5756,
      "theoretical_loss": 3.3701352332792074,
      "tokens_seen": 2485059584
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025127272727272726,
      "loss": 1.4556,
      "theoretical_loss": 3.3701281354716057,
      "tokens_seen": 2485125120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002512525252525252,
      "loss": 1.6597,
      "theoretical_loss": 3.370121037903588,
      "tokens_seen": 2485190656
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002512323232323233,
      "loss": 1.6032,
      "theoretical_loss": 3.370113940575141,
      "tokens_seen": 2485256192
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025121212121212124,
      "loss": 1.5485,
      "theoretical_loss": 3.3701068434862487,
      "tokens_seen": 2485321728
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002511919191919192,
      "loss": 1.3958,
      "theoretical_loss": 3.3700997466368983,
      "tokens_seen": 2485387264
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025117171717171715,
      "loss": 1.5587,
      "theoretical_loss": 3.3700926500270745,
      "tokens_seen": 2485452800
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002511515151515151,
      "loss": 1.3468,
      "theoretical_loss": 3.3700855536567627,
      "tokens_seen": 2485518336
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002511313131313132,
      "loss": 1.3677,
      "theoretical_loss": 3.3700784575259495,
      "tokens_seen": 2485583872
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025111111111111113,
      "loss": 1.4325,
      "theoretical_loss": 3.3700713616346194,
      "tokens_seen": 2485649408
    },
    {
      "epoch": 0.5,
      "objective/train/aligned_token_count": 0.3698069453239441,
      "objective/train/avg_token_score": 0.006077800877392292,
      "objective/train/docs_used": 1398115,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3136042356491089,
      "objective/train/original_loss": 4.365874290466309,
      "objective/train/theoretical_loss": 3.3700642659827587,
      "objective/train/tokens_used": 844574176,
      "objective/train/unlikelihood_loss": 0.06717482954263687,
      "theoretical_loss": 3.3700642659827587,
      "tokens_seen": 2485714944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002510909090909091,
      "loss": 1.5392,
      "theoretical_loss": 3.3700642659827587,
      "tokens_seen": 2485714944
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025107070707070704,
      "loss": 1.451,
      "theoretical_loss": 3.370057170570353,
      "tokens_seen": 2485780480
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025105050505050506,
      "loss": 1.4344,
      "theoretical_loss": 3.3700500753973874,
      "tokens_seen": 2485846016
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025103030303030307,
      "loss": 1.4413,
      "theoretical_loss": 3.370042980463848,
      "tokens_seen": 2485911552
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000251010101010101,
      "loss": 1.2856,
      "theoretical_loss": 3.3700358857697204,
      "tokens_seen": 2485977088
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000250989898989899,
      "loss": 1.4005,
      "theoretical_loss": 3.3700287913149896,
      "tokens_seen": 2486042624
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025096969696969694,
      "loss": 1.5842,
      "theoretical_loss": 3.3700216970996415,
      "tokens_seen": 2486108160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025094949494949495,
      "loss": 1.3555,
      "theoretical_loss": 3.3700146031236624,
      "tokens_seen": 2486173696
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025092929292929296,
      "loss": 1.5611,
      "theoretical_loss": 3.370007509387037,
      "tokens_seen": 2486239232
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002509090909090909,
      "loss": 1.4248,
      "theoretical_loss": 3.3700004158897516,
      "tokens_seen": 2486304768
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002508888888888889,
      "loss": 1.3572,
      "theoretical_loss": 3.3699933226317915,
      "tokens_seen": 2486370304
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025086868686868683,
      "loss": 1.4804,
      "theoretical_loss": 3.3699862296131418,
      "tokens_seen": 2486435840
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002508484848484849,
      "loss": 1.3504,
      "theoretical_loss": 3.3699791368337895,
      "tokens_seen": 2486501376
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025082828282828285,
      "loss": 1.357,
      "theoretical_loss": 3.3699720442937187,
      "tokens_seen": 2486566912
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002508080808080808,
      "loss": 1.3974,
      "theoretical_loss": 3.369964951992916,
      "tokens_seen": 2486632448
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025078787878787877,
      "loss": 1.4239,
      "theoretical_loss": 3.3699578599313664,
      "tokens_seen": 2486697984
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002507676767676768,
      "loss": 1.5022,
      "theoretical_loss": 3.3699507681090557,
      "tokens_seen": 2486763520
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002507474747474748,
      "loss": 1.4465,
      "theoretical_loss": 3.36994367652597,
      "tokens_seen": 2486829056
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025072727272727275,
      "loss": 1.4352,
      "theoretical_loss": 3.369936585182095,
      "tokens_seen": 2486894592
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002507070707070707,
      "loss": 1.4818,
      "theoretical_loss": 3.3699294940774154,
      "tokens_seen": 2486960128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025068686868686866,
      "loss": 1.3959,
      "theoretical_loss": 3.369922403211917,
      "tokens_seen": 2487025664
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025066666666666667,
      "loss": 1.4059,
      "theoretical_loss": 3.369915312585586,
      "tokens_seen": 2487091200
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002506464646464647,
      "loss": 1.661,
      "theoretical_loss": 3.3699082221984082,
      "tokens_seen": 2487156736
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025062626262626264,
      "loss": 1.2793,
      "theoretical_loss": 3.3699011320503685,
      "tokens_seen": 2487222272
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002506060606060606,
      "loss": 1.5756,
      "theoretical_loss": 3.369894042141453,
      "tokens_seen": 2487287808
    },
    {
      "epoch": 0.5,
      "objective/train/aligned_token_count": 0.43114614486694336,
      "objective/train/avg_token_score": 0.011788773350417614,
      "objective/train/docs_used": 1398737,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4830436706542969,
      "objective/train/original_loss": 4.010889530181885,
      "objective/train/theoretical_loss": 3.369886952471647,
      "objective/train/tokens_used": 846212576,
      "objective/train/unlikelihood_loss": 0.07453753054141998,
      "theoretical_loss": 3.369886952471647,
      "tokens_seen": 2487353344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002505858585858586,
      "loss": 1.3821,
      "theoretical_loss": 3.369886952471647,
      "tokens_seen": 2487353344
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025056565656565656,
      "loss": 1.4644,
      "theoretical_loss": 3.3698798630409366,
      "tokens_seen": 2487418880
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002505454545454546,
      "loss": 1.3716,
      "theoretical_loss": 3.369872773849307,
      "tokens_seen": 2487484416
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025052525252525253,
      "loss": 1.3074,
      "theoretical_loss": 3.3698656848967437,
      "tokens_seen": 2487549952
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002505050505050505,
      "loss": 1.1893,
      "theoretical_loss": 3.3698585961832332,
      "tokens_seen": 2487615488
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002504848484848485,
      "loss": 1.3315,
      "theoretical_loss": 3.36985150770876,
      "tokens_seen": 2487681024
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025046464646464646,
      "loss": 1.3316,
      "theoretical_loss": 3.369844419473311,
      "tokens_seen": 2487746560
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025044444444444447,
      "loss": 1.497,
      "theoretical_loss": 3.369837331476871,
      "tokens_seen": 2487812096
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002504242424242424,
      "loss": 1.4152,
      "theoretical_loss": 3.3698302437194254,
      "tokens_seen": 2487877632
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025040404040404043,
      "loss": 1.4465,
      "theoretical_loss": 3.369823156200961,
      "tokens_seen": 2487943168
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002503838383838384,
      "loss": 1.2677,
      "theoretical_loss": 3.369816068921462,
      "tokens_seen": 2488008704
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025036363636363635,
      "loss": 1.441,
      "theoretical_loss": 3.369808981880915,
      "tokens_seen": 2488074240
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025034343434343436,
      "loss": 1.5473,
      "theoretical_loss": 3.3698018950793056,
      "tokens_seen": 2488139776
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002503232323232323,
      "loss": 1.5958,
      "theoretical_loss": 3.369794808516619,
      "tokens_seen": 2488205312
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002503030303030303,
      "loss": 1.3451,
      "theoretical_loss": 3.369787722192841,
      "tokens_seen": 2488270848
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002502828282828283,
      "loss": 1.4139,
      "theoretical_loss": 3.369780636107958,
      "tokens_seen": 2488336384
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025026262626262624,
      "loss": 1.2984,
      "theoretical_loss": 3.3697735502619546,
      "tokens_seen": 2488401920
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025024242424242425,
      "loss": 1.4429,
      "theoretical_loss": 3.369766464654817,
      "tokens_seen": 2488467456
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025022222222222226,
      "loss": 1.4052,
      "theoretical_loss": 3.369759379286531,
      "tokens_seen": 2488532992
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002502020202020202,
      "loss": 1.3773,
      "theoretical_loss": 3.3697522941570814,
      "tokens_seen": 2488598528
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002501818181818182,
      "loss": 1.1516,
      "theoretical_loss": 3.369745209266455,
      "tokens_seen": 2488664064
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025016161616161613,
      "loss": 1.5339,
      "theoretical_loss": 3.3697381246146363,
      "tokens_seen": 2488729600
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025014141414141414,
      "loss": 1.4547,
      "theoretical_loss": 3.369731040201612,
      "tokens_seen": 2488795136
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.00025012121212121216,
      "loss": 1.2949,
      "theoretical_loss": 3.369723956027367,
      "tokens_seen": 2488860672
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002501010101010101,
      "loss": 1.2072,
      "theoretical_loss": 3.369716872091888,
      "tokens_seen": 2488926208
    },
    {
      "epoch": 0.5,
      "objective/train/aligned_token_count": 0.4481304883956909,
      "objective/train/avg_token_score": 0.007207569666206837,
      "objective/train/docs_used": 1399945,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.283136248588562,
      "objective/train/original_loss": 4.074010372161865,
      "objective/train/theoretical_loss": 3.3697097883951597,
      "objective/train/tokens_used": 847850976,
      "objective/train/unlikelihood_loss": 0.08538097888231277,
      "theoretical_loss": 3.3697097883951597,
      "tokens_seen": 2488991744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00025008080808080807,
      "loss": 1.3625,
      "theoretical_loss": 3.3697097883951597,
      "tokens_seen": 2488991744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000250060606060606,
      "loss": 1.4357,
      "theoretical_loss": 3.3697027049371675,
      "tokens_seen": 2489057280
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002500404040404041,
      "loss": 1.3355,
      "theoretical_loss": 3.369695621717898,
      "tokens_seen": 2489122816
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00025002020202020205,
      "loss": 1.5294,
      "theoretical_loss": 3.369688538737337,
      "tokens_seen": 2489188352
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00025,
      "loss": 1.4611,
      "theoretical_loss": 3.369681455995469,
      "tokens_seen": 2489253888
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024997979797979796,
      "loss": 1.3802,
      "theoretical_loss": 3.3696743734922805,
      "tokens_seen": 2489319424
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000249959595959596,
      "loss": 1.5304,
      "theoretical_loss": 3.369667291227757,
      "tokens_seen": 2489384960
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024993939393939393,
      "loss": 1.414,
      "theoretical_loss": 3.369660209201884,
      "tokens_seen": 2489450496
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024991919191919194,
      "loss": 1.3748,
      "theoretical_loss": 3.3696531274146477,
      "tokens_seen": 2489516032
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002498989898989899,
      "loss": 1.5094,
      "theoretical_loss": 3.369646045866033,
      "tokens_seen": 2489581568
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024987878787878785,
      "loss": 1.2389,
      "theoretical_loss": 3.3696389645560263,
      "tokens_seen": 2489647104
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024985858585858587,
      "loss": 1.6678,
      "theoretical_loss": 3.369631883484613,
      "tokens_seen": 2489712640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002498383838383838,
      "loss": 1.3026,
      "theoretical_loss": 3.3696248026517788,
      "tokens_seen": 2489778176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024981818181818183,
      "loss": 1.5347,
      "theoretical_loss": 3.3696177220575096,
      "tokens_seen": 2489843712
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497979797979798,
      "loss": 1.5492,
      "theoretical_loss": 3.36961064170179,
      "tokens_seen": 2489909248
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024977777777777775,
      "loss": 1.3587,
      "theoretical_loss": 3.369603561584607,
      "tokens_seen": 2489974784
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024975757575757576,
      "loss": 1.6395,
      "theoretical_loss": 3.369596481705946,
      "tokens_seen": 2490040320
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497373737373737,
      "loss": 1.144,
      "theoretical_loss": 3.3695894020657926,
      "tokens_seen": 2490105856
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002497171717171717,
      "loss": 1.4156,
      "theoretical_loss": 3.3695823226641317,
      "tokens_seen": 2490171392
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496969696969697,
      "loss": 1.4021,
      "theoretical_loss": 3.3695752435009503,
      "tokens_seen": 2490236928
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496767676767677,
      "loss": 1.6072,
      "theoretical_loss": 3.369568164576233,
      "tokens_seen": 2490302464
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024965656565656565,
      "loss": 1.6537,
      "theoretical_loss": 3.3695610858899663,
      "tokens_seen": 2490368000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496363636363636,
      "loss": 1.3644,
      "theoretical_loss": 3.3695540074421353,
      "tokens_seen": 2490433536
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002496161616161616,
      "loss": 1.2885,
      "theoretical_loss": 3.369546929232726,
      "tokens_seen": 2490499072
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002495959595959596,
      "loss": 1.2928,
      "theoretical_loss": 3.369539851261724,
      "tokens_seen": 2490564608
    },
    {
      "epoch": 0.51,
      "objective/train/aligned_token_count": 0.4950513243675232,
      "objective/train/avg_token_score": 0.009924127720296383,
      "objective/train/docs_used": 1400696,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4836986064910889,
      "objective/train/original_loss": 4.051167964935303,
      "objective/train/theoretical_loss": 3.369532773529115,
      "objective/train/tokens_used": 849489376,
      "objective/train/unlikelihood_loss": 0.04935472086071968,
      "theoretical_loss": 3.369532773529115,
      "tokens_seen": 2490630144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002495757575757576,
      "loss": 1.4431,
      "theoretical_loss": 3.369532773529115,
      "tokens_seen": 2490630144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024955555555555554,
      "loss": 1.5281,
      "theoretical_loss": 3.369525696034885,
      "tokens_seen": 2490695680
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024953535353535355,
      "loss": 1.1206,
      "theoretical_loss": 3.3695186187790194,
      "tokens_seen": 2490761216
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002495151515151515,
      "loss": 1.3028,
      "theoretical_loss": 3.3695115417615034,
      "tokens_seen": 2490826752
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002494949494949495,
      "loss": 1.3349,
      "theoretical_loss": 3.369504464982324,
      "tokens_seen": 2490892288
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002494747474747475,
      "loss": 1.6027,
      "theoretical_loss": 3.3694973884414656,
      "tokens_seen": 2490957824
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024945454545454544,
      "loss": 1.4238,
      "theoretical_loss": 3.3694903121389146,
      "tokens_seen": 2491023360
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024943434343434345,
      "loss": 1.7674,
      "theoretical_loss": 3.3694832360746565,
      "tokens_seen": 2491088896
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002494141414141414,
      "loss": 1.5696,
      "theoretical_loss": 3.369476160248677,
      "tokens_seen": 2491154432
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002493939393939394,
      "loss": 1.4335,
      "theoretical_loss": 3.369469084660962,
      "tokens_seen": 2491219968
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024937373737373737,
      "loss": 1.3739,
      "theoretical_loss": 3.369462009311497,
      "tokens_seen": 2491285504
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002493535353535354,
      "loss": 1.3713,
      "theoretical_loss": 3.3694549342002675,
      "tokens_seen": 2491351040
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024933333333333334,
      "loss": 1.332,
      "theoretical_loss": 3.36944785932726,
      "tokens_seen": 2491416576
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024931313131313135,
      "loss": 1.3876,
      "theoretical_loss": 3.3694407846924594,
      "tokens_seen": 2491482112
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492929292929293,
      "loss": 1.5077,
      "theoretical_loss": 3.369433710295852,
      "tokens_seen": 2491547648
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024927272727272727,
      "loss": 1.4509,
      "theoretical_loss": 3.3694266361374225,
      "tokens_seen": 2491613184
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002492525252525253,
      "loss": 1.5522,
      "theoretical_loss": 3.3694195622171583,
      "tokens_seen": 2491678720
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024923232323232323,
      "loss": 1.3013,
      "theoretical_loss": 3.3694124885350436,
      "tokens_seen": 2491744256
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024921212121212124,
      "loss": 1.599,
      "theoretical_loss": 3.369405415091065,
      "tokens_seen": 2491809792
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491919191919192,
      "loss": 1.3319,
      "theoretical_loss": 3.3693983418852076,
      "tokens_seen": 2491875328
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491717171717172,
      "loss": 1.4808,
      "theoretical_loss": 3.3693912689174574,
      "tokens_seen": 2491940864
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024915151515151517,
      "loss": 1.479,
      "theoretical_loss": 3.3693841961878004,
      "tokens_seen": 2492006400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002491313131313131,
      "loss": 1.297,
      "theoretical_loss": 3.3693771236962218,
      "tokens_seen": 2492071936
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024911111111111114,
      "loss": 1.547,
      "theoretical_loss": 3.369370051442708,
      "tokens_seen": 2492137472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002490909090909091,
      "loss": 1.3939,
      "theoretical_loss": 3.369362979427244,
      "tokens_seen": 2492203008
    },
    {
      "epoch": 0.51,
      "objective/train/aligned_token_count": 0.31983137130737305,
      "objective/train/avg_token_score": 0.014625142328441143,
      "objective/train/docs_used": 1402287,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0876846313476562,
      "objective/train/original_loss": 5.1158318519592285,
      "objective/train/theoretical_loss": 3.369355907649816,
      "objective/train/tokens_used": 851127776,
      "objective/train/unlikelihood_loss": 0.05833510309457779,
      "theoretical_loss": 3.369355907649816,
      "tokens_seen": 2492268544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002490707070707071,
      "loss": 1.5166,
      "theoretical_loss": 3.369355907649816,
      "tokens_seen": 2492268544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024905050505050506,
      "loss": 1.4668,
      "theoretical_loss": 3.3693488361104094,
      "tokens_seen": 2492334080
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000249030303030303,
      "loss": 1.2985,
      "theoretical_loss": 3.3693417648090103,
      "tokens_seen": 2492399616
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024901010101010103,
      "loss": 1.4333,
      "theoretical_loss": 3.3693346937456043,
      "tokens_seen": 2492465152
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000248989898989899,
      "loss": 1.4046,
      "theoretical_loss": 3.369327622920177,
      "tokens_seen": 2492530688
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000248969696969697,
      "loss": 1.0686,
      "theoretical_loss": 3.369320552332714,
      "tokens_seen": 2492596224
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024894949494949495,
      "loss": 1.4414,
      "theoretical_loss": 3.3693134819832014,
      "tokens_seen": 2492661760
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002489292929292929,
      "loss": 1.1459,
      "theoretical_loss": 3.3693064118716247,
      "tokens_seen": 2492727296
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002489090909090909,
      "loss": 1.3525,
      "theoretical_loss": 3.36929934199797,
      "tokens_seen": 2492792832
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488888888888889,
      "loss": 1.6536,
      "theoretical_loss": 3.3692922723622223,
      "tokens_seen": 2492858368
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488686868686869,
      "loss": 1.5337,
      "theoretical_loss": 3.3692852029643685,
      "tokens_seen": 2492923904
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024884848484848485,
      "loss": 1.4864,
      "theoretical_loss": 3.369278133804393,
      "tokens_seen": 2492989440
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488282828282828,
      "loss": 1.5196,
      "theoretical_loss": 3.3692710648822826,
      "tokens_seen": 2493054976
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002488080808080808,
      "loss": 1.7233,
      "theoretical_loss": 3.369263996198022,
      "tokens_seen": 2493120512
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024878787878787877,
      "loss": 1.4277,
      "theoretical_loss": 3.3692569277515982,
      "tokens_seen": 2493186048
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487676767676768,
      "loss": 1.3315,
      "theoretical_loss": 3.3692498595429963,
      "tokens_seen": 2493251584
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024874747474747474,
      "loss": 1.3936,
      "theoretical_loss": 3.369242791572202,
      "tokens_seen": 2493317120
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487272727272727,
      "loss": 1.3622,
      "theoretical_loss": 3.369235723839201,
      "tokens_seen": 2493382656
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002487070707070707,
      "loss": 1.3661,
      "theoretical_loss": 3.369228656343979,
      "tokens_seen": 2493448192
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024868686868686866,
      "loss": 1.2716,
      "theoretical_loss": 3.3692215890865223,
      "tokens_seen": 2493513728
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486666666666667,
      "loss": 1.4645,
      "theoretical_loss": 3.369214522066816,
      "tokens_seen": 2493579264
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024864646464646463,
      "loss": 1.5887,
      "theoretical_loss": 3.3692074552848457,
      "tokens_seen": 2493644800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024862626262626264,
      "loss": 1.6292,
      "theoretical_loss": 3.369200388740598,
      "tokens_seen": 2493710336
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002486060606060606,
      "loss": 1.4688,
      "theoretical_loss": 3.3691933224340587,
      "tokens_seen": 2493775872
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024858585858585856,
      "loss": 1.4441,
      "theoretical_loss": 3.3691862563652126,
      "tokens_seen": 2493841408
    },
    {
      "epoch": 0.51,
      "objective/train/aligned_token_count": 0.43658357858657837,
      "objective/train/avg_token_score": 0.014538480900228024,
      "objective/train/docs_used": 1402946,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.445590615272522,
      "objective/train/original_loss": 5.081567287445068,
      "objective/train/theoretical_loss": 3.369179190534046,
      "objective/train/tokens_used": 852766176,
      "objective/train/unlikelihood_loss": 0.030677707865834236,
      "theoretical_loss": 3.369179190534046,
      "tokens_seen": 2493906944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024856565656565657,
      "loss": 1.6349,
      "theoretical_loss": 3.369179190534046,
      "tokens_seen": 2493906944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002485454545454545,
      "loss": 1.3772,
      "theoretical_loss": 3.3691721249405444,
      "tokens_seen": 2493972480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024852525252525254,
      "loss": 1.4035,
      "theoretical_loss": 3.3691650595846943,
      "tokens_seen": 2494038016
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002485050505050505,
      "loss": 1.4056,
      "theoretical_loss": 3.3691579944664802,
      "tokens_seen": 2494103552
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002484848484848485,
      "loss": 1.3458,
      "theoretical_loss": 3.3691509295858895,
      "tokens_seen": 2494169088
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024846464646464646,
      "loss": 1.4873,
      "theoretical_loss": 3.3691438649429064,
      "tokens_seen": 2494234624
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024844444444444447,
      "loss": 1.4553,
      "theoretical_loss": 3.3691368005375173,
      "tokens_seen": 2494300160
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024842424242424243,
      "loss": 1.3286,
      "theoretical_loss": 3.3691297363697084,
      "tokens_seen": 2494365696
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002484040404040404,
      "loss": 1.4596,
      "theoretical_loss": 3.369122672439465,
      "tokens_seen": 2494431232
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002483838383838384,
      "loss": 1.4079,
      "theoretical_loss": 3.3691156087467724,
      "tokens_seen": 2494496768
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024836363636363635,
      "loss": 1.1054,
      "theoretical_loss": 3.3691085452916174,
      "tokens_seen": 2494562304
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024834343434343436,
      "loss": 1.3578,
      "theoretical_loss": 3.369101482073985,
      "tokens_seen": 2494627840
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002483232323232323,
      "loss": 1.4659,
      "theoretical_loss": 3.3690944190938614,
      "tokens_seen": 2494693376
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024830303030303033,
      "loss": 1.3168,
      "theoretical_loss": 3.3690873563512325,
      "tokens_seen": 2494758912
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002482828282828283,
      "loss": 1.562,
      "theoretical_loss": 3.3690802938460833,
      "tokens_seen": 2494824448
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024826262626262625,
      "loss": 1.5357,
      "theoretical_loss": 3.3690732315784,
      "tokens_seen": 2494889984
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024824242424242426,
      "loss": 1.6102,
      "theoretical_loss": 3.369066169548169,
      "tokens_seen": 2494955520
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002482222222222222,
      "loss": 1.2489,
      "theoretical_loss": 3.3690591077553753,
      "tokens_seen": 2495021056
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002482020202020202,
      "loss": 1.4169,
      "theoretical_loss": 3.3690520462000046,
      "tokens_seen": 2495086592
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481818181818182,
      "loss": 1.3514,
      "theoretical_loss": 3.369044984882043,
      "tokens_seen": 2495152128
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481616161616162,
      "loss": 1.3214,
      "theoretical_loss": 3.369037923801477,
      "tokens_seen": 2495217664
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024814141414141415,
      "loss": 1.3114,
      "theoretical_loss": 3.369030862958291,
      "tokens_seen": 2495283200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024812121212121216,
      "loss": 1.6165,
      "theoretical_loss": 3.369023802352472,
      "tokens_seen": 2495348736
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002481010101010101,
      "loss": 1.367,
      "theoretical_loss": 3.3690167419840047,
      "tokens_seen": 2495414272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002480808080808081,
      "loss": 1.1903,
      "theoretical_loss": 3.3690096818528756,
      "tokens_seen": 2495479808
    },
    {
      "epoch": 0.51,
      "objective/train/aligned_token_count": 0.2968597114086151,
      "objective/train/avg_token_score": 0.014619704335927963,
      "objective/train/docs_used": 1404179,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9112189412117004,
      "objective/train/original_loss": 4.859391212463379,
      "objective/train/theoretical_loss": 3.36900262195907,
      "objective/train/tokens_used": 854404576,
      "objective/train/unlikelihood_loss": 0.05202590674161911,
      "theoretical_loss": 3.36900262195907,
      "tokens_seen": 2495545344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002480606060606061,
      "loss": 1.4674,
      "theoretical_loss": 3.36900262195907,
      "tokens_seen": 2495545344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024804040404040404,
      "loss": 1.4478,
      "theoretical_loss": 3.3689955623025747,
      "tokens_seen": 2495610880
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024802020202020205,
      "loss": 1.1885,
      "theoretical_loss": 3.368988502883374,
      "tokens_seen": 2495676416
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000248,
      "loss": 1.48,
      "theoretical_loss": 3.368981443701455,
      "tokens_seen": 2495741952
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024797979797979797,
      "loss": 1.4643,
      "theoretical_loss": 3.368974384756803,
      "tokens_seen": 2495807488
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000247959595959596,
      "loss": 1.498,
      "theoretical_loss": 3.3689673260494035,
      "tokens_seen": 2495873024
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024793939393939394,
      "loss": 1.2376,
      "theoretical_loss": 3.3689602675792427,
      "tokens_seen": 2495938560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024791919191919195,
      "loss": 1.4132,
      "theoretical_loss": 3.368953209346306,
      "tokens_seen": 2496004096
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002478989898989899,
      "loss": 1.3149,
      "theoretical_loss": 3.36894615135058,
      "tokens_seen": 2496069632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024787878787878786,
      "loss": 1.4376,
      "theoretical_loss": 3.36893909359205,
      "tokens_seen": 2496135168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024785858585858587,
      "loss": 1.4545,
      "theoretical_loss": 3.368932036070701,
      "tokens_seen": 2496200704
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024783838383838383,
      "loss": 1.3839,
      "theoretical_loss": 3.36892497878652,
      "tokens_seen": 2496266240
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024781818181818184,
      "loss": 1.3785,
      "theoretical_loss": 3.3689179217394924,
      "tokens_seen": 2496331776
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002477979797979798,
      "loss": 1.1703,
      "theoretical_loss": 3.368910864929604,
      "tokens_seen": 2496397312
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002477777777777778,
      "loss": 1.4633,
      "theoretical_loss": 3.3689038083568406,
      "tokens_seen": 2496462848
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024775757575757576,
      "loss": 1.3636,
      "theoretical_loss": 3.368896752021188,
      "tokens_seen": 2496528384
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002477373737373737,
      "loss": 1.452,
      "theoretical_loss": 3.3688896959226318,
      "tokens_seen": 2496593920
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024771717171717173,
      "loss": 1.2071,
      "theoretical_loss": 3.368882640061158,
      "tokens_seen": 2496659456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476969696969697,
      "loss": 1.272,
      "theoretical_loss": 3.3688755844367524,
      "tokens_seen": 2496724992
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476767676767677,
      "loss": 1.3351,
      "theoretical_loss": 3.368868529049401,
      "tokens_seen": 2496790528
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024765656565656566,
      "loss": 1.278,
      "theoretical_loss": 3.3688614738990896,
      "tokens_seen": 2496856064
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476363636363636,
      "loss": 1.3449,
      "theoretical_loss": 3.368854418985804,
      "tokens_seen": 2496921600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002476161616161616,
      "loss": 1.3531,
      "theoretical_loss": 3.3688473643095294,
      "tokens_seen": 2496987136
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475959595959596,
      "loss": 1.5048,
      "theoretical_loss": 3.368840309870252,
      "tokens_seen": 2497052672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475757575757576,
      "loss": 1.5558,
      "theoretical_loss": 3.3688332556679583,
      "tokens_seen": 2497118208
    },
    {
      "epoch": 0.51,
      "objective/train/aligned_token_count": 0.45747801661491394,
      "objective/train/avg_token_score": 0.027921805158257484,
      "objective/train/docs_used": 1405043,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4367603063583374,
      "objective/train/original_loss": 4.823309898376465,
      "objective/train/theoretical_loss": 3.368826201702633,
      "objective/train/tokens_used": 856042976,
      "objective/train/unlikelihood_loss": 0.06313079595565796,
      "theoretical_loss": 3.368826201702633,
      "tokens_seen": 2497183744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024755555555555555,
      "loss": 1.5369,
      "theoretical_loss": 3.368826201702633,
      "tokens_seen": 2497183744
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475353535353535,
      "loss": 1.3028,
      "theoretical_loss": 3.368819147974263,
      "tokens_seen": 2497249280
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002475151515151515,
      "loss": 1.3345,
      "theoretical_loss": 3.3688120944828333,
      "tokens_seen": 2497314816
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474949494949495,
      "loss": 1.4315,
      "theoretical_loss": 3.36880504122833,
      "tokens_seen": 2497380352
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474747474747475,
      "loss": 1.4055,
      "theoretical_loss": 3.368797988210739,
      "tokens_seen": 2497445888
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024745454545454544,
      "loss": 1.4877,
      "theoretical_loss": 3.3687909354300456,
      "tokens_seen": 2497511424
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024743434343434345,
      "loss": 1.5272,
      "theoretical_loss": 3.3687838828862366,
      "tokens_seen": 2497576960
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002474141414141414,
      "loss": 1.39,
      "theoretical_loss": 3.368776830579297,
      "tokens_seen": 2497642496
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024739393939393937,
      "loss": 1.3589,
      "theoretical_loss": 3.368769778509213,
      "tokens_seen": 2497708032
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002473737373737374,
      "loss": 1.5127,
      "theoretical_loss": 3.3687627266759703,
      "tokens_seen": 2497773568
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024735353535353533,
      "loss": 1.5,
      "theoretical_loss": 3.368755675079555,
      "tokens_seen": 2497839104
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024733333333333335,
      "loss": 1.3771,
      "theoretical_loss": 3.3687486237199527,
      "tokens_seen": 2497904640
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002473131313131313,
      "loss": 1.5251,
      "theoretical_loss": 3.3687415725971492,
      "tokens_seen": 2497970176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002472929292929293,
      "loss": 1.3833,
      "theoretical_loss": 3.36873452171113,
      "tokens_seen": 2498035712
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024727272727272727,
      "loss": 1.3837,
      "theoretical_loss": 3.368727471061882,
      "tokens_seen": 2498101248
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002472525252525253,
      "loss": 1.4871,
      "theoretical_loss": 3.3687204206493906,
      "tokens_seen": 2498166784
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024723232323232324,
      "loss": 1.1555,
      "theoretical_loss": 3.3687133704736407,
      "tokens_seen": 2498232320
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002472121212121212,
      "loss": 1.3607,
      "theoretical_loss": 3.368706320534619,
      "tokens_seen": 2498297856
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002471919191919192,
      "loss": 1.6154,
      "theoretical_loss": 3.3686992708323116,
      "tokens_seen": 2498363392
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024717171717171716,
      "loss": 1.299,
      "theoretical_loss": 3.3686922213667034,
      "tokens_seen": 2498428928
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002471515151515152,
      "loss": 1.2941,
      "theoretical_loss": 3.368685172137781,
      "tokens_seen": 2498494464
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024713131313131313,
      "loss": 1.452,
      "theoretical_loss": 3.3686781231455303,
      "tokens_seen": 2498560000
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024711111111111114,
      "loss": 1.3157,
      "theoretical_loss": 3.3686710743899364,
      "tokens_seen": 2498625536
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002470909090909091,
      "loss": 1.5576,
      "theoretical_loss": 3.368664025870986,
      "tokens_seen": 2498691072
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002470707070707071,
      "loss": 1.2198,
      "theoretical_loss": 3.3686569775886643,
      "tokens_seen": 2498756608
    },
    {
      "epoch": 0.51,
      "objective/train/aligned_token_count": 0.48075512051582336,
      "objective/train/avg_token_score": 0.012280876748263836,
      "objective/train/docs_used": 1405984,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5566829442977905,
      "objective/train/original_loss": 4.972794055938721,
      "objective/train/theoretical_loss": 3.3686499295429577,
      "objective/train/tokens_used": 857681376,
      "objective/train/unlikelihood_loss": 0.03961954265832901,
      "theoretical_loss": 3.3686499295429577,
      "tokens_seen": 2498822144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024705050505050507,
      "loss": 1.2842,
      "theoretical_loss": 3.3686499295429577,
      "tokens_seen": 2498822144
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000247030303030303,
      "loss": 1.5872,
      "theoretical_loss": 3.3686428817338516,
      "tokens_seen": 2498887680
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024701010101010104,
      "loss": 1.548,
      "theoretical_loss": 3.368635834161332,
      "tokens_seen": 2498953216
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000246989898989899,
      "loss": 1.2442,
      "theoretical_loss": 3.3686287868253846,
      "tokens_seen": 2499018752
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000246969696969697,
      "loss": 1.3154,
      "theoretical_loss": 3.3686217397259957,
      "tokens_seen": 2499084288
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024694949494949496,
      "loss": 1.4313,
      "theoretical_loss": 3.368614692863151,
      "tokens_seen": 2499149824
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024692929292929297,
      "loss": 1.2399,
      "theoretical_loss": 3.368607646236836,
      "tokens_seen": 2499215360
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024690909090909093,
      "loss": 1.5227,
      "theoretical_loss": 3.3686005998470367,
      "tokens_seen": 2499280896
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002468888888888889,
      "loss": 1.2764,
      "theoretical_loss": 3.3685935536937395,
      "tokens_seen": 2499346432
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002468686868686869,
      "loss": 1.4164,
      "theoretical_loss": 3.3685865077769295,
      "tokens_seen": 2499411968
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024684848484848485,
      "loss": 1.3448,
      "theoretical_loss": 3.368579462096593,
      "tokens_seen": 2499477504
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024682828282828286,
      "loss": 1.6273,
      "theoretical_loss": 3.368572416652716,
      "tokens_seen": 2499543040
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002468080808080808,
      "loss": 1.1861,
      "theoretical_loss": 3.3685653714452837,
      "tokens_seen": 2499608576
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002467878787878788,
      "loss": 1.2842,
      "theoretical_loss": 3.3685583264742824,
      "tokens_seen": 2499674112
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002467676767676768,
      "loss": 1.5467,
      "theoretical_loss": 3.3685512817396983,
      "tokens_seen": 2499739648
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024674747474747475,
      "loss": 1.4684,
      "theoretical_loss": 3.3685442372415166,
      "tokens_seen": 2499805184
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024672727272727276,
      "loss": 1.3063,
      "theoretical_loss": 3.3685371929797236,
      "tokens_seen": 2499870720
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002467070707070707,
      "loss": 1.5055,
      "theoretical_loss": 3.368530148954305,
      "tokens_seen": 2499936256
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024668686868686867,
      "loss": 1.4102,
      "theoretical_loss": 3.368523105165247,
      "tokens_seen": 2500001792
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002466666666666667,
      "loss": 1.2494,
      "theoretical_loss": 3.3685160616125347,
      "tokens_seen": 2500067328
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024664646464646464,
      "loss": 1.6077,
      "theoretical_loss": 3.368509018296155,
      "tokens_seen": 2500132864
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024662626262626265,
      "loss": 1.521,
      "theoretical_loss": 3.368501975216093,
      "tokens_seen": 2500198400
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002466060606060606,
      "loss": 1.3695,
      "theoretical_loss": 3.368494932372335,
      "tokens_seen": 2500263936
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024658585858585856,
      "loss": 1.568,
      "theoretical_loss": 3.3684878897648662,
      "tokens_seen": 2500329472
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002465656565656566,
      "loss": 1.3616,
      "theoretical_loss": 3.3684808473936734,
      "tokens_seen": 2500395008
    },
    {
      "epoch": 0.51,
      "objective/train/aligned_token_count": 0.3404814302921295,
      "objective/train/avg_token_score": 0.01239077840000391,
      "objective/train/docs_used": 1406694,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2070962190628052,
      "objective/train/original_loss": 5.5607171058654785,
      "objective/train/theoretical_loss": 3.368473805258742,
      "objective/train/tokens_used": 859319776,
      "objective/train/unlikelihood_loss": 0.03527798876166344,
      "theoretical_loss": 3.368473805258742,
      "tokens_seen": 2500460544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024654545454545453,
      "loss": 1.3817,
      "theoretical_loss": 3.368473805258742,
      "tokens_seen": 2500460544
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024652525252525254,
      "loss": 1.485,
      "theoretical_loss": 3.3684667633600585,
      "tokens_seen": 2500526080
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002465050505050505,
      "loss": 1.5298,
      "theoretical_loss": 3.3684597216976075,
      "tokens_seen": 2500591616
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024648484848484846,
      "loss": 1.4577,
      "theoretical_loss": 3.3684526802713757,
      "tokens_seen": 2500657152
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024646464646464647,
      "loss": 1.4272,
      "theoretical_loss": 3.368445639081349,
      "tokens_seen": 2500722688
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002464444444444444,
      "loss": 1.4931,
      "theoretical_loss": 3.3684385981275136,
      "tokens_seen": 2500788224
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024642424242424243,
      "loss": 1.467,
      "theoretical_loss": 3.3684315574098544,
      "tokens_seen": 2500853760
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002464040404040404,
      "loss": 1.3092,
      "theoretical_loss": 3.368424516928358,
      "tokens_seen": 2500919296
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002463838383838384,
      "loss": 1.5952,
      "theoretical_loss": 3.3684174766830104,
      "tokens_seen": 2500984832
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024636363636363636,
      "loss": 1.2555,
      "theoretical_loss": 3.3684104366737975,
      "tokens_seen": 2501050368
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002463434343434343,
      "loss": 1.4893,
      "theoretical_loss": 3.368403396900704,
      "tokens_seen": 2501115904
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024632323232323233,
      "loss": 1.3307,
      "theoretical_loss": 3.368396357363718,
      "tokens_seen": 2501181440
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002463030303030303,
      "loss": 1.4538,
      "theoretical_loss": 3.368389318062823,
      "tokens_seen": 2501246976
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002462828282828283,
      "loss": 1.3046,
      "theoretical_loss": 3.3683822789980065,
      "tokens_seen": 2501312512
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024626262626262625,
      "loss": 1.535,
      "theoretical_loss": 3.3683752401692537,
      "tokens_seen": 2501378048
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024624242424242426,
      "loss": 1.2071,
      "theoretical_loss": 3.3683682015765513,
      "tokens_seen": 2501443584
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002462222222222222,
      "loss": 1.1556,
      "theoretical_loss": 3.368361163219884,
      "tokens_seen": 2501509120
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024620202020202023,
      "loss": 1.3729,
      "theoretical_loss": 3.3683541250992386,
      "tokens_seen": 2501574656
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002461818181818182,
      "loss": 1.4409,
      "theoretical_loss": 3.368347087214601,
      "tokens_seen": 2501640192
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024616161616161614,
      "loss": 1.3922,
      "theoretical_loss": 3.368340049565956,
      "tokens_seen": 2501705728
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024614141414141416,
      "loss": 1.1961,
      "theoretical_loss": 3.368333012153291,
      "tokens_seen": 2501771264
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002461212121212121,
      "loss": 1.2812,
      "theoretical_loss": 3.368325974976591,
      "tokens_seen": 2501836800
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002461010101010101,
      "loss": 1.4841,
      "theoretical_loss": 3.3683189380358423,
      "tokens_seen": 2501902336
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002460808080808081,
      "loss": 1.5192,
      "theoretical_loss": 3.3683119013310305,
      "tokens_seen": 2501967872
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002460606060606061,
      "loss": 1.6896,
      "theoretical_loss": 3.368304864862142,
      "tokens_seen": 2502033408
    },
    {
      "epoch": 0.51,
      "objective/train/aligned_token_count": 0.4784335196018219,
      "objective/train/avg_token_score": 0.008035477250814438,
      "objective/train/docs_used": 1408162,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6085033416748047,
      "objective/train/original_loss": 4.389880180358887,
      "objective/train/theoretical_loss": 3.3682978286291623,
      "objective/train/tokens_used": 860958176,
      "objective/train/unlikelihood_loss": 0.05494939535856247,
      "theoretical_loss": 3.3682978286291623,
      "tokens_seen": 2502098944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024604040404040405,
      "loss": 1.5126,
      "theoretical_loss": 3.3682978286291623,
      "tokens_seen": 2502098944
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024602020202020206,
      "loss": 1.5901,
      "theoretical_loss": 3.3682907926320773,
      "tokens_seen": 2502164480
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000246,
      "loss": 1.3441,
      "theoretical_loss": 3.3682837568708726,
      "tokens_seen": 2502230016
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000245979797979798,
      "loss": 1.1168,
      "theoretical_loss": 3.368276721345535,
      "tokens_seen": 2502295552
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.000245959595959596,
      "loss": 1.4756,
      "theoretical_loss": 3.36826968605605,
      "tokens_seen": 2502361088
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024593939393939394,
      "loss": 1.3987,
      "theoretical_loss": 3.368262651002403,
      "tokens_seen": 2502426624
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024591919191919195,
      "loss": 1.3039,
      "theoretical_loss": 3.368255616184581,
      "tokens_seen": 2502492160
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002458989898989899,
      "loss": 1.3966,
      "theoretical_loss": 3.3682485816025687,
      "tokens_seen": 2502557696
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002458787878787879,
      "loss": 1.259,
      "theoretical_loss": 3.368241547256353,
      "tokens_seen": 2502623232
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002458585858585859,
      "loss": 1.3117,
      "theoretical_loss": 3.3682345131459193,
      "tokens_seen": 2502688768
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024583838383838383,
      "loss": 1.3671,
      "theoretical_loss": 3.3682274792712534,
      "tokens_seen": 2502754304
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024581818181818184,
      "loss": 1.4786,
      "theoretical_loss": 3.3682204456323417,
      "tokens_seen": 2502819840
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002457979797979798,
      "loss": 1.4666,
      "theoretical_loss": 3.36821341222917,
      "tokens_seen": 2502885376
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002457777777777778,
      "loss": 1.4423,
      "theoretical_loss": 3.368206379061724,
      "tokens_seen": 2502950912
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024575757575757577,
      "loss": 1.5892,
      "theoretical_loss": 3.36819934612999,
      "tokens_seen": 2503016448
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002457373737373737,
      "loss": 1.3087,
      "theoretical_loss": 3.3681923134339535,
      "tokens_seen": 2503081984
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024571717171717174,
      "loss": 1.5505,
      "theoretical_loss": 3.3681852809736004,
      "tokens_seen": 2503147520
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002456969696969697,
      "loss": 1.5905,
      "theoretical_loss": 3.368178248748917,
      "tokens_seen": 2503213056
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002456767676767677,
      "loss": 1.6489,
      "theoretical_loss": 3.3681712167598894,
      "tokens_seen": 2503278592
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024565656565656566,
      "loss": 1.5148,
      "theoretical_loss": 3.3681641850065027,
      "tokens_seen": 2503344128
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002456363636363636,
      "loss": 1.317,
      "theoretical_loss": 3.3681571534887436,
      "tokens_seen": 2503409664
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024561616161616163,
      "loss": 1.2316,
      "theoretical_loss": 3.3681501222065977,
      "tokens_seen": 2503475200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455959595959596,
      "loss": 1.5978,
      "theoretical_loss": 3.3681430911600514,
      "tokens_seen": 2503540736
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455757575757576,
      "loss": 1.4677,
      "theoretical_loss": 3.36813606034909,
      "tokens_seen": 2503606272
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024555555555555556,
      "loss": 1.4279,
      "theoretical_loss": 3.3681290297736997,
      "tokens_seen": 2503671808
    },
    {
      "epoch": 0.51,
      "objective/train/aligned_token_count": 0.34811827540397644,
      "objective/train/avg_token_score": 0.023100661113858223,
      "objective/train/docs_used": 1408714,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.042260766029358,
      "objective/train/original_loss": 4.947227954864502,
      "objective/train/theoretical_loss": 3.3681219994338663,
      "objective/train/tokens_used": 862596576,
      "objective/train/unlikelihood_loss": 0.06410616636276245,
      "theoretical_loss": 3.3681219994338663,
      "tokens_seen": 2503737344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455353535353535,
      "loss": 1.1423,
      "theoretical_loss": 3.3681219994338663,
      "tokens_seen": 2503737344
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002455151515151515,
      "loss": 1.3859,
      "theoretical_loss": 3.368114969329576,
      "tokens_seen": 2503802880
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454949494949495,
      "loss": 1.3371,
      "theoretical_loss": 3.3681079394608147,
      "tokens_seen": 2503868416
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454747474747475,
      "loss": 1.4635,
      "theoretical_loss": 3.368100909827568,
      "tokens_seen": 2503933952
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024545454545454545,
      "loss": 1.5162,
      "theoretical_loss": 3.3680938804298224,
      "tokens_seen": 2503999488
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454343434343434,
      "loss": 1.3081,
      "theoretical_loss": 3.3680868512675635,
      "tokens_seen": 2504065024
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002454141414141414,
      "loss": 1.6097,
      "theoretical_loss": 3.3680798223407775,
      "tokens_seen": 2504130560
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024539393939393937,
      "loss": 1.4297,
      "theoretical_loss": 3.36807279364945,
      "tokens_seen": 2504196096
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002453737373737374,
      "loss": 1.2523,
      "theoretical_loss": 3.368065765193567,
      "tokens_seen": 2504261632
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024535353535353534,
      "loss": 1.3043,
      "theoretical_loss": 3.3680587369731145,
      "tokens_seen": 2504327168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024533333333333335,
      "loss": 1.4625,
      "theoretical_loss": 3.3680517089880784,
      "tokens_seen": 2504392704
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002453131313131313,
      "loss": 1.3976,
      "theoretical_loss": 3.3680446812384455,
      "tokens_seen": 2504458240
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024529292929292927,
      "loss": 1.4195,
      "theoretical_loss": 3.3680376537242003,
      "tokens_seen": 2504523776
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002452727272727273,
      "loss": 1.5716,
      "theoretical_loss": 3.36803062644533,
      "tokens_seen": 2504589312
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024525252525252523,
      "loss": 1.3063,
      "theoretical_loss": 3.3680235994018197,
      "tokens_seen": 2504654848
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024523232323232324,
      "loss": 1.4859,
      "theoretical_loss": 3.368016572593656,
      "tokens_seen": 2504720384
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002452121212121212,
      "loss": 1.4174,
      "theoretical_loss": 3.3680095460208244,
      "tokens_seen": 2504785920
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002451919191919192,
      "loss": 1.2372,
      "theoretical_loss": 3.3680025196833108,
      "tokens_seen": 2504851456
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024517171717171717,
      "loss": 1.4963,
      "theoretical_loss": 3.367995493581102,
      "tokens_seen": 2504916992
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002451515151515152,
      "loss": 1.5618,
      "theoretical_loss": 3.3679884677141825,
      "tokens_seen": 2504982528
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024513131313131314,
      "loss": 1.4353,
      "theoretical_loss": 3.36798144208254,
      "tokens_seen": 2505048064
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002451111111111111,
      "loss": 1.2059,
      "theoretical_loss": 3.3679744166861587,
      "tokens_seen": 2505113600
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002450909090909091,
      "loss": 1.4318,
      "theoretical_loss": 3.367967391525026,
      "tokens_seen": 2505179136
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024507070707070706,
      "loss": 1.561,
      "theoretical_loss": 3.367960366599127,
      "tokens_seen": 2505244672
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00024505050505050507,
      "loss": 1.4972,
      "theoretical_loss": 3.3679533419084486,
      "tokens_seen": 2505310208
    },
    {
      "epoch": 0.51,
      "objective/train/aligned_token_count": 0.3787267804145813,
      "objective/train/avg_token_score": 0.046518296003341675,
      "objective/train/docs_used": 1409820,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.226811170578003,
      "objective/train/original_loss": 4.707108020782471,
      "objective/train/theoretical_loss": 3.367946317452976,
      "objective/train/tokens_used": 864234976,
      "objective/train/unlikelihood_loss": 0.050696276128292084,
      "theoretical_loss": 3.367946317452976,
      "tokens_seen": 2505375744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024503030303030303,
      "loss": 1.4178,
      "theoretical_loss": 3.367946317452976,
      "tokens_seen": 2505375744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024501010101010104,
      "loss": 1.6504,
      "theoretical_loss": 3.367939293232695,
      "tokens_seen": 2505441280
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000244989898989899,
      "loss": 1.5345,
      "theoretical_loss": 3.367932269247592,
      "tokens_seen": 2505506816
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024496969696969695,
      "loss": 1.263,
      "theoretical_loss": 3.367925245497653,
      "tokens_seen": 2505572352
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024494949494949497,
      "loss": 1.3678,
      "theoretical_loss": 3.3679182219828636,
      "tokens_seen": 2505637888
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002449292929292929,
      "loss": 1.8001,
      "theoretical_loss": 3.36791119870321,
      "tokens_seen": 2505703424
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024490909090909093,
      "loss": 1.4425,
      "theoretical_loss": 3.3679041756586785,
      "tokens_seen": 2505768960
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448888888888889,
      "loss": 1.284,
      "theoretical_loss": 3.3678971528492547,
      "tokens_seen": 2505834496
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448686868686869,
      "loss": 1.3788,
      "theoretical_loss": 3.3678901302749247,
      "tokens_seen": 2505900032
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024484848484848486,
      "loss": 1.6165,
      "theoretical_loss": 3.3678831079356746,
      "tokens_seen": 2505965568
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024482828282828287,
      "loss": 1.3843,
      "theoretical_loss": 3.3678760858314902,
      "tokens_seen": 2506031104
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002448080808080808,
      "loss": 1.4714,
      "theoretical_loss": 3.367869063962357,
      "tokens_seen": 2506096640
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002447878787878788,
      "loss": 1.3841,
      "theoretical_loss": 3.3678620423282624,
      "tokens_seen": 2506162176
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002447676767676768,
      "loss": 1.4803,
      "theoretical_loss": 3.367855020929191,
      "tokens_seen": 2506227712
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024474747474747475,
      "loss": 1.2718,
      "theoretical_loss": 3.367847999765129,
      "tokens_seen": 2506293248
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024472727272727276,
      "loss": 1.42,
      "theoretical_loss": 3.367840978836063,
      "tokens_seen": 2506358784
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002447070707070707,
      "loss": 1.5132,
      "theoretical_loss": 3.3678339581419787,
      "tokens_seen": 2506424320
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002446868686868687,
      "loss": 1.3202,
      "theoretical_loss": 3.367826937682862,
      "tokens_seen": 2506489856
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002446666666666667,
      "loss": 1.4561,
      "theoretical_loss": 3.367819917458699,
      "tokens_seen": 2506555392
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024464646464646464,
      "loss": 1.4049,
      "theoretical_loss": 3.3678128974694754,
      "tokens_seen": 2506620928
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024462626262626265,
      "loss": 1.3307,
      "theoretical_loss": 3.367805877715178,
      "tokens_seen": 2506686464
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002446060606060606,
      "loss": 1.5328,
      "theoretical_loss": 3.367798858195792,
      "tokens_seen": 2506752000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024458585858585857,
      "loss": 1.4522,
      "theoretical_loss": 3.3677918389113035,
      "tokens_seen": 2506817536
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002445656565656566,
      "loss": 1.2992,
      "theoretical_loss": 3.3677848198616984,
      "tokens_seen": 2506883072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024454545454545454,
      "loss": 1.2208,
      "theoretical_loss": 3.3677778010469632,
      "tokens_seen": 2506948608
    },
    {
      "epoch": 0.52,
      "objective/train/aligned_token_count": 0.33070626854896545,
      "objective/train/avg_token_score": 0.016669392585754395,
      "objective/train/docs_used": 1410261,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9412729144096375,
      "objective/train/original_loss": 4.611757755279541,
      "objective/train/theoretical_loss": 3.3677707824670837,
      "objective/train/tokens_used": 865873376,
      "objective/train/unlikelihood_loss": 0.045376867055892944,
      "theoretical_loss": 3.3677707824670837,
      "tokens_seen": 2507014144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024452525252525255,
      "loss": 1.2532,
      "theoretical_loss": 3.3677707824670837,
      "tokens_seen": 2507014144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002445050505050505,
      "loss": 1.5426,
      "theoretical_loss": 3.3677637641220457,
      "tokens_seen": 2507079680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024448484848484846,
      "loss": 1.2968,
      "theoretical_loss": 3.3677567460118354,
      "tokens_seen": 2507145216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024446464646464647,
      "loss": 1.3033,
      "theoretical_loss": 3.3677497281364386,
      "tokens_seen": 2507210752
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024444444444444443,
      "loss": 1.3379,
      "theoretical_loss": 3.3677427104958415,
      "tokens_seen": 2507276288
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024442424242424244,
      "loss": 1.4095,
      "theoretical_loss": 3.3677356930900304,
      "tokens_seen": 2507341824
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002444040404040404,
      "loss": 1.3569,
      "theoretical_loss": 3.3677286759189906,
      "tokens_seen": 2507407360
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002443838383838384,
      "loss": 1.2472,
      "theoretical_loss": 3.367721658982709,
      "tokens_seen": 2507472896
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024436363636363636,
      "loss": 1.4649,
      "theoretical_loss": 3.3677146422811703,
      "tokens_seen": 2507538432
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002443434343434343,
      "loss": 1.2672,
      "theoretical_loss": 3.3677076258143614,
      "tokens_seen": 2507603968
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024432323232323233,
      "loss": 1.3987,
      "theoretical_loss": 3.367700609582269,
      "tokens_seen": 2507669504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002443030303030303,
      "loss": 1.4397,
      "theoretical_loss": 3.3676935935848777,
      "tokens_seen": 2507735040
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002442828282828283,
      "loss": 1.4465,
      "theoretical_loss": 3.367686577822174,
      "tokens_seen": 2507800576
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024426262626262626,
      "loss": 1.3351,
      "theoretical_loss": 3.3676795622941444,
      "tokens_seen": 2507866112
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002442424242424242,
      "loss": 1.4638,
      "theoretical_loss": 3.3676725470007747,
      "tokens_seen": 2507931648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002442222222222222,
      "loss": 1.4184,
      "theoretical_loss": 3.3676655319420505,
      "tokens_seen": 2507997184
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002442020202020202,
      "loss": 1.545,
      "theoretical_loss": 3.3676585171179583,
      "tokens_seen": 2508062720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002441818181818182,
      "loss": 1.336,
      "theoretical_loss": 3.3676515025284837,
      "tokens_seen": 2508128256
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024416161616161615,
      "loss": 1.3181,
      "theoretical_loss": 3.367644488173613,
      "tokens_seen": 2508193792
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024414141414141416,
      "loss": 1.3648,
      "theoretical_loss": 3.3676374740533324,
      "tokens_seen": 2508259328
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024412121212121215,
      "loss": 1.4751,
      "theoretical_loss": 3.3676304601676277,
      "tokens_seen": 2508324864
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002441010101010101,
      "loss": 1.4195,
      "theoretical_loss": 3.3676234465164847,
      "tokens_seen": 2508390400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024408080808080809,
      "loss": 1.278,
      "theoretical_loss": 3.3676164330998897,
      "tokens_seen": 2508455936
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024406060606060607,
      "loss": 1.433,
      "theoretical_loss": 3.367609419917829,
      "tokens_seen": 2508521472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024404040404040403,
      "loss": 1.3142,
      "theoretical_loss": 3.367602406970288,
      "tokens_seen": 2508587008
    },
    {
      "epoch": 0.52,
      "objective/train/aligned_token_count": 0.4558895528316498,
      "objective/train/avg_token_score": 0.0033793437760323286,
      "objective/train/docs_used": 1411438,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.389076828956604,
      "objective/train/original_loss": 3.864685535430908,
      "objective/train/theoretical_loss": 3.367595394257253,
      "objective/train/tokens_used": 867511776,
      "objective/train/unlikelihood_loss": 0.07558061182498932,
      "theoretical_loss": 3.367595394257253,
      "tokens_seen": 2508652544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024402020202020204,
      "loss": 1.4676,
      "theoretical_loss": 3.367595394257253,
      "tokens_seen": 2508652544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000244,
      "loss": 1.4483,
      "theoretical_loss": 3.3675883817787104,
      "tokens_seen": 2508718080
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000243979797979798,
      "loss": 1.252,
      "theoretical_loss": 3.367581369534646,
      "tokens_seen": 2508783616
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024395959595959596,
      "loss": 1.6486,
      "theoretical_loss": 3.3675743575250454,
      "tokens_seen": 2508849152
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024393939393939392,
      "loss": 1.3788,
      "theoretical_loss": 3.367567345749895,
      "tokens_seen": 2508914688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024391919191919193,
      "loss": 1.5776,
      "theoretical_loss": 3.3675603342091813,
      "tokens_seen": 2508980224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002438989898989899,
      "loss": 1.4365,
      "theoretical_loss": 3.3675533229028893,
      "tokens_seen": 2509045760
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002438787878787879,
      "loss": 1.3308,
      "theoretical_loss": 3.367546311831006,
      "tokens_seen": 2509111296
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024385858585858586,
      "loss": 1.2505,
      "theoretical_loss": 3.3675393009935166,
      "tokens_seen": 2509176832
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024383838383838384,
      "loss": 1.3065,
      "theoretical_loss": 3.367532290390408,
      "tokens_seen": 2509242368
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024381818181818182,
      "loss": 1.3789,
      "theoretical_loss": 3.3675252800216655,
      "tokens_seen": 2509307904
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002437979797979798,
      "loss": 1.3043,
      "theoretical_loss": 3.3675182698872757,
      "tokens_seen": 2509373440
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002437777777777778,
      "loss": 1.4755,
      "theoretical_loss": 3.3675112599872246,
      "tokens_seen": 2509438976
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024375757575757575,
      "loss": 1.7508,
      "theoretical_loss": 3.367504250321498,
      "tokens_seen": 2509504512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024373737373737373,
      "loss": 1.3431,
      "theoretical_loss": 3.367497240890082,
      "tokens_seen": 2509570048
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024371717171717172,
      "loss": 1.1796,
      "theoretical_loss": 3.3674902316929622,
      "tokens_seen": 2509635584
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002436969696969697,
      "loss": 1.5486,
      "theoretical_loss": 3.3674832227301255,
      "tokens_seen": 2509701120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024367676767676768,
      "loss": 1.5071,
      "theoretical_loss": 3.3674762140015577,
      "tokens_seen": 2509766656
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024365656565656567,
      "loss": 1.3409,
      "theoretical_loss": 3.3674692055072444,
      "tokens_seen": 2509832192
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024363636363636362,
      "loss": 1.2663,
      "theoretical_loss": 3.367462197247172,
      "tokens_seen": 2509897728
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024361616161616164,
      "loss": 1.3766,
      "theoretical_loss": 3.367455189221327,
      "tokens_seen": 2509963264
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002435959595959596,
      "loss": 1.2853,
      "theoretical_loss": 3.367448181429695,
      "tokens_seen": 2510028800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024357575757575758,
      "loss": 1.446,
      "theoretical_loss": 3.3674411738722614,
      "tokens_seen": 2510094336
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024355555555555556,
      "loss": 1.4328,
      "theoretical_loss": 3.367434166549013,
      "tokens_seen": 2510159872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024353535353535352,
      "loss": 1.4627,
      "theoretical_loss": 3.367427159459936,
      "tokens_seen": 2510225408
    },
    {
      "epoch": 0.52,
      "objective/train/aligned_token_count": 0.5127077102661133,
      "objective/train/avg_token_score": 0.005105176474899054,
      "objective/train/docs_used": 1412115,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5258616209030151,
      "objective/train/original_loss": 3.8747799396514893,
      "objective/train/theoretical_loss": 3.3674201526050163,
      "objective/train/tokens_used": 869150176,
      "objective/train/unlikelihood_loss": 0.0522550605237484,
      "theoretical_loss": 3.3674201526050163,
      "tokens_seen": 2510290944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024351515151515153,
      "loss": 1.3539,
      "theoretical_loss": 3.3674201526050163,
      "tokens_seen": 2510290944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024349494949494949,
      "loss": 1.675,
      "theoretical_loss": 3.36741314598424,
      "tokens_seen": 2510356480
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002434747474747475,
      "loss": 1.2322,
      "theoretical_loss": 3.3674061395975925,
      "tokens_seen": 2510422016
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024345454545454545,
      "loss": 1.2633,
      "theoretical_loss": 3.3673991334450606,
      "tokens_seen": 2510487552
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024343434343434346,
      "loss": 1.4148,
      "theoretical_loss": 3.3673921275266303,
      "tokens_seen": 2510553088
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024341414141414142,
      "loss": 1.4231,
      "theoretical_loss": 3.367385121842288,
      "tokens_seen": 2510618624
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024339393939393938,
      "loss": 1.3749,
      "theoretical_loss": 3.3673781163920187,
      "tokens_seen": 2510684160
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002433737373737374,
      "loss": 1.4009,
      "theoretical_loss": 3.3673711111758093,
      "tokens_seen": 2510749696
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024335353535353535,
      "loss": 1.3862,
      "theoretical_loss": 3.3673641061936457,
      "tokens_seen": 2510815232
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024333333333333336,
      "loss": 1.3169,
      "theoretical_loss": 3.3673571014455135,
      "tokens_seen": 2510880768
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024331313131313131,
      "loss": 1.3848,
      "theoretical_loss": 3.3673500969314,
      "tokens_seen": 2510946304
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002432929292929293,
      "loss": 1.4875,
      "theoretical_loss": 3.3673430926512897,
      "tokens_seen": 2511011840
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024327272727272728,
      "loss": 1.4247,
      "theoretical_loss": 3.3673360886051698,
      "tokens_seen": 2511077376
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024325252525252527,
      "loss": 1.4595,
      "theoretical_loss": 3.367329084793026,
      "tokens_seen": 2511142912
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024323232323232325,
      "loss": 1.3541,
      "theoretical_loss": 3.3673220812148443,
      "tokens_seen": 2511208448
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002432121212121212,
      "loss": 1.3122,
      "theoretical_loss": 3.3673150778706114,
      "tokens_seen": 2511273984
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002431919191919192,
      "loss": 1.4181,
      "theoretical_loss": 3.367308074760312,
      "tokens_seen": 2511339520
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024317171717171717,
      "loss": 1.2577,
      "theoretical_loss": 3.367301071883934,
      "tokens_seen": 2511405056
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024315151515151516,
      "loss": 1.3505,
      "theoretical_loss": 3.3672940692414617,
      "tokens_seen": 2511470592
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024313131313131314,
      "loss": 1.3927,
      "theoretical_loss": 3.3672870668328825,
      "tokens_seen": 2511536128
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024311111111111113,
      "loss": 1.4479,
      "theoretical_loss": 3.367280064658182,
      "tokens_seen": 2511601664
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024309090909090908,
      "loss": 1.3151,
      "theoretical_loss": 3.367273062717346,
      "tokens_seen": 2511667200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002430707070707071,
      "loss": 1.4769,
      "theoretical_loss": 3.367266061010361,
      "tokens_seen": 2511732736
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024305050505050505,
      "loss": 1.2022,
      "theoretical_loss": 3.367259059537213,
      "tokens_seen": 2511798272
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024303030303030304,
      "loss": 1.3218,
      "theoretical_loss": 3.367252058297888,
      "tokens_seen": 2511863808
    },
    {
      "epoch": 0.52,
      "objective/train/aligned_token_count": 0.42173752188682556,
      "objective/train/avg_token_score": 0.012256687507033348,
      "objective/train/docs_used": 1412775,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6062207221984863,
      "objective/train/original_loss": 4.752782344818115,
      "objective/train/theoretical_loss": 3.3672450572923722,
      "objective/train/tokens_used": 870788576,
      "objective/train/unlikelihood_loss": 0.04705429449677467,
      "theoretical_loss": 3.3672450572923722,
      "tokens_seen": 2511929344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024301010101010102,
      "loss": 1.6329,
      "theoretical_loss": 3.3672450572923722,
      "tokens_seen": 2511929344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024298989898989898,
      "loss": 1.5765,
      "theoretical_loss": 3.3672380565206517,
      "tokens_seen": 2511994880
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000242969696969697,
      "loss": 1.5772,
      "theoretical_loss": 3.367231055982712,
      "tokens_seen": 2512060416
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024294949494949494,
      "loss": 1.4845,
      "theoretical_loss": 3.36722405567854,
      "tokens_seen": 2512125952
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024292929292929296,
      "loss": 1.3391,
      "theoretical_loss": 3.367217055608122,
      "tokens_seen": 2512191488
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002429090909090909,
      "loss": 1.2706,
      "theoretical_loss": 3.367210055771443,
      "tokens_seen": 2512257024
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024288888888888887,
      "loss": 1.5173,
      "theoretical_loss": 3.3672030561684902,
      "tokens_seen": 2512322560
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024286868686868688,
      "loss": 1.3846,
      "theoretical_loss": 3.3671960567992487,
      "tokens_seen": 2512388096
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024284848484848484,
      "loss": 1.6499,
      "theoretical_loss": 3.3671890576637056,
      "tokens_seen": 2512453632
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024282828282828285,
      "loss": 1.3318,
      "theoretical_loss": 3.367182058761846,
      "tokens_seen": 2512519168
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002428080808080808,
      "loss": 1.2871,
      "theoretical_loss": 3.367175060093657,
      "tokens_seen": 2512584704
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002427878787878788,
      "loss": 1.3309,
      "theoretical_loss": 3.367168061659124,
      "tokens_seen": 2512650240
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024276767676767677,
      "loss": 1.2197,
      "theoretical_loss": 3.3671610634582336,
      "tokens_seen": 2512715776
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024274747474747476,
      "loss": 1.6569,
      "theoretical_loss": 3.367154065490971,
      "tokens_seen": 2512781312
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024272727272727274,
      "loss": 1.4308,
      "theoretical_loss": 3.3671470677573234,
      "tokens_seen": 2512846848
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002427070707070707,
      "loss": 1.5251,
      "theoretical_loss": 3.3671400702572765,
      "tokens_seen": 2512912384
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024268686868686868,
      "loss": 1.3585,
      "theoretical_loss": 3.367133072990816,
      "tokens_seen": 2512977920
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024266666666666667,
      "loss": 1.2903,
      "theoretical_loss": 3.3671260759579287,
      "tokens_seen": 2513043456
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024264646464646465,
      "loss": 1.4704,
      "theoretical_loss": 3.3671190791586003,
      "tokens_seen": 2513108992
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024262626262626263,
      "loss": 1.3653,
      "theoretical_loss": 3.3671120825928167,
      "tokens_seen": 2513174528
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024260606060606062,
      "loss": 1.2717,
      "theoretical_loss": 3.367105086260565,
      "tokens_seen": 2513240064
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002425858585858586,
      "loss": 1.4236,
      "theoretical_loss": 3.36709809016183,
      "tokens_seen": 2513305600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024256565656565659,
      "loss": 1.4043,
      "theoretical_loss": 3.367091094296599,
      "tokens_seen": 2513371136
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024254545454545454,
      "loss": 1.293,
      "theoretical_loss": 3.367084098664857,
      "tokens_seen": 2513436672
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024252525252525253,
      "loss": 1.5321,
      "theoretical_loss": 3.367077103266591,
      "tokens_seen": 2513502208
    },
    {
      "epoch": 0.52,
      "objective/train/aligned_token_count": 0.455278605222702,
      "objective/train/avg_token_score": 0.005221270490437746,
      "objective/train/docs_used": 1413871,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4556875228881836,
      "objective/train/original_loss": 3.908552408218384,
      "objective/train/theoretical_loss": 3.367070108101786,
      "objective/train/tokens_used": 872426976,
      "objective/train/unlikelihood_loss": 0.06845640391111374,
      "theoretical_loss": 3.367070108101786,
      "tokens_seen": 2513567744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002425050505050505,
      "loss": 1.6202,
      "theoretical_loss": 3.367070108101786,
      "tokens_seen": 2513567744
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002424848484848485,
      "loss": 1.3391,
      "theoretical_loss": 3.36706311317043,
      "tokens_seen": 2513633280
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024246464646464648,
      "loss": 1.4055,
      "theoretical_loss": 3.3670561184725076,
      "tokens_seen": 2513698816
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024244444444444443,
      "loss": 1.4749,
      "theoretical_loss": 3.3670491240080054,
      "tokens_seen": 2513764352
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024242424242424245,
      "loss": 1.3019,
      "theoretical_loss": 3.3670421297769093,
      "tokens_seen": 2513829888
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002424040404040404,
      "loss": 1.4109,
      "theoretical_loss": 3.367035135779206,
      "tokens_seen": 2513895424
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024238383838383841,
      "loss": 1.4825,
      "theoretical_loss": 3.367028142014881,
      "tokens_seen": 2513960960
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024236363636363637,
      "loss": 1.2524,
      "theoretical_loss": 3.3670211484839205,
      "tokens_seen": 2514026496
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024234343434343433,
      "loss": 1.6174,
      "theoretical_loss": 3.367014155186311,
      "tokens_seen": 2514092032
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024232323232323234,
      "loss": 1.4618,
      "theoretical_loss": 3.3670071621220385,
      "tokens_seen": 2514157568
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002423030303030303,
      "loss": 1.2389,
      "theoretical_loss": 3.367000169291089,
      "tokens_seen": 2514223104
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002422828282828283,
      "loss": 1.3485,
      "theoretical_loss": 3.366993176693449,
      "tokens_seen": 2514288640
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024226262626262626,
      "loss": 1.484,
      "theoretical_loss": 3.3669861843291042,
      "tokens_seen": 2514354176
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024224242424242425,
      "loss": 1.0516,
      "theoretical_loss": 3.3669791921980408,
      "tokens_seen": 2514419712
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024222222222222223,
      "loss": 1.4102,
      "theoretical_loss": 3.366972200300245,
      "tokens_seen": 2514485248
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002422020202020202,
      "loss": 1.6043,
      "theoretical_loss": 3.366965208635703,
      "tokens_seen": 2514550784
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002421818181818182,
      "loss": 1.3648,
      "theoretical_loss": 3.3669582172044006,
      "tokens_seen": 2514616320
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024216161616161616,
      "loss": 1.6342,
      "theoretical_loss": 3.3669512260063246,
      "tokens_seen": 2514681856
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024214141414141414,
      "loss": 1.2424,
      "theoretical_loss": 3.3669442350414602,
      "tokens_seen": 2514747392
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024212121212121212,
      "loss": 1.2164,
      "theoretical_loss": 3.3669372443097947,
      "tokens_seen": 2514812928
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002421010101010101,
      "loss": 1.3182,
      "theoretical_loss": 3.3669302538113137,
      "tokens_seen": 2514878464
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002420808080808081,
      "loss": 1.3191,
      "theoretical_loss": 3.366923263546003,
      "tokens_seen": 2514944000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024206060606060608,
      "loss": 1.271,
      "theoretical_loss": 3.366916273513849,
      "tokens_seen": 2515009536
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024204040404040403,
      "loss": 1.3631,
      "theoretical_loss": 3.366909283714838,
      "tokens_seen": 2515075072
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024202020202020202,
      "loss": 1.2594,
      "theoretical_loss": 3.3669022941489564,
      "tokens_seen": 2515140608
    },
    {
      "epoch": 0.52,
      "objective/train/aligned_token_count": 0.3737170100212097,
      "objective/train/avg_token_score": 0.021781224757432938,
      "objective/train/docs_used": 1414570,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2252116203308105,
      "objective/train/original_loss": 4.234691143035889,
      "objective/train/theoretical_loss": 3.3668953048161896,
      "objective/train/tokens_used": 874065376,
      "objective/train/unlikelihood_loss": 0.06306973844766617,
      "theoretical_loss": 3.3668953048161896,
      "tokens_seen": 2515206144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.000242,
      "loss": 1.3576,
      "theoretical_loss": 3.3668953048161896,
      "tokens_seen": 2515206144
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024197979797979798,
      "loss": 1.5681,
      "theoretical_loss": 3.3668883157165244,
      "tokens_seen": 2515271680
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024195959595959597,
      "loss": 1.4767,
      "theoretical_loss": 3.3668813268499465,
      "tokens_seen": 2515337216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024193939393939393,
      "loss": 1.6378,
      "theoretical_loss": 3.366874338216442,
      "tokens_seen": 2515402752
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024191919191919194,
      "loss": 1.4127,
      "theoretical_loss": 3.3668673498159976,
      "tokens_seen": 2515468288
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002418989898989899,
      "loss": 1.3826,
      "theoretical_loss": 3.3668603616485995,
      "tokens_seen": 2515533824
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002418787878787879,
      "loss": 1.4284,
      "theoretical_loss": 3.366853373714233,
      "tokens_seen": 2515599360
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024185858585858586,
      "loss": 1.225,
      "theoretical_loss": 3.366846386012885,
      "tokens_seen": 2515664896
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024183838383838382,
      "loss": 1.3522,
      "theoretical_loss": 3.3668393985445415,
      "tokens_seen": 2515730432
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024181818181818183,
      "loss": 1.4349,
      "theoretical_loss": 3.3668324113091885,
      "tokens_seen": 2515795968
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024179797979797979,
      "loss": 1.4467,
      "theoretical_loss": 3.3668254243068123,
      "tokens_seen": 2515861504
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002417777777777778,
      "loss": 1.572,
      "theoretical_loss": 3.366818437537399,
      "tokens_seen": 2515927040
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024175757575757575,
      "loss": 1.4718,
      "theoretical_loss": 3.366811451000935,
      "tokens_seen": 2515992576
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024173737373737376,
      "loss": 1.308,
      "theoretical_loss": 3.366804464697406,
      "tokens_seen": 2516058112
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024171717171717172,
      "loss": 1.3263,
      "theoretical_loss": 3.3667974786267987,
      "tokens_seen": 2516123648
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002416969696969697,
      "loss": 1.4597,
      "theoretical_loss": 3.366790492789099,
      "tokens_seen": 2516189184
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002416767676767677,
      "loss": 1.5087,
      "theoretical_loss": 3.3667835071842926,
      "tokens_seen": 2516254720
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024165656565656565,
      "loss": 1.2749,
      "theoretical_loss": 3.3667765218123664,
      "tokens_seen": 2516320256
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024163636363636366,
      "loss": 1.4824,
      "theoretical_loss": 3.3667695366733064,
      "tokens_seen": 2516385792
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024161616161616161,
      "loss": 1.2914,
      "theoretical_loss": 3.3667625517670987,
      "tokens_seen": 2516451328
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002415959595959596,
      "loss": 1.5026,
      "theoretical_loss": 3.3667555670937297,
      "tokens_seen": 2516516864
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024157575757575758,
      "loss": 1.3381,
      "theoretical_loss": 3.3667485826531847,
      "tokens_seen": 2516582400
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024155555555555557,
      "loss": 1.475,
      "theoretical_loss": 3.3667415984454507,
      "tokens_seen": 2516647936
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024153535353535355,
      "loss": 1.2454,
      "theoretical_loss": 3.3667346144705137,
      "tokens_seen": 2516713472
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024151515151515153,
      "loss": 1.4488,
      "theoretical_loss": 3.3667276307283602,
      "tokens_seen": 2516779008
    },
    {
      "epoch": 0.52,
      "objective/train/aligned_token_count": 0.33571603894233704,
      "objective/train/avg_token_score": 0.011416483670473099,
      "objective/train/docs_used": 1415934,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1242914199829102,
      "objective/train/original_loss": 4.375868320465088,
      "objective/train/theoretical_loss": 3.3667206472189757,
      "objective/train/tokens_used": 875703776,
      "objective/train/unlikelihood_loss": 0.07855267822742462,
      "theoretical_loss": 3.3667206472189757,
      "tokens_seen": 2516844544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002414949494949495,
      "loss": 1.3085,
      "theoretical_loss": 3.3667206472189757,
      "tokens_seen": 2516844544
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024147474747474748,
      "loss": 1.0466,
      "theoretical_loss": 3.366713663942347,
      "tokens_seen": 2516910080
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024145454545454546,
      "loss": 1.4732,
      "theoretical_loss": 3.3667066808984596,
      "tokens_seen": 2516975616
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024143434343434344,
      "loss": 1.4002,
      "theoretical_loss": 3.3666996980873005,
      "tokens_seen": 2517041152
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024141414141414143,
      "loss": 1.3301,
      "theoretical_loss": 3.3666927155088553,
      "tokens_seen": 2517106688
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024139393939393938,
      "loss": 1.3412,
      "theoretical_loss": 3.36668573316311,
      "tokens_seen": 2517172224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002413737373737374,
      "loss": 1.1726,
      "theoretical_loss": 3.3666787510500518,
      "tokens_seen": 2517237760
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024135353535353535,
      "loss": 1.3825,
      "theoretical_loss": 3.366671769169666,
      "tokens_seen": 2517303296
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024133333333333334,
      "loss": 1.3728,
      "theoretical_loss": 3.366664787521939,
      "tokens_seen": 2517368832
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024131313131313132,
      "loss": 1.5187,
      "theoretical_loss": 3.366657806106857,
      "tokens_seen": 2517434368
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024129292929292928,
      "loss": 1.2452,
      "theoretical_loss": 3.3666508249244056,
      "tokens_seen": 2517499904
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002412727272727273,
      "loss": 1.3651,
      "theoretical_loss": 3.3666438439745723,
      "tokens_seen": 2517565440
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024125252525252524,
      "loss": 1.3626,
      "theoretical_loss": 3.3666368632573427,
      "tokens_seen": 2517630976
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024123232323232326,
      "loss": 1.3217,
      "theoretical_loss": 3.366629882772702,
      "tokens_seen": 2517696512
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002412121212121212,
      "loss": 1.4661,
      "theoretical_loss": 3.3666229025206382,
      "tokens_seen": 2517762048
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002411919191919192,
      "loss": 1.5451,
      "theoretical_loss": 3.3666159225011363,
      "tokens_seen": 2517827584
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024117171717171718,
      "loss": 1.3307,
      "theoretical_loss": 3.3666089427141825,
      "tokens_seen": 2517893120
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024115151515151514,
      "loss": 1.2271,
      "theoretical_loss": 3.3666019631597637,
      "tokens_seen": 2517958656
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024113131313131315,
      "loss": 1.3378,
      "theoretical_loss": 3.3665949838378655,
      "tokens_seen": 2518024192
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002411111111111111,
      "loss": 1.3709,
      "theoretical_loss": 3.3665880047484746,
      "tokens_seen": 2518089728
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002410909090909091,
      "loss": 1.5162,
      "theoretical_loss": 3.3665810258915765,
      "tokens_seen": 2518155264
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024107070707070707,
      "loss": 1.2728,
      "theoretical_loss": 3.3665740472671577,
      "tokens_seen": 2518220800
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024105050505050506,
      "loss": 1.3841,
      "theoretical_loss": 3.366567068875205,
      "tokens_seen": 2518286336
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024103030303030304,
      "loss": 1.6191,
      "theoretical_loss": 3.366560090715703,
      "tokens_seen": 2518351872
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024101010101010102,
      "loss": 1.4387,
      "theoretical_loss": 3.3665531127886403,
      "tokens_seen": 2518417408
    },
    {
      "epoch": 0.52,
      "objective/train/aligned_token_count": 0.4742790758609772,
      "objective/train/avg_token_score": 0.004780460149049759,
      "objective/train/docs_used": 1416375,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4100862741470337,
      "objective/train/original_loss": 3.823479652404785,
      "objective/train/theoretical_loss": 3.3665461350940014,
      "objective/train/tokens_used": 877342176,
      "objective/train/unlikelihood_loss": 0.08516652882099152,
      "theoretical_loss": 3.3665461350940014,
      "tokens_seen": 2518482944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024098989898989898,
      "loss": 1.3547,
      "theoretical_loss": 3.3665461350940014,
      "tokens_seen": 2518482944
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024096969696969697,
      "loss": 1.3634,
      "theoretical_loss": 3.3665391576317725,
      "tokens_seen": 2518548480
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024094949494949495,
      "loss": 1.3878,
      "theoretical_loss": 3.366532180401941,
      "tokens_seen": 2518614016
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024092929292929293,
      "loss": 1.5374,
      "theoretical_loss": 3.3665252034044917,
      "tokens_seen": 2518679552
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024090909090909092,
      "loss": 1.2322,
      "theoretical_loss": 3.3665182266394114,
      "tokens_seen": 2518745088
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002408888888888889,
      "loss": 1.5286,
      "theoretical_loss": 3.366511250106687,
      "tokens_seen": 2518810624
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024086868686868689,
      "loss": 1.3369,
      "theoretical_loss": 3.366504273806304,
      "tokens_seen": 2518876160
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024084848484848484,
      "loss": 1.4983,
      "theoretical_loss": 3.366497297738248,
      "tokens_seen": 2518941696
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024082828282828285,
      "loss": 1.2667,
      "theoretical_loss": 3.3664903219025066,
      "tokens_seen": 2519007232
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002408080808080808,
      "loss": 1.3147,
      "theoretical_loss": 3.366483346299065,
      "tokens_seen": 2519072768
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002407878787878788,
      "loss": 1.2108,
      "theoretical_loss": 3.36647637092791,
      "tokens_seen": 2519138304
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024076767676767678,
      "loss": 1.3861,
      "theoretical_loss": 3.3664693957890277,
      "tokens_seen": 2519203840
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024074747474747474,
      "loss": 1.3387,
      "theoretical_loss": 3.366462420882404,
      "tokens_seen": 2519269376
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024072727272727275,
      "loss": 1.4831,
      "theoretical_loss": 3.3664554462080254,
      "tokens_seen": 2519334912
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002407070707070707,
      "loss": 1.5804,
      "theoretical_loss": 3.366448471765878,
      "tokens_seen": 2519400448
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024068686868686871,
      "loss": 1.2084,
      "theoretical_loss": 3.3664414975559485,
      "tokens_seen": 2519465984
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024066666666666667,
      "loss": 1.3454,
      "theoretical_loss": 3.366434523578222,
      "tokens_seen": 2519531520
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024064646464646463,
      "loss": 1.1946,
      "theoretical_loss": 3.366427549832686,
      "tokens_seen": 2519597056
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024062626262626264,
      "loss": 1.2836,
      "theoretical_loss": 3.3664205763193262,
      "tokens_seen": 2519662592
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002406060606060606,
      "loss": 1.5758,
      "theoretical_loss": 3.3664136030381284,
      "tokens_seen": 2519728128
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002405858585858586,
      "loss": 1.6876,
      "theoretical_loss": 3.3664066299890796,
      "tokens_seen": 2519793664
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024056565656565656,
      "loss": 1.3768,
      "theoretical_loss": 3.366399657172166,
      "tokens_seen": 2519859200
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024054545454545455,
      "loss": 1.3757,
      "theoretical_loss": 3.366392684587373,
      "tokens_seen": 2519924736
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024052525252525253,
      "loss": 1.2754,
      "theoretical_loss": 3.3663857122346874,
      "tokens_seen": 2519990272
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024050505050505052,
      "loss": 1.3078,
      "theoretical_loss": 3.3663787401140954,
      "tokens_seen": 2520055808
    },
    {
      "epoch": 0.52,
      "objective/train/aligned_token_count": 0.41825512051582336,
      "objective/train/avg_token_score": 0.021243063732981682,
      "objective/train/docs_used": 1417108,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.04865300655365,
      "objective/train/original_loss": 4.158831596374512,
      "objective/train/theoretical_loss": 3.3663717682255836,
      "objective/train/tokens_used": 878980576,
      "objective/train/unlikelihood_loss": 0.03333920240402222,
      "theoretical_loss": 3.3663717682255836,
      "tokens_seen": 2520121344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002404848484848485,
      "loss": 1.2431,
      "theoretical_loss": 3.3663717682255836,
      "tokens_seen": 2520121344
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024046464646464646,
      "loss": 1.4491,
      "theoretical_loss": 3.3663647965691377,
      "tokens_seen": 2520186880
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024044444444444444,
      "loss": 1.2109,
      "theoretical_loss": 3.3663578251447435,
      "tokens_seen": 2520252416
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024042424242424242,
      "loss": 1.4568,
      "theoretical_loss": 3.3663508539523885,
      "tokens_seen": 2520317952
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002404040404040404,
      "loss": 1.3635,
      "theoretical_loss": 3.3663438829920582,
      "tokens_seen": 2520383488
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002403838383838384,
      "loss": 1.4415,
      "theoretical_loss": 3.3663369122637388,
      "tokens_seen": 2520449024
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024036363636363638,
      "loss": 1.3648,
      "theoretical_loss": 3.366329941767417,
      "tokens_seen": 2520514560
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024034343434343433,
      "loss": 1.3435,
      "theoretical_loss": 3.3663229715030787,
      "tokens_seen": 2520580096
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024032323232323234,
      "loss": 1.303,
      "theoretical_loss": 3.3663160014707096,
      "tokens_seen": 2520645632
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002403030303030303,
      "loss": 1.4612,
      "theoretical_loss": 3.3663090316702973,
      "tokens_seen": 2520711168
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024028282828282828,
      "loss": 1.419,
      "theoretical_loss": 3.366302062101827,
      "tokens_seen": 2520776704
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024026262626262627,
      "loss": 1.1854,
      "theoretical_loss": 3.366295092765285,
      "tokens_seen": 2520842240
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024024242424242423,
      "loss": 1.398,
      "theoretical_loss": 3.366288123660658,
      "tokens_seen": 2520907776
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024022222222222224,
      "loss": 1.4828,
      "theoretical_loss": 3.3662811547879317,
      "tokens_seen": 2520973312
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002402020202020202,
      "loss": 1.3074,
      "theoretical_loss": 3.366274186147093,
      "tokens_seen": 2521038848
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002401818181818182,
      "loss": 1.3746,
      "theoretical_loss": 3.366267217738128,
      "tokens_seen": 2521104384
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024016161616161616,
      "loss": 1.4158,
      "theoretical_loss": 3.366260249561023,
      "tokens_seen": 2521169920
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024014141414141415,
      "loss": 1.394,
      "theoretical_loss": 3.3662532816157635,
      "tokens_seen": 2521235456
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024012121212121213,
      "loss": 1.4981,
      "theoretical_loss": 3.3662463139023364,
      "tokens_seen": 2521300992
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024010101010101009,
      "loss": 1.3501,
      "theoretical_loss": 3.3662393464207283,
      "tokens_seen": 2521366528
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002400808080808081,
      "loss": 1.2519,
      "theoretical_loss": 3.366232379170925,
      "tokens_seen": 2521432064
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024006060606060605,
      "loss": 1.5198,
      "theoretical_loss": 3.3662254121529123,
      "tokens_seen": 2521497600
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024004040404040407,
      "loss": 1.4918,
      "theoretical_loss": 3.3662184453666777,
      "tokens_seen": 2521563136
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024002020202020202,
      "loss": 1.3204,
      "theoretical_loss": 3.366211478812206,
      "tokens_seen": 2521628672
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00024,
      "loss": 1.5226,
      "theoretical_loss": 3.3662045124894844,
      "tokens_seen": 2521694208
    },
    {
      "epoch": 0.52,
      "objective/train/aligned_token_count": 0.3615591526031494,
      "objective/train/avg_token_score": 0.01131171453744173,
      "objective/train/docs_used": 1417967,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.198206901550293,
      "objective/train/original_loss": 4.193903923034668,
      "objective/train/theoretical_loss": 3.3661975463984994,
      "objective/train/tokens_used": 880618976,
      "objective/train/unlikelihood_loss": 0.07620194554328918,
      "theoretical_loss": 3.3661975463984994,
      "tokens_seen": 2521759744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000239979797979798,
      "loss": 1.2923,
      "theoretical_loss": 3.3661975463984994,
      "tokens_seen": 2521759744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023995959595959597,
      "loss": 1.3621,
      "theoretical_loss": 3.3661905805392367,
      "tokens_seen": 2521825280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023993939393939396,
      "loss": 1.3763,
      "theoretical_loss": 3.3661836149116824,
      "tokens_seen": 2521890816
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023991919191919191,
      "loss": 1.4208,
      "theoretical_loss": 3.3661766495158236,
      "tokens_seen": 2521956352
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002398989898989899,
      "loss": 1.1068,
      "theoretical_loss": 3.3661696843516458,
      "tokens_seen": 2522021888
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023987878787878788,
      "loss": 1.3197,
      "theoretical_loss": 3.3661627194191355,
      "tokens_seen": 2522087424
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023985858585858587,
      "loss": 1.3354,
      "theoretical_loss": 3.366155754718279,
      "tokens_seen": 2522152960
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023983838383838385,
      "loss": 1.273,
      "theoretical_loss": 3.3661487902490626,
      "tokens_seen": 2522218496
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023981818181818183,
      "loss": 1.2473,
      "theoretical_loss": 3.3661418260114724,
      "tokens_seen": 2522284032
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002397979797979798,
      "loss": 1.2004,
      "theoretical_loss": 3.366134862005495,
      "tokens_seen": 2522349568
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002397777777777778,
      "loss": 1.2216,
      "theoretical_loss": 3.3661278982311167,
      "tokens_seen": 2522415104
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023975757575757576,
      "loss": 1.4573,
      "theoretical_loss": 3.366120934688323,
      "tokens_seen": 2522480640
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023973737373737374,
      "loss": 1.4574,
      "theoretical_loss": 3.3661139713771018,
      "tokens_seen": 2522546176
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023971717171717173,
      "loss": 1.3306,
      "theoretical_loss": 3.366107008297438,
      "tokens_seen": 2522611712
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023969696969696968,
      "loss": 1.4671,
      "theoretical_loss": 3.3661000454493175,
      "tokens_seen": 2522677248
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002396767676767677,
      "loss": 1.5073,
      "theoretical_loss": 3.366093082832728,
      "tokens_seen": 2522742784
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023965656565656565,
      "loss": 1.3777,
      "theoretical_loss": 3.3660861204476555,
      "tokens_seen": 2522808320
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023963636363636366,
      "loss": 1.2265,
      "theoretical_loss": 3.3660791582940854,
      "tokens_seen": 2522873856
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023961616161616162,
      "loss": 1.2516,
      "theoretical_loss": 3.3660721963720044,
      "tokens_seen": 2522939392
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023959595959595958,
      "loss": 1.3629,
      "theoretical_loss": 3.366065234681399,
      "tokens_seen": 2523004928
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002395757575757576,
      "loss": 1.2301,
      "theoretical_loss": 3.3660582732222553,
      "tokens_seen": 2523070464
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023955555555555554,
      "loss": 1.3838,
      "theoretical_loss": 3.36605131199456,
      "tokens_seen": 2523136000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023953535353535356,
      "loss": 1.5114,
      "theoretical_loss": 3.366044350998299,
      "tokens_seen": 2523201536
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002395151515151515,
      "loss": 1.0342,
      "theoretical_loss": 3.3660373902334584,
      "tokens_seen": 2523267072
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002394949494949495,
      "loss": 1.419,
      "theoretical_loss": 3.366030429700025,
      "tokens_seen": 2523332608
    },
    {
      "epoch": 0.53,
      "objective/train/aligned_token_count": 0.47360703349113464,
      "objective/train/avg_token_score": 0.02000376209616661,
      "objective/train/docs_used": 1418616,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2760837078094482,
      "objective/train/original_loss": 3.738672971725464,
      "objective/train/theoretical_loss": 3.3660234693979847,
      "objective/train/tokens_used": 882257376,
      "objective/train/unlikelihood_loss": 0.0662318766117096,
      "theoretical_loss": 3.3660234693979847,
      "tokens_seen": 2523398144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023947474747474748,
      "loss": 1.4039,
      "theoretical_loss": 3.3660234693979847,
      "tokens_seen": 2523398144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023945454545454546,
      "loss": 1.2362,
      "theoretical_loss": 3.366016509327324,
      "tokens_seen": 2523463680
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023943434343434345,
      "loss": 1.228,
      "theoretical_loss": 3.3660095494880293,
      "tokens_seen": 2523529216
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002394141414141414,
      "loss": 1.3998,
      "theoretical_loss": 3.366002589880087,
      "tokens_seen": 2523594752
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002393939393939394,
      "loss": 1.4257,
      "theoretical_loss": 3.3659956305034826,
      "tokens_seen": 2523660288
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023937373737373737,
      "loss": 1.5887,
      "theoretical_loss": 3.365988671358203,
      "tokens_seen": 2523725824
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023935353535353536,
      "loss": 1.3713,
      "theoretical_loss": 3.365981712444235,
      "tokens_seen": 2523791360
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023933333333333334,
      "loss": 1.316,
      "theoretical_loss": 3.365974753761564,
      "tokens_seen": 2523856896
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023931313131313133,
      "loss": 1.4875,
      "theoretical_loss": 3.3659677953101763,
      "tokens_seen": 2523922432
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023929292929292928,
      "loss": 1.3174,
      "theoretical_loss": 3.3659608370900593,
      "tokens_seen": 2523987968
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002392727272727273,
      "loss": 1.5757,
      "theoretical_loss": 3.3659538791011983,
      "tokens_seen": 2524053504
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023925252525252525,
      "loss": 1.4418,
      "theoretical_loss": 3.3659469213435798,
      "tokens_seen": 2524119040
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023923232323232323,
      "loss": 1.4328,
      "theoretical_loss": 3.3659399638171905,
      "tokens_seen": 2524184576
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023921212121212122,
      "loss": 1.2283,
      "theoretical_loss": 3.365933006522016,
      "tokens_seen": 2524250112
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002391919191919192,
      "loss": 1.356,
      "theoretical_loss": 3.3659260494580434,
      "tokens_seen": 2524315648
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023917171717171719,
      "loss": 1.3044,
      "theoretical_loss": 3.3659190926252585,
      "tokens_seen": 2524381184
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023915151515151514,
      "loss": 1.3185,
      "theoretical_loss": 3.365912136023648,
      "tokens_seen": 2524446720
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023913131313131315,
      "loss": 1.3288,
      "theoretical_loss": 3.365905179653198,
      "tokens_seen": 2524512256
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002391111111111111,
      "loss": 1.2988,
      "theoretical_loss": 3.365898223513894,
      "tokens_seen": 2524577792
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023909090909090912,
      "loss": 1.4164,
      "theoretical_loss": 3.3658912676057238,
      "tokens_seen": 2524643328
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023907070707070708,
      "loss": 1.2836,
      "theoretical_loss": 3.3658843119286734,
      "tokens_seen": 2524708864
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023905050505050504,
      "loss": 1.4192,
      "theoretical_loss": 3.365877356482728,
      "tokens_seen": 2524774400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023903030303030305,
      "loss": 1.1823,
      "theoretical_loss": 3.365870401267875,
      "tokens_seen": 2524839936
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000239010101010101,
      "loss": 1.2264,
      "theoretical_loss": 3.365863446284101,
      "tokens_seen": 2524905472
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023898989898989901,
      "loss": 1.4412,
      "theoretical_loss": 3.365856491531391,
      "tokens_seen": 2524971008
    },
    {
      "epoch": 0.53,
      "objective/train/aligned_token_count": 0.42155423760414124,
      "objective/train/avg_token_score": 0.010902047157287598,
      "objective/train/docs_used": 1419908,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2965419292449951,
      "objective/train/original_loss": 4.07478141784668,
      "objective/train/theoretical_loss": 3.3658495370097326,
      "objective/train/tokens_used": 883895776,
      "objective/train/unlikelihood_loss": 0.07022913545370102,
      "theoretical_loss": 3.3658495370097326,
      "tokens_seen": 2525036544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023896969696969697,
      "loss": 1.4122,
      "theoretical_loss": 3.3658495370097326,
      "tokens_seen": 2525036544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023894949494949496,
      "loss": 1.1614,
      "theoretical_loss": 3.365842582719111,
      "tokens_seen": 2525102080
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023892929292929294,
      "loss": 1.3263,
      "theoretical_loss": 3.3658356286595135,
      "tokens_seen": 2525167616
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002389090909090909,
      "loss": 1.4605,
      "theoretical_loss": 3.365828674830926,
      "tokens_seen": 2525233152
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002388888888888889,
      "loss": 1.5766,
      "theoretical_loss": 3.365821721233335,
      "tokens_seen": 2525298688
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023886868686868686,
      "loss": 1.5962,
      "theoretical_loss": 3.3658147678667265,
      "tokens_seen": 2525364224
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023884848484848485,
      "loss": 1.6716,
      "theoretical_loss": 3.365807814731087,
      "tokens_seen": 2525429760
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023882828282828283,
      "loss": 1.3234,
      "theoretical_loss": 3.3658008618264033,
      "tokens_seen": 2525495296
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023880808080808082,
      "loss": 1.2495,
      "theoretical_loss": 3.3657939091526607,
      "tokens_seen": 2525560832
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002387878787878788,
      "loss": 1.3122,
      "theoretical_loss": 3.3657869567098464,
      "tokens_seen": 2525626368
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023876767676767678,
      "loss": 1.3493,
      "theoretical_loss": 3.365780004497947,
      "tokens_seen": 2525691904
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023874747474747474,
      "loss": 1.5368,
      "theoretical_loss": 3.3657730525169476,
      "tokens_seen": 2525757440
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023872727272727272,
      "loss": 1.2713,
      "theoretical_loss": 3.3657661007668356,
      "tokens_seen": 2525822976
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002387070707070707,
      "loss": 1.2382,
      "theoretical_loss": 3.365759149247597,
      "tokens_seen": 2525888512
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002386868686868687,
      "loss": 1.3102,
      "theoretical_loss": 3.3657521979592184,
      "tokens_seen": 2525954048
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023866666666666668,
      "loss": 1.5375,
      "theoretical_loss": 3.3657452469016853,
      "tokens_seen": 2526019584
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023864646464646463,
      "loss": 1.4349,
      "theoretical_loss": 3.3657382960749853,
      "tokens_seen": 2526085120
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023862626262626264,
      "loss": 1.4548,
      "theoretical_loss": 3.365731345479104,
      "tokens_seen": 2526150656
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002386060606060606,
      "loss": 1.3199,
      "theoretical_loss": 3.3657243951140274,
      "tokens_seen": 2526216192
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002385858585858586,
      "loss": 1.4634,
      "theoretical_loss": 3.365717444979743,
      "tokens_seen": 2526281728
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023856565656565657,
      "loss": 1.3643,
      "theoretical_loss": 3.3657104950762355,
      "tokens_seen": 2526347264
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023854545454545453,
      "loss": 1.4539,
      "theoretical_loss": 3.365703545403493,
      "tokens_seen": 2526412800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023852525252525254,
      "loss": 1.3225,
      "theoretical_loss": 3.3656965959615004,
      "tokens_seen": 2526478336
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002385050505050505,
      "loss": 1.4547,
      "theoretical_loss": 3.365689646750245,
      "tokens_seen": 2526543872
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002384848484848485,
      "loss": 1.0691,
      "theoretical_loss": 3.365682697769713,
      "tokens_seen": 2526609408
    },
    {
      "epoch": 0.53,
      "objective/train/aligned_token_count": 0.35807672142982483,
      "objective/train/avg_token_score": 0.023771250620484352,
      "objective/train/docs_used": 1420528,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1180223226547241,
      "objective/train/original_loss": 4.581310749053955,
      "objective/train/theoretical_loss": 3.3656757490198905,
      "objective/train/tokens_used": 885534176,
      "objective/train/unlikelihood_loss": 0.07055738568305969,
      "theoretical_loss": 3.3656757490198905,
      "tokens_seen": 2526674944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023846464646464646,
      "loss": 1.4428,
      "theoretical_loss": 3.3656757490198905,
      "tokens_seen": 2526674944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023844444444444445,
      "loss": 1.3245,
      "theoretical_loss": 3.365668800500764,
      "tokens_seen": 2526740480
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023842424242424243,
      "loss": 1.535,
      "theoretical_loss": 3.3656618522123196,
      "tokens_seen": 2526806016
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023840404040404041,
      "loss": 1.4438,
      "theoretical_loss": 3.3656549041545443,
      "tokens_seen": 2526871552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002383838383838384,
      "loss": 1.3259,
      "theoretical_loss": 3.365647956327424,
      "tokens_seen": 2526937088
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023836363636363635,
      "loss": 1.4423,
      "theoretical_loss": 3.3656410087309445,
      "tokens_seen": 2527002624
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023834343434343437,
      "loss": 1.3018,
      "theoretical_loss": 3.3656340613650935,
      "tokens_seen": 2527068160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023832323232323232,
      "loss": 1.3706,
      "theoretical_loss": 3.365627114229856,
      "tokens_seen": 2527133696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002383030303030303,
      "loss": 1.4118,
      "theoretical_loss": 3.3656201673252193,
      "tokens_seen": 2527199232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002382828282828283,
      "loss": 1.3509,
      "theoretical_loss": 3.3656132206511695,
      "tokens_seen": 2527264768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023826262626262627,
      "loss": 1.4662,
      "theoretical_loss": 3.3656062742076926,
      "tokens_seen": 2527330304
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023824242424242426,
      "loss": 1.3992,
      "theoretical_loss": 3.365599327994776,
      "tokens_seen": 2527395840
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023822222222222224,
      "loss": 1.5259,
      "theoretical_loss": 3.3655923820124047,
      "tokens_seen": 2527461376
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002382020202020202,
      "loss": 1.3707,
      "theoretical_loss": 3.3655854362605657,
      "tokens_seen": 2527526912
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023818181818181818,
      "loss": 1.2446,
      "theoretical_loss": 3.365578490739246,
      "tokens_seen": 2527592448
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023816161616161617,
      "loss": 1.3352,
      "theoretical_loss": 3.365571545448431,
      "tokens_seen": 2527657984
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023814141414141415,
      "loss": 1.4961,
      "theoretical_loss": 3.3655646003881072,
      "tokens_seen": 2527723520
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023812121212121214,
      "loss": 1.3827,
      "theoretical_loss": 3.365557655558262,
      "tokens_seen": 2527789056
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002381010101010101,
      "loss": 1.4193,
      "theoretical_loss": 3.36555071095888,
      "tokens_seen": 2527854592
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002380808080808081,
      "loss": 1.2644,
      "theoretical_loss": 3.3655437665899495,
      "tokens_seen": 2527920128
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023806060606060606,
      "loss": 1.1658,
      "theoretical_loss": 3.365536822451455,
      "tokens_seen": 2527985664
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023804040404040404,
      "loss": 1.4532,
      "theoretical_loss": 3.3655298785433847,
      "tokens_seen": 2528051200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023802020202020203,
      "loss": 1.4302,
      "theoretical_loss": 3.365522934865724,
      "tokens_seen": 2528116736
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023799999999999998,
      "loss": 1.4786,
      "theoretical_loss": 3.3655159914184587,
      "tokens_seen": 2528182272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000237979797979798,
      "loss": 1.3337,
      "theoretical_loss": 3.3655090482015764,
      "tokens_seen": 2528247808
    },
    {
      "epoch": 0.53,
      "objective/train/aligned_token_count": 0.46884164214134216,
      "objective/train/avg_token_score": 0.005133732222020626,
      "objective/train/docs_used": 1420951,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.575600266456604,
      "objective/train/original_loss": 4.162727355957031,
      "objective/train/theoretical_loss": 3.365502105215063,
      "objective/train/tokens_used": 887172576,
      "objective/train/unlikelihood_loss": 0.05329639092087746,
      "theoretical_loss": 3.365502105215063,
      "tokens_seen": 2528313344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023795959595959595,
      "loss": 1.5059,
      "theoretical_loss": 3.365502105215063,
      "tokens_seen": 2528313344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023793939393939396,
      "loss": 1.4078,
      "theoretical_loss": 3.3654951624589047,
      "tokens_seen": 2528378880
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023791919191919192,
      "loss": 1.411,
      "theoretical_loss": 3.365488219933088,
      "tokens_seen": 2528444416
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002378989898989899,
      "loss": 1.3363,
      "theoretical_loss": 3.3654812776376,
      "tokens_seen": 2528509952
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002378787878787879,
      "loss": 1.4295,
      "theoretical_loss": 3.3654743355724257,
      "tokens_seen": 2528575488
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023785858585858585,
      "loss": 1.4035,
      "theoretical_loss": 3.365467393737552,
      "tokens_seen": 2528641024
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023783838383838386,
      "loss": 1.5776,
      "theoretical_loss": 3.3654604521329663,
      "tokens_seen": 2528706560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002378181818181818,
      "loss": 1.476,
      "theoretical_loss": 3.3654535107586536,
      "tokens_seen": 2528772096
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002377979797979798,
      "loss": 1.4451,
      "theoretical_loss": 3.365446569614601,
      "tokens_seen": 2528837632
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023777777777777778,
      "loss": 1.549,
      "theoretical_loss": 3.365439628700795,
      "tokens_seen": 2528903168
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023775757575757577,
      "loss": 1.2331,
      "theoretical_loss": 3.365432688017221,
      "tokens_seen": 2528968704
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023773737373737375,
      "loss": 1.3207,
      "theoretical_loss": 3.3654257475638674,
      "tokens_seen": 2529034240
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023771717171717173,
      "loss": 1.5005,
      "theoretical_loss": 3.3654188073407187,
      "tokens_seen": 2529099776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002376969696969697,
      "loss": 1.2516,
      "theoretical_loss": 3.365411867347762,
      "tokens_seen": 2529165312
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023767676767676767,
      "loss": 1.3054,
      "theoretical_loss": 3.3654049275849838,
      "tokens_seen": 2529230848
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023765656565656566,
      "loss": 1.2936,
      "theoretical_loss": 3.36539798805237,
      "tokens_seen": 2529296384
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023763636363636364,
      "loss": 1.4882,
      "theoretical_loss": 3.365391048749908,
      "tokens_seen": 2529361920
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023761616161616163,
      "loss": 1.3886,
      "theoretical_loss": 3.365384109677583,
      "tokens_seen": 2529427456
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023759595959595958,
      "loss": 1.4674,
      "theoretical_loss": 3.365377170835382,
      "tokens_seen": 2529492992
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002375757575757576,
      "loss": 1.3054,
      "theoretical_loss": 3.3653702322232917,
      "tokens_seen": 2529558528
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023755555555555555,
      "loss": 1.4446,
      "theoretical_loss": 3.3653632938412983,
      "tokens_seen": 2529624064
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023753535353535356,
      "loss": 1.2631,
      "theoretical_loss": 3.3653563556893875,
      "tokens_seen": 2529689600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023751515151515152,
      "loss": 1.3245,
      "theoretical_loss": 3.365349417767547,
      "tokens_seen": 2529755136
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002374949494949495,
      "loss": 1.3694,
      "theoretical_loss": 3.3653424800757623,
      "tokens_seen": 2529820672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023747474747474749,
      "loss": 1.2076,
      "theoretical_loss": 3.3653355426140195,
      "tokens_seen": 2529886208
    },
    {
      "epoch": 0.53,
      "objective/train/aligned_token_count": 0.45747801661491394,
      "objective/train/avg_token_score": 0.006984468549489975,
      "objective/train/docs_used": 1422170,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5191912651062012,
      "objective/train/original_loss": 4.257667541503906,
      "objective/train/theoretical_loss": 3.3653286053823064,
      "objective/train/tokens_used": 888810976,
      "objective/train/unlikelihood_loss": 0.053750962018966675,
      "theoretical_loss": 3.3653286053823064,
      "tokens_seen": 2529951744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023745454545454544,
      "loss": 1.4101,
      "theoretical_loss": 3.3653286053823064,
      "tokens_seen": 2529951744
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023743434343434345,
      "loss": 1.3409,
      "theoretical_loss": 3.365321668380608,
      "tokens_seen": 2530017280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002374141414141414,
      "loss": 1.4591,
      "theoretical_loss": 3.365314731608912,
      "tokens_seen": 2530082816
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023739393939393942,
      "loss": 1.4588,
      "theoretical_loss": 3.3653077950672032,
      "tokens_seen": 2530148352
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023737373737373738,
      "loss": 1.3746,
      "theoretical_loss": 3.3653008587554694,
      "tokens_seen": 2530213888
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023735353535353534,
      "loss": 1.1923,
      "theoretical_loss": 3.3652939226736964,
      "tokens_seen": 2530279424
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023733333333333335,
      "loss": 1.1716,
      "theoretical_loss": 3.3652869868218707,
      "tokens_seen": 2530344960
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002373131313131313,
      "loss": 1.2397,
      "theoretical_loss": 3.3652800511999788,
      "tokens_seen": 2530410496
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023729292929292931,
      "loss": 1.4692,
      "theoretical_loss": 3.3652731158080074,
      "tokens_seen": 2530476032
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023727272727272727,
      "loss": 1.3429,
      "theoretical_loss": 3.3652661806459423,
      "tokens_seen": 2530541568
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023725252525252526,
      "loss": 1.3391,
      "theoretical_loss": 3.36525924571377,
      "tokens_seen": 2530607104
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023723232323232324,
      "loss": 1.2558,
      "theoretical_loss": 3.365252311011478,
      "tokens_seen": 2530672640
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023721212121212122,
      "loss": 1.182,
      "theoretical_loss": 3.3652453765390513,
      "tokens_seen": 2530738176
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002371919191919192,
      "loss": 1.1944,
      "theoretical_loss": 3.365238442296477,
      "tokens_seen": 2530803712
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023717171717171716,
      "loss": 1.2652,
      "theoretical_loss": 3.3652315082837414,
      "tokens_seen": 2530869248
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023715151515151515,
      "loss": 1.3032,
      "theoretical_loss": 3.3652245745008313,
      "tokens_seen": 2530934784
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023713131313131313,
      "loss": 1.1521,
      "theoretical_loss": 3.3652176409477326,
      "tokens_seen": 2531000320
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023711111111111112,
      "loss": 1.3609,
      "theoretical_loss": 3.3652107076244318,
      "tokens_seen": 2531065856
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002370909090909091,
      "loss": 1.4652,
      "theoretical_loss": 3.3652037745309156,
      "tokens_seen": 2531131392
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023707070707070708,
      "loss": 1.0591,
      "theoretical_loss": 3.3651968416671707,
      "tokens_seen": 2531196928
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023705050505050504,
      "loss": 1.3319,
      "theoretical_loss": 3.3651899090331825,
      "tokens_seen": 2531262464
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023703030303030305,
      "loss": 1.4888,
      "theoretical_loss": 3.3651829766289385,
      "tokens_seen": 2531328000
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000237010101010101,
      "loss": 1.0604,
      "theoretical_loss": 3.365176044454425,
      "tokens_seen": 2531393536
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000236989898989899,
      "loss": 1.3491,
      "theoretical_loss": 3.3651691125096277,
      "tokens_seen": 2531459072
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023696969696969698,
      "loss": 1.4157,
      "theoretical_loss": 3.3651621807945338,
      "tokens_seen": 2531524608
    },
    {
      "epoch": 0.53,
      "objective/train/aligned_token_count": 0.4508797526359558,
      "objective/train/avg_token_score": 0.021863192319869995,
      "objective/train/docs_used": 1422669,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4423658847808838,
      "objective/train/original_loss": 4.249288082122803,
      "objective/train/theoretical_loss": 3.3651552493091295,
      "objective/train/tokens_used": 890449376,
      "objective/train/unlikelihood_loss": 0.07330141961574554,
      "theoretical_loss": 3.3651552493091295,
      "tokens_seen": 2531590144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023694949494949493,
      "loss": 1.2622,
      "theoretical_loss": 3.3651552493091295,
      "tokens_seen": 2531590144
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023692929292929294,
      "loss": 1.3388,
      "theoretical_loss": 3.365148318053401,
      "tokens_seen": 2531655680
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002369090909090909,
      "loss": 1.3962,
      "theoretical_loss": 3.365141387027335,
      "tokens_seen": 2531721216
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002368888888888889,
      "loss": 1.3432,
      "theoretical_loss": 3.365134456230918,
      "tokens_seen": 2531786752
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023686868686868687,
      "loss": 1.3954,
      "theoretical_loss": 3.365127525664136,
      "tokens_seen": 2531852288
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023684848484848485,
      "loss": 1.3687,
      "theoretical_loss": 3.365120595326976,
      "tokens_seen": 2531917824
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023682828282828284,
      "loss": 1.1154,
      "theoretical_loss": 3.365113665219424,
      "tokens_seen": 2531983360
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002368080808080808,
      "loss": 1.4281,
      "theoretical_loss": 3.365106735341467,
      "tokens_seen": 2532048896
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002367878787878788,
      "loss": 1.2594,
      "theoretical_loss": 3.365099805693091,
      "tokens_seen": 2532114432
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023676767676767676,
      "loss": 1.3483,
      "theoretical_loss": 3.365092876274283,
      "tokens_seen": 2532179968
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023674747474747475,
      "loss": 1.5195,
      "theoretical_loss": 3.3650859470850287,
      "tokens_seen": 2532245504
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023672727272727273,
      "loss": 1.2874,
      "theoretical_loss": 3.3650790181253147,
      "tokens_seen": 2532311040
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023670707070707071,
      "loss": 1.8443,
      "theoretical_loss": 3.3650720893951283,
      "tokens_seen": 2532376576
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002366868686868687,
      "loss": 1.1462,
      "theoretical_loss": 3.3650651608944546,
      "tokens_seen": 2532442112
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023666666666666668,
      "loss": 1.1726,
      "theoretical_loss": 3.365058232623281,
      "tokens_seen": 2532507648
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023664646464646464,
      "loss": 1.4761,
      "theoretical_loss": 3.365051304581594,
      "tokens_seen": 2532573184
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023662626262626262,
      "loss": 1.2317,
      "theoretical_loss": 3.3650443767693794,
      "tokens_seen": 2532638720
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002366060606060606,
      "loss": 1.4504,
      "theoretical_loss": 3.365037449186624,
      "tokens_seen": 2532704256
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002365858585858586,
      "loss": 1.2658,
      "theoretical_loss": 3.3650305218333147,
      "tokens_seen": 2532769792
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023656565656565657,
      "loss": 1.2804,
      "theoretical_loss": 3.365023594709437,
      "tokens_seen": 2532835328
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023654545454545456,
      "loss": 1.4286,
      "theoretical_loss": 3.3650166678149787,
      "tokens_seen": 2532900864
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023652525252525254,
      "loss": 1.4159,
      "theoretical_loss": 3.365009741149925,
      "tokens_seen": 2532966400
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002365050505050505,
      "loss": 1.0982,
      "theoretical_loss": 3.365002814714263,
      "tokens_seen": 2533031936
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002364848484848485,
      "loss": 1.5199,
      "theoretical_loss": 3.3649958885079796,
      "tokens_seen": 2533097472
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023646464646464647,
      "loss": 1.1683,
      "theoretical_loss": 3.36498896253106,
      "tokens_seen": 2533163008
    },
    {
      "epoch": 0.53,
      "objective/train/aligned_token_count": 0.3450024425983429,
      "objective/train/avg_token_score": 0.02846650592982769,
      "objective/train/docs_used": 1423865,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9891325235366821,
      "objective/train/original_loss": 4.681108474731445,
      "objective/train/theoretical_loss": 3.3649820367834917,
      "objective/train/tokens_used": 892087776,
      "objective/train/unlikelihood_loss": 0.05223282054066658,
      "theoretical_loss": 3.3649820367834917,
      "tokens_seen": 2533228544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023644444444444445,
      "loss": 1.3064,
      "theoretical_loss": 3.3649820367834917,
      "tokens_seen": 2533228544
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023642424242424244,
      "loss": 1.5889,
      "theoretical_loss": 3.364975111265261,
      "tokens_seen": 2533294080
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002364040404040404,
      "loss": 1.3304,
      "theoretical_loss": 3.3649681859763536,
      "tokens_seen": 2533359616
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002363838383838384,
      "loss": 1.375,
      "theoretical_loss": 3.3649612609167567,
      "tokens_seen": 2533425152
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023636363636363636,
      "loss": 1.0669,
      "theoretical_loss": 3.3649543360864573,
      "tokens_seen": 2533490688
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023634343434343437,
      "loss": 1.2553,
      "theoretical_loss": 3.3649474114854407,
      "tokens_seen": 2533556224
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023632323232323233,
      "loss": 1.4688,
      "theoretical_loss": 3.3649404871136945,
      "tokens_seen": 2533621760
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023630303030303028,
      "loss": 1.2267,
      "theoretical_loss": 3.364933562971204,
      "tokens_seen": 2533687296
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002362828282828283,
      "loss": 1.4181,
      "theoretical_loss": 3.364926639057957,
      "tokens_seen": 2533752832
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023626262626262625,
      "loss": 1.3469,
      "theoretical_loss": 3.3649197153739383,
      "tokens_seen": 2533818368
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023624242424242426,
      "loss": 1.3599,
      "theoretical_loss": 3.364912791919136,
      "tokens_seen": 2533883904
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023622222222222222,
      "loss": 1.3668,
      "theoretical_loss": 3.364905868693536,
      "tokens_seen": 2533949440
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002362020202020202,
      "loss": 1.3455,
      "theoretical_loss": 3.3648989456971243,
      "tokens_seen": 2534014976
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002361818181818182,
      "loss": 1.4009,
      "theoretical_loss": 3.3648920229298884,
      "tokens_seen": 2534080512
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023616161616161617,
      "loss": 1.2751,
      "theoretical_loss": 3.364885100391814,
      "tokens_seen": 2534146048
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023614141414141416,
      "loss": 1.3546,
      "theoretical_loss": 3.3648781780828876,
      "tokens_seen": 2534211584
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002361212121212121,
      "loss": 1.3789,
      "theoretical_loss": 3.364871256003096,
      "tokens_seen": 2534277120
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002361010101010101,
      "loss": 1.1034,
      "theoretical_loss": 3.3648643341524256,
      "tokens_seen": 2534342656
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023608080808080808,
      "loss": 1.2934,
      "theoretical_loss": 3.3648574125308626,
      "tokens_seen": 2534408192
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023606060606060607,
      "loss": 0.9571,
      "theoretical_loss": 3.3648504911383936,
      "tokens_seen": 2534473728
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023604040404040405,
      "loss": 1.2133,
      "theoretical_loss": 3.364843569975006,
      "tokens_seen": 2534539264
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023602020202020203,
      "loss": 1.3099,
      "theoretical_loss": 3.3648366490406847,
      "tokens_seen": 2534604800
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000236,
      "loss": 1.3508,
      "theoretical_loss": 3.3648297283354176,
      "tokens_seen": 2534670336
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000235979797979798,
      "loss": 1.3205,
      "theoretical_loss": 3.3648228078591904,
      "tokens_seen": 2534735872
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023595959595959596,
      "loss": 1.3099,
      "theoretical_loss": 3.3648158876119902,
      "tokens_seen": 2534801408
    },
    {
      "epoch": 0.53,
      "objective/train/aligned_token_count": 0.43511730432510376,
      "objective/train/avg_token_score": 0.013051752001047134,
      "objective/train/docs_used": 1424480,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5836750268936157,
      "objective/train/original_loss": 4.756967067718506,
      "objective/train/theoretical_loss": 3.3648089675938024,
      "objective/train/tokens_used": 893726176,
      "objective/train/unlikelihood_loss": 0.05763814598321915,
      "theoretical_loss": 3.3648089675938024,
      "tokens_seen": 2534866944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023593939393939394,
      "loss": 1.4208,
      "theoretical_loss": 3.3648089675938024,
      "tokens_seen": 2534866944
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023591919191919193,
      "loss": 1.4503,
      "theoretical_loss": 3.364802047804615,
      "tokens_seen": 2534932480
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023589898989898988,
      "loss": 1.2158,
      "theoretical_loss": 3.3647951282444133,
      "tokens_seen": 2534998016
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002358787878787879,
      "loss": 1.3258,
      "theoretical_loss": 3.3647882089131844,
      "tokens_seen": 2535063552
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023585858585858585,
      "loss": 1.3832,
      "theoretical_loss": 3.3647812898109146,
      "tokens_seen": 2535129088
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023583838383838386,
      "loss": 1.4235,
      "theoretical_loss": 3.36477437093759,
      "tokens_seen": 2535194624
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023581818181818182,
      "loss": 1.2754,
      "theoretical_loss": 3.364767452293198,
      "tokens_seen": 2535260160
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002357979797979798,
      "loss": 1.3073,
      "theoretical_loss": 3.3647605338777247,
      "tokens_seen": 2535325696
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002357777777777778,
      "loss": 1.3074,
      "theoretical_loss": 3.3647536156911566,
      "tokens_seen": 2535391232
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023575757575757574,
      "loss": 1.3893,
      "theoretical_loss": 3.36474669773348,
      "tokens_seen": 2535456768
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023573737373737375,
      "loss": 1.0148,
      "theoretical_loss": 3.364739780004682,
      "tokens_seen": 2535522304
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002357171717171717,
      "loss": 1.3868,
      "theoretical_loss": 3.364732862504748,
      "tokens_seen": 2535587840
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023569696969696972,
      "loss": 1.2674,
      "theoretical_loss": 3.3647259452336655,
      "tokens_seen": 2535653376
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023567676767676768,
      "loss": 1.2535,
      "theoretical_loss": 3.3647190281914208,
      "tokens_seen": 2535718912
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023565656565656566,
      "loss": 1.3859,
      "theoretical_loss": 3.3647121113780005,
      "tokens_seen": 2535784448
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023563636363636365,
      "loss": 1.325,
      "theoretical_loss": 3.364705194793391,
      "tokens_seen": 2535849984
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002356161616161616,
      "loss": 1.6488,
      "theoretical_loss": 3.3646982784375785,
      "tokens_seen": 2535915520
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023559595959595962,
      "loss": 1.4031,
      "theoretical_loss": 3.36469136231055,
      "tokens_seen": 2535981056
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023557575757575757,
      "loss": 1.4973,
      "theoretical_loss": 3.3646844464122916,
      "tokens_seen": 2536046592
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023555555555555556,
      "loss": 1.3711,
      "theoretical_loss": 3.36467753074279,
      "tokens_seen": 2536112128
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023553535353535354,
      "loss": 1.3144,
      "theoretical_loss": 3.364670615302032,
      "tokens_seen": 2536177664
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023551515151515152,
      "loss": 1.3715,
      "theoretical_loss": 3.364663700090004,
      "tokens_seen": 2536243200
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002354949494949495,
      "loss": 1.3104,
      "theoretical_loss": 3.3646567851066917,
      "tokens_seen": 2536308736
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002354747474747475,
      "loss": 1.4137,
      "theoretical_loss": 3.364649870352083,
      "tokens_seen": 2536374272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023545454545454545,
      "loss": 1.3086,
      "theoretical_loss": 3.3646429558261635,
      "tokens_seen": 2536439808
    },
    {
      "epoch": 0.53,
      "objective/train/aligned_token_count": 0.3492790758609772,
      "objective/train/avg_token_score": 0.005822951905429363,
      "objective/train/docs_used": 1425841,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.127997875213623,
      "objective/train/original_loss": 5.317699432373047,
      "objective/train/theoretical_loss": 3.36463604152892,
      "objective/train/tokens_used": 895364576,
      "objective/train/unlikelihood_loss": 0.04770021140575409,
      "theoretical_loss": 3.36463604152892,
      "tokens_seen": 2536505344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023543434343434343,
      "loss": 1.2168,
      "theoretical_loss": 3.36463604152892,
      "tokens_seen": 2536505344
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023541414141414142,
      "loss": 1.4928,
      "theoretical_loss": 3.364629127460339,
      "tokens_seen": 2536570880
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002353939393939394,
      "loss": 1.2984,
      "theoretical_loss": 3.3646222136204074,
      "tokens_seen": 2536636416
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023537373737373738,
      "loss": 1.2086,
      "theoretical_loss": 3.364615300009111,
      "tokens_seen": 2536701952
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023535353535353534,
      "loss": 1.3636,
      "theoretical_loss": 3.364608386626437,
      "tokens_seen": 2536767488
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023533333333333335,
      "loss": 1.3955,
      "theoretical_loss": 3.3646014734723715,
      "tokens_seen": 2536833024
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002353131313131313,
      "loss": 1.2319,
      "theoretical_loss": 3.3645945605469008,
      "tokens_seen": 2536898560
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023529292929292932,
      "loss": 1.4339,
      "theoretical_loss": 3.3645876478500125,
      "tokens_seen": 2536964096
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023527272727272728,
      "loss": 1.2921,
      "theoretical_loss": 3.364580735381692,
      "tokens_seen": 2537029632
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023525252525252523,
      "loss": 1.4688,
      "theoretical_loss": 3.364573823141926,
      "tokens_seen": 2537095168
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023523232323232325,
      "loss": 1.3263,
      "theoretical_loss": 3.3645669111307024,
      "tokens_seen": 2537160704
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002352121212121212,
      "loss": 1.3955,
      "theoretical_loss": 3.364559999348006,
      "tokens_seen": 2537226240
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002351919191919192,
      "loss": 1.2073,
      "theoretical_loss": 3.364553087793824,
      "tokens_seen": 2537291776
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023517171717171717,
      "loss": 1.3209,
      "theoretical_loss": 3.364546176468143,
      "tokens_seen": 2537357312
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023515151515151515,
      "loss": 1.1212,
      "theoretical_loss": 3.3645392653709494,
      "tokens_seen": 2537422848
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023513131313131314,
      "loss": 1.4721,
      "theoretical_loss": 3.36453235450223,
      "tokens_seen": 2537488384
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023511111111111112,
      "loss": 1.336,
      "theoretical_loss": 3.364525443861971,
      "tokens_seen": 2537553920
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002350909090909091,
      "loss": 1.1446,
      "theoretical_loss": 3.3645185334501595,
      "tokens_seen": 2537619456
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023507070707070706,
      "loss": 1.5687,
      "theoretical_loss": 3.3645116232667815,
      "tokens_seen": 2537684992
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023505050505050505,
      "loss": 1.2372,
      "theoretical_loss": 3.364504713311824,
      "tokens_seen": 2537750528
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023503030303030303,
      "loss": 1.2258,
      "theoretical_loss": 3.3644978035852726,
      "tokens_seen": 2537816064
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023501010101010101,
      "loss": 1.2778,
      "theoretical_loss": 3.364490894087115,
      "tokens_seen": 2537881600
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.000234989898989899,
      "loss": 1.4013,
      "theoretical_loss": 3.3644839848173373,
      "tokens_seen": 2537947136
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023496969696969698,
      "loss": 1.286,
      "theoretical_loss": 3.364477075775926,
      "tokens_seen": 2538012672
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00023494949494949494,
      "loss": 1.5532,
      "theoretical_loss": 3.364470166962868,
      "tokens_seen": 2538078208
    },
    {
      "epoch": 0.53,
      "objective/train/aligned_token_count": 0.4157502353191376,
      "objective/train/avg_token_score": 0.005650267470628023,
      "objective/train/docs_used": 1426240,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1659504175186157,
      "objective/train/original_loss": 4.030059337615967,
      "objective/train/theoretical_loss": 3.364463258378149,
      "objective/train/tokens_used": 897002976,
      "objective/train/unlikelihood_loss": 0.055918820202350616,
      "theoretical_loss": 3.364463258378149,
      "tokens_seen": 2538143744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023492929292929295,
      "loss": 1.2625,
      "theoretical_loss": 3.364463258378149,
      "tokens_seen": 2538143744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002349090909090909,
      "loss": 1.3739,
      "theoretical_loss": 3.364456350021756,
      "tokens_seen": 2538209280
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002348888888888889,
      "loss": 1.3657,
      "theoretical_loss": 3.3644494418936763,
      "tokens_seen": 2538274816
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023486868686868688,
      "loss": 1.3929,
      "theoretical_loss": 3.3644425339938957,
      "tokens_seen": 2538340352
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023484848484848486,
      "loss": 1.4061,
      "theoretical_loss": 3.364435626322401,
      "tokens_seen": 2538405888
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023482828282828284,
      "loss": 1.3678,
      "theoretical_loss": 3.3644287188791786,
      "tokens_seen": 2538471424
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002348080808080808,
      "loss": 1.3516,
      "theoretical_loss": 3.364421811664215,
      "tokens_seen": 2538536960
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002347878787878788,
      "loss": 1.4054,
      "theoretical_loss": 3.3644149046774965,
      "tokens_seen": 2538602496
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023476767676767677,
      "loss": 1.214,
      "theoretical_loss": 3.3644079979190105,
      "tokens_seen": 2538668032
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023474747474747475,
      "loss": 1.5873,
      "theoretical_loss": 3.364401091388743,
      "tokens_seen": 2538733568
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023472727272727274,
      "loss": 1.4743,
      "theoretical_loss": 3.364394185086681,
      "tokens_seen": 2538799104
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002347070707070707,
      "loss": 1.5385,
      "theoretical_loss": 3.3643872790128104,
      "tokens_seen": 2538864640
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002346868686868687,
      "loss": 1.2918,
      "theoretical_loss": 3.3643803731671182,
      "tokens_seen": 2538930176
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023466666666666666,
      "loss": 1.5171,
      "theoretical_loss": 3.364373467549591,
      "tokens_seen": 2538995712
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023464646464646467,
      "loss": 1.3441,
      "theoretical_loss": 3.3643665621602157,
      "tokens_seen": 2539061248
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023462626262626263,
      "loss": 1.0152,
      "theoretical_loss": 3.3643596569989778,
      "tokens_seen": 2539126784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002346060606060606,
      "loss": 1.3226,
      "theoretical_loss": 3.3643527520658645,
      "tokens_seen": 2539192320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002345858585858586,
      "loss": 1.5369,
      "theoretical_loss": 3.3643458473608625,
      "tokens_seen": 2539257856
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023456565656565655,
      "loss": 1.0837,
      "theoretical_loss": 3.3643389428839585,
      "tokens_seen": 2539323392
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023454545454545456,
      "loss": 1.3365,
      "theoretical_loss": 3.3643320386351387,
      "tokens_seen": 2539388928
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023452525252525252,
      "loss": 1.2933,
      "theoretical_loss": 3.36432513461439,
      "tokens_seen": 2539454464
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002345050505050505,
      "loss": 1.4137,
      "theoretical_loss": 3.3643182308216986,
      "tokens_seen": 2539520000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002344848484848485,
      "loss": 1.4326,
      "theoretical_loss": 3.3643113272570515,
      "tokens_seen": 2539585536
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023446464646464647,
      "loss": 1.5344,
      "theoretical_loss": 3.3643044239204345,
      "tokens_seen": 2539651072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023444444444444446,
      "loss": 1.3597,
      "theoretical_loss": 3.3642975208118355,
      "tokens_seen": 2539716608
    },
    {
      "epoch": 0.54,
      "objective/train/aligned_token_count": 0.3921676576137543,
      "objective/train/avg_token_score": 0.021439138799905777,
      "objective/train/docs_used": 1426896,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1117773056030273,
      "objective/train/original_loss": 4.441211223602295,
      "objective/train/theoretical_loss": 3.36429061793124,
      "objective/train/tokens_used": 898641376,
      "objective/train/unlikelihood_loss": 0.060393474996089935,
      "theoretical_loss": 3.36429061793124,
      "tokens_seen": 2539782144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023442424242424244,
      "loss": 1.2237,
      "theoretical_loss": 3.36429061793124,
      "tokens_seen": 2539782144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002344040404040404,
      "loss": 1.333,
      "theoretical_loss": 3.364283715278635,
      "tokens_seen": 2539847680
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023438383838383838,
      "loss": 1.5269,
      "theoretical_loss": 3.364276812854007,
      "tokens_seen": 2539913216
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023436363636363637,
      "loss": 1.4377,
      "theoretical_loss": 3.3642699106573426,
      "tokens_seen": 2539978752
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023434343434343435,
      "loss": 1.5252,
      "theoretical_loss": 3.3642630086886287,
      "tokens_seen": 2540044288
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023432323232323233,
      "loss": 1.259,
      "theoretical_loss": 3.364256106947851,
      "tokens_seen": 2540109824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002343030303030303,
      "loss": 1.4344,
      "theoretical_loss": 3.364249205434997,
      "tokens_seen": 2540175360
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002342828282828283,
      "loss": 1.3917,
      "theoretical_loss": 3.364242304150053,
      "tokens_seen": 2540240896
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023426262626262626,
      "loss": 1.2939,
      "theoretical_loss": 3.3642354030930055,
      "tokens_seen": 2540306432
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023424242424242427,
      "loss": 1.4491,
      "theoretical_loss": 3.3642285022638414,
      "tokens_seen": 2540371968
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023422222222222223,
      "loss": 1.2389,
      "theoretical_loss": 3.3642216016625466,
      "tokens_seen": 2540437504
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023420202020202018,
      "loss": 1.3132,
      "theoretical_loss": 3.3642147012891086,
      "tokens_seen": 2540503040
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002341818181818182,
      "loss": 1.3341,
      "theoretical_loss": 3.3642078011435133,
      "tokens_seen": 2540568576
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023416161616161615,
      "loss": 1.1726,
      "theoretical_loss": 3.3642009012257477,
      "tokens_seen": 2540634112
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023414141414141416,
      "loss": 1.2437,
      "theoretical_loss": 3.364194001535798,
      "tokens_seen": 2540699648
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023412121212121212,
      "loss": 1.4737,
      "theoretical_loss": 3.3641871020736516,
      "tokens_seen": 2540765184
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002341010101010101,
      "loss": 1.3121,
      "theoretical_loss": 3.364180202839294,
      "tokens_seen": 2540830720
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002340808080808081,
      "loss": 1.4017,
      "theoretical_loss": 3.3641733038327124,
      "tokens_seen": 2540896256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023406060606060604,
      "loss": 1.4592,
      "theoretical_loss": 3.3641664050538935,
      "tokens_seen": 2540961792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023404040404040405,
      "loss": 1.3779,
      "theoretical_loss": 3.364159506502824,
      "tokens_seen": 2541027328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000234020202020202,
      "loss": 1.371,
      "theoretical_loss": 3.36415260817949,
      "tokens_seen": 2541092864
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023400000000000002,
      "loss": 1.1932,
      "theoretical_loss": 3.3641457100838785,
      "tokens_seen": 2541158400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023397979797979798,
      "loss": 1.2663,
      "theoretical_loss": 3.3641388122159763,
      "tokens_seen": 2541223936
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023395959595959596,
      "loss": 1.3529,
      "theoretical_loss": 3.3641319145757693,
      "tokens_seen": 2541289472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023393939393939395,
      "loss": 1.566,
      "theoretical_loss": 3.3641250171632446,
      "tokens_seen": 2541355008
    },
    {
      "epoch": 0.54,
      "objective/train/aligned_token_count": 0.37982648611068726,
      "objective/train/avg_token_score": 0.03235911577939987,
      "objective/train/docs_used": 1428293,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1447076797485352,
      "objective/train/original_loss": 5.111784934997559,
      "objective/train/theoretical_loss": 3.364118119978389,
      "objective/train/tokens_used": 900279776,
      "objective/train/unlikelihood_loss": 0.036657728254795074,
      "theoretical_loss": 3.364118119978389,
      "tokens_seen": 2541420544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023391919191919193,
      "loss": 1.2365,
      "theoretical_loss": 3.364118119978389,
      "tokens_seen": 2541420544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023389898989898992,
      "loss": 1.4999,
      "theoretical_loss": 3.3641112230211885,
      "tokens_seen": 2541486080
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023387878787878787,
      "loss": 1.0834,
      "theoretical_loss": 3.3641043262916304,
      "tokens_seen": 2541551616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023385858585858586,
      "loss": 1.3696,
      "theoretical_loss": 3.3640974297897004,
      "tokens_seen": 2541617152
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023383838383838384,
      "loss": 1.1965,
      "theoretical_loss": 3.364090533515386,
      "tokens_seen": 2541682688
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023381818181818182,
      "loss": 1.272,
      "theoretical_loss": 3.364083637468674,
      "tokens_seen": 2541748224
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002337979797979798,
      "loss": 1.2462,
      "theoretical_loss": 3.36407674164955,
      "tokens_seen": 2541813760
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002337777777777778,
      "loss": 1.4855,
      "theoretical_loss": 3.364069846058001,
      "tokens_seen": 2541879296
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023375757575757575,
      "loss": 1.3372,
      "theoretical_loss": 3.3640629506940147,
      "tokens_seen": 2541944832
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023373737373737376,
      "loss": 1.4431,
      "theoretical_loss": 3.364056055557576,
      "tokens_seen": 2542010368
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023371717171717172,
      "loss": 1.3309,
      "theoretical_loss": 3.3640491606486727,
      "tokens_seen": 2542075904
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002336969696969697,
      "loss": 1.4414,
      "theoretical_loss": 3.364042265967291,
      "tokens_seen": 2542141440
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023367676767676768,
      "loss": 1.429,
      "theoretical_loss": 3.3640353715134172,
      "tokens_seen": 2542206976
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023365656565656564,
      "loss": 1.3093,
      "theoretical_loss": 3.364028477287039,
      "tokens_seen": 2542272512
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023363636363636365,
      "loss": 1.468,
      "theoretical_loss": 3.364021583288142,
      "tokens_seen": 2542338048
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002336161616161616,
      "loss": 1.2566,
      "theoretical_loss": 3.364014689516713,
      "tokens_seen": 2542403584
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023359595959595962,
      "loss": 1.4161,
      "theoretical_loss": 3.3640077959727392,
      "tokens_seen": 2542469120
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023357575757575758,
      "loss": 1.1462,
      "theoretical_loss": 3.3640009026562065,
      "tokens_seen": 2542534656
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023355555555555556,
      "loss": 1.1945,
      "theoretical_loss": 3.363994009567102,
      "tokens_seen": 2542600192
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023353535353535355,
      "loss": 1.4743,
      "theoretical_loss": 3.3639871167054123,
      "tokens_seen": 2542665728
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002335151515151515,
      "loss": 1.5498,
      "theoretical_loss": 3.3639802240711236,
      "tokens_seen": 2542731264
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002334949494949495,
      "loss": 1.4684,
      "theoretical_loss": 3.3639733316642233,
      "tokens_seen": 2542796800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023347474747474747,
      "loss": 1.347,
      "theoretical_loss": 3.3639664394846975,
      "tokens_seen": 2542862336
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023345454545454545,
      "loss": 1.2627,
      "theoretical_loss": 3.363959547532533,
      "tokens_seen": 2542927872
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023343434343434344,
      "loss": 1.4122,
      "theoretical_loss": 3.363952655807716,
      "tokens_seen": 2542993408
    },
    {
      "epoch": 0.54,
      "objective/train/aligned_token_count": 0.396871954202652,
      "objective/train/avg_token_score": 0.0038444444071501493,
      "objective/train/docs_used": 1429761,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.442063808441162,
      "objective/train/original_loss": 4.91703987121582,
      "objective/train/theoretical_loss": 3.363945764310234,
      "objective/train/tokens_used": 901918176,
      "objective/train/unlikelihood_loss": 0.0629727765917778,
      "theoretical_loss": 3.363945764310234,
      "tokens_seen": 2543058944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023341414141414142,
      "loss": 1.4731,
      "theoretical_loss": 3.363945764310234,
      "tokens_seen": 2543058944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002333939393939394,
      "loss": 1.2243,
      "theoretical_loss": 3.363938873040073,
      "tokens_seen": 2543124480
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002333737373737374,
      "loss": 1.3155,
      "theoretical_loss": 3.36393198199722,
      "tokens_seen": 2543190016
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023335353535353535,
      "loss": 1.3094,
      "theoretical_loss": 3.3639250911816614,
      "tokens_seen": 2543255552
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023333333333333333,
      "loss": 1.244,
      "theoretical_loss": 3.3639182005933836,
      "tokens_seen": 2543321088
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023331313131313131,
      "loss": 1.3326,
      "theoretical_loss": 3.3639113102323743,
      "tokens_seen": 2543386624
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002332929292929293,
      "loss": 1.0834,
      "theoretical_loss": 3.363904420098619,
      "tokens_seen": 2543452160
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023327272727272728,
      "loss": 1.2495,
      "theoretical_loss": 3.3638975301921046,
      "tokens_seen": 2543517696
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023325252525252524,
      "loss": 1.4094,
      "theoretical_loss": 3.363890640512818,
      "tokens_seen": 2543583232
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023323232323232325,
      "loss": 1.2475,
      "theoretical_loss": 3.3638837510607456,
      "tokens_seen": 2543648768
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002332121212121212,
      "loss": 1.618,
      "theoretical_loss": 3.3638768618358745,
      "tokens_seen": 2543714304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002331919191919192,
      "loss": 1.1673,
      "theoretical_loss": 3.363869972838191,
      "tokens_seen": 2543779840
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023317171717171718,
      "loss": 1.3532,
      "theoretical_loss": 3.363863084067682,
      "tokens_seen": 2543845376
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023315151515151516,
      "loss": 1.3522,
      "theoretical_loss": 3.363856195524334,
      "tokens_seen": 2543910912
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023313131313131314,
      "loss": 1.3618,
      "theoretical_loss": 3.3638493072081332,
      "tokens_seen": 2543976448
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002331111111111111,
      "loss": 1.2268,
      "theoretical_loss": 3.363842419119067,
      "tokens_seen": 2544041984
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002330909090909091,
      "loss": 1.3149,
      "theoretical_loss": 3.3638355312571218,
      "tokens_seen": 2544107520
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023307070707070707,
      "loss": 1.4428,
      "theoretical_loss": 3.363828643622284,
      "tokens_seen": 2544173056
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023305050505050508,
      "loss": 1.34,
      "theoretical_loss": 3.363821756214541,
      "tokens_seen": 2544238592
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023303030303030304,
      "loss": 1.2618,
      "theoretical_loss": 3.3638148690338783,
      "tokens_seen": 2544304128
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000233010101010101,
      "loss": 1.417,
      "theoretical_loss": 3.3638079820802838,
      "tokens_seen": 2544369664
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000232989898989899,
      "loss": 1.112,
      "theoretical_loss": 3.363801095353743,
      "tokens_seen": 2544435200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023296969696969696,
      "loss": 1.3139,
      "theoretical_loss": 3.3637942088542436,
      "tokens_seen": 2544500736
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023294949494949497,
      "loss": 1.3199,
      "theoretical_loss": 3.3637873225817714,
      "tokens_seen": 2544566272
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023292929292929293,
      "loss": 1.3685,
      "theoretical_loss": 3.363780436536314,
      "tokens_seen": 2544631808
    },
    {
      "epoch": 0.54,
      "objective/train/aligned_token_count": 0.5119134783744812,
      "objective/train/avg_token_score": 0.004078378900885582,
      "objective/train/docs_used": 1430428,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.622788906097412,
      "objective/train/original_loss": 4.2214531898498535,
      "objective/train/theoretical_loss": 3.363773550717857,
      "objective/train/tokens_used": 903556576,
      "objective/train/unlikelihood_loss": 0.05783025547862053,
      "theoretical_loss": 3.363773550717857,
      "tokens_seen": 2544697344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002329090909090909,
      "loss": 1.4896,
      "theoretical_loss": 3.363773550717857,
      "tokens_seen": 2544697344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002328888888888889,
      "loss": 1.2815,
      "theoretical_loss": 3.3637666651263882,
      "tokens_seen": 2544762880
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023286868686868688,
      "loss": 1.7805,
      "theoretical_loss": 3.363759779761893,
      "tokens_seen": 2544828416
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023284848484848486,
      "loss": 1.2819,
      "theoretical_loss": 3.3637528946243593,
      "tokens_seen": 2544893952
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023282828282828282,
      "loss": 1.5301,
      "theoretical_loss": 3.3637460097137732,
      "tokens_seen": 2544959488
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002328080808080808,
      "loss": 1.2635,
      "theoretical_loss": 3.363739125030121,
      "tokens_seen": 2545025024
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002327878787878788,
      "loss": 1.4609,
      "theoretical_loss": 3.3637322405733903,
      "tokens_seen": 2545090560
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023276767676767677,
      "loss": 1.3277,
      "theoretical_loss": 3.363725356343567,
      "tokens_seen": 2545156096
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023274747474747476,
      "loss": 1.3934,
      "theoretical_loss": 3.363718472340638,
      "tokens_seen": 2545221632
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023272727272727274,
      "loss": 1.2605,
      "theoretical_loss": 3.3637115885645903,
      "tokens_seen": 2545287168
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002327070707070707,
      "loss": 1.1483,
      "theoretical_loss": 3.3637047050154103,
      "tokens_seen": 2545352704
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002326868686868687,
      "loss": 1.5009,
      "theoretical_loss": 3.363697821693085,
      "tokens_seen": 2545418240
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023266666666666667,
      "loss": 1.2392,
      "theoretical_loss": 3.3636909385976,
      "tokens_seen": 2545483776
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023264646464646465,
      "loss": 1.5498,
      "theoretical_loss": 3.3636840557289434,
      "tokens_seen": 2545549312
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023262626262626263,
      "loss": 1.2821,
      "theoretical_loss": 3.3636771730871007,
      "tokens_seen": 2545614848
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002326060606060606,
      "loss": 1.4102,
      "theoretical_loss": 3.3636702906720592,
      "tokens_seen": 2545680384
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002325858585858586,
      "loss": 1.16,
      "theoretical_loss": 3.3636634084838057,
      "tokens_seen": 2545745920
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023256565656565656,
      "loss": 1.426,
      "theoretical_loss": 3.363656526522327,
      "tokens_seen": 2545811456
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023254545454545457,
      "loss": 1.4644,
      "theoretical_loss": 3.363649644787609,
      "tokens_seen": 2545876992
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023252525252525253,
      "loss": 1.3648,
      "theoretical_loss": 3.363642763279639,
      "tokens_seen": 2545942528
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023250505050505048,
      "loss": 1.1824,
      "theoretical_loss": 3.3636358819984036,
      "tokens_seen": 2546008064
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002324848484848485,
      "loss": 1.4295,
      "theoretical_loss": 3.36362900094389,
      "tokens_seen": 2546073600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023246464646464645,
      "loss": 1.2138,
      "theoretical_loss": 3.3636221201160836,
      "tokens_seen": 2546139136
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023244444444444446,
      "loss": 1.3155,
      "theoretical_loss": 3.3636152395149725,
      "tokens_seen": 2546204672
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023242424242424242,
      "loss": 1.4892,
      "theoretical_loss": 3.3636083591405423,
      "tokens_seen": 2546270208
    },
    {
      "epoch": 0.54,
      "objective/train/aligned_token_count": 0.363575279712677,
      "objective/train/avg_token_score": 0.03922693803906441,
      "objective/train/docs_used": 1431073,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2787057161331177,
      "objective/train/original_loss": 4.806171894073486,
      "objective/train/theoretical_loss": 3.36360147899278,
      "objective/train/tokens_used": 905194976,
      "objective/train/unlikelihood_loss": 0.046168092638254166,
      "theoretical_loss": 3.36360147899278,
      "tokens_seen": 2546335744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002324040404040404,
      "loss": 1.3137,
      "theoretical_loss": 3.36360147899278,
      "tokens_seen": 2546335744
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002323838383838384,
      "loss": 1.2218,
      "theoretical_loss": 3.363594599071673,
      "tokens_seen": 2546401280
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023236363636363637,
      "loss": 1.1898,
      "theoretical_loss": 3.3635877193772075,
      "tokens_seen": 2546466816
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023234343434343436,
      "loss": 1.5328,
      "theoretical_loss": 3.36358083990937,
      "tokens_seen": 2546532352
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002323232323232323,
      "loss": 1.2323,
      "theoretical_loss": 3.363573960668147,
      "tokens_seen": 2546597888
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023230303030303032,
      "loss": 1.3354,
      "theoretical_loss": 3.363567081653526,
      "tokens_seen": 2546663424
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023228282828282828,
      "loss": 1.3616,
      "theoretical_loss": 3.363560202865493,
      "tokens_seen": 2546728960
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023226262626262626,
      "loss": 1.4183,
      "theoretical_loss": 3.363553324304035,
      "tokens_seen": 2546794496
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023224242424242425,
      "loss": 1.4082,
      "theoretical_loss": 3.363546445969139,
      "tokens_seen": 2546860032
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023222222222222223,
      "loss": 1.4266,
      "theoretical_loss": 3.363539567860791,
      "tokens_seen": 2546925568
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023220202020202022,
      "loss": 1.4354,
      "theoretical_loss": 3.363532689978978,
      "tokens_seen": 2546991104
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002321818181818182,
      "loss": 1.501,
      "theoretical_loss": 3.3635258123236875,
      "tokens_seen": 2547056640
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023216161616161616,
      "loss": 1.3073,
      "theoretical_loss": 3.363518934894905,
      "tokens_seen": 2547122176
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023214141414141414,
      "loss": 1.342,
      "theoretical_loss": 3.363512057692618,
      "tokens_seen": 2547187712
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023212121212121212,
      "loss": 1.3375,
      "theoretical_loss": 3.3635051807168126,
      "tokens_seen": 2547253248
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002321010101010101,
      "loss": 1.2146,
      "theoretical_loss": 3.363498303967476,
      "tokens_seen": 2547318784
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002320808080808081,
      "loss": 1.3824,
      "theoretical_loss": 3.363491427444595,
      "tokens_seen": 2547384320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023206060606060605,
      "loss": 1.3289,
      "theoretical_loss": 3.363484551148156,
      "tokens_seen": 2547449856
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023204040404040406,
      "loss": 1.0459,
      "theoretical_loss": 3.363477675078146,
      "tokens_seen": 2547515392
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023202020202020202,
      "loss": 1.3472,
      "theoretical_loss": 3.363470799234551,
      "tokens_seen": 2547580928
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023200000000000003,
      "loss": 1.1775,
      "theoretical_loss": 3.3634639236173585,
      "tokens_seen": 2547646464
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023197979797979799,
      "loss": 1.4495,
      "theoretical_loss": 3.363457048226555,
      "tokens_seen": 2547712000
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023195959595959594,
      "loss": 1.4478,
      "theoretical_loss": 3.3634501730621276,
      "tokens_seen": 2547777536
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023193939393939395,
      "loss": 1.3947,
      "theoretical_loss": 3.363443298124062,
      "tokens_seen": 2547843072
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002319191919191919,
      "loss": 1.2028,
      "theoretical_loss": 3.3634364234123457,
      "tokens_seen": 2547908608
    },
    {
      "epoch": 0.54,
      "objective/train/aligned_token_count": 0.4233871102333069,
      "objective/train/avg_token_score": 0.023195581510663033,
      "objective/train/docs_used": 1432125,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.348218560218811,
      "objective/train/original_loss": 4.791736602783203,
      "objective/train/theoretical_loss": 3.363429548926965,
      "objective/train/tokens_used": 906833376,
      "objective/train/unlikelihood_loss": 0.05622988939285278,
      "theoretical_loss": 3.363429548926965,
      "tokens_seen": 2547974144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023189898989898992,
      "loss": 1.2827,
      "theoretical_loss": 3.363429548926965,
      "tokens_seen": 2547974144
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023187878787878788,
      "loss": 1.606,
      "theoretical_loss": 3.3634226746679077,
      "tokens_seen": 2548039680
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023185858585858586,
      "loss": 1.4382,
      "theoretical_loss": 3.3634158006351593,
      "tokens_seen": 2548105216
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023183838383838385,
      "loss": 1.3817,
      "theoretical_loss": 3.363408926828707,
      "tokens_seen": 2548170752
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023181818181818183,
      "loss": 1.3059,
      "theoretical_loss": 3.3634020532485374,
      "tokens_seen": 2548236288
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023179797979797981,
      "loss": 1.3484,
      "theoretical_loss": 3.3633951798946375,
      "tokens_seen": 2548301824
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023177777777777777,
      "loss": 1.5243,
      "theoretical_loss": 3.3633883067669936,
      "tokens_seen": 2548367360
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023175757575757575,
      "loss": 1.4001,
      "theoretical_loss": 3.3633814338655927,
      "tokens_seen": 2548432896
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023173737373737374,
      "loss": 1.302,
      "theoretical_loss": 3.363374561190422,
      "tokens_seen": 2548498432
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023171717171717172,
      "loss": 1.2268,
      "theoretical_loss": 3.363367688741467,
      "tokens_seen": 2548563968
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002316969696969697,
      "loss": 1.3169,
      "theoretical_loss": 3.363360816518716,
      "tokens_seen": 2548629504
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002316767676767677,
      "loss": 1.2963,
      "theoretical_loss": 3.3633539445221543,
      "tokens_seen": 2548695040
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023165656565656565,
      "loss": 1.401,
      "theoretical_loss": 3.3633470727517696,
      "tokens_seen": 2548760576
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023163636363636366,
      "loss": 1.1259,
      "theoretical_loss": 3.3633402012075484,
      "tokens_seen": 2548826112
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023161616161616162,
      "loss": 1.3877,
      "theoretical_loss": 3.363333329889477,
      "tokens_seen": 2548891648
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002315959595959596,
      "loss": 1.3932,
      "theoretical_loss": 3.363326458797543,
      "tokens_seen": 2548957184
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023157575757575758,
      "loss": 1.3357,
      "theoretical_loss": 3.363319587931732,
      "tokens_seen": 2549022720
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023155555555555554,
      "loss": 1.4703,
      "theoretical_loss": 3.3633127172920316,
      "tokens_seen": 2549088256
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023153535353535355,
      "loss": 1.2537,
      "theoretical_loss": 3.3633058468784287,
      "tokens_seen": 2549153792
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002315151515151515,
      "loss": 1.2171,
      "theoretical_loss": 3.3632989766909094,
      "tokens_seen": 2549219328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023149494949494952,
      "loss": 1.3298,
      "theoretical_loss": 3.3632921067294608,
      "tokens_seen": 2549284864
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023147474747474748,
      "loss": 1.415,
      "theoretical_loss": 3.3632852369940696,
      "tokens_seen": 2549350400
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023145454545454546,
      "loss": 1.6099,
      "theoretical_loss": 3.3632783674847224,
      "tokens_seen": 2549415936
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023143434343434344,
      "loss": 1.1941,
      "theoretical_loss": 3.363271498201406,
      "tokens_seen": 2549481472
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002314141414141414,
      "loss": 1.3006,
      "theoretical_loss": 3.3632646291441075,
      "tokens_seen": 2549547008
    },
    {
      "epoch": 0.54,
      "objective/train/aligned_token_count": 0.5,
      "objective/train/avg_token_score": 0.006855877581983805,
      "objective/train/docs_used": 1432724,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6697540283203125,
      "objective/train/original_loss": 4.018422603607178,
      "objective/train/theoretical_loss": 3.363257760312813,
      "objective/train/tokens_used": 908471776,
      "objective/train/unlikelihood_loss": 0.06775180250406265,
      "theoretical_loss": 3.363257760312813,
      "tokens_seen": 2549612544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002313939393939394,
      "loss": 1.6237,
      "theoretical_loss": 3.363257760312813,
      "tokens_seen": 2549612544
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023137373737373737,
      "loss": 1.2947,
      "theoretical_loss": 3.36325089170751,
      "tokens_seen": 2549678080
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023135353535353538,
      "loss": 1.0799,
      "theoretical_loss": 3.3632440233281846,
      "tokens_seen": 2549743616
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023133333333333334,
      "loss": 1.2328,
      "theoretical_loss": 3.3632371551748244,
      "tokens_seen": 2549809152
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023131313131313132,
      "loss": 1.0739,
      "theoretical_loss": 3.3632302872474154,
      "tokens_seen": 2549874688
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002312929292929293,
      "loss": 1.5212,
      "theoretical_loss": 3.3632234195459443,
      "tokens_seen": 2549940224
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023127272727272726,
      "loss": 1.1652,
      "theoretical_loss": 3.3632165520703983,
      "tokens_seen": 2550005760
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023125252525252527,
      "loss": 1.5173,
      "theoretical_loss": 3.363209684820764,
      "tokens_seen": 2550071296
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023123232323232323,
      "loss": 1.3937,
      "theoretical_loss": 3.363202817797028,
      "tokens_seen": 2550136832
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002312121212121212,
      "loss": 1.3795,
      "theoretical_loss": 3.3631959509991773,
      "tokens_seen": 2550202368
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002311919191919192,
      "loss": 1.5669,
      "theoretical_loss": 3.3631890844271988,
      "tokens_seen": 2550267904
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023117171717171718,
      "loss": 1.4479,
      "theoretical_loss": 3.363182218081079,
      "tokens_seen": 2550333440
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023115151515151517,
      "loss": 1.2207,
      "theoretical_loss": 3.3631753519608045,
      "tokens_seen": 2550398976
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023113131313131315,
      "loss": 1.3834,
      "theoretical_loss": 3.3631684860663627,
      "tokens_seen": 2550464512
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002311111111111111,
      "loss": 1.3767,
      "theoretical_loss": 3.3631616203977392,
      "tokens_seen": 2550530048
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002310909090909091,
      "loss": 1.2897,
      "theoretical_loss": 3.3631547549549223,
      "tokens_seen": 2550595584
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023107070707070707,
      "loss": 1.2453,
      "theoretical_loss": 3.3631478897378977,
      "tokens_seen": 2550661120
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023105050505050506,
      "loss": 1.404,
      "theoretical_loss": 3.3631410247466524,
      "tokens_seen": 2550726656
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023103030303030304,
      "loss": 1.3963,
      "theoretical_loss": 3.3631341599811737,
      "tokens_seen": 2550792192
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000231010101010101,
      "loss": 1.505,
      "theoretical_loss": 3.3631272954414473,
      "tokens_seen": 2550857728
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000230989898989899,
      "loss": 1.5567,
      "theoretical_loss": 3.3631204311274607,
      "tokens_seen": 2550923264
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023096969696969697,
      "loss": 1.5564,
      "theoretical_loss": 3.3631135670392007,
      "tokens_seen": 2550988800
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023094949494949498,
      "loss": 1.3869,
      "theoretical_loss": 3.363106703176654,
      "tokens_seen": 2551054336
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023092929292929293,
      "loss": 1.6974,
      "theoretical_loss": 3.3630998395398075,
      "tokens_seen": 2551119872
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002309090909090909,
      "loss": 1.4823,
      "theoretical_loss": 3.3630929761286477,
      "tokens_seen": 2551185408
    },
    {
      "epoch": 0.54,
      "objective/train/aligned_token_count": 0.5094085931777954,
      "objective/train/avg_token_score": 0.011134840548038483,
      "objective/train/docs_used": 1434145,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4722883701324463,
      "objective/train/original_loss": 3.574592351913452,
      "objective/train/theoretical_loss": 3.3630861129431615,
      "objective/train/tokens_used": 910110176,
      "objective/train/unlikelihood_loss": 0.07794718444347382,
      "theoretical_loss": 3.3630861129431615,
      "tokens_seen": 2551250944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002308888888888889,
      "loss": 1.5271,
      "theoretical_loss": 3.3630861129431615,
      "tokens_seen": 2551250944
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023086868686868686,
      "loss": 1.3358,
      "theoretical_loss": 3.3630792499833357,
      "tokens_seen": 2551316480
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023084848484848487,
      "loss": 1.1616,
      "theoretical_loss": 3.363072387249157,
      "tokens_seen": 2551382016
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023082828282828283,
      "loss": 1.2007,
      "theoretical_loss": 3.363065524740612,
      "tokens_seen": 2551447552
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002308080808080808,
      "loss": 1.1377,
      "theoretical_loss": 3.3630586624576884,
      "tokens_seen": 2551513088
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002307878787878788,
      "loss": 1.5403,
      "theoretical_loss": 3.3630518004003718,
      "tokens_seen": 2551578624
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023076767676767675,
      "loss": 1.4805,
      "theoretical_loss": 3.3630449385686494,
      "tokens_seen": 2551644160
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023074747474747476,
      "loss": 1.4498,
      "theoretical_loss": 3.3630380769625083,
      "tokens_seen": 2551709696
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023072727272727272,
      "loss": 1.4264,
      "theoretical_loss": 3.3630312155819353,
      "tokens_seen": 2551775232
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002307070707070707,
      "loss": 1.3958,
      "theoretical_loss": 3.3630243544269165,
      "tokens_seen": 2551840768
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002306868686868687,
      "loss": 1.5007,
      "theoretical_loss": 3.3630174934974395,
      "tokens_seen": 2551906304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023066666666666667,
      "loss": 1.3486,
      "theoretical_loss": 3.363010632793491,
      "tokens_seen": 2551971840
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023064646464646466,
      "loss": 1.49,
      "theoretical_loss": 3.363003772315057,
      "tokens_seen": 2552037376
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023062626262626264,
      "loss": 1.5477,
      "theoretical_loss": 3.3629969120621257,
      "tokens_seen": 2552102912
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002306060606060606,
      "loss": 1.0366,
      "theoretical_loss": 3.3629900520346823,
      "tokens_seen": 2552168448
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023058585858585858,
      "loss": 1.1594,
      "theoretical_loss": 3.362983192232715,
      "tokens_seen": 2552233984
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023056565656565656,
      "loss": 1.274,
      "theoretical_loss": 3.3629763326562094,
      "tokens_seen": 2552299520
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023054545454545455,
      "loss": 1.242,
      "theoretical_loss": 3.362969473305153,
      "tokens_seen": 2552365056
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023052525252525253,
      "loss": 1.3357,
      "theoretical_loss": 3.3629626141795326,
      "tokens_seen": 2552430592
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023050505050505052,
      "loss": 1.3715,
      "theoretical_loss": 3.3629557552793345,
      "tokens_seen": 2552496128
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002304848484848485,
      "loss": 1.394,
      "theoretical_loss": 3.3629488966045464,
      "tokens_seen": 2552561664
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023046464646464646,
      "loss": 1.4691,
      "theoretical_loss": 3.362942038155154,
      "tokens_seen": 2552627200
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023044444444444447,
      "loss": 1.2398,
      "theoretical_loss": 3.362935179931145,
      "tokens_seen": 2552692736
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023042424242424243,
      "loss": 1.4378,
      "theoretical_loss": 3.3629283219325057,
      "tokens_seen": 2552758272
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002304040404040404,
      "loss": 1.4034,
      "theoretical_loss": 3.3629214641592236,
      "tokens_seen": 2552823808
    },
    {
      "epoch": 0.54,
      "objective/train/aligned_token_count": 0.41287878155708313,
      "objective/train/avg_token_score": 0.015444948337972164,
      "objective/train/docs_used": 1434830,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4251757860183716,
      "objective/train/original_loss": 4.5141730308532715,
      "objective/train/theoretical_loss": 3.3629146066112847,
      "objective/train/tokens_used": 911748576,
      "objective/train/unlikelihood_loss": 0.06007169559597969,
      "theoretical_loss": 3.3629146066112847,
      "tokens_seen": 2552889344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002303838383838384,
      "loss": 1.6714,
      "theoretical_loss": 3.3629146066112847,
      "tokens_seen": 2552889344
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023036363636363635,
      "loss": 1.2285,
      "theoretical_loss": 3.3629077492886763,
      "tokens_seen": 2552954880
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023034343434343436,
      "loss": 1.2507,
      "theoretical_loss": 3.3629008921913845,
      "tokens_seen": 2553020416
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023032323232323232,
      "loss": 1.1415,
      "theoretical_loss": 3.362894035319397,
      "tokens_seen": 2553085952
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023030303030303033,
      "loss": 1.3939,
      "theoretical_loss": 3.3628871786727004,
      "tokens_seen": 2553151488
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023028282828282829,
      "loss": 1.2428,
      "theoretical_loss": 3.362880322251281,
      "tokens_seen": 2553217024
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023026262626262627,
      "loss": 1.6603,
      "theoretical_loss": 3.362873466055126,
      "tokens_seen": 2553282560
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023024242424242425,
      "loss": 1.4643,
      "theoretical_loss": 3.3628666100842226,
      "tokens_seen": 2553348096
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002302222222222222,
      "loss": 1.2683,
      "theoretical_loss": 3.3628597543385568,
      "tokens_seen": 2553413632
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023020202020202022,
      "loss": 1.5272,
      "theoretical_loss": 3.362852898818116,
      "tokens_seen": 2553479168
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023018181818181818,
      "loss": 1.4084,
      "theoretical_loss": 3.362846043522887,
      "tokens_seen": 2553544704
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023016161616161616,
      "loss": 1.6246,
      "theoretical_loss": 3.3628391884528566,
      "tokens_seen": 2553610240
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023014141414141415,
      "loss": 1.3586,
      "theoretical_loss": 3.362832333608011,
      "tokens_seen": 2553675776
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023012121212121213,
      "loss": 1.5447,
      "theoretical_loss": 3.362825478988338,
      "tokens_seen": 2553741312
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023010101010101011,
      "loss": 1.4111,
      "theoretical_loss": 3.3628186245938236,
      "tokens_seen": 2553806848
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002300808080808081,
      "loss": 1.2537,
      "theoretical_loss": 3.362811770424455,
      "tokens_seen": 2553872384
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023006060606060606,
      "loss": 1.2581,
      "theoretical_loss": 3.3628049164802194,
      "tokens_seen": 2553937920
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023004040404040404,
      "loss": 1.4993,
      "theoretical_loss": 3.362798062761103,
      "tokens_seen": 2554003456
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023002020202020202,
      "loss": 1.3936,
      "theoretical_loss": 3.362791209267093,
      "tokens_seen": 2554068992
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00023,
      "loss": 1.2764,
      "theoretical_loss": 3.3627843559981763,
      "tokens_seen": 2554134528
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000229979797979798,
      "loss": 1.465,
      "theoretical_loss": 3.362777502954339,
      "tokens_seen": 2554200064
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022995959595959595,
      "loss": 1.419,
      "theoretical_loss": 3.3627706501355688,
      "tokens_seen": 2554265600
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022993939393939396,
      "loss": 1.6566,
      "theoretical_loss": 3.362763797541852,
      "tokens_seen": 2554331136
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.00022991919191919192,
      "loss": 1.3825,
      "theoretical_loss": 3.3627569451731754,
      "tokens_seen": 2554396672
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002298989898989899,
      "loss": 1.509,
      "theoretical_loss": 3.3627500930295264,
      "tokens_seen": 2554462208
    },
    {
      "epoch": 0.54,
      "objective/train/aligned_token_count": 0.23582600057125092,
      "objective/train/avg_token_score": 0.04061037302017212,
      "objective/train/docs_used": 1435894,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.7687034606933594,
      "objective/train/original_loss": 4.823043346405029,
      "objective/train/theoretical_loss": 3.3627432411108917,
      "objective/train/tokens_used": 913386976,
      "objective/train/unlikelihood_loss": 0.068697988986969,
      "theoretical_loss": 3.3627432411108917,
      "tokens_seen": 2554527744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022987878787878788,
      "loss": 1.3657,
      "theoretical_loss": 3.3627432411108917,
      "tokens_seen": 2554527744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022985858585858584,
      "loss": 1.46,
      "theoretical_loss": 3.3627363894172575,
      "tokens_seen": 2554593280
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022983838383838385,
      "loss": 1.2664,
      "theoretical_loss": 3.3627295379486117,
      "tokens_seen": 2554658816
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002298181818181818,
      "loss": 1.445,
      "theoretical_loss": 3.36272268670494,
      "tokens_seen": 2554724352
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022979797979797982,
      "loss": 1.4319,
      "theoretical_loss": 3.36271583568623,
      "tokens_seen": 2554789888
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022977777777777778,
      "loss": 1.4993,
      "theoretical_loss": 3.362708984892468,
      "tokens_seen": 2554855424
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022975757575757576,
      "loss": 1.4903,
      "theoretical_loss": 3.3627021343236416,
      "tokens_seen": 2554920960
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022973737373737374,
      "loss": 1.1226,
      "theoretical_loss": 3.362695283979737,
      "tokens_seen": 2554986496
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002297171717171717,
      "loss": 1.4298,
      "theoretical_loss": 3.362688433860741,
      "tokens_seen": 2555052032
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002296969696969697,
      "loss": 1.5307,
      "theoretical_loss": 3.362681583966641,
      "tokens_seen": 2555117568
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022967676767676767,
      "loss": 1.3938,
      "theoretical_loss": 3.3626747342974235,
      "tokens_seen": 2555183104
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022965656565656568,
      "loss": 1.2544,
      "theoretical_loss": 3.3626678848530753,
      "tokens_seen": 2555248640
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022963636363636364,
      "loss": 1.4669,
      "theoretical_loss": 3.3626610356335833,
      "tokens_seen": 2555314176
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022961616161616162,
      "loss": 1.4838,
      "theoretical_loss": 3.3626541866389346,
      "tokens_seen": 2555379712
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002295959595959596,
      "loss": 1.4086,
      "theoretical_loss": 3.3626473378691157,
      "tokens_seen": 2555445248
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002295757575757576,
      "loss": 1.3895,
      "theoretical_loss": 3.3626404893241135,
      "tokens_seen": 2555510784
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022955555555555557,
      "loss": 1.5278,
      "theoretical_loss": 3.362633641003915,
      "tokens_seen": 2555576320
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022953535353535353,
      "loss": 1.4359,
      "theoretical_loss": 3.3626267929085074,
      "tokens_seen": 2555641856
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022951515151515151,
      "loss": 1.4274,
      "theoretical_loss": 3.362619945037877,
      "tokens_seen": 2555707392
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002294949494949495,
      "loss": 1.4629,
      "theoretical_loss": 3.3626130973920105,
      "tokens_seen": 2555772928
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022947474747474748,
      "loss": 1.4836,
      "theoretical_loss": 3.362606249970895,
      "tokens_seen": 2555838464
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022945454545454547,
      "loss": 1.3898,
      "theoretical_loss": 3.362599402774518,
      "tokens_seen": 2555904000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022943434343434345,
      "loss": 1.4167,
      "theoretical_loss": 3.3625925558028653,
      "tokens_seen": 2555969536
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002294141414141414,
      "loss": 1.3711,
      "theoretical_loss": 3.362585709055925,
      "tokens_seen": 2556035072
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022939393939393942,
      "loss": 1.4329,
      "theoretical_loss": 3.3625788625336823,
      "tokens_seen": 2556100608
    },
    {
      "epoch": 0.55,
      "objective/train/aligned_token_count": 0.48032745718955994,
      "objective/train/avg_token_score": 0.011688020080327988,
      "objective/train/docs_used": 1436568,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5360291004180908,
      "objective/train/original_loss": 3.6795661449432373,
      "objective/train/theoretical_loss": 3.3625720162361254,
      "objective/train/tokens_used": 915025376,
      "objective/train/unlikelihood_loss": 0.0791502445936203,
      "theoretical_loss": 3.3625720162361254,
      "tokens_seen": 2556166144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022937373737373737,
      "loss": 1.5355,
      "theoretical_loss": 3.3625720162361254,
      "tokens_seen": 2556166144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022935353535353536,
      "loss": 1.2649,
      "theoretical_loss": 3.362565170163241,
      "tokens_seen": 2556231680
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022933333333333334,
      "loss": 1.146,
      "theoretical_loss": 3.3625583243150157,
      "tokens_seen": 2556297216
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002293131313131313,
      "loss": 1.3385,
      "theoretical_loss": 3.362551478691436,
      "tokens_seen": 2556362752
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002292929292929293,
      "loss": 1.623,
      "theoretical_loss": 3.3625446332924893,
      "tokens_seen": 2556428288
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022927272727272727,
      "loss": 1.3034,
      "theoretical_loss": 3.362537788118163,
      "tokens_seen": 2556493824
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022925252525252528,
      "loss": 1.3444,
      "theoretical_loss": 3.3625309431684425,
      "tokens_seen": 2556559360
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022923232323232323,
      "loss": 1.5194,
      "theoretical_loss": 3.362524098443316,
      "tokens_seen": 2556624896
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002292121212121212,
      "loss": 1.4392,
      "theoretical_loss": 3.3625172539427695,
      "tokens_seen": 2556690432
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002291919191919192,
      "loss": 1.3645,
      "theoretical_loss": 3.36251040966679,
      "tokens_seen": 2556755968
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022917171717171716,
      "loss": 1.1228,
      "theoretical_loss": 3.362503565615365,
      "tokens_seen": 2556821504
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022915151515151517,
      "loss": 1.3271,
      "theoretical_loss": 3.362496721788481,
      "tokens_seen": 2556887040
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022913131313131313,
      "loss": 1.4267,
      "theoretical_loss": 3.3624898781861248,
      "tokens_seen": 2556952576
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002291111111111111,
      "loss": 1.5699,
      "theoretical_loss": 3.3624830348082835,
      "tokens_seen": 2557018112
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002290909090909091,
      "loss": 1.1996,
      "theoretical_loss": 3.3624761916549435,
      "tokens_seen": 2557083648
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022907070707070708,
      "loss": 1.4509,
      "theoretical_loss": 3.3624693487260924,
      "tokens_seen": 2557149184
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022905050505050506,
      "loss": 1.3668,
      "theoretical_loss": 3.3624625060217164,
      "tokens_seen": 2557214720
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022903030303030302,
      "loss": 1.4903,
      "theoretical_loss": 3.3624556635418026,
      "tokens_seen": 2557280256
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000229010101010101,
      "loss": 1.2235,
      "theoretical_loss": 3.362448821286338,
      "tokens_seen": 2557345792
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000228989898989899,
      "loss": 1.3416,
      "theoretical_loss": 3.3624419792553097,
      "tokens_seen": 2557411328
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022896969696969697,
      "loss": 1.4914,
      "theoretical_loss": 3.3624351374487045,
      "tokens_seen": 2557476864
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022894949494949496,
      "loss": 1.5042,
      "theoretical_loss": 3.3624282958665086,
      "tokens_seen": 2557542400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022892929292929294,
      "loss": 1.3269,
      "theoretical_loss": 3.3624214545087097,
      "tokens_seen": 2557607936
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002289090909090909,
      "loss": 1.2632,
      "theoretical_loss": 3.3624146133752943,
      "tokens_seen": 2557673472
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002288888888888889,
      "loss": 1.2275,
      "theoretical_loss": 3.3624077724662493,
      "tokens_seen": 2557739008
    },
    {
      "epoch": 0.55,
      "objective/train/aligned_token_count": 0.37188416719436646,
      "objective/train/avg_token_score": 0.014824194833636284,
      "objective/train/docs_used": 1437925,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1733684539794922,
      "objective/train/original_loss": 4.127971649169922,
      "objective/train/theoretical_loss": 3.362400931781562,
      "objective/train/tokens_used": 916663776,
      "objective/train/unlikelihood_loss": 0.05518443137407303,
      "theoretical_loss": 3.362400931781562,
      "tokens_seen": 2557804544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022886868686868686,
      "loss": 1.4124,
      "theoretical_loss": 3.362400931781562,
      "tokens_seen": 2557804544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022884848484848485,
      "loss": 1.2996,
      "theoretical_loss": 3.3623940913212187,
      "tokens_seen": 2557870080
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022882828282828283,
      "loss": 1.3661,
      "theoretical_loss": 3.362387251085207,
      "tokens_seen": 2557935616
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022880808080808082,
      "loss": 1.2778,
      "theoretical_loss": 3.362380411073513,
      "tokens_seen": 2558001152
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002287878787878788,
      "loss": 1.4389,
      "theoretical_loss": 3.3623735712861236,
      "tokens_seen": 2558066688
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022876767676767676,
      "loss": 1.5497,
      "theoretical_loss": 3.362366731723027,
      "tokens_seen": 2558132224
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022874747474747477,
      "loss": 1.3928,
      "theoretical_loss": 3.3623598923842084,
      "tokens_seen": 2558197760
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022872727272727273,
      "loss": 1.2432,
      "theoretical_loss": 3.3623530532696555,
      "tokens_seen": 2558263296
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022870707070707074,
      "loss": 1.4252,
      "theoretical_loss": 3.362346214379355,
      "tokens_seen": 2558328832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002286868686868687,
      "loss": 1.4917,
      "theoretical_loss": 3.3623393757132947,
      "tokens_seen": 2558394368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022866666666666665,
      "loss": 1.2814,
      "theoretical_loss": 3.3623325372714605,
      "tokens_seen": 2558459904
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022864646464646466,
      "loss": 1.5131,
      "theoretical_loss": 3.362325699053839,
      "tokens_seen": 2558525440
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022862626262626262,
      "loss": 1.2469,
      "theoretical_loss": 3.3623188610604187,
      "tokens_seen": 2558590976
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022860606060606063,
      "loss": 1.5001,
      "theoretical_loss": 3.3623120232911847,
      "tokens_seen": 2558656512
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022858585858585859,
      "loss": 1.5342,
      "theoretical_loss": 3.3623051857461252,
      "tokens_seen": 2558722048
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022856565656565657,
      "loss": 1.3737,
      "theoretical_loss": 3.362298348425226,
      "tokens_seen": 2558787584
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022854545454545455,
      "loss": 1.2087,
      "theoretical_loss": 3.362291511328475,
      "tokens_seen": 2558853120
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022852525252525254,
      "loss": 1.5076,
      "theoretical_loss": 3.3622846744558585,
      "tokens_seen": 2558918656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022850505050505052,
      "loss": 1.4688,
      "theoretical_loss": 3.362277837807364,
      "tokens_seen": 2558984192
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022848484848484848,
      "loss": 1.3602,
      "theoretical_loss": 3.362271001382978,
      "tokens_seen": 2559049728
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022846464646464646,
      "loss": 1.3616,
      "theoretical_loss": 3.3622641651826872,
      "tokens_seen": 2559115264
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022844444444444445,
      "loss": 1.2779,
      "theoretical_loss": 3.362257329206479,
      "tokens_seen": 2559180800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022842424242424243,
      "loss": 1.4712,
      "theoretical_loss": 3.36225049345434,
      "tokens_seen": 2559246336
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022840404040404041,
      "loss": 1.368,
      "theoretical_loss": 3.362243657926257,
      "tokens_seen": 2559311872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002283838383838384,
      "loss": 1.4695,
      "theoretical_loss": 3.3622368226222177,
      "tokens_seen": 2559377408
    },
    {
      "epoch": 0.55,
      "objective/train/aligned_token_count": 0.3752443790435791,
      "objective/train/avg_token_score": 0.038568634539842606,
      "objective/train/docs_used": 1438686,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0415372848510742,
      "objective/train/original_loss": 4.369557857513428,
      "objective/train/theoretical_loss": 3.362229987542208,
      "objective/train/tokens_used": 918302176,
      "objective/train/unlikelihood_loss": 0.060740191489458084,
      "theoretical_loss": 3.362229987542208,
      "tokens_seen": 2559442944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022836363636363636,
      "loss": 1.4255,
      "theoretical_loss": 3.362229987542208,
      "tokens_seen": 2559442944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022834343434343437,
      "loss": 1.5292,
      "theoretical_loss": 3.3622231526862154,
      "tokens_seen": 2559508480
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022832323232323232,
      "loss": 1.34,
      "theoretical_loss": 3.3622163180542266,
      "tokens_seen": 2559574016
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002283030303030303,
      "loss": 1.2139,
      "theoretical_loss": 3.362209483646229,
      "tokens_seen": 2559639552
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002282828282828283,
      "loss": 1.3883,
      "theoretical_loss": 3.3622026494622084,
      "tokens_seen": 2559705088
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022826262626262625,
      "loss": 1.6315,
      "theoretical_loss": 3.362195815502153,
      "tokens_seen": 2559770624
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022824242424242426,
      "loss": 1.2125,
      "theoretical_loss": 3.362188981766049,
      "tokens_seen": 2559836160
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022822222222222222,
      "loss": 1.2238,
      "theoretical_loss": 3.3621821482538836,
      "tokens_seen": 2559901696
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022820202020202023,
      "loss": 1.5105,
      "theoretical_loss": 3.362175314965644,
      "tokens_seen": 2559967232
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022818181818181818,
      "loss": 1.2185,
      "theoretical_loss": 3.3621684819013167,
      "tokens_seen": 2560032768
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022816161616161614,
      "loss": 1.6552,
      "theoretical_loss": 3.3621616490608885,
      "tokens_seen": 2560098304
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022814141414141415,
      "loss": 1.2106,
      "theoretical_loss": 3.362154816444346,
      "tokens_seen": 2560163840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002281212121212121,
      "loss": 1.4675,
      "theoretical_loss": 3.3621479840516777,
      "tokens_seen": 2560229376
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022810101010101012,
      "loss": 1.3589,
      "theoretical_loss": 3.362141151882869,
      "tokens_seen": 2560294912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022808080808080808,
      "loss": 1.4814,
      "theoretical_loss": 3.3621343199379075,
      "tokens_seen": 2560360448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022806060606060606,
      "loss": 1.3369,
      "theoretical_loss": 3.36212748821678,
      "tokens_seen": 2560425984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022804040404040404,
      "loss": 1.2484,
      "theoretical_loss": 3.3621206567194735,
      "tokens_seen": 2560491520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022802020202020203,
      "loss": 1.3331,
      "theoretical_loss": 3.3621138254459746,
      "tokens_seen": 2560557056
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000228,
      "loss": 1.4567,
      "theoretical_loss": 3.3621069943962705,
      "tokens_seen": 2560622592
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022797979797979797,
      "loss": 1.4096,
      "theoretical_loss": 3.3621001635703482,
      "tokens_seen": 2560688128
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022795959595959598,
      "loss": 1.3297,
      "theoretical_loss": 3.3620933329681946,
      "tokens_seen": 2560753664
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022793939393939394,
      "loss": 1.5448,
      "theoretical_loss": 3.362086502589797,
      "tokens_seen": 2560819200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022791919191919192,
      "loss": 1.5913,
      "theoretical_loss": 3.3620796724351414,
      "tokens_seen": 2560884736
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002278989898989899,
      "loss": 1.5539,
      "theoretical_loss": 3.3620728425042157,
      "tokens_seen": 2560950272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002278787878787879,
      "loss": 1.3648,
      "theoretical_loss": 3.3620660127970066,
      "tokens_seen": 2561015808
    },
    {
      "epoch": 0.55,
      "objective/train/aligned_token_count": 0.579362154006958,
      "objective/train/avg_token_score": 0.0255513284355402,
      "objective/train/docs_used": 1440126,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.871814489364624,
      "objective/train/original_loss": 4.019387245178223,
      "objective/train/theoretical_loss": 3.3620591833135003,
      "objective/train/tokens_used": 919940576,
      "objective/train/unlikelihood_loss": 0.05169510841369629,
      "theoretical_loss": 3.3620591833135003,
      "tokens_seen": 2561081344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022785858585858587,
      "loss": 1.7205,
      "theoretical_loss": 3.3620591833135003,
      "tokens_seen": 2561081344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022783838383838386,
      "loss": 1.2061,
      "theoretical_loss": 3.362052354053685,
      "tokens_seen": 2561146880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022781818181818181,
      "loss": 1.7068,
      "theoretical_loss": 3.3620455250175465,
      "tokens_seen": 2561212416
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002277979797979798,
      "loss": 1.4347,
      "theoretical_loss": 3.3620386962050723,
      "tokens_seen": 2561277952
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022777777777777778,
      "loss": 1.2191,
      "theoretical_loss": 3.36203186761625,
      "tokens_seen": 2561343488
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022775757575757577,
      "loss": 1.3796,
      "theoretical_loss": 3.362025039251065,
      "tokens_seen": 2561409024
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022773737373737375,
      "loss": 1.2604,
      "theoretical_loss": 3.362018211109506,
      "tokens_seen": 2561474560
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002277171717171717,
      "loss": 1.3621,
      "theoretical_loss": 3.3620113831915583,
      "tokens_seen": 2561540096
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022769696969696972,
      "loss": 1.3977,
      "theoretical_loss": 3.36200455549721,
      "tokens_seen": 2561605632
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022767676767676767,
      "loss": 1.3626,
      "theoretical_loss": 3.3619977280264477,
      "tokens_seen": 2561671168
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022765656565656569,
      "loss": 1.2967,
      "theoretical_loss": 3.361990900779258,
      "tokens_seen": 2561736704
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022763636363636364,
      "loss": 1.4302,
      "theoretical_loss": 3.361984073755629,
      "tokens_seen": 2561802240
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002276161616161616,
      "loss": 1.4123,
      "theoretical_loss": 3.3619772469555462,
      "tokens_seen": 2561867776
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002275959595959596,
      "loss": 1.5327,
      "theoretical_loss": 3.3619704203789973,
      "tokens_seen": 2561933312
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022757575757575757,
      "loss": 1.4474,
      "theoretical_loss": 3.361963594025969,
      "tokens_seen": 2561998848
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022755555555555558,
      "loss": 1.2205,
      "theoretical_loss": 3.361956767896449,
      "tokens_seen": 2562064384
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022753535353535354,
      "loss": 1.4211,
      "theoretical_loss": 3.3619499419904235,
      "tokens_seen": 2562129920
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022751515151515152,
      "loss": 1.4474,
      "theoretical_loss": 3.3619431163078795,
      "tokens_seen": 2562195456
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002274949494949495,
      "loss": 1.3674,
      "theoretical_loss": 3.3619362908488046,
      "tokens_seen": 2562260992
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022747474747474746,
      "loss": 1.2325,
      "theoretical_loss": 3.361929465613185,
      "tokens_seen": 2562326528
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022745454545454547,
      "loss": 1.3094,
      "theoretical_loss": 3.3619226406010077,
      "tokens_seen": 2562392064
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022743434343434343,
      "loss": 1.4347,
      "theoretical_loss": 3.36191581581226,
      "tokens_seen": 2562457600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002274141414141414,
      "loss": 1.2948,
      "theoretical_loss": 3.3619089912469295,
      "tokens_seen": 2562523136
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002273939393939394,
      "loss": 1.535,
      "theoretical_loss": 3.3619021669050024,
      "tokens_seen": 2562588672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022737373737373738,
      "loss": 1.4059,
      "theoretical_loss": 3.3618953427864655,
      "tokens_seen": 2562654208
    },
    {
      "epoch": 0.55,
      "objective/train/aligned_token_count": 0.2995479106903076,
      "objective/train/avg_token_score": 0.011286714114248753,
      "objective/train/docs_used": 1440814,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9203324913978577,
      "objective/train/original_loss": 4.943639755249023,
      "objective/train/theoretical_loss": 3.361888518891306,
      "objective/train/tokens_used": 921578976,
      "objective/train/unlikelihood_loss": 0.07500756531953812,
      "theoretical_loss": 3.361888518891306,
      "tokens_seen": 2562719744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022735353535353536,
      "loss": 1.5129,
      "theoretical_loss": 3.361888518891306,
      "tokens_seen": 2562719744
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022733333333333335,
      "loss": 1.4371,
      "theoretical_loss": 3.3618816952195107,
      "tokens_seen": 2562785280
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002273131313131313,
      "loss": 1.3672,
      "theoretical_loss": 3.361874871771067,
      "tokens_seen": 2562850816
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002272929292929293,
      "loss": 1.3486,
      "theoretical_loss": 3.361868048545962,
      "tokens_seen": 2562916352
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022727272727272727,
      "loss": 1.48,
      "theoretical_loss": 3.3618612255441827,
      "tokens_seen": 2562981888
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022725252525252526,
      "loss": 1.1488,
      "theoretical_loss": 3.361854402765715,
      "tokens_seen": 2563047424
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022723232323232324,
      "loss": 1.3915,
      "theoretical_loss": 3.3618475802105467,
      "tokens_seen": 2563112960
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002272121212121212,
      "loss": 1.2992,
      "theoretical_loss": 3.361840757878665,
      "tokens_seen": 2563178496
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002271919191919192,
      "loss": 1.3888,
      "theoretical_loss": 3.361833935770057,
      "tokens_seen": 2563244032
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022717171717171717,
      "loss": 1.5402,
      "theoretical_loss": 3.3618271138847082,
      "tokens_seen": 2563309568
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022715151515151518,
      "loss": 1.5832,
      "theoretical_loss": 3.3618202922226073,
      "tokens_seen": 2563375104
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022713131313131313,
      "loss": 1.2737,
      "theoretical_loss": 3.3618134707837406,
      "tokens_seen": 2563440640
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022711111111111112,
      "loss": 1.2994,
      "theoretical_loss": 3.361806649568095,
      "tokens_seen": 2563506176
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002270909090909091,
      "loss": 1.2562,
      "theoretical_loss": 3.361799828575658,
      "tokens_seen": 2563571712
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022707070707070706,
      "loss": 1.28,
      "theoretical_loss": 3.361793007806416,
      "tokens_seen": 2563637248
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022705050505050507,
      "loss": 1.2709,
      "theoretical_loss": 3.361786187260356,
      "tokens_seen": 2563702784
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022703030303030303,
      "loss": 1.409,
      "theoretical_loss": 3.3617793669374656,
      "tokens_seen": 2563768320
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022701010101010104,
      "loss": 1.3718,
      "theoretical_loss": 3.361772546837731,
      "tokens_seen": 2563833856
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000226989898989899,
      "loss": 1.3223,
      "theoretical_loss": 3.36176572696114,
      "tokens_seen": 2563899392
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022696969696969698,
      "loss": 1.2964,
      "theoretical_loss": 3.361758907307679,
      "tokens_seen": 2563964928
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022694949494949496,
      "loss": 1.3252,
      "theoretical_loss": 3.361752087877335,
      "tokens_seen": 2564030464
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022692929292929292,
      "loss": 1.3896,
      "theoretical_loss": 3.3617452686700955,
      "tokens_seen": 2564096000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022690909090909093,
      "loss": 1.4559,
      "theoretical_loss": 3.361738449685947,
      "tokens_seen": 2564161536
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002268888888888889,
      "loss": 1.3468,
      "theoretical_loss": 3.3617316309248766,
      "tokens_seen": 2564227072
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022686868686868687,
      "loss": 1.4077,
      "theoretical_loss": 3.3617248123868713,
      "tokens_seen": 2564292608
    },
    {
      "epoch": 0.55,
      "objective/train/aligned_token_count": 0.5260874629020691,
      "objective/train/avg_token_score": 0.002698588417842984,
      "objective/train/docs_used": 1442376,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6721302270889282,
      "objective/train/original_loss": 4.183875560760498,
      "objective/train/theoretical_loss": 3.3617179940719186,
      "objective/train/tokens_used": 923217376,
      "objective/train/unlikelihood_loss": 0.05283377319574356,
      "theoretical_loss": 3.3617179940719186,
      "tokens_seen": 2564358144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022684848484848485,
      "loss": 1.4639,
      "theoretical_loss": 3.3617179940719186,
      "tokens_seen": 2564358144
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022682828282828284,
      "loss": 1.0705,
      "theoretical_loss": 3.361711175980005,
      "tokens_seen": 2564423680
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022680808080808082,
      "loss": 1.255,
      "theoretical_loss": 3.3617043581111172,
      "tokens_seen": 2564489216
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002267878787878788,
      "loss": 1.3195,
      "theoretical_loss": 3.361697540465243,
      "tokens_seen": 2564554752
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022676767676767676,
      "loss": 1.4238,
      "theoretical_loss": 3.361690723042369,
      "tokens_seen": 2564620288
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022674747474747475,
      "loss": 1.422,
      "theoretical_loss": 3.361683905842482,
      "tokens_seen": 2564685824
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022672727272727273,
      "loss": 1.4761,
      "theoretical_loss": 3.3616770888655694,
      "tokens_seen": 2564751360
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022670707070707072,
      "loss": 1.7043,
      "theoretical_loss": 3.361670272111618,
      "tokens_seen": 2564816896
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002266868686868687,
      "loss": 1.3114,
      "theoretical_loss": 3.3616634555806146,
      "tokens_seen": 2564882432
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022666666666666666,
      "loss": 1.3793,
      "theoretical_loss": 3.3616566392725464,
      "tokens_seen": 2564947968
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022664646464646467,
      "loss": 1.2354,
      "theoretical_loss": 3.361649823187401,
      "tokens_seen": 2565013504
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022662626262626262,
      "loss": 1.5416,
      "theoretical_loss": 3.3616430073251644,
      "tokens_seen": 2565079040
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002266060606060606,
      "loss": 1.1954,
      "theoretical_loss": 3.3616361916858244,
      "tokens_seen": 2565144576
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002265858585858586,
      "loss": 1.1302,
      "theoretical_loss": 3.3616293762693674,
      "tokens_seen": 2565210112
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022656565656565655,
      "loss": 1.3906,
      "theoretical_loss": 3.361622561075781,
      "tokens_seen": 2565275648
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022654545454545456,
      "loss": 1.437,
      "theoretical_loss": 3.361615746105052,
      "tokens_seen": 2565341184
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022652525252525252,
      "loss": 1.2345,
      "theoretical_loss": 3.361608931357167,
      "tokens_seen": 2565406720
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022650505050505053,
      "loss": 1.5688,
      "theoretical_loss": 3.3616021168321137,
      "tokens_seen": 2565472256
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022648484848484848,
      "loss": 1.5349,
      "theoretical_loss": 3.361595302529879,
      "tokens_seen": 2565537792
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022646464646464647,
      "loss": 1.3423,
      "theoretical_loss": 3.361588488450449,
      "tokens_seen": 2565603328
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022644444444444445,
      "loss": 1.3888,
      "theoretical_loss": 3.361581674593812,
      "tokens_seen": 2565668864
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002264242424242424,
      "loss": 1.5953,
      "theoretical_loss": 3.3615748609599545,
      "tokens_seen": 2565734400
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022640404040404042,
      "loss": 1.516,
      "theoretical_loss": 3.361568047548863,
      "tokens_seen": 2565799936
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022638383838383838,
      "loss": 1.1586,
      "theoretical_loss": 3.3615612343605257,
      "tokens_seen": 2565865472
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022636363636363636,
      "loss": 1.6158,
      "theoretical_loss": 3.3615544213949287,
      "tokens_seen": 2565931008
    },
    {
      "epoch": 0.55,
      "objective/train/aligned_token_count": 0.6025782227516174,
      "objective/train/avg_token_score": 0.0036794268526136875,
      "objective/train/docs_used": 1443131,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8535332679748535,
      "objective/train/original_loss": 3.6706316471099854,
      "objective/train/theoretical_loss": 3.3615476086520593,
      "objective/train/tokens_used": 924855776,
      "objective/train/unlikelihood_loss": 0.06745962798595428,
      "theoretical_loss": 3.3615476086520593,
      "tokens_seen": 2565996544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022634343434343435,
      "loss": 1.4177,
      "theoretical_loss": 3.3615476086520593,
      "tokens_seen": 2565996544
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022632323232323233,
      "loss": 1.3662,
      "theoretical_loss": 3.3615407961319046,
      "tokens_seen": 2566062080
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002263030303030303,
      "loss": 1.4661,
      "theoretical_loss": 3.3615339838344513,
      "tokens_seen": 2566127616
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002262828282828283,
      "loss": 1.5265,
      "theoretical_loss": 3.361527171759687,
      "tokens_seen": 2566193152
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022626262626262628,
      "loss": 1.4004,
      "theoretical_loss": 3.3615203599075985,
      "tokens_seen": 2566258688
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022624242424242424,
      "loss": 1.2842,
      "theoretical_loss": 3.361513548278172,
      "tokens_seen": 2566324224
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022622222222222222,
      "loss": 1.3792,
      "theoretical_loss": 3.3615067368713962,
      "tokens_seen": 2566389760
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002262020202020202,
      "loss": 1.3795,
      "theoretical_loss": 3.361499925687257,
      "tokens_seen": 2566455296
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002261818181818182,
      "loss": 1.4436,
      "theoretical_loss": 3.361493114725741,
      "tokens_seen": 2566520832
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022616161616161617,
      "loss": 1.3799,
      "theoretical_loss": 3.361486303986837,
      "tokens_seen": 2566586368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022614141414141416,
      "loss": 1.306,
      "theoretical_loss": 3.36147949347053,
      "tokens_seen": 2566651904
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022612121212121211,
      "loss": 1.4318,
      "theoretical_loss": 3.3614726831768085,
      "tokens_seen": 2566717440
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022610101010101013,
      "loss": 1.3781,
      "theoretical_loss": 3.361465873105659,
      "tokens_seen": 2566782976
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022608080808080808,
      "loss": 1.5055,
      "theoretical_loss": 3.361459063257068,
      "tokens_seen": 2566848512
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022606060606060607,
      "loss": 1.4953,
      "theoretical_loss": 3.3614522536310236,
      "tokens_seen": 2566914048
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022604040404040405,
      "loss": 1.5113,
      "theoretical_loss": 3.3614454442275123,
      "tokens_seen": 2566979584
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000226020202020202,
      "loss": 1.4992,
      "theoretical_loss": 3.3614386350465213,
      "tokens_seen": 2567045120
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022600000000000002,
      "loss": 1.5504,
      "theoretical_loss": 3.3614318260880376,
      "tokens_seen": 2567110656
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022597979797979798,
      "loss": 1.275,
      "theoretical_loss": 3.361425017352048,
      "tokens_seen": 2567176192
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022595959595959599,
      "loss": 1.5595,
      "theoretical_loss": 3.36141820883854,
      "tokens_seen": 2567241728
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022593939393939394,
      "loss": 1.5299,
      "theoretical_loss": 3.3614114005475004,
      "tokens_seen": 2567307264
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002259191919191919,
      "loss": 1.503,
      "theoretical_loss": 3.361404592478916,
      "tokens_seen": 2567372800
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002258989898989899,
      "loss": 1.1726,
      "theoretical_loss": 3.361397784632774,
      "tokens_seen": 2567438336
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022587878787878787,
      "loss": 1.48,
      "theoretical_loss": 3.361390977009062,
      "tokens_seen": 2567503872
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022585858585858588,
      "loss": 1.4114,
      "theoretical_loss": 3.361384169607766,
      "tokens_seen": 2567569408
    },
    {
      "epoch": 0.55,
      "objective/train/aligned_token_count": 0.5413001179695129,
      "objective/train/avg_token_score": 0.007238256745040417,
      "objective/train/docs_used": 1444461,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7451589107513428,
      "objective/train/original_loss": 4.241446018218994,
      "objective/train/theoretical_loss": 3.3613773624288745,
      "objective/train/tokens_used": 926494176,
      "objective/train/unlikelihood_loss": 0.06344182044267654,
      "theoretical_loss": 3.3613773624288745,
      "tokens_seen": 2567634944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022583838383838384,
      "loss": 1.6232,
      "theoretical_loss": 3.3613773624288745,
      "tokens_seen": 2567634944
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022581818181818182,
      "loss": 1.5748,
      "theoretical_loss": 3.3613705554723734,
      "tokens_seen": 2567700480
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002257979797979798,
      "loss": 1.5055,
      "theoretical_loss": 3.36136374873825,
      "tokens_seen": 2567766016
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002257777777777778,
      "loss": 1.5435,
      "theoretical_loss": 3.3613569422264913,
      "tokens_seen": 2567831552
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022575757575757577,
      "loss": 1.5402,
      "theoretical_loss": 3.361350135937085,
      "tokens_seen": 2567897088
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022573737373737373,
      "loss": 1.3733,
      "theoretical_loss": 3.3613433298700173,
      "tokens_seen": 2567962624
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002257171717171717,
      "loss": 1.3124,
      "theoretical_loss": 3.3613365240252757,
      "tokens_seen": 2568028160
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002256969696969697,
      "loss": 1.1937,
      "theoretical_loss": 3.361329718402847,
      "tokens_seen": 2568093696
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022567676767676768,
      "loss": 1.2627,
      "theoretical_loss": 3.361322913002719,
      "tokens_seen": 2568159232
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022565656565656566,
      "loss": 1.537,
      "theoretical_loss": 3.361316107824878,
      "tokens_seen": 2568224768
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022563636363636365,
      "loss": 1.1733,
      "theoretical_loss": 3.3613093028693113,
      "tokens_seen": 2568290304
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002256161616161616,
      "loss": 1.4266,
      "theoretical_loss": 3.3613024981360056,
      "tokens_seen": 2568355840
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022559595959595962,
      "loss": 1.3788,
      "theoretical_loss": 3.3612956936249487,
      "tokens_seen": 2568421376
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022557575757575757,
      "loss": 1.5365,
      "theoretical_loss": 3.3612888893361275,
      "tokens_seen": 2568486912
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022555555555555556,
      "loss": 1.5599,
      "theoretical_loss": 3.3612820852695284,
      "tokens_seen": 2568552448
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022553535353535354,
      "loss": 1.6316,
      "theoretical_loss": 3.361275281425139,
      "tokens_seen": 2568617984
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002255151515151515,
      "loss": 1.5105,
      "theoretical_loss": 3.361268477802947,
      "tokens_seen": 2568683520
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002254949494949495,
      "loss": 1.3412,
      "theoretical_loss": 3.361261674402938,
      "tokens_seen": 2568749056
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022547474747474747,
      "loss": 1.3537,
      "theoretical_loss": 3.3612548712251002,
      "tokens_seen": 2568814592
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022545454545454548,
      "loss": 1.5087,
      "theoretical_loss": 3.3612480682694206,
      "tokens_seen": 2568880128
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022543434343434343,
      "loss": 1.2073,
      "theoretical_loss": 3.3612412655358854,
      "tokens_seen": 2568945664
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022541414141414144,
      "loss": 1.337,
      "theoretical_loss": 3.361234463024483,
      "tokens_seen": 2569011200
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002253939393939394,
      "loss": 1.3827,
      "theoretical_loss": 3.3612276607351994,
      "tokens_seen": 2569076736
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022537373737373736,
      "loss": 1.2977,
      "theoretical_loss": 3.3612208586680223,
      "tokens_seen": 2569142272
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022535353535353537,
      "loss": 1.3955,
      "theoretical_loss": 3.3612140568229383,
      "tokens_seen": 2569207808
    },
    {
      "epoch": 0.55,
      "objective/train/aligned_token_count": 0.5076979398727417,
      "objective/train/avg_token_score": 0.006365778390318155,
      "objective/train/docs_used": 1445073,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6958489418029785,
      "objective/train/original_loss": 4.13284969329834,
      "objective/train/theoretical_loss": 3.3612072551999352,
      "objective/train/tokens_used": 928132576,
      "objective/train/unlikelihood_loss": 0.05602798983454704,
      "theoretical_loss": 3.3612072551999352,
      "tokens_seen": 2569273344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022533333333333333,
      "loss": 1.413,
      "theoretical_loss": 3.3612072551999352,
      "tokens_seen": 2569273344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022531313131313134,
      "loss": 1.6071,
      "theoretical_loss": 3.361200453798999,
      "tokens_seen": 2569338880
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002252929292929293,
      "loss": 1.2984,
      "theoretical_loss": 3.361193652620118,
      "tokens_seen": 2569404416
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022527272727272728,
      "loss": 1.4737,
      "theoretical_loss": 3.361186851663278,
      "tokens_seen": 2569469952
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022525252525252526,
      "loss": 1.3457,
      "theoretical_loss": 3.3611800509284677,
      "tokens_seen": 2569535488
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022523232323232325,
      "loss": 1.1265,
      "theoretical_loss": 3.3611732504156726,
      "tokens_seen": 2569601024
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022521212121212123,
      "loss": 1.3491,
      "theoretical_loss": 3.361166450124881,
      "tokens_seen": 2569666560
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002251919191919192,
      "loss": 1.2513,
      "theoretical_loss": 3.361159650056079,
      "tokens_seen": 2569732096
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022517171717171717,
      "loss": 1.1886,
      "theoretical_loss": 3.361152850209254,
      "tokens_seen": 2569797632
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022515151515151515,
      "loss": 1.4985,
      "theoretical_loss": 3.3611460505843933,
      "tokens_seen": 2569863168
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022513131313131314,
      "loss": 1.4487,
      "theoretical_loss": 3.361139251181484,
      "tokens_seen": 2569928704
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022511111111111112,
      "loss": 1.3437,
      "theoretical_loss": 3.361132452000514,
      "tokens_seen": 2569994240
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002250909090909091,
      "loss": 1.3046,
      "theoretical_loss": 3.3611256530414684,
      "tokens_seen": 2570059776
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022507070707070706,
      "loss": 1.5022,
      "theoretical_loss": 3.361118854304336,
      "tokens_seen": 2570125312
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022505050505050507,
      "loss": 1.2068,
      "theoretical_loss": 3.361112055789103,
      "tokens_seen": 2570190848
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022503030303030303,
      "loss": 1.3943,
      "theoretical_loss": 3.361105257495757,
      "tokens_seen": 2570256384
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022501010101010102,
      "loss": 1.3968,
      "theoretical_loss": 3.361098459424285,
      "tokens_seen": 2570321920
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.000224989898989899,
      "loss": 1.347,
      "theoretical_loss": 3.3610916615746738,
      "tokens_seen": 2570387456
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022496969696969696,
      "loss": 1.4111,
      "theoretical_loss": 3.361084863946911,
      "tokens_seen": 2570452992
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022494949494949497,
      "loss": 1.4655,
      "theoretical_loss": 3.3610780665409834,
      "tokens_seen": 2570518528
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022492929292929292,
      "loss": 1.4423,
      "theoretical_loss": 3.361071269356878,
      "tokens_seen": 2570584064
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022490909090909094,
      "loss": 1.3561,
      "theoretical_loss": 3.361064472394582,
      "tokens_seen": 2570649600
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002248888888888889,
      "loss": 1.417,
      "theoretical_loss": 3.3610576756540826,
      "tokens_seen": 2570715136
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022486868686868685,
      "loss": 1.4114,
      "theoretical_loss": 3.361050879135367,
      "tokens_seen": 2570780672
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00022484848484848486,
      "loss": 1.4252,
      "theoretical_loss": 3.361044082838422,
      "tokens_seen": 2570846208
    },
    {
      "epoch": 0.55,
      "objective/train/aligned_token_count": 0.3723118305206299,
      "objective/train/avg_token_score": 0.03076431155204773,
      "objective/train/docs_used": 1446627,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1633058786392212,
      "objective/train/original_loss": 4.927713394165039,
      "objective/train/theoretical_loss": 3.361037286763235,
      "objective/train/tokens_used": 929770976,
      "objective/train/unlikelihood_loss": 0.054030515253543854,
      "theoretical_loss": 3.361037286763235,
      "tokens_seen": 2570911744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022482828282828282,
      "loss": 1.4558,
      "theoretical_loss": 3.361037286763235,
      "tokens_seen": 2570911744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022480808080808083,
      "loss": 1.2384,
      "theoretical_loss": 3.361030490909793,
      "tokens_seen": 2570977280
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022478787878787878,
      "loss": 1.3341,
      "theoretical_loss": 3.361023695278083,
      "tokens_seen": 2571042816
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022476767676767677,
      "loss": 1.5952,
      "theoretical_loss": 3.3610168998680923,
      "tokens_seen": 2571108352
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022474747474747475,
      "loss": 1.3186,
      "theoretical_loss": 3.3610101046798078,
      "tokens_seen": 2571173888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022472727272727274,
      "loss": 1.449,
      "theoretical_loss": 3.3610033097132166,
      "tokens_seen": 2571239424
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022470707070707072,
      "loss": 1.3722,
      "theoretical_loss": 3.3609965149683063,
      "tokens_seen": 2571304960
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022468686868686868,
      "loss": 1.5489,
      "theoretical_loss": 3.3609897204450636,
      "tokens_seen": 2571370496
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022466666666666666,
      "loss": 1.5714,
      "theoretical_loss": 3.3609829261434756,
      "tokens_seen": 2571436032
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022464646464646465,
      "loss": 1.2953,
      "theoretical_loss": 3.36097613206353,
      "tokens_seen": 2571501568
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022462626262626263,
      "loss": 1.4574,
      "theoretical_loss": 3.360969338205213,
      "tokens_seen": 2571567104
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002246060606060606,
      "loss": 1.5515,
      "theoretical_loss": 3.3609625445685123,
      "tokens_seen": 2571632640
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002245858585858586,
      "loss": 1.4051,
      "theoretical_loss": 3.3609557511534147,
      "tokens_seen": 2571698176
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022456565656565658,
      "loss": 1.4488,
      "theoretical_loss": 3.3609489579599074,
      "tokens_seen": 2571763712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022454545454545457,
      "loss": 1.6168,
      "theoretical_loss": 3.360942164987978,
      "tokens_seen": 2571829248
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022452525252525252,
      "loss": 1.3144,
      "theoretical_loss": 3.360935372237613,
      "tokens_seen": 2571894784
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002245050505050505,
      "loss": 1.4169,
      "theoretical_loss": 3.3609285797088004,
      "tokens_seen": 2571960320
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002244848484848485,
      "loss": 1.4572,
      "theoretical_loss": 3.3609217874015265,
      "tokens_seen": 2572025856
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022446464646464647,
      "loss": 1.5298,
      "theoretical_loss": 3.3609149953157784,
      "tokens_seen": 2572091392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022444444444444446,
      "loss": 1.1723,
      "theoretical_loss": 3.360908203451543,
      "tokens_seen": 2572156928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022442424242424241,
      "loss": 1.3438,
      "theoretical_loss": 3.3609014118088085,
      "tokens_seen": 2572222464
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022440404040404043,
      "loss": 1.2535,
      "theoretical_loss": 3.3608946203875614,
      "tokens_seen": 2572288000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022438383838383838,
      "loss": 1.3805,
      "theoretical_loss": 3.360887829187789,
      "tokens_seen": 2572353536
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002243636363636364,
      "loss": 1.5625,
      "theoretical_loss": 3.3608810382094783,
      "tokens_seen": 2572419072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022434343434343435,
      "loss": 1.4425,
      "theoretical_loss": 3.3608742474526165,
      "tokens_seen": 2572484608
    },
    {
      "epoch": 0.56,
      "objective/train/aligned_token_count": 0.3367546498775482,
      "objective/train/avg_token_score": 0.029736477881669998,
      "objective/train/docs_used": 1447125,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1136897802352905,
      "objective/train/original_loss": 4.943774700164795,
      "objective/train/theoretical_loss": 3.36086745691719,
      "objective/train/tokens_used": 931409376,
      "objective/train/unlikelihood_loss": 0.04741252213716507,
      "theoretical_loss": 3.36086745691719,
      "tokens_seen": 2572550144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002243232323232323,
      "loss": 1.4259,
      "theoretical_loss": 3.36086745691719,
      "tokens_seen": 2572550144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022430303030303032,
      "loss": 1.1761,
      "theoretical_loss": 3.3608606666031875,
      "tokens_seen": 2572615680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022428282828282828,
      "loss": 1.1985,
      "theoretical_loss": 3.360853876510595,
      "tokens_seen": 2572681216
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002242626262626263,
      "loss": 1.3555,
      "theoretical_loss": 3.3608470866394,
      "tokens_seen": 2572746752
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022424242424242424,
      "loss": 1.6411,
      "theoretical_loss": 3.360840296989589,
      "tokens_seen": 2572812288
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022422222222222223,
      "loss": 1.6482,
      "theoretical_loss": 3.3608335075611504,
      "tokens_seen": 2572877824
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002242020202020202,
      "loss": 1.4181,
      "theoretical_loss": 3.3608267183540703,
      "tokens_seen": 2572943360
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022418181818181817,
      "loss": 1.2735,
      "theoretical_loss": 3.3608199293683363,
      "tokens_seen": 2573008896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022416161616161618,
      "loss": 1.4303,
      "theoretical_loss": 3.360813140603935,
      "tokens_seen": 2573074432
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022414141414141414,
      "loss": 1.5768,
      "theoretical_loss": 3.3608063520608544,
      "tokens_seen": 2573139968
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022412121212121212,
      "loss": 1.2446,
      "theoretical_loss": 3.3607995637390813,
      "tokens_seen": 2573205504
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002241010101010101,
      "loss": 1.266,
      "theoretical_loss": 3.3607927756386027,
      "tokens_seen": 2573271040
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002240808080808081,
      "loss": 1.3275,
      "theoretical_loss": 3.3607859877594057,
      "tokens_seen": 2573336576
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022406060606060607,
      "loss": 1.4812,
      "theoretical_loss": 3.3607792001014776,
      "tokens_seen": 2573402112
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022404040404040406,
      "loss": 1.507,
      "theoretical_loss": 3.3607724126648058,
      "tokens_seen": 2573467648
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000224020202020202,
      "loss": 1.1501,
      "theoretical_loss": 3.360765625449377,
      "tokens_seen": 2573533184
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000224,
      "loss": 1.2942,
      "theoretical_loss": 3.3607588384551788,
      "tokens_seen": 2573598720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022397979797979798,
      "loss": 1.4081,
      "theoretical_loss": 3.3607520516821974,
      "tokens_seen": 2573664256
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022395959595959596,
      "loss": 1.298,
      "theoretical_loss": 3.3607452651304213,
      "tokens_seen": 2573729792
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022393939393939395,
      "loss": 1.6518,
      "theoretical_loss": 3.3607384787998367,
      "tokens_seen": 2573795328
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002239191919191919,
      "loss": 1.5337,
      "theoretical_loss": 3.360731692690431,
      "tokens_seen": 2573860864
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022389898989898992,
      "loss": 1.5273,
      "theoretical_loss": 3.360724906802192,
      "tokens_seen": 2573926400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022387878787878787,
      "loss": 1.4185,
      "theoretical_loss": 3.3607181211351054,
      "tokens_seen": 2573991936
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022385858585858588,
      "loss": 1.4221,
      "theoretical_loss": 3.36071133568916,
      "tokens_seen": 2574057472
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022383838383838384,
      "loss": 1.1312,
      "theoretical_loss": 3.360704550464342,
      "tokens_seen": 2574123008
    },
    {
      "epoch": 0.56,
      "objective/train/aligned_token_count": 0.5105693936347961,
      "objective/train/avg_token_score": 0.005291319452226162,
      "objective/train/docs_used": 1447717,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6546176671981812,
      "objective/train/original_loss": 3.752079725265503,
      "objective/train/theoretical_loss": 3.3606977654606385,
      "objective/train/tokens_used": 933047776,
      "objective/train/unlikelihood_loss": 0.06454199552536011,
      "theoretical_loss": 3.3606977654606385,
      "tokens_seen": 2574188544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002238181818181818,
      "loss": 1.3505,
      "theoretical_loss": 3.3606977654606385,
      "tokens_seen": 2574188544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002237979797979798,
      "loss": 1.4475,
      "theoretical_loss": 3.360690980678037,
      "tokens_seen": 2574254080
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022377777777777777,
      "loss": 1.209,
      "theoretical_loss": 3.360684196116525,
      "tokens_seen": 2574319616
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022375757575757578,
      "loss": 1.2548,
      "theoretical_loss": 3.3606774117760887,
      "tokens_seen": 2574385152
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022373737373737373,
      "loss": 1.2896,
      "theoretical_loss": 3.360670627656716,
      "tokens_seen": 2574450688
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022371717171717172,
      "loss": 1.24,
      "theoretical_loss": 3.3606638437583944,
      "tokens_seen": 2574516224
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002236969696969697,
      "loss": 1.4289,
      "theoretical_loss": 3.36065706008111,
      "tokens_seen": 2574581760
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022367676767676769,
      "loss": 1.378,
      "theoretical_loss": 3.3606502766248507,
      "tokens_seen": 2574647296
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022365656565656567,
      "loss": 1.3765,
      "theoretical_loss": 3.360643493389604,
      "tokens_seen": 2574712832
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022363636363636363,
      "loss": 1.3329,
      "theoretical_loss": 3.3606367103753563,
      "tokens_seen": 2574778368
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022361616161616164,
      "loss": 1.2627,
      "theoretical_loss": 3.360629927582095,
      "tokens_seen": 2574843904
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002235959595959596,
      "loss": 1.4574,
      "theoretical_loss": 3.360623145009807,
      "tokens_seen": 2574909440
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022357575757575758,
      "loss": 1.3087,
      "theoretical_loss": 3.3606163626584804,
      "tokens_seen": 2574974976
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022355555555555556,
      "loss": 1.4547,
      "theoretical_loss": 3.3606095805281018,
      "tokens_seen": 2575040512
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022353535353535355,
      "loss": 1.406,
      "theoretical_loss": 3.3606027986186584,
      "tokens_seen": 2575106048
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022351515151515153,
      "loss": 1.3002,
      "theoretical_loss": 3.360596016930137,
      "tokens_seen": 2575171584
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022349494949494951,
      "loss": 1.2865,
      "theoretical_loss": 3.3605892354625255,
      "tokens_seen": 2575237120
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022347474747474747,
      "loss": 1.5481,
      "theoretical_loss": 3.3605824542158107,
      "tokens_seen": 2575302656
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022345454545454546,
      "loss": 1.4987,
      "theoretical_loss": 3.3605756731899796,
      "tokens_seen": 2575368192
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022343434343434344,
      "loss": 1.099,
      "theoretical_loss": 3.3605688923850194,
      "tokens_seen": 2575433728
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022341414141414142,
      "loss": 1.481,
      "theoretical_loss": 3.3605621118009177,
      "tokens_seen": 2575499264
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002233939393939394,
      "loss": 1.2602,
      "theoretical_loss": 3.360555331437662,
      "tokens_seen": 2575564800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022337373737373736,
      "loss": 1.3933,
      "theoretical_loss": 3.3605485512952384,
      "tokens_seen": 2575630336
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022335353535353537,
      "loss": 1.313,
      "theoretical_loss": 3.3605417713736347,
      "tokens_seen": 2575695872
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022333333333333333,
      "loss": 1.4967,
      "theoretical_loss": 3.360534991672838,
      "tokens_seen": 2575761408
    },
    {
      "epoch": 0.56,
      "objective/train/aligned_token_count": 0.374450147151947,
      "objective/train/avg_token_score": 0.01964675262570381,
      "objective/train/docs_used": 1449217,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1246662139892578,
      "objective/train/original_loss": 4.803574085235596,
      "objective/train/theoretical_loss": 3.3605282121928357,
      "objective/train/tokens_used": 934686176,
      "objective/train/unlikelihood_loss": 0.08262909948825836,
      "theoretical_loss": 3.3605282121928357,
      "tokens_seen": 2575826944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022331313131313132,
      "loss": 1.1786,
      "theoretical_loss": 3.3605282121928357,
      "tokens_seen": 2575826944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002232929292929293,
      "loss": 1.3228,
      "theoretical_loss": 3.3605214329336146,
      "tokens_seen": 2575892480
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022327272727272726,
      "loss": 1.1503,
      "theoretical_loss": 3.3605146538951627,
      "tokens_seen": 2575958016
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022325252525252527,
      "loss": 1.452,
      "theoretical_loss": 3.360507875077466,
      "tokens_seen": 2576023552
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022323232323232322,
      "loss": 1.2441,
      "theoretical_loss": 3.3605010964805127,
      "tokens_seen": 2576089088
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022321212121212124,
      "loss": 1.4426,
      "theoretical_loss": 3.3604943181042897,
      "tokens_seen": 2576154624
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002231919191919192,
      "loss": 1.299,
      "theoretical_loss": 3.3604875399487835,
      "tokens_seen": 2576220160
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022317171717171718,
      "loss": 1.2237,
      "theoretical_loss": 3.360480762013982,
      "tokens_seen": 2576285696
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022315151515151516,
      "loss": 1.4432,
      "theoretical_loss": 3.3604739842998725,
      "tokens_seen": 2576351232
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022313131313131312,
      "loss": 1.3979,
      "theoretical_loss": 3.360467206806442,
      "tokens_seen": 2576416768
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022311111111111113,
      "loss": 1.3286,
      "theoretical_loss": 3.3604604295336777,
      "tokens_seen": 2576482304
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022309090909090909,
      "loss": 1.2064,
      "theoretical_loss": 3.360453652481567,
      "tokens_seen": 2576547840
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022307070707070707,
      "loss": 1.3541,
      "theoretical_loss": 3.360446875650097,
      "tokens_seen": 2576613376
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022305050505050505,
      "loss": 1.3501,
      "theoretical_loss": 3.3604400990392547,
      "tokens_seen": 2576678912
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022303030303030304,
      "loss": 1.6645,
      "theoretical_loss": 3.360433322649027,
      "tokens_seen": 2576744448
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022301010101010102,
      "loss": 1.4107,
      "theoretical_loss": 3.3604265464794016,
      "tokens_seen": 2576809984
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000222989898989899,
      "loss": 1.3029,
      "theoretical_loss": 3.360419770530366,
      "tokens_seen": 2576875520
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022296969696969696,
      "loss": 1.4655,
      "theoretical_loss": 3.360412994801907,
      "tokens_seen": 2576941056
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022294949494949495,
      "loss": 1.2675,
      "theoretical_loss": 3.3604062192940116,
      "tokens_seen": 2577006592
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022292929292929293,
      "loss": 1.4365,
      "theoretical_loss": 3.360399444006667,
      "tokens_seen": 2577072128
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022290909090909091,
      "loss": 1.2802,
      "theoretical_loss": 3.3603926689398613,
      "tokens_seen": 2577137664
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002228888888888889,
      "loss": 1.2442,
      "theoretical_loss": 3.360385894093581,
      "tokens_seen": 2577203200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022286868686868685,
      "loss": 1.1822,
      "theoretical_loss": 3.360379119467813,
      "tokens_seen": 2577268736
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022284848484848487,
      "loss": 1.6455,
      "theoretical_loss": 3.3603723450625456,
      "tokens_seen": 2577334272
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022282828282828282,
      "loss": 1.2621,
      "theoretical_loss": 3.3603655708777644,
      "tokens_seen": 2577399808
    },
    {
      "epoch": 0.56,
      "objective/train/aligned_token_count": 0.31885385513305664,
      "objective/train/avg_token_score": 0.007736933883279562,
      "objective/train/docs_used": 1449913,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1018469333648682,
      "objective/train/original_loss": 5.282415390014648,
      "objective/train/theoretical_loss": 3.3603587969134585,
      "objective/train/tokens_used": 936324576,
      "objective/train/unlikelihood_loss": 0.052301522344350815,
      "theoretical_loss": 3.3603587969134585,
      "tokens_seen": 2577465344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022280808080808083,
      "loss": 1.4588,
      "theoretical_loss": 3.3603587969134585,
      "tokens_seen": 2577465344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002227878787878788,
      "loss": 1.0516,
      "theoretical_loss": 3.3603520231696136,
      "tokens_seen": 2577530880
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022276767676767677,
      "loss": 1.29,
      "theoretical_loss": 3.3603452496462176,
      "tokens_seen": 2577596416
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022274747474747476,
      "loss": 1.4947,
      "theoretical_loss": 3.3603384763432578,
      "tokens_seen": 2577661952
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022272727272727272,
      "loss": 1.3545,
      "theoretical_loss": 3.360331703260721,
      "tokens_seen": 2577727488
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022270707070707073,
      "loss": 1.2736,
      "theoretical_loss": 3.3603249303985945,
      "tokens_seen": 2577793024
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022268686868686868,
      "loss": 1.288,
      "theoretical_loss": 3.360318157756866,
      "tokens_seen": 2577858560
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002226666666666667,
      "loss": 1.6162,
      "theoretical_loss": 3.3603113853355224,
      "tokens_seen": 2577924096
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022264646464646465,
      "loss": 1.1718,
      "theoretical_loss": 3.360304613134551,
      "tokens_seen": 2577989632
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002226262626262626,
      "loss": 1.258,
      "theoretical_loss": 3.3602978411539386,
      "tokens_seen": 2578055168
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022260606060606062,
      "loss": 1.1874,
      "theoretical_loss": 3.360291069393673,
      "tokens_seen": 2578120704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022258585858585858,
      "loss": 1.4589,
      "theoretical_loss": 3.3602842978537413,
      "tokens_seen": 2578186240
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002225656565656566,
      "loss": 1.4761,
      "theoretical_loss": 3.360277526534131,
      "tokens_seen": 2578251776
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022254545454545454,
      "loss": 1.2871,
      "theoretical_loss": 3.3602707554348283,
      "tokens_seen": 2578317312
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022252525252525253,
      "loss": 1.2251,
      "theoretical_loss": 3.360263984555821,
      "tokens_seen": 2578382848
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002225050505050505,
      "loss": 1.2989,
      "theoretical_loss": 3.360257213897097,
      "tokens_seen": 2578448384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002224848484848485,
      "loss": 1.5024,
      "theoretical_loss": 3.360250443458643,
      "tokens_seen": 2578513920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022246464646464648,
      "loss": 1.373,
      "theoretical_loss": 3.360243673240446,
      "tokens_seen": 2578579456
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022244444444444444,
      "loss": 1.1986,
      "theoretical_loss": 3.3602369032424937,
      "tokens_seen": 2578644992
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022242424242424242,
      "loss": 1.3787,
      "theoretical_loss": 3.3602301334647726,
      "tokens_seen": 2578710528
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002224040404040404,
      "loss": 1.1235,
      "theoretical_loss": 3.360223363907271,
      "tokens_seen": 2578776064
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002223838383838384,
      "loss": 1.2505,
      "theoretical_loss": 3.360216594569975,
      "tokens_seen": 2578841600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022236363636363637,
      "loss": 1.3667,
      "theoretical_loss": 3.3602098254528725,
      "tokens_seen": 2578907136
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022234343434343436,
      "loss": 1.3972,
      "theoretical_loss": 3.3602030565559513,
      "tokens_seen": 2578972672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002223232323232323,
      "loss": 1.4805,
      "theoretical_loss": 3.360196287879197,
      "tokens_seen": 2579038208
    },
    {
      "epoch": 0.56,
      "objective/train/aligned_token_count": 0.36003178358078003,
      "objective/train/avg_token_score": 0.012265270575881004,
      "objective/train/docs_used": 1451054,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.212681770324707,
      "objective/train/original_loss": 4.321657657623291,
      "objective/train/theoretical_loss": 3.360189519422599,
      "objective/train/tokens_used": 937962976,
      "objective/train/unlikelihood_loss": 0.07353614270687103,
      "theoretical_loss": 3.360189519422599,
      "tokens_seen": 2579103744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022230303030303032,
      "loss": 1.2299,
      "theoretical_loss": 3.360189519422599,
      "tokens_seen": 2579103744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022228282828282828,
      "loss": 1.4827,
      "theoretical_loss": 3.3601827511861426,
      "tokens_seen": 2579169280
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022226262626262626,
      "loss": 1.2953,
      "theoretical_loss": 3.3601759831698157,
      "tokens_seen": 2579234816
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022224242424242425,
      "loss": 1.3292,
      "theoretical_loss": 3.3601692153736065,
      "tokens_seen": 2579300352
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002222222222222222,
      "loss": 1.2338,
      "theoretical_loss": 3.3601624477975007,
      "tokens_seen": 2579365888
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022220202020202022,
      "loss": 1.4546,
      "theoretical_loss": 3.3601556804414865,
      "tokens_seen": 2579431424
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022218181818181817,
      "loss": 1.4757,
      "theoretical_loss": 3.360148913305551,
      "tokens_seen": 2579496960
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022216161616161618,
      "loss": 1.2787,
      "theoretical_loss": 3.3601421463896815,
      "tokens_seen": 2579562496
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022214141414141414,
      "loss": 1.2455,
      "theoretical_loss": 3.3601353796938644,
      "tokens_seen": 2579628032
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022212121212121213,
      "loss": 1.1283,
      "theoretical_loss": 3.3601286132180883,
      "tokens_seen": 2579693568
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002221010101010101,
      "loss": 1.2599,
      "theoretical_loss": 3.36012184696234,
      "tokens_seen": 2579759104
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022208080808080807,
      "loss": 1.3811,
      "theoretical_loss": 3.3601150809266063,
      "tokens_seen": 2579824640
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022206060606060608,
      "loss": 1.5556,
      "theoretical_loss": 3.360108315110875,
      "tokens_seen": 2579890176
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022204040404040403,
      "loss": 1.3794,
      "theoretical_loss": 3.360101549515133,
      "tokens_seen": 2579955712
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022202020202020202,
      "loss": 1.3777,
      "theoretical_loss": 3.360094784139368,
      "tokens_seen": 2580021248
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000222,
      "loss": 1.3035,
      "theoretical_loss": 3.360088018983566,
      "tokens_seen": 2580086784
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022197979797979799,
      "loss": 1.457,
      "theoretical_loss": 3.3600812540477163,
      "tokens_seen": 2580152320
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022195959595959597,
      "loss": 1.0994,
      "theoretical_loss": 3.3600744893318044,
      "tokens_seen": 2580217856
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022193939393939395,
      "loss": 1.5001,
      "theoretical_loss": 3.3600677248358184,
      "tokens_seen": 2580283392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022191919191919194,
      "loss": 1.4878,
      "theoretical_loss": 3.3600609605597453,
      "tokens_seen": 2580348928
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002218989898989899,
      "loss": 1.3555,
      "theoretical_loss": 3.3600541965035724,
      "tokens_seen": 2580414464
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022187878787878788,
      "loss": 1.2467,
      "theoretical_loss": 3.360047432667287,
      "tokens_seen": 2580480000
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022185858585858586,
      "loss": 1.5221,
      "theoretical_loss": 3.3600406690508766,
      "tokens_seen": 2580545536
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022183838383838385,
      "loss": 1.3442,
      "theoretical_loss": 3.360033905654328,
      "tokens_seen": 2580611072
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022181818181818183,
      "loss": 1.3576,
      "theoretical_loss": 3.360027142477629,
      "tokens_seen": 2580676608
    },
    {
      "epoch": 0.56,
      "objective/train/aligned_token_count": 0.467253178358078,
      "objective/train/avg_token_score": 0.0198719073086977,
      "objective/train/docs_used": 1451835,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5594650506973267,
      "objective/train/original_loss": 4.650272369384766,
      "objective/train/theoretical_loss": 3.3600203795207664,
      "objective/train/tokens_used": 939601376,
      "objective/train/unlikelihood_loss": 0.04788258671760559,
      "theoretical_loss": 3.3600203795207664,
      "tokens_seen": 2580742144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022179797979797981,
      "loss": 1.3465,
      "theoretical_loss": 3.3600203795207664,
      "tokens_seen": 2580742144
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022177777777777777,
      "loss": 1.2599,
      "theoretical_loss": 3.360013616783728,
      "tokens_seen": 2580807680
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022175757575757576,
      "loss": 1.2361,
      "theoretical_loss": 3.3600068542665005,
      "tokens_seen": 2580873216
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022173737373737374,
      "loss": 1.4571,
      "theoretical_loss": 3.3600000919690713,
      "tokens_seen": 2580938752
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022171717171717172,
      "loss": 1.4269,
      "theoretical_loss": 3.3599933298914277,
      "tokens_seen": 2581004288
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002216969696969697,
      "loss": 1.4532,
      "theoretical_loss": 3.359986568033557,
      "tokens_seen": 2581069824
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022167676767676766,
      "loss": 1.2831,
      "theoretical_loss": 3.3599798063954474,
      "tokens_seen": 2581135360
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022165656565656568,
      "loss": 1.3904,
      "theoretical_loss": 3.3599730449770844,
      "tokens_seen": 2581200896
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022163636363636363,
      "loss": 1.3269,
      "theoretical_loss": 3.3599662837784567,
      "tokens_seen": 2581266432
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022161616161616164,
      "loss": 1.3338,
      "theoretical_loss": 3.359959522799551,
      "tokens_seen": 2581331968
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002215959595959596,
      "loss": 1.4558,
      "theoretical_loss": 3.359952762040354,
      "tokens_seen": 2581397504
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022157575757575756,
      "loss": 1.4323,
      "theoretical_loss": 3.359946001500855,
      "tokens_seen": 2581463040
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022155555555555557,
      "loss": 1.1677,
      "theoretical_loss": 3.359939241181039,
      "tokens_seen": 2581528576
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022153535353535352,
      "loss": 1.4649,
      "theoretical_loss": 3.3599324810808944,
      "tokens_seen": 2581594112
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022151515151515154,
      "loss": 1.2148,
      "theoretical_loss": 3.3599257212004083,
      "tokens_seen": 2581659648
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002214949494949495,
      "loss": 1.2817,
      "theoretical_loss": 3.3599189615395675,
      "tokens_seen": 2581725184
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022147474747474748,
      "loss": 1.33,
      "theoretical_loss": 3.3599122020983607,
      "tokens_seen": 2581790720
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022145454545454546,
      "loss": 1.3766,
      "theoretical_loss": 3.359905442876774,
      "tokens_seen": 2581856256
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022143434343434344,
      "loss": 1.448,
      "theoretical_loss": 3.3598986838747944,
      "tokens_seen": 2581921792
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022141414141414143,
      "loss": 1.4007,
      "theoretical_loss": 3.35989192509241,
      "tokens_seen": 2581987328
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022139393939393939,
      "loss": 1.3056,
      "theoretical_loss": 3.359885166529608,
      "tokens_seen": 2582052864
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022137373737373737,
      "loss": 1.2278,
      "theoretical_loss": 3.3598784081863755,
      "tokens_seen": 2582118400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022135353535353535,
      "loss": 1.3436,
      "theoretical_loss": 3.3598716500626997,
      "tokens_seen": 2582183936
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022133333333333334,
      "loss": 1.419,
      "theoretical_loss": 3.359864892158568,
      "tokens_seen": 2582249472
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022131313131313132,
      "loss": 1.388,
      "theoretical_loss": 3.359858134473968,
      "tokens_seen": 2582315008
    },
    {
      "epoch": 0.56,
      "objective/train/aligned_token_count": 0.39821603894233704,
      "objective/train/avg_token_score": 0.018147850409150124,
      "objective/train/docs_used": 1453267,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2614893913269043,
      "objective/train/original_loss": 4.818430423736572,
      "objective/train/theoretical_loss": 3.3598513770088863,
      "objective/train/tokens_used": 941239776,
      "objective/train/unlikelihood_loss": 0.04437202587723732,
      "theoretical_loss": 3.3598513770088863,
      "tokens_seen": 2582380544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002212929292929293,
      "loss": 1.4903,
      "theoretical_loss": 3.3598513770088863,
      "tokens_seen": 2582380544
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022127272727272726,
      "loss": 1.2972,
      "theoretical_loss": 3.3598446197633107,
      "tokens_seen": 2582446080
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022125252525252527,
      "loss": 1.4752,
      "theoretical_loss": 3.3598378627372285,
      "tokens_seen": 2582511616
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022123232323232323,
      "loss": 1.4921,
      "theoretical_loss": 3.359831105930627,
      "tokens_seen": 2582577152
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022121212121212121,
      "loss": 1.0996,
      "theoretical_loss": 3.3598243493434934,
      "tokens_seen": 2582642688
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002211919191919192,
      "loss": 1.5738,
      "theoretical_loss": 3.359817592975815,
      "tokens_seen": 2582708224
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022117171717171715,
      "loss": 1.4451,
      "theoretical_loss": 3.359810836827579,
      "tokens_seen": 2582773760
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022115151515151517,
      "loss": 1.4438,
      "theoretical_loss": 3.359804080898773,
      "tokens_seen": 2582839296
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022113131313131312,
      "loss": 1.3545,
      "theoretical_loss": 3.3597973251893842,
      "tokens_seen": 2582904832
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022111111111111113,
      "loss": 1.4429,
      "theoretical_loss": 3.3597905696993995,
      "tokens_seen": 2582970368
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002210909090909091,
      "loss": 1.1898,
      "theoretical_loss": 3.359783814428807,
      "tokens_seen": 2583035904
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002210707070707071,
      "loss": 1.5361,
      "theoretical_loss": 3.359777059377593,
      "tokens_seen": 2583101440
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022105050505050506,
      "loss": 1.4624,
      "theoretical_loss": 3.3597703045457457,
      "tokens_seen": 2583166976
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022103030303030302,
      "loss": 1.5441,
      "theoretical_loss": 3.359763549933252,
      "tokens_seen": 2583232512
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022101010101010103,
      "loss": 1.3542,
      "theoretical_loss": 3.359756795540099,
      "tokens_seen": 2583298048
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022098989898989898,
      "loss": 1.3278,
      "theoretical_loss": 3.359750041366275,
      "tokens_seen": 2583363584
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.000220969696969697,
      "loss": 1.5481,
      "theoretical_loss": 3.359743287411766,
      "tokens_seen": 2583429120
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022094949494949495,
      "loss": 1.3825,
      "theoretical_loss": 3.35973653367656,
      "tokens_seen": 2583494656
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022092929292929294,
      "loss": 1.4765,
      "theoretical_loss": 3.3597297801606443,
      "tokens_seen": 2583560192
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022090909090909092,
      "loss": 1.3255,
      "theoretical_loss": 3.3597230268640064,
      "tokens_seen": 2583625728
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022088888888888888,
      "loss": 1.4902,
      "theoretical_loss": 3.359716273786633,
      "tokens_seen": 2583691264
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002208686868686869,
      "loss": 1.3058,
      "theoretical_loss": 3.3597095209285124,
      "tokens_seen": 2583756800
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022084848484848484,
      "loss": 1.1516,
      "theoretical_loss": 3.3597027682896305,
      "tokens_seen": 2583822336
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022082828282828283,
      "loss": 1.4403,
      "theoretical_loss": 3.359696015869976,
      "tokens_seen": 2583887872
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002208080808080808,
      "loss": 1.3537,
      "theoretical_loss": 3.3596892636695355,
      "tokens_seen": 2583953408
    },
    {
      "epoch": 0.56,
      "objective/train/aligned_token_count": 0.3715175986289978,
      "objective/train/avg_token_score": 0.010973823256790638,
      "objective/train/docs_used": 1453825,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9919140338897705,
      "objective/train/original_loss": 4.099684238433838,
      "objective/train/theoretical_loss": 3.3596825116882965,
      "objective/train/tokens_used": 942878176,
      "objective/train/unlikelihood_loss": 0.06531152874231339,
      "theoretical_loss": 3.3596825116882965,
      "tokens_seen": 2584018944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002207878787878788,
      "loss": 1.3124,
      "theoretical_loss": 3.3596825116882965,
      "tokens_seen": 2584018944
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022076767676767678,
      "loss": 1.3673,
      "theoretical_loss": 3.3596757599262466,
      "tokens_seen": 2584084480
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022074747474747476,
      "loss": 1.5433,
      "theoretical_loss": 3.3596690083833725,
      "tokens_seen": 2584150016
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022072727272727272,
      "loss": 1.3628,
      "theoretical_loss": 3.3596622570596617,
      "tokens_seen": 2584215552
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002207070707070707,
      "loss": 1.3922,
      "theoretical_loss": 3.3596555059551023,
      "tokens_seen": 2584281088
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002206868686868687,
      "loss": 1.2585,
      "theoretical_loss": 3.3596487550696805,
      "tokens_seen": 2584346624
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022066666666666667,
      "loss": 1.4225,
      "theoretical_loss": 3.3596420044033843,
      "tokens_seen": 2584412160
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022064646464646466,
      "loss": 1.265,
      "theoretical_loss": 3.359635253956201,
      "tokens_seen": 2584477696
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002206262626262626,
      "loss": 1.2734,
      "theoretical_loss": 3.3596285037281177,
      "tokens_seen": 2584543232
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022060606060606062,
      "loss": 1.205,
      "theoretical_loss": 3.3596217537191215,
      "tokens_seen": 2584608768
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022058585858585858,
      "loss": 1.2135,
      "theoretical_loss": 3.3596150039292008,
      "tokens_seen": 2584674304
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002205656565656566,
      "loss": 1.4511,
      "theoretical_loss": 3.359608254358342,
      "tokens_seen": 2584739840
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022054545454545455,
      "loss": 1.184,
      "theoretical_loss": 3.359601505006532,
      "tokens_seen": 2584805376
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002205252525252525,
      "loss": 1.3055,
      "theoretical_loss": 3.3595947558737596,
      "tokens_seen": 2584870912
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022050505050505052,
      "loss": 1.3172,
      "theoretical_loss": 3.359588006960011,
      "tokens_seen": 2584936448
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022048484848484847,
      "loss": 1.3577,
      "theoretical_loss": 3.359581258265274,
      "tokens_seen": 2585001984
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022046464646464649,
      "loss": 1.3392,
      "theoretical_loss": 3.3595745097895358,
      "tokens_seen": 2585067520
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022044444444444444,
      "loss": 1.281,
      "theoretical_loss": 3.3595677615327837,
      "tokens_seen": 2585133056
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022042424242424243,
      "loss": 0.9519,
      "theoretical_loss": 3.359561013495005,
      "tokens_seen": 2585198592
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002204040404040404,
      "loss": 1.2053,
      "theoretical_loss": 3.359554265676187,
      "tokens_seen": 2585264128
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002203838383838384,
      "loss": 1.2398,
      "theoretical_loss": 3.3595475180763175,
      "tokens_seen": 2585329664
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022036363636363638,
      "loss": 1.2149,
      "theoretical_loss": 3.3595407706953835,
      "tokens_seen": 2585395200
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022034343434343433,
      "loss": 1.3925,
      "theoretical_loss": 3.3595340235333726,
      "tokens_seen": 2585460736
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022032323232323232,
      "loss": 1.3149,
      "theoretical_loss": 3.3595272765902715,
      "tokens_seen": 2585526272
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002203030303030303,
      "loss": 1.3965,
      "theoretical_loss": 3.3595205298660678,
      "tokens_seen": 2585591808
    },
    {
      "epoch": 0.56,
      "objective/train/aligned_token_count": 0.3884408473968506,
      "objective/train/avg_token_score": 0.014901697635650635,
      "objective/train/docs_used": 1454456,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3513553142547607,
      "objective/train/original_loss": 5.131579399108887,
      "objective/train/theoretical_loss": 3.3595137833607494,
      "objective/train/tokens_used": 944516576,
      "objective/train/unlikelihood_loss": 0.041818611323833466,
      "theoretical_loss": 3.3595137833607494,
      "tokens_seen": 2585657344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002202828282828283,
      "loss": 1.5559,
      "theoretical_loss": 3.3595137833607494,
      "tokens_seen": 2585657344
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022026262626262627,
      "loss": 1.1281,
      "theoretical_loss": 3.3595070370743034,
      "tokens_seen": 2585722880
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022024242424242425,
      "loss": 1.3685,
      "theoretical_loss": 3.3595002910067167,
      "tokens_seen": 2585788416
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022022222222222224,
      "loss": 1.3789,
      "theoretical_loss": 3.359493545157977,
      "tokens_seen": 2585853952
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022020202020202022,
      "loss": 1.2819,
      "theoretical_loss": 3.3594867995280717,
      "tokens_seen": 2585919488
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022018181818181818,
      "loss": 1.3423,
      "theoretical_loss": 3.3594800541169882,
      "tokens_seen": 2585985024
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022016161616161616,
      "loss": 1.3769,
      "theoretical_loss": 3.3594733089247137,
      "tokens_seen": 2586050560
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022014141414141415,
      "loss": 1.2973,
      "theoretical_loss": 3.3594665639512358,
      "tokens_seen": 2586116096
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022012121212121213,
      "loss": 1.3739,
      "theoretical_loss": 3.359459819196541,
      "tokens_seen": 2586181632
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022010101010101012,
      "loss": 1.2586,
      "theoretical_loss": 3.359453074660618,
      "tokens_seen": 2586247168
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022008080808080807,
      "loss": 1.3217,
      "theoretical_loss": 3.359446330343453,
      "tokens_seen": 2586312704
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022006060606060608,
      "loss": 1.1279,
      "theoretical_loss": 3.3594395862450344,
      "tokens_seen": 2586378240
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022004040404040404,
      "loss": 1.281,
      "theoretical_loss": 3.3594328423653486,
      "tokens_seen": 2586443776
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022002020202020202,
      "loss": 1.1394,
      "theoretical_loss": 3.3594260987043834,
      "tokens_seen": 2586509312
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00022,
      "loss": 1.2573,
      "theoretical_loss": 3.3594193552621263,
      "tokens_seen": 2586574848
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021997979797979796,
      "loss": 1.2163,
      "theoretical_loss": 3.3594126120385646,
      "tokens_seen": 2586640384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021995959595959598,
      "loss": 1.3334,
      "theoretical_loss": 3.359405869033685,
      "tokens_seen": 2586705920
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021993939393939393,
      "loss": 1.0991,
      "theoretical_loss": 3.359399126247476,
      "tokens_seen": 2586771456
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021991919191919194,
      "loss": 1.3944,
      "theoretical_loss": 3.359392383679924,
      "tokens_seen": 2586836992
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002198989898989899,
      "loss": 1.5423,
      "theoretical_loss": 3.3593856413310172,
      "tokens_seen": 2586902528
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021987878787878788,
      "loss": 1.3778,
      "theoretical_loss": 3.3593788992007423,
      "tokens_seen": 2586968064
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021985858585858587,
      "loss": 1.2533,
      "theoretical_loss": 3.359372157289087,
      "tokens_seen": 2587033600
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021983838383838383,
      "loss": 1.2668,
      "theoretical_loss": 3.3593654155960384,
      "tokens_seen": 2587099136
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021981818181818184,
      "loss": 1.3771,
      "theoretical_loss": 3.359358674121584,
      "tokens_seen": 2587164672
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002197979797979798,
      "loss": 1.291,
      "theoretical_loss": 3.3593519328657115,
      "tokens_seen": 2587230208
    },
    {
      "epoch": 0.56,
      "objective/train/aligned_token_count": 0.47904446721076965,
      "objective/train/avg_token_score": 0.014366981573402882,
      "objective/train/docs_used": 1455631,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.684888482093811,
      "objective/train/original_loss": 4.289407253265381,
      "objective/train/theoretical_loss": 3.359345191828408,
      "objective/train/tokens_used": 946154976,
      "objective/train/unlikelihood_loss": 0.07344711571931839,
      "theoretical_loss": 3.359345191828408,
      "tokens_seen": 2587295744
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00021977777777777778,
      "loss": 1.5358,
      "theoretical_loss": 3.359345191828408,
      "tokens_seen": 2587295744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021975757575757576,
      "loss": 1.4078,
      "theoretical_loss": 3.359338451009661,
      "tokens_seen": 2587361280
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021973737373737375,
      "loss": 1.2837,
      "theoretical_loss": 3.3593317104094575,
      "tokens_seen": 2587426816
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021971717171717173,
      "loss": 1.4053,
      "theoretical_loss": 3.359324970027785,
      "tokens_seen": 2587492352
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002196969696969697,
      "loss": 1.5701,
      "theoretical_loss": 3.359318229864632,
      "tokens_seen": 2587557888
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021967676767676767,
      "loss": 1.2656,
      "theoretical_loss": 3.3593114899199836,
      "tokens_seen": 2587623424
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021965656565656565,
      "loss": 1.0733,
      "theoretical_loss": 3.3593047501938296,
      "tokens_seen": 2587688960
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021963636363636364,
      "loss": 1.2583,
      "theoretical_loss": 3.359298010686156,
      "tokens_seen": 2587754496
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021961616161616162,
      "loss": 1.2138,
      "theoretical_loss": 3.35929127139695,
      "tokens_seen": 2587820032
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002195959595959596,
      "loss": 1.488,
      "theoretical_loss": 3.3592845323261997,
      "tokens_seen": 2587885568
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021957575757575756,
      "loss": 1.4552,
      "theoretical_loss": 3.3592777934738924,
      "tokens_seen": 2587951104
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021955555555555557,
      "loss": 1.2135,
      "theoretical_loss": 3.359271054840015,
      "tokens_seen": 2588016640
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021953535353535353,
      "loss": 1.0854,
      "theoretical_loss": 3.3592643164245555,
      "tokens_seen": 2588082176
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021951515151515154,
      "loss": 1.5081,
      "theoretical_loss": 3.359257578227501,
      "tokens_seen": 2588147712
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002194949494949495,
      "loss": 1.1398,
      "theoretical_loss": 3.3592508402488384,
      "tokens_seen": 2588213248
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021947474747474746,
      "loss": 1.4246,
      "theoretical_loss": 3.3592441024885558,
      "tokens_seen": 2588278784
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021945454545454547,
      "loss": 1.2815,
      "theoretical_loss": 3.3592373649466407,
      "tokens_seen": 2588344320
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021943434343434342,
      "loss": 1.5321,
      "theoretical_loss": 3.35923062762308,
      "tokens_seen": 2588409856
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021941414141414143,
      "loss": 1.3999,
      "theoretical_loss": 3.3592238905178613,
      "tokens_seen": 2588475392
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002193939393939394,
      "loss": 1.4635,
      "theoretical_loss": 3.359217153630972,
      "tokens_seen": 2588540928
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002193737373737374,
      "loss": 1.5719,
      "theoretical_loss": 3.3592104169623997,
      "tokens_seen": 2588606464
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021935353535353536,
      "loss": 1.5592,
      "theoretical_loss": 3.359203680512131,
      "tokens_seen": 2588672000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021933333333333332,
      "loss": 1.0947,
      "theoretical_loss": 3.3591969442801544,
      "tokens_seen": 2588737536
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021931313131313133,
      "loss": 1.5534,
      "theoretical_loss": 3.3591902082664564,
      "tokens_seen": 2588803072
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021929292929292928,
      "loss": 1.3052,
      "theoretical_loss": 3.359183472471025,
      "tokens_seen": 2588868608
    },
    {
      "epoch": 0.57,
      "objective/train/aligned_token_count": 0.24315738677978516,
      "objective/train/avg_token_score": 0.02543279156088829,
      "objective/train/docs_used": 1456190,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.7437964677810669,
      "objective/train/original_loss": 5.034529685974121,
      "objective/train/theoretical_loss": 3.359176736893847,
      "objective/train/tokens_used": 947793376,
      "objective/train/unlikelihood_loss": 0.04849932715296745,
      "theoretical_loss": 3.359176736893847,
      "tokens_seen": 2588934144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002192727272727273,
      "loss": 1.2037,
      "theoretical_loss": 3.359176736893847,
      "tokens_seen": 2588934144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021925252525252525,
      "loss": 1.3412,
      "theoretical_loss": 3.3591700015349106,
      "tokens_seen": 2588999680
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021923232323232324,
      "loss": 1.5354,
      "theoretical_loss": 3.3591632663942024,
      "tokens_seen": 2589065216
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021921212121212122,
      "loss": 1.3035,
      "theoretical_loss": 3.3591565314717107,
      "tokens_seen": 2589130752
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002191919191919192,
      "loss": 1.2688,
      "theoretical_loss": 3.3591497967674218,
      "tokens_seen": 2589196288
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002191717171717172,
      "loss": 1.48,
      "theoretical_loss": 3.3591430622813236,
      "tokens_seen": 2589261824
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021915151515151514,
      "loss": 1.4111,
      "theoretical_loss": 3.3591363280134043,
      "tokens_seen": 2589327360
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021913131313131313,
      "loss": 1.3819,
      "theoretical_loss": 3.35912959396365,
      "tokens_seen": 2589392896
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002191111111111111,
      "loss": 1.2459,
      "theoretical_loss": 3.3591228601320493,
      "tokens_seen": 2589458432
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002190909090909091,
      "loss": 1.4422,
      "theoretical_loss": 3.3591161265185887,
      "tokens_seen": 2589523968
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021907070707070708,
      "loss": 1.2243,
      "theoretical_loss": 3.359109393123256,
      "tokens_seen": 2589589504
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021905050505050506,
      "loss": 1.5655,
      "theoretical_loss": 3.3591026599460383,
      "tokens_seen": 2589655040
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021903030303030302,
      "loss": 1.3098,
      "theoretical_loss": 3.3590959269869236,
      "tokens_seen": 2589720576
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021901010101010103,
      "loss": 1.289,
      "theoretical_loss": 3.359089194245899,
      "tokens_seen": 2589786112
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000218989898989899,
      "loss": 1.3578,
      "theoretical_loss": 3.3590824617229513,
      "tokens_seen": 2589851648
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021896969696969697,
      "loss": 1.2007,
      "theoretical_loss": 3.3590757294180693,
      "tokens_seen": 2589917184
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021894949494949496,
      "loss": 1.2668,
      "theoretical_loss": 3.3590689973312395,
      "tokens_seen": 2589982720
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021892929292929291,
      "loss": 1.3851,
      "theoretical_loss": 3.359062265462449,
      "tokens_seen": 2590048256
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021890909090909092,
      "loss": 1.2519,
      "theoretical_loss": 3.359055533811686,
      "tokens_seen": 2590113792
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021888888888888888,
      "loss": 1.3249,
      "theoretical_loss": 3.359048802378938,
      "tokens_seen": 2590179328
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002188686868686869,
      "loss": 1.291,
      "theoretical_loss": 3.3590420711641915,
      "tokens_seen": 2590244864
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021884848484848485,
      "loss": 1.5196,
      "theoretical_loss": 3.3590353401674347,
      "tokens_seen": 2590310400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021882828282828283,
      "loss": 1.4226,
      "theoretical_loss": 3.3590286093886546,
      "tokens_seen": 2590375936
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021880808080808082,
      "loss": 1.225,
      "theoretical_loss": 3.359021878827839,
      "tokens_seen": 2590441472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021878787878787877,
      "loss": 1.4193,
      "theoretical_loss": 3.359015148484975,
      "tokens_seen": 2590507008
    },
    {
      "epoch": 0.57,
      "objective/train/aligned_token_count": 0.4172165095806122,
      "objective/train/avg_token_score": 0.01105334423482418,
      "objective/train/docs_used": 1457612,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3442898988723755,
      "objective/train/original_loss": 4.390530109405518,
      "objective/train/theoretical_loss": 3.35900841836005,
      "objective/train/tokens_used": 949431776,
      "objective/train/unlikelihood_loss": 0.05024366453289986,
      "theoretical_loss": 3.35900841836005,
      "tokens_seen": 2590572544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021876767676767679,
      "loss": 1.3533,
      "theoretical_loss": 3.35900841836005,
      "tokens_seen": 2590572544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021874747474747474,
      "loss": 1.5174,
      "theoretical_loss": 3.359001688453052,
      "tokens_seen": 2590638080
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021872727272727273,
      "loss": 1.2364,
      "theoretical_loss": 3.358994958763968,
      "tokens_seen": 2590703616
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002187070707070707,
      "loss": 1.34,
      "theoretical_loss": 3.3589882292927853,
      "tokens_seen": 2590769152
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002186868686868687,
      "loss": 1.2486,
      "theoretical_loss": 3.3589815000394916,
      "tokens_seen": 2590834688
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021866666666666668,
      "loss": 1.4645,
      "theoretical_loss": 3.358974771004074,
      "tokens_seen": 2590900224
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021864646464646466,
      "loss": 1.433,
      "theoretical_loss": 3.35896804218652,
      "tokens_seen": 2590965760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021862626262626262,
      "loss": 1.3359,
      "theoretical_loss": 3.3589613135868177,
      "tokens_seen": 2591031296
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002186060606060606,
      "loss": 1.359,
      "theoretical_loss": 3.3589545852049536,
      "tokens_seen": 2591096832
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002185858585858586,
      "loss": 1.2666,
      "theoretical_loss": 3.358947857040916,
      "tokens_seen": 2591162368
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021856565656565657,
      "loss": 1.3051,
      "theoretical_loss": 3.3589411290946916,
      "tokens_seen": 2591227904
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021854545454545455,
      "loss": 1.5853,
      "theoretical_loss": 3.3589344013662683,
      "tokens_seen": 2591293440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021852525252525254,
      "loss": 1.2348,
      "theoretical_loss": 3.358927673855633,
      "tokens_seen": 2591358976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021850505050505052,
      "loss": 1.2914,
      "theoretical_loss": 3.3589209465627743,
      "tokens_seen": 2591424512
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021848484848484848,
      "loss": 1.4705,
      "theoretical_loss": 3.358914219487678,
      "tokens_seen": 2591490048
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021846464646464646,
      "loss": 1.3132,
      "theoretical_loss": 3.3589074926303333,
      "tokens_seen": 2591555584
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021844444444444445,
      "loss": 1.3526,
      "theoretical_loss": 3.358900765990726,
      "tokens_seen": 2591621120
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021842424242424243,
      "loss": 1.457,
      "theoretical_loss": 3.358894039568845,
      "tokens_seen": 2591686656
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021840404040404042,
      "loss": 1.2059,
      "theoretical_loss": 3.3588873133646766,
      "tokens_seen": 2591752192
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021838383838383837,
      "loss": 1.2819,
      "theoretical_loss": 3.358880587378209,
      "tokens_seen": 2591817728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021836363636363638,
      "loss": 1.6304,
      "theoretical_loss": 3.358873861609429,
      "tokens_seen": 2591883264
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021834343434343434,
      "loss": 1.4551,
      "theoretical_loss": 3.3588671360583247,
      "tokens_seen": 2591948800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021832323232323235,
      "loss": 1.3465,
      "theoretical_loss": 3.358860410724883,
      "tokens_seen": 2592014336
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002183030303030303,
      "loss": 1.3775,
      "theoretical_loss": 3.358853685609092,
      "tokens_seen": 2592079872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021828282828282827,
      "loss": 1.2211,
      "theoretical_loss": 3.358846960710938,
      "tokens_seen": 2592145408
    },
    {
      "epoch": 0.57,
      "objective/train/aligned_token_count": 0.31634896993637085,
      "objective/train/avg_token_score": 0.017007457092404366,
      "objective/train/docs_used": 1458174,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0611191987991333,
      "objective/train/original_loss": 4.791499137878418,
      "objective/train/theoretical_loss": 3.35884023603041,
      "objective/train/tokens_used": 951070176,
      "objective/train/unlikelihood_loss": 0.06705103069543839,
      "theoretical_loss": 3.35884023603041,
      "tokens_seen": 2592210944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021826262626262628,
      "loss": 1.4101,
      "theoretical_loss": 3.35884023603041,
      "tokens_seen": 2592210944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021824242424242423,
      "loss": 1.3224,
      "theoretical_loss": 3.3588335115674943,
      "tokens_seen": 2592276480
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021822222222222224,
      "loss": 1.6009,
      "theoretical_loss": 3.358826787322179,
      "tokens_seen": 2592342016
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002182020202020202,
      "loss": 1.1774,
      "theoretical_loss": 3.358820063294451,
      "tokens_seen": 2592407552
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021818181818181818,
      "loss": 1.3524,
      "theoretical_loss": 3.3588133394842985,
      "tokens_seen": 2592473088
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021816161616161617,
      "loss": 1.6764,
      "theoretical_loss": 3.358806615891708,
      "tokens_seen": 2592538624
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021814141414141415,
      "loss": 1.1397,
      "theoretical_loss": 3.3587998925166676,
      "tokens_seen": 2592604160
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021812121212121214,
      "loss": 1.4575,
      "theoretical_loss": 3.358793169359165,
      "tokens_seen": 2592669696
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002181010101010101,
      "loss": 1.4108,
      "theoretical_loss": 3.3587864464191868,
      "tokens_seen": 2592735232
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021808080808080808,
      "loss": 1.2293,
      "theoretical_loss": 3.358779723696721,
      "tokens_seen": 2592800768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021806060606060606,
      "loss": 1.3669,
      "theoretical_loss": 3.3587730011917554,
      "tokens_seen": 2592866304
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021804040404040405,
      "loss": 1.3535,
      "theoretical_loss": 3.3587662789042767,
      "tokens_seen": 2592931840
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021802020202020203,
      "loss": 1.3117,
      "theoretical_loss": 3.358759556834273,
      "tokens_seen": 2592997376
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000218,
      "loss": 1.4087,
      "theoretical_loss": 3.3587528349817313,
      "tokens_seen": 2593062912
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021797979797979797,
      "loss": 1.4706,
      "theoretical_loss": 3.3587461133466396,
      "tokens_seen": 2593128448
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021795959595959598,
      "loss": 1.3848,
      "theoretical_loss": 3.358739391928985,
      "tokens_seen": 2593193984
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021793939393939394,
      "loss": 1.3509,
      "theoretical_loss": 3.358732670728755,
      "tokens_seen": 2593259520
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021791919191919192,
      "loss": 1.3338,
      "theoretical_loss": 3.3587259497459367,
      "tokens_seen": 2593325056
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002178989898989899,
      "loss": 1.4972,
      "theoretical_loss": 3.358719228980519,
      "tokens_seen": 2593390592
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021787878787878786,
      "loss": 1.3487,
      "theoretical_loss": 3.3587125084324874,
      "tokens_seen": 2593456128
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021785858585858587,
      "loss": 1.3665,
      "theoretical_loss": 3.3587057881018305,
      "tokens_seen": 2593521664
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021783838383838383,
      "loss": 1.3324,
      "theoretical_loss": 3.358699067988536,
      "tokens_seen": 2593587200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021781818181818184,
      "loss": 1.541,
      "theoretical_loss": 3.3586923480925908,
      "tokens_seen": 2593652736
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002177979797979798,
      "loss": 1.3524,
      "theoretical_loss": 3.358685628413982,
      "tokens_seen": 2593718272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021777777777777776,
      "loss": 1.2963,
      "theoretical_loss": 3.3586789089526983,
      "tokens_seen": 2593783808
    },
    {
      "epoch": 0.57,
      "objective/train/aligned_token_count": 0.5134408473968506,
      "objective/train/avg_token_score": 0.031915463507175446,
      "objective/train/docs_used": 1459433,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5547882318496704,
      "objective/train/original_loss": 4.175063133239746,
      "objective/train/theoretical_loss": 3.3586721897087264,
      "objective/train/tokens_used": 952708576,
      "objective/train/unlikelihood_loss": 0.05858859792351723,
      "theoretical_loss": 3.3586721897087264,
      "tokens_seen": 2593849344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021775757575757577,
      "loss": 1.4772,
      "theoretical_loss": 3.3586721897087264,
      "tokens_seen": 2593849344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021773737373737372,
      "loss": 1.3547,
      "theoretical_loss": 3.358665470682054,
      "tokens_seen": 2593914880
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021771717171717173,
      "loss": 1.0891,
      "theoretical_loss": 3.3586587518726683,
      "tokens_seen": 2593980416
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002176969696969697,
      "loss": 1.3506,
      "theoretical_loss": 3.358652033280557,
      "tokens_seen": 2594045952
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021767676767676768,
      "loss": 1.5601,
      "theoretical_loss": 3.3586453149057074,
      "tokens_seen": 2594111488
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021765656565656566,
      "loss": 1.3283,
      "theoretical_loss": 3.3586385967481074,
      "tokens_seen": 2594177024
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021763636363636364,
      "loss": 1.4169,
      "theoretical_loss": 3.3586318788077443,
      "tokens_seen": 2594242560
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021761616161616163,
      "loss": 1.1453,
      "theoretical_loss": 3.358625161084605,
      "tokens_seen": 2594308096
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021759595959595958,
      "loss": 1.3817,
      "theoretical_loss": 3.358618443578678,
      "tokens_seen": 2594373632
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002175757575757576,
      "loss": 1.3484,
      "theoretical_loss": 3.3586117262899506,
      "tokens_seen": 2594439168
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021755555555555555,
      "loss": 1.3529,
      "theoretical_loss": 3.3586050092184094,
      "tokens_seen": 2594504704
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021753535353535354,
      "loss": 1.3904,
      "theoretical_loss": 3.358598292364043,
      "tokens_seen": 2594570240
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021751515151515152,
      "loss": 1.2999,
      "theoretical_loss": 3.3585915757268374,
      "tokens_seen": 2594635776
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002174949494949495,
      "loss": 1.3008,
      "theoretical_loss": 3.3585848593067817,
      "tokens_seen": 2594701312
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002174747474747475,
      "loss": 1.5631,
      "theoretical_loss": 3.358578143103863,
      "tokens_seen": 2594766848
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021745454545454547,
      "loss": 1.3745,
      "theoretical_loss": 3.3585714271180684,
      "tokens_seen": 2594832384
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021743434343434343,
      "loss": 1.1211,
      "theoretical_loss": 3.3585647113493855,
      "tokens_seen": 2594897920
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002174141414141414,
      "loss": 1.2901,
      "theoretical_loss": 3.358557995797802,
      "tokens_seen": 2594963456
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002173939393939394,
      "loss": 1.2097,
      "theoretical_loss": 3.358551280463305,
      "tokens_seen": 2595028992
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021737373737373738,
      "loss": 1.2824,
      "theoretical_loss": 3.3585445653458823,
      "tokens_seen": 2595094528
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021735353535353536,
      "loss": 1.3835,
      "theoretical_loss": 3.3585378504455212,
      "tokens_seen": 2595160064
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021733333333333332,
      "loss": 1.6532,
      "theoretical_loss": 3.3585311357622096,
      "tokens_seen": 2595225600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021731313131313133,
      "loss": 1.3808,
      "theoretical_loss": 3.3585244212959346,
      "tokens_seen": 2595291136
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002172929292929293,
      "loss": 1.3653,
      "theoretical_loss": 3.3585177070466843,
      "tokens_seen": 2595356672
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002172727272727273,
      "loss": 1.4747,
      "theoretical_loss": 3.3585109930144452,
      "tokens_seen": 2595422208
    },
    {
      "epoch": 0.57,
      "objective/train/aligned_token_count": 0.4124511182308197,
      "objective/train/avg_token_score": 0.040364500135183334,
      "objective/train/docs_used": 1459948,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.230784296989441,
      "objective/train/original_loss": 4.126612186431885,
      "objective/train/theoretical_loss": 3.358504279199206,
      "objective/train/tokens_used": 954346976,
      "objective/train/unlikelihood_loss": 0.06743073463439941,
      "theoretical_loss": 3.358504279199206,
      "tokens_seen": 2595487744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021725252525252526,
      "loss": 1.3638,
      "theoretical_loss": 3.358504279199206,
      "tokens_seen": 2595487744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021723232323232321,
      "loss": 1.3353,
      "theoretical_loss": 3.358497565600953,
      "tokens_seen": 2595553280
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021721212121212123,
      "loss": 1.3227,
      "theoretical_loss": 3.358490852219674,
      "tokens_seen": 2595618816
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021719191919191918,
      "loss": 1.2911,
      "theoretical_loss": 3.3584841390553577,
      "tokens_seen": 2595684352
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002171717171717172,
      "loss": 1.3121,
      "theoretical_loss": 3.3584774261079904,
      "tokens_seen": 2595749888
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021715151515151515,
      "loss": 1.5319,
      "theoretical_loss": 3.3584707133775598,
      "tokens_seen": 2595815424
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021713131313131313,
      "loss": 1.3214,
      "theoretical_loss": 3.3584640008640534,
      "tokens_seen": 2595880960
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021711111111111112,
      "loss": 1.3757,
      "theoretical_loss": 3.3584572885674593,
      "tokens_seen": 2595946496
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002170909090909091,
      "loss": 1.4901,
      "theoretical_loss": 3.358450576487764,
      "tokens_seen": 2596012032
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021707070707070709,
      "loss": 1.3803,
      "theoretical_loss": 3.358443864624956,
      "tokens_seen": 2596077568
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021705050505050504,
      "loss": 1.3382,
      "theoretical_loss": 3.3584371529790222,
      "tokens_seen": 2596143104
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021703030303030303,
      "loss": 1.3136,
      "theoretical_loss": 3.3584304415499506,
      "tokens_seen": 2596208640
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000217010101010101,
      "loss": 1.3465,
      "theoretical_loss": 3.3584237303377282,
      "tokens_seen": 2596274176
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000216989898989899,
      "loss": 1.4562,
      "theoretical_loss": 3.3584170193423426,
      "tokens_seen": 2596339712
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021696969696969698,
      "loss": 1.2964,
      "theoretical_loss": 3.3584103085637818,
      "tokens_seen": 2596405248
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021694949494949496,
      "loss": 1.4633,
      "theoretical_loss": 3.358403598002033,
      "tokens_seen": 2596470784
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021692929292929292,
      "loss": 1.5747,
      "theoretical_loss": 3.3583968876570833,
      "tokens_seen": 2596536320
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021690909090909093,
      "loss": 1.5363,
      "theoretical_loss": 3.3583901775289213,
      "tokens_seen": 2596601856
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002168888888888889,
      "loss": 1.4326,
      "theoretical_loss": 3.3583834676175335,
      "tokens_seen": 2596667392
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021686868686868687,
      "loss": 1.4947,
      "theoretical_loss": 3.358376757922908,
      "tokens_seen": 2596732928
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021684848484848486,
      "loss": 1.2305,
      "theoretical_loss": 3.358370048445032,
      "tokens_seen": 2596798464
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002168282828282828,
      "loss": 1.2866,
      "theoretical_loss": 3.3583633391838927,
      "tokens_seen": 2596864000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021680808080808082,
      "loss": 1.265,
      "theoretical_loss": 3.3583566301394785,
      "tokens_seen": 2596929536
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021678787878787878,
      "loss": 1.5307,
      "theoretical_loss": 3.3583499213117767,
      "tokens_seen": 2596995072
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002167676767676768,
      "loss": 1.2012,
      "theoretical_loss": 3.358343212700774,
      "tokens_seen": 2597060608
    },
    {
      "epoch": 0.57,
      "objective/train/aligned_token_count": 0.5315859913825989,
      "objective/train/avg_token_score": 0.007840541191399097,
      "objective/train/docs_used": 1461323,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8087257146835327,
      "objective/train/original_loss": 4.097517967224121,
      "objective/train/theoretical_loss": 3.3583365043064592,
      "objective/train/tokens_used": 955985376,
      "objective/train/unlikelihood_loss": 0.045627664774656296,
      "theoretical_loss": 3.3583365043064592,
      "tokens_seen": 2597126144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021674747474747475,
      "loss": 1.5813,
      "theoretical_loss": 3.3583365043064592,
      "tokens_seen": 2597126144
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021672727272727273,
      "loss": 1.5443,
      "theoretical_loss": 3.358329796128819,
      "tokens_seen": 2597191680
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021670707070707072,
      "loss": 1.2449,
      "theoretical_loss": 3.3583230881678414,
      "tokens_seen": 2597257216
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021668686868686867,
      "loss": 1.6791,
      "theoretical_loss": 3.358316380423513,
      "tokens_seen": 2597322752
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021666666666666668,
      "loss": 1.2313,
      "theoretical_loss": 3.358309672895823,
      "tokens_seen": 2597388288
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021664646464646464,
      "loss": 1.4496,
      "theoretical_loss": 3.358302965584757,
      "tokens_seen": 2597453824
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021662626262626265,
      "loss": 1.4316,
      "theoretical_loss": 3.358296258490304,
      "tokens_seen": 2597519360
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002166060606060606,
      "loss": 1.1555,
      "theoretical_loss": 3.358289551612451,
      "tokens_seen": 2597584896
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002165858585858586,
      "loss": 1.334,
      "theoretical_loss": 3.3582828449511855,
      "tokens_seen": 2597650432
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021656565656565658,
      "loss": 1.4231,
      "theoretical_loss": 3.358276138506495,
      "tokens_seen": 2597715968
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021654545454545453,
      "loss": 1.2779,
      "theoretical_loss": 3.3582694322783673,
      "tokens_seen": 2597781504
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021652525252525254,
      "loss": 1.353,
      "theoretical_loss": 3.3582627262667897,
      "tokens_seen": 2597847040
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002165050505050505,
      "loss": 1.1906,
      "theoretical_loss": 3.35825602047175,
      "tokens_seen": 2597912576
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021648484848484849,
      "loss": 1.2822,
      "theoretical_loss": 3.3582493148932357,
      "tokens_seen": 2597978112
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021646464646464647,
      "loss": 1.4391,
      "theoretical_loss": 3.358242609531234,
      "tokens_seen": 2598043648
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021644444444444445,
      "loss": 1.1344,
      "theoretical_loss": 3.3582359043857326,
      "tokens_seen": 2598109184
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021642424242424244,
      "loss": 1.7116,
      "theoretical_loss": 3.3582291994567193,
      "tokens_seen": 2598174720
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021640404040404042,
      "loss": 1.3158,
      "theoretical_loss": 3.3582224947441817,
      "tokens_seen": 2598240256
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021638383838383838,
      "loss": 1.6071,
      "theoretical_loss": 3.3582157902481065,
      "tokens_seen": 2598305792
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021636363636363636,
      "loss": 1.4056,
      "theoretical_loss": 3.3582090859684826,
      "tokens_seen": 2598371328
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021634343434343435,
      "loss": 1.3901,
      "theoretical_loss": 3.358202381905296,
      "tokens_seen": 2598436864
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021632323232323233,
      "loss": 1.3458,
      "theoretical_loss": 3.358195678058536,
      "tokens_seen": 2598502400
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021630303030303031,
      "loss": 1.1323,
      "theoretical_loss": 3.358188974428189,
      "tokens_seen": 2598567936
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021628282828282827,
      "loss": 1.4147,
      "theoretical_loss": 3.3581822710142424,
      "tokens_seen": 2598633472
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021626262626262628,
      "loss": 1.3904,
      "theoretical_loss": 3.3581755678166845,
      "tokens_seen": 2598699008
    },
    {
      "epoch": 0.57,
      "objective/train/aligned_token_count": 0.4615102708339691,
      "objective/train/avg_token_score": 0.021151669323444366,
      "objective/train/docs_used": 1462170,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3311375379562378,
      "objective/train/original_loss": 4.147708415985107,
      "objective/train/theoretical_loss": 3.358168864835503,
      "objective/train/tokens_used": 957623776,
      "objective/train/unlikelihood_loss": 0.05435161665081978,
      "theoretical_loss": 3.358168864835503,
      "tokens_seen": 2598764544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021624242424242424,
      "loss": 1.425,
      "theoretical_loss": 3.358168864835503,
      "tokens_seen": 2598764544
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021622222222222225,
      "loss": 1.2209,
      "theoretical_loss": 3.358162162070684,
      "tokens_seen": 2598830080
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002162020202020202,
      "loss": 1.2242,
      "theoretical_loss": 3.3581554595222167,
      "tokens_seen": 2598895616
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021618181818181816,
      "loss": 1.3914,
      "theoretical_loss": 3.3581487571900874,
      "tokens_seen": 2598961152
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021616161616161617,
      "loss": 1.5226,
      "theoretical_loss": 3.3581420550742846,
      "tokens_seen": 2599026688
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021614141414141413,
      "loss": 1.5255,
      "theoretical_loss": 3.358135353174796,
      "tokens_seen": 2599092224
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021612121212121214,
      "loss": 1.2809,
      "theoretical_loss": 3.358128651491608,
      "tokens_seen": 2599157760
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002161010101010101,
      "loss": 1.5158,
      "theoretical_loss": 3.3581219500247093,
      "tokens_seen": 2599223296
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021608080808080808,
      "loss": 1.5276,
      "theoretical_loss": 3.358115248774087,
      "tokens_seen": 2599288832
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021606060606060607,
      "loss": 1.4594,
      "theoretical_loss": 3.3581085477397283,
      "tokens_seen": 2599354368
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021604040404040402,
      "loss": 1.2933,
      "theoretical_loss": 3.3581018469216213,
      "tokens_seen": 2599419904
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021602020202020204,
      "loss": 1.3379,
      "theoretical_loss": 3.3580951463197537,
      "tokens_seen": 2599485440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000216,
      "loss": 1.3946,
      "theoretical_loss": 3.3580884459341123,
      "tokens_seen": 2599550976
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021597979797979798,
      "loss": 1.2913,
      "theoretical_loss": 3.358081745764686,
      "tokens_seen": 2599616512
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021595959595959596,
      "loss": 1.2264,
      "theoretical_loss": 3.3580750458114608,
      "tokens_seen": 2599682048
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021593939393939394,
      "loss": 1.4468,
      "theoretical_loss": 3.3580683460744254,
      "tokens_seen": 2599747584
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021591919191919193,
      "loss": 1.4513,
      "theoretical_loss": 3.3580616465535664,
      "tokens_seen": 2599813120
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002158989898989899,
      "loss": 1.6856,
      "theoretical_loss": 3.3580549472488728,
      "tokens_seen": 2599878656
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002158787878787879,
      "loss": 1.3683,
      "theoretical_loss": 3.3580482481603307,
      "tokens_seen": 2599944192
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021585858585858585,
      "loss": 1.3212,
      "theoretical_loss": 3.358041549287929,
      "tokens_seen": 2600009728
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021583838383838384,
      "loss": 1.2861,
      "theoretical_loss": 3.358034850631654,
      "tokens_seen": 2600075264
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021581818181818182,
      "loss": 1.2426,
      "theoretical_loss": 3.358028152191494,
      "tokens_seen": 2600140800
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002157979797979798,
      "loss": 1.2526,
      "theoretical_loss": 3.3580214539674365,
      "tokens_seen": 2600206336
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002157777777777778,
      "loss": 1.5176,
      "theoretical_loss": 3.358014755959469,
      "tokens_seen": 2600271872
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021575757575757577,
      "loss": 1.4039,
      "theoretical_loss": 3.358008058167579,
      "tokens_seen": 2600337408
    },
    {
      "epoch": 0.57,
      "objective/train/aligned_token_count": 0.44654202461242676,
      "objective/train/avg_token_score": 0.02163173444569111,
      "objective/train/docs_used": 1463432,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2779566049575806,
      "objective/train/original_loss": 4.358915328979492,
      "objective/train/theoretical_loss": 3.3580013605917545,
      "objective/train/tokens_used": 959262176,
      "objective/train/unlikelihood_loss": 0.05929029360413551,
      "theoretical_loss": 3.3580013605917545,
      "tokens_seen": 2600402944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021573737373737373,
      "loss": 1.2609,
      "theoretical_loss": 3.3580013605917545,
      "tokens_seen": 2600402944
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021571717171717174,
      "loss": 1.4129,
      "theoretical_loss": 3.357994663231983,
      "tokens_seen": 2600468480
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002156969696969697,
      "loss": 1.1968,
      "theoretical_loss": 3.3579879660882512,
      "tokens_seen": 2600534016
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021567676767676768,
      "loss": 1.522,
      "theoretical_loss": 3.3579812691605477,
      "tokens_seen": 2600599552
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021565656565656567,
      "loss": 1.4362,
      "theoretical_loss": 3.35797457244886,
      "tokens_seen": 2600665088
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021563636363636362,
      "loss": 1.3525,
      "theoretical_loss": 3.3579678759531753,
      "tokens_seen": 2600730624
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021561616161616163,
      "loss": 1.2217,
      "theoretical_loss": 3.357961179673481,
      "tokens_seen": 2600796160
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002155959595959596,
      "loss": 1.4272,
      "theoretical_loss": 3.357954483609765,
      "tokens_seen": 2600861696
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002155757575757576,
      "loss": 1.3614,
      "theoretical_loss": 3.3579477877620154,
      "tokens_seen": 2600927232
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021555555555555556,
      "loss": 1.6053,
      "theoretical_loss": 3.3579410921302193,
      "tokens_seen": 2600992768
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021553535353535354,
      "loss": 1.5931,
      "theoretical_loss": 3.3579343967143642,
      "tokens_seen": 2601058304
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021551515151515153,
      "loss": 1.3021,
      "theoretical_loss": 3.3579277015144378,
      "tokens_seen": 2601123840
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021549494949494948,
      "loss": 1.4421,
      "theoretical_loss": 3.3579210065304275,
      "tokens_seen": 2601189376
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002154747474747475,
      "loss": 1.5111,
      "theoretical_loss": 3.357914311762321,
      "tokens_seen": 2601254912
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021545454545454545,
      "loss": 1.5904,
      "theoretical_loss": 3.357907617210106,
      "tokens_seen": 2601320448
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021543434343434343,
      "loss": 1.3674,
      "theoretical_loss": 3.3579009228737706,
      "tokens_seen": 2601385984
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021541414141414142,
      "loss": 1.5266,
      "theoretical_loss": 3.3578942287533016,
      "tokens_seen": 2601451520
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002153939393939394,
      "loss": 1.5894,
      "theoretical_loss": 3.3578875348486865,
      "tokens_seen": 2601517056
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021537373737373739,
      "loss": 1.3366,
      "theoretical_loss": 3.357880841159914,
      "tokens_seen": 2601582592
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021535353535353537,
      "loss": 1.1848,
      "theoretical_loss": 3.3578741476869705,
      "tokens_seen": 2601648128
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021533333333333333,
      "loss": 1.6015,
      "theoretical_loss": 3.357867454429844,
      "tokens_seen": 2601713664
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002153131313131313,
      "loss": 1.2988,
      "theoretical_loss": 3.3578607613885225,
      "tokens_seen": 2601779200
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002152929292929293,
      "loss": 1.2118,
      "theoretical_loss": 3.357854068562993,
      "tokens_seen": 2601844736
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021527272727272728,
      "loss": 1.153,
      "theoretical_loss": 3.357847375953244,
      "tokens_seen": 2601910272
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021525252525252526,
      "loss": 1.4061,
      "theoretical_loss": 3.357840683559262,
      "tokens_seen": 2601975808
    },
    {
      "epoch": 0.57,
      "objective/train/aligned_token_count": 0.29154446721076965,
      "objective/train/avg_token_score": 0.047269534319639206,
      "objective/train/docs_used": 1464290,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9880658388137817,
      "objective/train/original_loss": 5.830506324768066,
      "objective/train/theoretical_loss": 3.3578339913810353,
      "objective/train/tokens_used": 960900576,
      "objective/train/unlikelihood_loss": 0.03633600473403931,
      "theoretical_loss": 3.3578339913810353,
      "tokens_seen": 2602041344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021523232323232322,
      "loss": 1.2346,
      "theoretical_loss": 3.3578339913810353,
      "tokens_seen": 2602041344
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021521212121212123,
      "loss": 1.2877,
      "theoretical_loss": 3.3578272994185516,
      "tokens_seen": 2602106880
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002151919191919192,
      "loss": 1.295,
      "theoretical_loss": 3.357820607671798,
      "tokens_seen": 2602172416
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021517171717171717,
      "loss": 1.3274,
      "theoretical_loss": 3.3578139161407625,
      "tokens_seen": 2602237952
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021515151515151516,
      "loss": 1.1818,
      "theoretical_loss": 3.3578072248254327,
      "tokens_seen": 2602303488
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002151313131313131,
      "loss": 1.4246,
      "theoretical_loss": 3.3578005337257957,
      "tokens_seen": 2602369024
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021511111111111112,
      "loss": 1.2396,
      "theoretical_loss": 3.35779384284184,
      "tokens_seen": 2602434560
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021509090909090908,
      "loss": 1.2555,
      "theoretical_loss": 3.357787152173552,
      "tokens_seen": 2602500096
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002150707070707071,
      "loss": 1.3347,
      "theoretical_loss": 3.3577804617209206,
      "tokens_seen": 2602565632
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021505050505050505,
      "loss": 1.5824,
      "theoretical_loss": 3.3577737714839326,
      "tokens_seen": 2602631168
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021503030303030306,
      "loss": 1.3836,
      "theoretical_loss": 3.357767081462576,
      "tokens_seen": 2602696704
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021501010101010102,
      "loss": 1.4631,
      "theoretical_loss": 3.357760391656839,
      "tokens_seen": 2602762240
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021498989898989897,
      "loss": 1.4506,
      "theoretical_loss": 3.3577537020667076,
      "tokens_seen": 2602827776
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021496969696969698,
      "loss": 1.4767,
      "theoretical_loss": 3.3577470126921707,
      "tokens_seen": 2602893312
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021494949494949494,
      "loss": 1.3972,
      "theoretical_loss": 3.3577403235332155,
      "tokens_seen": 2602958848
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021492929292929295,
      "loss": 1.3204,
      "theoretical_loss": 3.3577336345898297,
      "tokens_seen": 2603024384
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002149090909090909,
      "loss": 1.2921,
      "theoretical_loss": 3.3577269458620007,
      "tokens_seen": 2603089920
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002148888888888889,
      "loss": 1.247,
      "theoretical_loss": 3.357720257349717,
      "tokens_seen": 2603155456
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021486868686868688,
      "loss": 1.646,
      "theoretical_loss": 3.357713569052965,
      "tokens_seen": 2603220992
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021484848484848486,
      "loss": 1.3193,
      "theoretical_loss": 3.357706880971733,
      "tokens_seen": 2603286528
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021482828282828284,
      "loss": 1.3295,
      "theoretical_loss": 3.357700193106009,
      "tokens_seen": 2603352064
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002148080808080808,
      "loss": 1.2071,
      "theoretical_loss": 3.3576935054557797,
      "tokens_seen": 2603417600
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021478787878787879,
      "loss": 1.3611,
      "theoretical_loss": 3.3576868180210333,
      "tokens_seen": 2603483136
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021476767676767677,
      "loss": 1.2309,
      "theoretical_loss": 3.357680130801757,
      "tokens_seen": 2603548672
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021474747474747475,
      "loss": 1.1919,
      "theoretical_loss": 3.3576734437979394,
      "tokens_seen": 2603614208
    },
    {
      "epoch": 0.57,
      "objective/train/aligned_token_count": 0.4323069453239441,
      "objective/train/avg_token_score": 0.004492130596190691,
      "objective/train/docs_used": 1465580,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1928822994232178,
      "objective/train/original_loss": 3.6169724464416504,
      "objective/train/theoretical_loss": 3.357666757009567,
      "objective/train/tokens_used": 962538976,
      "objective/train/unlikelihood_loss": 0.07897266000509262,
      "theoretical_loss": 3.357666757009567,
      "tokens_seen": 2603679744
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.00021472727272727274,
      "loss": 1.3402,
      "theoretical_loss": 3.357666757009567,
      "tokens_seen": 2603679744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021470707070707072,
      "loss": 1.3743,
      "theoretical_loss": 3.357660070436628,
      "tokens_seen": 2603745280
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021468686868686868,
      "loss": 1.4964,
      "theoretical_loss": 3.3576533840791107,
      "tokens_seen": 2603810816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002146666666666667,
      "loss": 1.452,
      "theoretical_loss": 3.3576466979370014,
      "tokens_seen": 2603876352
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021464646464646465,
      "loss": 1.3126,
      "theoretical_loss": 3.3576400120102883,
      "tokens_seen": 2603941888
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021462626262626263,
      "loss": 1.2248,
      "theoretical_loss": 3.3576333262989593,
      "tokens_seen": 2604007424
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021460606060606061,
      "loss": 1.4077,
      "theoretical_loss": 3.3576266408030015,
      "tokens_seen": 2604072960
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021458585858585857,
      "loss": 1.4388,
      "theoretical_loss": 3.3576199555224036,
      "tokens_seen": 2604138496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021456565656565658,
      "loss": 1.4122,
      "theoretical_loss": 3.357613270457152,
      "tokens_seen": 2604204032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021454545454545454,
      "loss": 1.4525,
      "theoretical_loss": 3.357606585607235,
      "tokens_seen": 2604269568
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021452525252525255,
      "loss": 1.3263,
      "theoretical_loss": 3.35759990097264,
      "tokens_seen": 2604335104
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002145050505050505,
      "loss": 1.4464,
      "theoretical_loss": 3.3575932165533544,
      "tokens_seen": 2604400640
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021448484848484846,
      "loss": 1.4808,
      "theoretical_loss": 3.3575865323493668,
      "tokens_seen": 2604466176
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021446464646464647,
      "loss": 1.3716,
      "theoretical_loss": 3.357579848360664,
      "tokens_seen": 2604531712
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021444444444444443,
      "loss": 1.3457,
      "theoretical_loss": 3.357573164587234,
      "tokens_seen": 2604597248
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021442424242424244,
      "loss": 1.5119,
      "theoretical_loss": 3.3575664810290644,
      "tokens_seen": 2604662784
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002144040404040404,
      "loss": 1.3457,
      "theoretical_loss": 3.3575597976861427,
      "tokens_seen": 2604728320
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021438383838383838,
      "loss": 1.2551,
      "theoretical_loss": 3.3575531145584567,
      "tokens_seen": 2604793856
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021436363636363637,
      "loss": 1.4475,
      "theoretical_loss": 3.3575464316459938,
      "tokens_seen": 2604859392
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021434343434343435,
      "loss": 1.3966,
      "theoretical_loss": 3.357539748948742,
      "tokens_seen": 2604924928
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021432323232323234,
      "loss": 1.4623,
      "theoretical_loss": 3.357533066466689,
      "tokens_seen": 2604990464
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002143030303030303,
      "loss": 1.3354,
      "theoretical_loss": 3.357526384199822,
      "tokens_seen": 2605056000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021428282828282828,
      "loss": 1.2399,
      "theoretical_loss": 3.3575197021481293,
      "tokens_seen": 2605121536
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021426262626262626,
      "loss": 1.3517,
      "theoretical_loss": 3.357513020311598,
      "tokens_seen": 2605187072
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021424242424242424,
      "loss": 1.2032,
      "theoretical_loss": 3.357506338690216,
      "tokens_seen": 2605252608
    },
    {
      "epoch": 0.58,
      "objective/train/aligned_token_count": 0.42320382595062256,
      "objective/train/avg_token_score": 0.02012099139392376,
      "objective/train/docs_used": 1466226,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2806178331375122,
      "objective/train/original_loss": 4.641231060028076,
      "objective/train/theoretical_loss": 3.3574996572839706,
      "objective/train/tokens_used": 964177376,
      "objective/train/unlikelihood_loss": 0.05308978259563446,
      "theoretical_loss": 3.3574996572839706,
      "tokens_seen": 2605318144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021422222222222223,
      "loss": 1.3561,
      "theoretical_loss": 3.3574996572839706,
      "tokens_seen": 2605318144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002142020202020202,
      "loss": 1.3392,
      "theoretical_loss": 3.3574929760928502,
      "tokens_seen": 2605383680
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002141818181818182,
      "loss": 1.336,
      "theoretical_loss": 3.357486295116842,
      "tokens_seen": 2605449216
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021416161616161618,
      "loss": 1.35,
      "theoretical_loss": 3.3574796143559333,
      "tokens_seen": 2605514752
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021414141414141414,
      "loss": 1.3032,
      "theoretical_loss": 3.3574729338101124,
      "tokens_seen": 2605580288
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021412121212121212,
      "loss": 1.4413,
      "theoretical_loss": 3.357466253479367,
      "tokens_seen": 2605645824
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002141010101010101,
      "loss": 1.4285,
      "theoretical_loss": 3.357459573363684,
      "tokens_seen": 2605711360
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002140808080808081,
      "loss": 1.5546,
      "theoretical_loss": 3.357452893463052,
      "tokens_seen": 2605776896
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021406060606060607,
      "loss": 1.2387,
      "theoretical_loss": 3.357446213777458,
      "tokens_seen": 2605842432
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021404040404040403,
      "loss": 1.4645,
      "theoretical_loss": 3.3574395343068897,
      "tokens_seen": 2605907968
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021402020202020204,
      "loss": 1.7914,
      "theoretical_loss": 3.3574328550513353,
      "tokens_seen": 2605973504
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000214,
      "loss": 1.2983,
      "theoretical_loss": 3.357426176010782,
      "tokens_seen": 2606039040
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000213979797979798,
      "loss": 1.4041,
      "theoretical_loss": 3.3574194971852176,
      "tokens_seen": 2606104576
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021395959595959597,
      "loss": 1.4686,
      "theoretical_loss": 3.35741281857463,
      "tokens_seen": 2606170112
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021393939393939392,
      "loss": 1.3861,
      "theoretical_loss": 3.3574061401790067,
      "tokens_seen": 2606235648
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021391919191919193,
      "loss": 1.0558,
      "theoretical_loss": 3.357399461998335,
      "tokens_seen": 2606301184
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002138989898989899,
      "loss": 1.3143,
      "theoretical_loss": 3.357392784032603,
      "tokens_seen": 2606366720
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002138787878787879,
      "loss": 1.3194,
      "theoretical_loss": 3.3573861062817985,
      "tokens_seen": 2606432256
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021385858585858586,
      "loss": 1.2522,
      "theoretical_loss": 3.357379428745909,
      "tokens_seen": 2606497792
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021383838383838384,
      "loss": 1.1263,
      "theoretical_loss": 3.357372751424922,
      "tokens_seen": 2606563328
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021381818181818183,
      "loss": 1.2486,
      "theoretical_loss": 3.357366074318825,
      "tokens_seen": 2606628864
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002137979797979798,
      "loss": 1.301,
      "theoretical_loss": 3.3573593974276066,
      "tokens_seen": 2606694400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002137777777777778,
      "loss": 1.2183,
      "theoretical_loss": 3.357352720751254,
      "tokens_seen": 2606759936
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021375757575757575,
      "loss": 1.383,
      "theoretical_loss": 3.3573460442897542,
      "tokens_seen": 2606825472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021373737373737373,
      "loss": 1.5173,
      "theoretical_loss": 3.3573393680430956,
      "tokens_seen": 2606891008
    },
    {
      "epoch": 0.58,
      "objective/train/aligned_token_count": 0.3191593289375305,
      "objective/train/avg_token_score": 0.02344459481537342,
      "objective/train/docs_used": 1467591,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.037183403968811,
      "objective/train/original_loss": 5.32392692565918,
      "objective/train/theoretical_loss": 3.357332692011266,
      "objective/train/tokens_used": 965815776,
      "objective/train/unlikelihood_loss": 0.050909653306007385,
      "theoretical_loss": 3.357332692011266,
      "tokens_seen": 2606956544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021371717171717172,
      "loss": 1.478,
      "theoretical_loss": 3.357332692011266,
      "tokens_seen": 2606956544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002136969696969697,
      "loss": 1.4068,
      "theoretical_loss": 3.3573260161942526,
      "tokens_seen": 2607022080
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002136767676767677,
      "loss": 1.556,
      "theoretical_loss": 3.357319340592044,
      "tokens_seen": 2607087616
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021365656565656567,
      "loss": 1.4475,
      "theoretical_loss": 3.3573126652046263,
      "tokens_seen": 2607153152
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021363636363636363,
      "loss": 1.4153,
      "theoretical_loss": 3.3573059900319886,
      "tokens_seen": 2607218688
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021361616161616164,
      "loss": 1.4735,
      "theoretical_loss": 3.357299315074118,
      "tokens_seen": 2607284224
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002135959595959596,
      "loss": 1.5313,
      "theoretical_loss": 3.357292640331002,
      "tokens_seen": 2607349760
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021357575757575758,
      "loss": 1.4833,
      "theoretical_loss": 3.357285965802629,
      "tokens_seen": 2607415296
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021355555555555556,
      "loss": 1.5261,
      "theoretical_loss": 3.357279291488986,
      "tokens_seen": 2607480832
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021353535353535352,
      "loss": 1.472,
      "theoretical_loss": 3.3572726173900613,
      "tokens_seen": 2607546368
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021351515151515153,
      "loss": 1.205,
      "theoretical_loss": 3.3572659435058423,
      "tokens_seen": 2607611904
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002134949494949495,
      "loss": 1.3966,
      "theoretical_loss": 3.3572592698363164,
      "tokens_seen": 2607677440
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002134747474747475,
      "loss": 1.4726,
      "theoretical_loss": 3.3572525963814717,
      "tokens_seen": 2607742976
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021345454545454546,
      "loss": 1.3805,
      "theoretical_loss": 3.3572459231412957,
      "tokens_seen": 2607808512
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002134343434343434,
      "loss": 1.4421,
      "theoretical_loss": 3.357239250115776,
      "tokens_seen": 2607874048
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021341414141414142,
      "loss": 1.3233,
      "theoretical_loss": 3.3572325773049005,
      "tokens_seen": 2607939584
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021339393939393938,
      "loss": 1.4273,
      "theoretical_loss": 3.357225904708657,
      "tokens_seen": 2608005120
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002133737373737374,
      "loss": 1.3597,
      "theoretical_loss": 3.3572192323270333,
      "tokens_seen": 2608070656
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021335353535353535,
      "loss": 1.3185,
      "theoretical_loss": 3.3572125601600167,
      "tokens_seen": 2608136192
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021333333333333336,
      "loss": 1.4114,
      "theoretical_loss": 3.357205888207595,
      "tokens_seen": 2608201728
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021331313131313132,
      "loss": 1.3622,
      "theoretical_loss": 3.357199216469756,
      "tokens_seen": 2608267264
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002132929292929293,
      "loss": 1.3085,
      "theoretical_loss": 3.3571925449464874,
      "tokens_seen": 2608332800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021327272727272728,
      "loss": 1.5422,
      "theoretical_loss": 3.3571858736377767,
      "tokens_seen": 2608398336
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021325252525252524,
      "loss": 1.6023,
      "theoretical_loss": 3.357179202543612,
      "tokens_seen": 2608463872
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021323232323232325,
      "loss": 1.4142,
      "theoretical_loss": 3.3571725316639807,
      "tokens_seen": 2608529408
    },
    {
      "epoch": 0.58,
      "objective/train/aligned_token_count": 0.44855815172195435,
      "objective/train/avg_token_score": 0.007639321032911539,
      "objective/train/docs_used": 1468389,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3646167516708374,
      "objective/train/original_loss": 4.245997428894043,
      "objective/train/theoretical_loss": 3.357165860998871,
      "objective/train/tokens_used": 967454176,
      "objective/train/unlikelihood_loss": 0.060800593346357346,
      "theoretical_loss": 3.357165860998871,
      "tokens_seen": 2608594944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002132121212121212,
      "loss": 1.406,
      "theoretical_loss": 3.357165860998871,
      "tokens_seen": 2608594944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002131919191919192,
      "loss": 1.2053,
      "theoretical_loss": 3.35715919054827,
      "tokens_seen": 2608660480
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021317171717171718,
      "loss": 1.4236,
      "theoretical_loss": 3.3571525203121655,
      "tokens_seen": 2608726016
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021315151515151516,
      "loss": 1.368,
      "theoretical_loss": 3.3571458502905456,
      "tokens_seen": 2608791552
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021313131313131315,
      "loss": 1.4901,
      "theoretical_loss": 3.3571391804833977,
      "tokens_seen": 2608857088
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021311111111111113,
      "loss": 1.3443,
      "theoretical_loss": 3.3571325108907097,
      "tokens_seen": 2608922624
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021309090909090909,
      "loss": 1.4031,
      "theoretical_loss": 3.357125841512469,
      "tokens_seen": 2608988160
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021307070707070707,
      "loss": 1.4341,
      "theoretical_loss": 3.3571191723486637,
      "tokens_seen": 2609053696
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021305050505050505,
      "loss": 1.3173,
      "theoretical_loss": 3.3571125033992812,
      "tokens_seen": 2609119232
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021303030303030304,
      "loss": 1.4814,
      "theoretical_loss": 3.35710583466431,
      "tokens_seen": 2609184768
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021301010101010102,
      "loss": 1.5373,
      "theoretical_loss": 3.357099166143737,
      "tokens_seen": 2609250304
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021298989898989898,
      "loss": 1.5163,
      "theoretical_loss": 3.3570924978375496,
      "tokens_seen": 2609315840
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000212969696969697,
      "loss": 1.2187,
      "theoretical_loss": 3.357085829745736,
      "tokens_seen": 2609381376
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021294949494949495,
      "loss": 1.3565,
      "theoretical_loss": 3.3570791618682843,
      "tokens_seen": 2609446912
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021292929292929296,
      "loss": 1.5133,
      "theoretical_loss": 3.357072494205182,
      "tokens_seen": 2609512448
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021290909090909091,
      "loss": 1.2125,
      "theoretical_loss": 3.3570658267564166,
      "tokens_seen": 2609577984
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021288888888888887,
      "loss": 1.4464,
      "theoretical_loss": 3.357059159521976,
      "tokens_seen": 2609643520
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021286868686868688,
      "loss": 1.334,
      "theoretical_loss": 3.3570524925018477,
      "tokens_seen": 2609709056
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021284848484848484,
      "loss": 1.6701,
      "theoretical_loss": 3.35704582569602,
      "tokens_seen": 2609774592
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021282828282828285,
      "loss": 1.4288,
      "theoretical_loss": 3.3570391591044797,
      "tokens_seen": 2609840128
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002128080808080808,
      "loss": 1.4533,
      "theoretical_loss": 3.3570324927272157,
      "tokens_seen": 2609905664
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002127878787878788,
      "loss": 1.6025,
      "theoretical_loss": 3.3570258265642146,
      "tokens_seen": 2609971200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021276767676767678,
      "loss": 1.1925,
      "theoretical_loss": 3.3570191606154647,
      "tokens_seen": 2610036736
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021274747474747473,
      "loss": 1.4942,
      "theoretical_loss": 3.3570124948809537,
      "tokens_seen": 2610102272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021272727272727274,
      "loss": 1.4432,
      "theoretical_loss": 3.3570058293606695,
      "tokens_seen": 2610167808
    },
    {
      "epoch": 0.58,
      "objective/train/aligned_token_count": 0.46890273690223694,
      "objective/train/avg_token_score": 0.0024247411638498306,
      "objective/train/docs_used": 1468957,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4003996849060059,
      "objective/train/original_loss": 3.858574867248535,
      "objective/train/theoretical_loss": 3.3569991640545993,
      "objective/train/tokens_used": 969092576,
      "objective/train/unlikelihood_loss": 0.06697901338338852,
      "theoretical_loss": 3.3569991640545993,
      "tokens_seen": 2610233344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002127070707070707,
      "loss": 1.3888,
      "theoretical_loss": 3.3569991640545993,
      "tokens_seen": 2610233344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021268686868686868,
      "loss": 1.554,
      "theoretical_loss": 3.3569924989627316,
      "tokens_seen": 2610298880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021266666666666667,
      "loss": 1.2196,
      "theoretical_loss": 3.3569858340850534,
      "tokens_seen": 2610364416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021264646464646465,
      "loss": 1.5336,
      "theoretical_loss": 3.3569791694215527,
      "tokens_seen": 2610429952
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021262626262626264,
      "loss": 1.2576,
      "theoretical_loss": 3.3569725049722177,
      "tokens_seen": 2610495488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021260606060606062,
      "loss": 1.3587,
      "theoretical_loss": 3.3569658407370353,
      "tokens_seen": 2610561024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021258585858585858,
      "loss": 1.6584,
      "theoretical_loss": 3.356959176715994,
      "tokens_seen": 2610626560
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021256565656565656,
      "loss": 1.3464,
      "theoretical_loss": 3.356952512909081,
      "tokens_seen": 2610692096
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021254545454545454,
      "loss": 1.3749,
      "theoretical_loss": 3.3569458493162845,
      "tokens_seen": 2610757632
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021252525252525253,
      "loss": 1.2721,
      "theoretical_loss": 3.356939185937592,
      "tokens_seen": 2610823168
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002125050505050505,
      "loss": 1.3321,
      "theoretical_loss": 3.356932522772991,
      "tokens_seen": 2610888704
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002124848484848485,
      "loss": 1.518,
      "theoretical_loss": 3.3569258598224696,
      "tokens_seen": 2610954240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021246464646464648,
      "loss": 1.2915,
      "theoretical_loss": 3.3569191970860155,
      "tokens_seen": 2611019776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021244444444444444,
      "loss": 1.4117,
      "theoretical_loss": 3.356912534563617,
      "tokens_seen": 2611085312
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021242424242424245,
      "loss": 1.4362,
      "theoretical_loss": 3.3569058722552603,
      "tokens_seen": 2611150848
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002124040404040404,
      "loss": 1.3365,
      "theoretical_loss": 3.3568992101609347,
      "tokens_seen": 2611216384
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002123838383838384,
      "loss": 1.2911,
      "theoretical_loss": 3.3568925482806273,
      "tokens_seen": 2611281920
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021236363636363637,
      "loss": 1.4173,
      "theoretical_loss": 3.3568858866143256,
      "tokens_seen": 2611347456
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021234343434343433,
      "loss": 1.2093,
      "theoretical_loss": 3.356879225162018,
      "tokens_seen": 2611412992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021232323232323234,
      "loss": 1.3055,
      "theoretical_loss": 3.356872563923692,
      "tokens_seen": 2611478528
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002123030303030303,
      "loss": 1.4817,
      "theoretical_loss": 3.356865902899335,
      "tokens_seen": 2611544064
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002122828282828283,
      "loss": 1.6152,
      "theoretical_loss": 3.356859242088935,
      "tokens_seen": 2611609600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021226262626262627,
      "loss": 1.2855,
      "theoretical_loss": 3.35685258149248,
      "tokens_seen": 2611675136
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021224242424242425,
      "loss": 1.3635,
      "theoretical_loss": 3.3568459211099575,
      "tokens_seen": 2611740672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021222222222222223,
      "loss": 1.417,
      "theoretical_loss": 3.356839260941355,
      "tokens_seen": 2611806208
    },
    {
      "epoch": 0.58,
      "objective/train/aligned_token_count": 0.4670698940753937,
      "objective/train/avg_token_score": 0.006864127237349749,
      "objective/train/docs_used": 1469548,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5752390623092651,
      "objective/train/original_loss": 4.4475297927856445,
      "objective/train/theoretical_loss": 3.356832600986661,
      "objective/train/tokens_used": 970730976,
      "objective/train/unlikelihood_loss": 0.0499972365796566,
      "theoretical_loss": 3.356832600986661,
      "tokens_seen": 2611871744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002122020202020202,
      "loss": 1.3298,
      "theoretical_loss": 3.356832600986661,
      "tokens_seen": 2611871744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002121818181818182,
      "loss": 1.2641,
      "theoretical_loss": 3.356825941245863,
      "tokens_seen": 2611937280
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021216161616161616,
      "loss": 1.2784,
      "theoretical_loss": 3.3568192817189484,
      "tokens_seen": 2612002816
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021214141414141414,
      "loss": 1.4041,
      "theoretical_loss": 3.356812622405905,
      "tokens_seen": 2612068352
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021212121212121213,
      "loss": 1.3153,
      "theoretical_loss": 3.356805963306721,
      "tokens_seen": 2612133888
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002121010101010101,
      "loss": 1.433,
      "theoretical_loss": 3.3567993044213837,
      "tokens_seen": 2612199424
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002120808080808081,
      "loss": 1.4181,
      "theoretical_loss": 3.3567926457498816,
      "tokens_seen": 2612264960
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021206060606060608,
      "loss": 1.292,
      "theoretical_loss": 3.3567859872922012,
      "tokens_seen": 2612330496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021204040404040404,
      "loss": 1.3528,
      "theoretical_loss": 3.356779329048331,
      "tokens_seen": 2612396032
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021202020202020202,
      "loss": 1.1887,
      "theoretical_loss": 3.3567726710182595,
      "tokens_seen": 2612461568
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000212,
      "loss": 1.4483,
      "theoretical_loss": 3.356766013201973,
      "tokens_seen": 2612527104
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000211979797979798,
      "loss": 1.2927,
      "theoretical_loss": 3.356759355599461,
      "tokens_seen": 2612592640
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021195959595959597,
      "loss": 1.5262,
      "theoretical_loss": 3.35675269821071,
      "tokens_seen": 2612658176
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021193939393939393,
      "loss": 1.2226,
      "theoretical_loss": 3.3567460410357075,
      "tokens_seen": 2612723712
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021191919191919194,
      "loss": 1.3089,
      "theoretical_loss": 3.356739384074442,
      "tokens_seen": 2612789248
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002118989898989899,
      "loss": 1.3629,
      "theoretical_loss": 3.3567327273269014,
      "tokens_seen": 2612854784
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021187878787878788,
      "loss": 1.3925,
      "theoretical_loss": 3.356726070793073,
      "tokens_seen": 2612920320
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021185858585858586,
      "loss": 1.326,
      "theoretical_loss": 3.356719414472945,
      "tokens_seen": 2612985856
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021183838383838382,
      "loss": 1.2909,
      "theoretical_loss": 3.3567127583665046,
      "tokens_seen": 2613051392
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021181818181818183,
      "loss": 1.5706,
      "theoretical_loss": 3.3567061024737406,
      "tokens_seen": 2613116928
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002117979797979798,
      "loss": 1.2355,
      "theoretical_loss": 3.3566994467946394,
      "tokens_seen": 2613182464
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002117777777777778,
      "loss": 1.3385,
      "theoretical_loss": 3.35669279132919,
      "tokens_seen": 2613248000
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021175757575757576,
      "loss": 1.5274,
      "theoretical_loss": 3.3566861360773794,
      "tokens_seen": 2613313536
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021173737373737374,
      "loss": 1.412,
      "theoretical_loss": 3.356679481039196,
      "tokens_seen": 2613379072
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021171717171717172,
      "loss": 1.4567,
      "theoretical_loss": 3.356672826214627,
      "tokens_seen": 2613444608
    },
    {
      "epoch": 0.58,
      "objective/train/aligned_token_count": 0.3836754560470581,
      "objective/train/avg_token_score": 0.010408344678580761,
      "objective/train/docs_used": 1470931,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2197434902191162,
      "objective/train/original_loss": 4.681241989135742,
      "objective/train/theoretical_loss": 3.3566661716036608,
      "objective/train/tokens_used": 972369376,
      "objective/train/unlikelihood_loss": 0.049052443355321884,
      "theoretical_loss": 3.3566661716036608,
      "tokens_seen": 2613510144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021169696969696968,
      "loss": 1.3361,
      "theoretical_loss": 3.3566661716036608,
      "tokens_seen": 2613510144
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002116767676767677,
      "loss": 1.4014,
      "theoretical_loss": 3.3566595172062845,
      "tokens_seen": 2613575680
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021165656565656565,
      "loss": 1.419,
      "theoretical_loss": 3.3566528630224863,
      "tokens_seen": 2613641216
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021163636363636366,
      "loss": 1.4001,
      "theoretical_loss": 3.356646209052254,
      "tokens_seen": 2613706752
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021161616161616162,
      "loss": 1.3095,
      "theoretical_loss": 3.3566395552955752,
      "tokens_seen": 2613772288
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002115959595959596,
      "loss": 1.2892,
      "theoretical_loss": 3.356632901752438,
      "tokens_seen": 2613837824
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021157575757575758,
      "loss": 1.5649,
      "theoretical_loss": 3.3566262484228298,
      "tokens_seen": 2613903360
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021155555555555557,
      "loss": 1.2848,
      "theoretical_loss": 3.356619595306739,
      "tokens_seen": 2613968896
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021153535353535355,
      "loss": 1.4149,
      "theoretical_loss": 3.356612942404152,
      "tokens_seen": 2614034432
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002115151515151515,
      "loss": 1.5162,
      "theoretical_loss": 3.3566062897150584,
      "tokens_seen": 2614099968
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002114949494949495,
      "loss": 1.4206,
      "theoretical_loss": 3.3565996372394453,
      "tokens_seen": 2614165504
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021147474747474748,
      "loss": 1.4726,
      "theoretical_loss": 3.3565929849772997,
      "tokens_seen": 2614231040
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021145454545454546,
      "loss": 1.4792,
      "theoretical_loss": 3.356586332928611,
      "tokens_seen": 2614296576
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021143434343434345,
      "loss": 1.1203,
      "theoretical_loss": 3.3565796810933652,
      "tokens_seen": 2614362112
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021141414141414143,
      "loss": 1.2479,
      "theoretical_loss": 3.3565730294715515,
      "tokens_seen": 2614427648
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021139393939393939,
      "loss": 1.1926,
      "theoretical_loss": 3.356566378063157,
      "tokens_seen": 2614493184
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002113737373737374,
      "loss": 1.4652,
      "theoretical_loss": 3.35655972686817,
      "tokens_seen": 2614558720
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021135353535353535,
      "loss": 1.305,
      "theoretical_loss": 3.3565530758865774,
      "tokens_seen": 2614624256
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021133333333333334,
      "loss": 1.3315,
      "theoretical_loss": 3.356546425118368,
      "tokens_seen": 2614689792
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021131313131313132,
      "loss": 1.2284,
      "theoretical_loss": 3.356539774563529,
      "tokens_seen": 2614755328
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021129292929292928,
      "loss": 1.2934,
      "theoretical_loss": 3.3565331242220484,
      "tokens_seen": 2614820864
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002112727272727273,
      "loss": 1.2576,
      "theoretical_loss": 3.356526474093914,
      "tokens_seen": 2614886400
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021125252525252525,
      "loss": 1.3499,
      "theoretical_loss": 3.356519824179114,
      "tokens_seen": 2614951936
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021123232323232326,
      "loss": 1.4609,
      "theoretical_loss": 3.3565131744776355,
      "tokens_seen": 2615017472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021121212121212121,
      "loss": 1.271,
      "theoretical_loss": 3.3565065249894666,
      "tokens_seen": 2615083008
    },
    {
      "epoch": 0.58,
      "objective/train/aligned_token_count": 0.4337121248245239,
      "objective/train/avg_token_score": 0.009157887659966946,
      "objective/train/docs_used": 1471590,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4694011211395264,
      "objective/train/original_loss": 4.450631141662598,
      "objective/train/theoretical_loss": 3.3564998757145954,
      "objective/train/tokens_used": 974007776,
      "objective/train/unlikelihood_loss": 0.06697653979063034,
      "theoretical_loss": 3.3564998757145954,
      "tokens_seen": 2615148544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021119191919191917,
      "loss": 1.5458,
      "theoretical_loss": 3.3564998757145954,
      "tokens_seen": 2615148544
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021117171717171718,
      "loss": 1.177,
      "theoretical_loss": 3.356493226653009,
      "tokens_seen": 2615214080
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021115151515151514,
      "loss": 1.5081,
      "theoretical_loss": 3.356486577804696,
      "tokens_seen": 2615279616
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021113131313131315,
      "loss": 1.388,
      "theoretical_loss": 3.356479929169644,
      "tokens_seen": 2615345152
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002111111111111111,
      "loss": 1.6333,
      "theoretical_loss": 3.356473280747841,
      "tokens_seen": 2615410688
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002110909090909091,
      "loss": 1.3731,
      "theoretical_loss": 3.3564666325392736,
      "tokens_seen": 2615476224
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021107070707070708,
      "loss": 1.2721,
      "theoretical_loss": 3.3564599845439314,
      "tokens_seen": 2615541760
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021105050505050506,
      "loss": 1.2808,
      "theoretical_loss": 3.356453336761801,
      "tokens_seen": 2615607296
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021103030303030304,
      "loss": 1.5756,
      "theoretical_loss": 3.356446689192871,
      "tokens_seen": 2615672832
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000211010101010101,
      "loss": 1.1885,
      "theoretical_loss": 3.356440041837128,
      "tokens_seen": 2615738368
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021098989898989898,
      "loss": 1.3096,
      "theoretical_loss": 3.356433394694561,
      "tokens_seen": 2615803904
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021096969696969697,
      "loss": 1.4681,
      "theoretical_loss": 3.3564267477651577,
      "tokens_seen": 2615869440
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021094949494949495,
      "loss": 1.4892,
      "theoretical_loss": 3.356420101048905,
      "tokens_seen": 2615934976
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021092929292929294,
      "loss": 1.5244,
      "theoretical_loss": 3.3564134545457924,
      "tokens_seen": 2616000512
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021090909090909092,
      "loss": 1.2861,
      "theoretical_loss": 3.356406808255806,
      "tokens_seen": 2616066048
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021088888888888888,
      "loss": 1.647,
      "theoretical_loss": 3.356400162178935,
      "tokens_seen": 2616131584
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002108686868686869,
      "loss": 1.4455,
      "theoretical_loss": 3.356393516315166,
      "tokens_seen": 2616197120
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021084848484848484,
      "loss": 1.2418,
      "theoretical_loss": 3.3563868706644877,
      "tokens_seen": 2616262656
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021082828282828283,
      "loss": 1.5844,
      "theoretical_loss": 3.3563802252268875,
      "tokens_seen": 2616328192
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002108080808080808,
      "loss": 1.3033,
      "theoretical_loss": 3.3563735800023533,
      "tokens_seen": 2616393728
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021078787878787877,
      "loss": 1.3736,
      "theoretical_loss": 3.356366934990873,
      "tokens_seen": 2616459264
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021076767676767678,
      "loss": 1.4112,
      "theoretical_loss": 3.3563602901924345,
      "tokens_seen": 2616524800
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021074747474747474,
      "loss": 1.3785,
      "theoretical_loss": 3.3563536456070255,
      "tokens_seen": 2616590336
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021072727272727275,
      "loss": 1.4807,
      "theoretical_loss": 3.356347001234634,
      "tokens_seen": 2616655872
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002107070707070707,
      "loss": 1.3307,
      "theoretical_loss": 3.356340357075248,
      "tokens_seen": 2616721408
    },
    {
      "epoch": 0.58,
      "objective/train/aligned_token_count": 0.3742057681083679,
      "objective/train/avg_token_score": 0.013741391710937023,
      "objective/train/docs_used": 1472953,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.219882607460022,
      "objective/train/original_loss": 4.356029510498047,
      "objective/train/theoretical_loss": 3.356333713128855,
      "objective/train/tokens_used": 975646176,
      "objective/train/unlikelihood_loss": 0.06335508078336716,
      "theoretical_loss": 3.356333713128855,
      "tokens_seen": 2616786944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021068686868686872,
      "loss": 1.2523,
      "theoretical_loss": 3.356333713128855,
      "tokens_seen": 2616786944
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021066666666666667,
      "loss": 1.4224,
      "theoretical_loss": 3.3563270693954426,
      "tokens_seen": 2616852480
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021064646464646463,
      "loss": 1.4187,
      "theoretical_loss": 3.356320425874999,
      "tokens_seen": 2616918016
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021062626262626264,
      "loss": 1.2807,
      "theoretical_loss": 3.3563137825675122,
      "tokens_seen": 2616983552
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002106060606060606,
      "loss": 1.294,
      "theoretical_loss": 3.3563071394729698,
      "tokens_seen": 2617049088
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002105858585858586,
      "loss": 1.4449,
      "theoretical_loss": 3.3563004965913597,
      "tokens_seen": 2617114624
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021056565656565657,
      "loss": 1.4617,
      "theoretical_loss": 3.3562938539226694,
      "tokens_seen": 2617180160
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021054545454545455,
      "loss": 1.3928,
      "theoretical_loss": 3.3562872114668876,
      "tokens_seen": 2617245696
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021052525252525253,
      "loss": 1.3074,
      "theoretical_loss": 3.356280569224001,
      "tokens_seen": 2617311232
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021050505050505052,
      "loss": 1.3789,
      "theoretical_loss": 3.3562739271939988,
      "tokens_seen": 2617376768
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002104848484848485,
      "loss": 1.4967,
      "theoretical_loss": 3.356267285376868,
      "tokens_seen": 2617442304
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021046464646464646,
      "loss": 1.5563,
      "theoretical_loss": 3.356260643772596,
      "tokens_seen": 2617507840
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021044444444444444,
      "loss": 1.4626,
      "theoretical_loss": 3.3562540023811716,
      "tokens_seen": 2617573376
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021042424242424243,
      "loss": 1.5841,
      "theoretical_loss": 3.356247361202582,
      "tokens_seen": 2617638912
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002104040404040404,
      "loss": 1.3993,
      "theoretical_loss": 3.356240720236815,
      "tokens_seen": 2617704448
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002103838383838384,
      "loss": 1.5271,
      "theoretical_loss": 3.356234079483859,
      "tokens_seen": 2617769984
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021036363636363638,
      "loss": 1.3582,
      "theoretical_loss": 3.3562274389437023,
      "tokens_seen": 2617835520
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021034343434343434,
      "loss": 1.2791,
      "theoretical_loss": 3.3562207986163313,
      "tokens_seen": 2617901056
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021032323232323232,
      "loss": 1.2044,
      "theoretical_loss": 3.356214158501735,
      "tokens_seen": 2617966592
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002103030303030303,
      "loss": 1.4556,
      "theoretical_loss": 3.3562075185999007,
      "tokens_seen": 2618032128
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002102828282828283,
      "loss": 1.3703,
      "theoretical_loss": 3.3562008789108164,
      "tokens_seen": 2618097664
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021026262626262627,
      "loss": 1.4076,
      "theoretical_loss": 3.35619423943447,
      "tokens_seen": 2618163200
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021024242424242423,
      "loss": 1.4059,
      "theoretical_loss": 3.3561876001708493,
      "tokens_seen": 2618228736
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021022222222222224,
      "loss": 1.2588,
      "theoretical_loss": 3.3561809611199425,
      "tokens_seen": 2618294272
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002102020202020202,
      "loss": 1.413,
      "theoretical_loss": 3.356174322281737,
      "tokens_seen": 2618359808
    },
    {
      "epoch": 0.58,
      "objective/train/aligned_token_count": 0.5461266040802002,
      "objective/train/avg_token_score": 0.048678524792194366,
      "objective/train/docs_used": 1473611,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4929711818695068,
      "objective/train/original_loss": 3.8963499069213867,
      "objective/train/theoretical_loss": 3.3561676836562206,
      "objective/train/tokens_used": 977284576,
      "objective/train/unlikelihood_loss": 0.07302477210760117,
      "theoretical_loss": 3.3561676836562206,
      "tokens_seen": 2618425344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002101818181818182,
      "loss": 1.3096,
      "theoretical_loss": 3.3561676836562206,
      "tokens_seen": 2618425344
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021016161616161616,
      "loss": 1.2665,
      "theoretical_loss": 3.3561610452433817,
      "tokens_seen": 2618490880
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021014141414141412,
      "loss": 1.2633,
      "theoretical_loss": 3.3561544070432077,
      "tokens_seen": 2618556416
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021012121212121213,
      "loss": 1.4666,
      "theoretical_loss": 3.3561477690556867,
      "tokens_seen": 2618621952
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002101010101010101,
      "loss": 1.5294,
      "theoretical_loss": 3.3561411312808067,
      "tokens_seen": 2618687488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002100808080808081,
      "loss": 1.3086,
      "theoretical_loss": 3.356134493718555,
      "tokens_seen": 2618753024
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021006060606060606,
      "loss": 1.4357,
      "theoretical_loss": 3.3561278563689196,
      "tokens_seen": 2618818560
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021004040404040404,
      "loss": 1.3226,
      "theoretical_loss": 3.356121219231889,
      "tokens_seen": 2618884096
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021002020202020202,
      "loss": 1.4922,
      "theoretical_loss": 3.3561145823074505,
      "tokens_seen": 2618949632
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00021,
      "loss": 1.5013,
      "theoretical_loss": 3.3561079455955922,
      "tokens_seen": 2619015168
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.000209979797979798,
      "loss": 1.3194,
      "theoretical_loss": 3.356101309096302,
      "tokens_seen": 2619080704
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020995959595959595,
      "loss": 1.393,
      "theoretical_loss": 3.3560946728095677,
      "tokens_seen": 2619146240
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020993939393939393,
      "loss": 1.3305,
      "theoretical_loss": 3.356088036735377,
      "tokens_seen": 2619211776
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020991919191919192,
      "loss": 1.5604,
      "theoretical_loss": 3.356081400873718,
      "tokens_seen": 2619277312
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002098989898989899,
      "loss": 1.5131,
      "theoretical_loss": 3.3560747652245784,
      "tokens_seen": 2619342848
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020987878787878789,
      "loss": 1.3284,
      "theoretical_loss": 3.356068129787946,
      "tokens_seen": 2619408384
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020985858585858587,
      "loss": 1.6372,
      "theoretical_loss": 3.3560614945638094,
      "tokens_seen": 2619473920
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020983838383838385,
      "loss": 1.3302,
      "theoretical_loss": 3.3560548595521555,
      "tokens_seen": 2619539456
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020981818181818184,
      "loss": 1.3934,
      "theoretical_loss": 3.356048224752973,
      "tokens_seen": 2619604992
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002097979797979798,
      "loss": 1.3698,
      "theoretical_loss": 3.356041590166249,
      "tokens_seen": 2619670528
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020977777777777778,
      "loss": 1.3306,
      "theoretical_loss": 3.356034955791972,
      "tokens_seen": 2619736064
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020975757575757576,
      "loss": 1.4595,
      "theoretical_loss": 3.3560283216301294,
      "tokens_seen": 2619801600
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020973737373737375,
      "loss": 1.3088,
      "theoretical_loss": 3.35602168768071,
      "tokens_seen": 2619867136
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00020971717171717173,
      "loss": 1.4508,
      "theoretical_loss": 3.3560150539437004,
      "tokens_seen": 2619932672
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002096969696969697,
      "loss": 1.5273,
      "theoretical_loss": 3.356008420419089,
      "tokens_seen": 2619998208
    },
    {
      "epoch": 0.58,
      "objective/train/aligned_token_count": 0.41849949955940247,
      "objective/train/avg_token_score": 0.009318639524281025,
      "objective/train/docs_used": 1474984,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3127321004867554,
      "objective/train/original_loss": 4.433729648590088,
      "objective/train/theoretical_loss": 3.356001787106864,
      "objective/train/tokens_used": 978922976,
      "objective/train/unlikelihood_loss": 0.0523061603307724,
      "theoretical_loss": 3.356001787106864,
      "tokens_seen": 2620063744
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002096767676767677,
      "loss": 1.5977,
      "theoretical_loss": 3.356001787106864,
      "tokens_seen": 2620063744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020965656565656565,
      "loss": 1.3472,
      "theoretical_loss": 3.3559951540070134,
      "tokens_seen": 2620129280
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020963636363636367,
      "loss": 1.3894,
      "theoretical_loss": 3.3559885211195244,
      "tokens_seen": 2620194816
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020961616161616162,
      "loss": 1.3711,
      "theoretical_loss": 3.3559818884443855,
      "tokens_seen": 2620260352
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020959595959595958,
      "loss": 1.2345,
      "theoretical_loss": 3.3559752559815843,
      "tokens_seen": 2620325888
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002095757575757576,
      "loss": 1.4332,
      "theoretical_loss": 3.355968623731109,
      "tokens_seen": 2620391424
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020955555555555555,
      "loss": 1.4709,
      "theoretical_loss": 3.3559619916929466,
      "tokens_seen": 2620456960
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020953535353535356,
      "loss": 1.4308,
      "theoretical_loss": 3.355955359867086,
      "tokens_seen": 2620522496
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020951515151515152,
      "loss": 1.3231,
      "theoretical_loss": 3.355948728253515,
      "tokens_seen": 2620588032
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002094949494949495,
      "loss": 1.42,
      "theoretical_loss": 3.355942096852221,
      "tokens_seen": 2620653568
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020947474747474748,
      "loss": 1.4275,
      "theoretical_loss": 3.355935465663192,
      "tokens_seen": 2620719104
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020945454545454544,
      "loss": 1.1465,
      "theoretical_loss": 3.355928834686416,
      "tokens_seen": 2620784640
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020943434343434345,
      "loss": 1.3998,
      "theoretical_loss": 3.3559222039218812,
      "tokens_seen": 2620850176
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002094141414141414,
      "loss": 1.3056,
      "theoretical_loss": 3.355915573369575,
      "tokens_seen": 2620915712
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002093939393939394,
      "loss": 1.3506,
      "theoretical_loss": 3.3559089430294855,
      "tokens_seen": 2620981248
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020937373737373738,
      "loss": 1.3148,
      "theoretical_loss": 3.3559023129016006,
      "tokens_seen": 2621046784
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020935353535353536,
      "loss": 1.6002,
      "theoretical_loss": 3.355895682985908,
      "tokens_seen": 2621112320
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020933333333333334,
      "loss": 1.4996,
      "theoretical_loss": 3.3558890532823966,
      "tokens_seen": 2621177856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020931313131313133,
      "loss": 1.4337,
      "theoretical_loss": 3.355882423791053,
      "tokens_seen": 2621243392
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020929292929292928,
      "loss": 1.3101,
      "theoretical_loss": 3.355875794511866,
      "tokens_seen": 2621308928
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020927272727272727,
      "loss": 1.5402,
      "theoretical_loss": 3.3558691654448225,
      "tokens_seen": 2621374464
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020925252525252525,
      "loss": 1.154,
      "theoretical_loss": 3.3558625365899117,
      "tokens_seen": 2621440000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020923232323232324,
      "loss": 1.379,
      "theoretical_loss": 3.3558559079471206,
      "tokens_seen": 2621505536
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020921212121212122,
      "loss": 1.4294,
      "theoretical_loss": 3.355849279516437,
      "tokens_seen": 2621571072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020919191919191918,
      "loss": 1.5242,
      "theoretical_loss": 3.35584265129785,
      "tokens_seen": 2621636608
    },
    {
      "epoch": 0.59,
      "objective/train/aligned_token_count": 0.3434750735759735,
      "objective/train/avg_token_score": 0.018536707386374474,
      "objective/train/docs_used": 1476237,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1067285537719727,
      "objective/train/original_loss": 4.488415718078613,
      "objective/train/theoretical_loss": 3.355836023291346,
      "objective/train/tokens_used": 980561376,
      "objective/train/unlikelihood_loss": 0.05490947142243385,
      "theoretical_loss": 3.355836023291346,
      "tokens_seen": 2621702144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002091717171717172,
      "loss": 1.2897,
      "theoretical_loss": 3.355836023291346,
      "tokens_seen": 2621702144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020915151515151515,
      "loss": 1.6053,
      "theoretical_loss": 3.355829395496914,
      "tokens_seen": 2621767680
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020913131313131316,
      "loss": 1.3678,
      "theoretical_loss": 3.355822767914541,
      "tokens_seen": 2621833216
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002091111111111111,
      "loss": 1.3428,
      "theoretical_loss": 3.355816140544216,
      "tokens_seen": 2621898752
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020909090909090907,
      "loss": 1.5762,
      "theoretical_loss": 3.3558095133859265,
      "tokens_seen": 2621964288
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020907070707070708,
      "loss": 1.5371,
      "theoretical_loss": 3.3558028864396596,
      "tokens_seen": 2622029824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020905050505050504,
      "loss": 1.4287,
      "theoretical_loss": 3.3557962597054045,
      "tokens_seen": 2622095360
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020903030303030305,
      "loss": 1.6138,
      "theoretical_loss": 3.3557896331831483,
      "tokens_seen": 2622160896
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000209010101010101,
      "loss": 1.3133,
      "theoretical_loss": 3.355783006872879,
      "tokens_seen": 2622226432
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020898989898989902,
      "loss": 1.4076,
      "theoretical_loss": 3.3557763807745844,
      "tokens_seen": 2622291968
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020896969696969697,
      "loss": 1.6379,
      "theoretical_loss": 3.355769754888253,
      "tokens_seen": 2622357504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020894949494949496,
      "loss": 1.5123,
      "theoretical_loss": 3.3557631292138725,
      "tokens_seen": 2622423040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020892929292929294,
      "loss": 1.5429,
      "theoretical_loss": 3.3557565037514303,
      "tokens_seen": 2622488576
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002089090909090909,
      "loss": 1.3943,
      "theoretical_loss": 3.355749878500915,
      "tokens_seen": 2622554112
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002088888888888889,
      "loss": 1.4153,
      "theoretical_loss": 3.3557432534623146,
      "tokens_seen": 2622619648
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020886868686868687,
      "loss": 1.6052,
      "theoretical_loss": 3.3557366286356163,
      "tokens_seen": 2622685184
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020884848484848485,
      "loss": 1.5151,
      "theoretical_loss": 3.355730004020808,
      "tokens_seen": 2622750720
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020882828282828283,
      "loss": 1.4433,
      "theoretical_loss": 3.3557233796178787,
      "tokens_seen": 2622816256
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020880808080808082,
      "loss": 1.4791,
      "theoretical_loss": 3.3557167554268155,
      "tokens_seen": 2622881792
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002087878787878788,
      "loss": 1.4445,
      "theoretical_loss": 3.3557101314476068,
      "tokens_seen": 2622947328
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020876767676767679,
      "loss": 1.5943,
      "theoretical_loss": 3.3557035076802397,
      "tokens_seen": 2623012864
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020874747474747474,
      "loss": 1.2724,
      "theoretical_loss": 3.355696884124703,
      "tokens_seen": 2623078400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020872727272727273,
      "loss": 1.3542,
      "theoretical_loss": 3.355690260780984,
      "tokens_seen": 2623143936
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002087070707070707,
      "loss": 1.1665,
      "theoretical_loss": 3.3556836376490713,
      "tokens_seen": 2623209472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002086868686868687,
      "loss": 1.5336,
      "theoretical_loss": 3.3556770147289523,
      "tokens_seen": 2623275008
    },
    {
      "epoch": 0.59,
      "objective/train/aligned_token_count": 0.35233381390571594,
      "objective/train/avg_token_score": 0.024509618058800697,
      "objective/train/docs_used": 1476787,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.083989143371582,
      "objective/train/original_loss": 5.154284477233887,
      "objective/train/theoretical_loss": 3.355670392020615,
      "objective/train/tokens_used": 982199776,
      "objective/train/unlikelihood_loss": 0.027292264625430107,
      "theoretical_loss": 3.355670392020615,
      "tokens_seen": 2623340544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020866666666666668,
      "loss": 1.3337,
      "theoretical_loss": 3.355670392020615,
      "tokens_seen": 2623340544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020864646464646464,
      "loss": 1.2625,
      "theoretical_loss": 3.3556637695240474,
      "tokens_seen": 2623406080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020862626262626265,
      "loss": 1.2077,
      "theoretical_loss": 3.3556571472392376,
      "tokens_seen": 2623471616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002086060606060606,
      "loss": 1.5105,
      "theoretical_loss": 3.3556505251661735,
      "tokens_seen": 2623537152
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002085858585858586,
      "loss": 1.2898,
      "theoretical_loss": 3.355643903304843,
      "tokens_seen": 2623602688
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020856565656565657,
      "loss": 1.45,
      "theoretical_loss": 3.355637281655234,
      "tokens_seen": 2623668224
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020854545454545453,
      "loss": 1.7133,
      "theoretical_loss": 3.3556306602173342,
      "tokens_seen": 2623733760
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020852525252525254,
      "loss": 1.3414,
      "theoretical_loss": 3.3556240389911323,
      "tokens_seen": 2623799296
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002085050505050505,
      "loss": 1.4692,
      "theoretical_loss": 3.355617417976615,
      "tokens_seen": 2623864832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002084848484848485,
      "loss": 1.515,
      "theoretical_loss": 3.355610797173772,
      "tokens_seen": 2623930368
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020846464646464646,
      "loss": 1.3807,
      "theoretical_loss": 3.355604176582589,
      "tokens_seen": 2623995904
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020844444444444445,
      "loss": 1.2578,
      "theoretical_loss": 3.355597556203056,
      "tokens_seen": 2624061440
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020842424242424243,
      "loss": 1.5305,
      "theoretical_loss": 3.35559093603516,
      "tokens_seen": 2624126976
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002084040404040404,
      "loss": 1.68,
      "theoretical_loss": 3.3555843160788887,
      "tokens_seen": 2624192512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002083838383838384,
      "loss": 1.3939,
      "theoretical_loss": 3.355577696334231,
      "tokens_seen": 2624258048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020836363636363636,
      "loss": 1.4967,
      "theoretical_loss": 3.3555710768011737,
      "tokens_seen": 2624323584
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020834343434343434,
      "loss": 1.3158,
      "theoretical_loss": 3.355564457479706,
      "tokens_seen": 2624389120
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020832323232323233,
      "loss": 1.4859,
      "theoretical_loss": 3.3555578383698146,
      "tokens_seen": 2624454656
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002083030303030303,
      "loss": 1.6239,
      "theoretical_loss": 3.3555512194714883,
      "tokens_seen": 2624520192
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002082828282828283,
      "loss": 1.3747,
      "theoretical_loss": 3.3555446007847145,
      "tokens_seen": 2624585728
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020826262626262628,
      "loss": 1.2913,
      "theoretical_loss": 3.3555379823094817,
      "tokens_seen": 2624651264
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020824242424242423,
      "loss": 1.5246,
      "theoretical_loss": 3.3555313640457776,
      "tokens_seen": 2624716800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020822222222222222,
      "loss": 1.6098,
      "theoretical_loss": 3.35552474599359,
      "tokens_seen": 2624782336
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002082020202020202,
      "loss": 1.4868,
      "theoretical_loss": 3.355518128152907,
      "tokens_seen": 2624847872
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020818181818181819,
      "loss": 1.4954,
      "theoretical_loss": 3.3555115105237165,
      "tokens_seen": 2624913408
    },
    {
      "epoch": 0.59,
      "objective/train/aligned_token_count": 0.47910556197166443,
      "objective/train/avg_token_score": 0.023119477555155754,
      "objective/train/docs_used": 1477458,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3300808668136597,
      "objective/train/original_loss": 3.7824625968933105,
      "objective/train/theoretical_loss": 3.3555048931060067,
      "objective/train/tokens_used": 983838176,
      "objective/train/unlikelihood_loss": 0.05748198553919792,
      "theoretical_loss": 3.3555048931060067,
      "tokens_seen": 2624978944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020816161616161617,
      "loss": 1.3083,
      "theoretical_loss": 3.3555048931060067,
      "tokens_seen": 2624978944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020814141414141415,
      "loss": 1.3562,
      "theoretical_loss": 3.3554982758997656,
      "tokens_seen": 2625044480
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020812121212121214,
      "loss": 1.2926,
      "theoretical_loss": 3.355491658904981,
      "tokens_seen": 2625110016
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002081010101010101,
      "loss": 1.3172,
      "theoretical_loss": 3.3554850421216402,
      "tokens_seen": 2625175552
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002080808080808081,
      "loss": 1.6026,
      "theoretical_loss": 3.3554784255497325,
      "tokens_seen": 2625241088
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020806060606060606,
      "loss": 1.3478,
      "theoretical_loss": 3.355471809189245,
      "tokens_seen": 2625306624
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020804040404040405,
      "loss": 1.3259,
      "theoretical_loss": 3.3554651930401658,
      "tokens_seen": 2625372160
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020802020202020203,
      "loss": 1.439,
      "theoretical_loss": 3.3554585771024827,
      "tokens_seen": 2625437696
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000208,
      "loss": 1.3429,
      "theoretical_loss": 3.3554519613761835,
      "tokens_seen": 2625503232
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000207979797979798,
      "loss": 1.4766,
      "theoretical_loss": 3.3554453458612574,
      "tokens_seen": 2625568768
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020795959595959596,
      "loss": 1.419,
      "theoretical_loss": 3.355438730557691,
      "tokens_seen": 2625634304
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020793939393939397,
      "loss": 1.3581,
      "theoretical_loss": 3.355432115465473,
      "tokens_seen": 2625699840
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020791919191919192,
      "loss": 1.4663,
      "theoretical_loss": 3.355425500584591,
      "tokens_seen": 2625765376
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020789898989898988,
      "loss": 1.1751,
      "theoretical_loss": 3.355418885915033,
      "tokens_seen": 2625830912
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002078787878787879,
      "loss": 1.4529,
      "theoretical_loss": 3.355412271456787,
      "tokens_seen": 2625896448
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020785858585858585,
      "loss": 1.4833,
      "theoretical_loss": 3.3554056572098414,
      "tokens_seen": 2625961984
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020783838383838386,
      "loss": 1.3371,
      "theoretical_loss": 3.355399043174184,
      "tokens_seen": 2626027520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020781818181818182,
      "loss": 1.3403,
      "theoretical_loss": 3.355392429349802,
      "tokens_seen": 2626093056
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002077979797979798,
      "loss": 1.181,
      "theoretical_loss": 3.3553858157366845,
      "tokens_seen": 2626158592
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020777777777777778,
      "loss": 1.3892,
      "theoretical_loss": 3.3553792023348192,
      "tokens_seen": 2626224128
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020775757575757577,
      "loss": 1.1858,
      "theoretical_loss": 3.3553725891441935,
      "tokens_seen": 2626289664
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020773737373737375,
      "loss": 1.4276,
      "theoretical_loss": 3.3553659761647956,
      "tokens_seen": 2626355200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002077171717171717,
      "loss": 1.3228,
      "theoretical_loss": 3.355359363396614,
      "tokens_seen": 2626420736
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002076969696969697,
      "loss": 1.5115,
      "theoretical_loss": 3.355352750839636,
      "tokens_seen": 2626486272
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020767676767676768,
      "loss": 1.4417,
      "theoretical_loss": 3.35534613849385,
      "tokens_seen": 2626551808
    },
    {
      "epoch": 0.59,
      "objective/train/aligned_token_count": 0.32300829887390137,
      "objective/train/avg_token_score": 0.014870766550302505,
      "objective/train/docs_used": 1478565,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8374523520469666,
      "objective/train/original_loss": 4.391622066497803,
      "objective/train/theoretical_loss": 3.355339526359244,
      "objective/train/tokens_used": 985476576,
      "objective/train/unlikelihood_loss": 0.08692853152751923,
      "theoretical_loss": 3.355339526359244,
      "tokens_seen": 2626617344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020765656565656566,
      "loss": 1.1759,
      "theoretical_loss": 3.355339526359244,
      "tokens_seen": 2626617344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020763636363636364,
      "loss": 1.5261,
      "theoretical_loss": 3.355332914435806,
      "tokens_seen": 2626682880
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020761616161616163,
      "loss": 1.5164,
      "theoretical_loss": 3.3553263027235234,
      "tokens_seen": 2626748416
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020759595959595959,
      "loss": 1.4282,
      "theoretical_loss": 3.355319691222385,
      "tokens_seen": 2626813952
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002075757575757576,
      "loss": 1.3933,
      "theoretical_loss": 3.3553130799323783,
      "tokens_seen": 2626879488
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020755555555555555,
      "loss": 1.4345,
      "theoretical_loss": 3.3553064688534913,
      "tokens_seen": 2626945024
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020753535353535354,
      "loss": 1.1727,
      "theoretical_loss": 3.3552998579857123,
      "tokens_seen": 2627010560
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020751515151515152,
      "loss": 1.3783,
      "theoretical_loss": 3.3552932473290293,
      "tokens_seen": 2627076096
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020749494949494948,
      "loss": 1.3202,
      "theoretical_loss": 3.35528663688343,
      "tokens_seen": 2627141632
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002074747474747475,
      "loss": 1.2756,
      "theoretical_loss": 3.3552800266489022,
      "tokens_seen": 2627207168
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020745454545454545,
      "loss": 1.277,
      "theoretical_loss": 3.3552734166254345,
      "tokens_seen": 2627272704
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020743434343434346,
      "loss": 1.5101,
      "theoretical_loss": 3.3552668068130145,
      "tokens_seen": 2627338240
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020741414141414141,
      "loss": 1.4179,
      "theoretical_loss": 3.35526019721163,
      "tokens_seen": 2627403776
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002073939393939394,
      "loss": 1.5823,
      "theoretical_loss": 3.3552535878212697,
      "tokens_seen": 2627469312
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020737373737373738,
      "loss": 1.0657,
      "theoretical_loss": 3.355246978641921,
      "tokens_seen": 2627534848
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020735353535353534,
      "loss": 1.3403,
      "theoretical_loss": 3.3552403696735724,
      "tokens_seen": 2627600384
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020733333333333335,
      "loss": 1.3816,
      "theoretical_loss": 3.355233760916211,
      "tokens_seen": 2627665920
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002073131313131313,
      "loss": 1.3156,
      "theoretical_loss": 3.355227152369826,
      "tokens_seen": 2627731456
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020729292929292932,
      "loss": 1.53,
      "theoretical_loss": 3.3552205440344043,
      "tokens_seen": 2627796992
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020727272727272727,
      "loss": 1.4294,
      "theoretical_loss": 3.355213935909935,
      "tokens_seen": 2627862528
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020725252525252526,
      "loss": 1.3041,
      "theoretical_loss": 3.355207327996405,
      "tokens_seen": 2627928064
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020723232323232324,
      "loss": 1.3808,
      "theoretical_loss": 3.355200720293803,
      "tokens_seen": 2627993600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020721212121212123,
      "loss": 1.3591,
      "theoretical_loss": 3.3551941128021165,
      "tokens_seen": 2628059136
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002071919191919192,
      "loss": 1.4964,
      "theoretical_loss": 3.355187505521334,
      "tokens_seen": 2628124672
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020717171717171717,
      "loss": 1.2441,
      "theoretical_loss": 3.3551808984514437,
      "tokens_seen": 2628190208
    },
    {
      "epoch": 0.59,
      "objective/train/aligned_token_count": 0.5675708651542664,
      "objective/train/avg_token_score": 0.004456192720681429,
      "objective/train/docs_used": 1479263,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7461999654769897,
      "objective/train/original_loss": 3.7948925495147705,
      "objective/train/theoretical_loss": 3.3551742915924327,
      "objective/train/tokens_used": 987114976,
      "objective/train/unlikelihood_loss": 0.0580318346619606,
      "theoretical_loss": 3.3551742915924327,
      "tokens_seen": 2628255744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020715151515151515,
      "loss": 1.5404,
      "theoretical_loss": 3.3551742915924327,
      "tokens_seen": 2628255744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020713131313131313,
      "loss": 1.2431,
      "theoretical_loss": 3.35516768494429,
      "tokens_seen": 2628321280
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020711111111111112,
      "loss": 1.463,
      "theoretical_loss": 3.3551610785070034,
      "tokens_seen": 2628386816
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002070909090909091,
      "loss": 1.6344,
      "theoretical_loss": 3.35515447228056,
      "tokens_seen": 2628452352
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002070707070707071,
      "loss": 1.3362,
      "theoretical_loss": 3.3551478662649488,
      "tokens_seen": 2628517888
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020705050505050504,
      "loss": 1.4962,
      "theoretical_loss": 3.3551412604601576,
      "tokens_seen": 2628583424
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020703030303030303,
      "loss": 1.3719,
      "theoretical_loss": 3.3551346548661742,
      "tokens_seen": 2628648960
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000207010101010101,
      "loss": 1.4754,
      "theoretical_loss": 3.3551280494829867,
      "tokens_seen": 2628714496
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000206989898989899,
      "loss": 1.2943,
      "theoretical_loss": 3.3551214443105835,
      "tokens_seen": 2628780032
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020696969696969698,
      "loss": 1.3453,
      "theoretical_loss": 3.3551148393489516,
      "tokens_seen": 2628845568
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020694949494949494,
      "loss": 1.4169,
      "theoretical_loss": 3.35510823459808,
      "tokens_seen": 2628911104
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020692929292929295,
      "loss": 1.2529,
      "theoretical_loss": 3.355101630057957,
      "tokens_seen": 2628976640
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002069090909090909,
      "loss": 1.4166,
      "theoretical_loss": 3.3550950257285694,
      "tokens_seen": 2629042176
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020688888888888892,
      "loss": 1.308,
      "theoretical_loss": 3.355088421609906,
      "tokens_seen": 2629107712
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020686868686868687,
      "loss": 1.5231,
      "theoretical_loss": 3.355081817701955,
      "tokens_seen": 2629173248
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020684848484848483,
      "loss": 1.3503,
      "theoretical_loss": 3.355075214004704,
      "tokens_seen": 2629238784
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020682828282828284,
      "loss": 1.5278,
      "theoretical_loss": 3.3550686105181406,
      "tokens_seen": 2629304320
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002068080808080808,
      "loss": 1.3295,
      "theoretical_loss": 3.355062007242254,
      "tokens_seen": 2629369856
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002067878787878788,
      "loss": 1.398,
      "theoretical_loss": 3.3550554041770315,
      "tokens_seen": 2629435392
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020676767676767676,
      "loss": 1.3577,
      "theoretical_loss": 3.355048801322461,
      "tokens_seen": 2629500928
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020674747474747475,
      "loss": 1.3705,
      "theoretical_loss": 3.3550421986785306,
      "tokens_seen": 2629566464
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020672727272727273,
      "loss": 1.3748,
      "theoretical_loss": 3.355035596245229,
      "tokens_seen": 2629632000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020670707070707072,
      "loss": 1.2951,
      "theoretical_loss": 3.3550289940225433,
      "tokens_seen": 2629697536
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002066868686868687,
      "loss": 1.2825,
      "theoretical_loss": 3.3550223920104623,
      "tokens_seen": 2629763072
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020666666666666666,
      "loss": 1.2308,
      "theoretical_loss": 3.3550157902089737,
      "tokens_seen": 2629828608
    },
    {
      "epoch": 0.59,
      "objective/train/aligned_token_count": 0.4327957034111023,
      "objective/train/avg_token_score": 0.019291020929813385,
      "objective/train/docs_used": 1480679,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1529117822647095,
      "objective/train/original_loss": 4.141412734985352,
      "objective/train/theoretical_loss": 3.355009188618065,
      "objective/train/tokens_used": 988753376,
      "objective/train/unlikelihood_loss": 0.054442428052425385,
      "theoretical_loss": 3.355009188618065,
      "tokens_seen": 2629894144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020664646464646464,
      "loss": 1.2946,
      "theoretical_loss": 3.355009188618065,
      "tokens_seen": 2629894144
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020662626262626263,
      "loss": 1.5198,
      "theoretical_loss": 3.3550025872377254,
      "tokens_seen": 2629959680
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002066060606060606,
      "loss": 1.5931,
      "theoretical_loss": 3.3549959860679417,
      "tokens_seen": 2630025216
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002065858585858586,
      "loss": 1.4359,
      "theoretical_loss": 3.354989385108703,
      "tokens_seen": 2630090752
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020656565656565658,
      "loss": 1.5714,
      "theoretical_loss": 3.3549827843599966,
      "tokens_seen": 2630156288
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020654545454545453,
      "loss": 1.1817,
      "theoretical_loss": 3.354976183821811,
      "tokens_seen": 2630221824
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020652525252525255,
      "loss": 1.2805,
      "theoretical_loss": 3.354969583494134,
      "tokens_seen": 2630287360
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002065050505050505,
      "loss": 1.264,
      "theoretical_loss": 3.3549629833769536,
      "tokens_seen": 2630352896
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020648484848484849,
      "loss": 1.4236,
      "theoretical_loss": 3.354956383470258,
      "tokens_seen": 2630418432
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020646464646464647,
      "loss": 1.3692,
      "theoretical_loss": 3.354949783774035,
      "tokens_seen": 2630483968
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020644444444444445,
      "loss": 1.513,
      "theoretical_loss": 3.354943184288273,
      "tokens_seen": 2630549504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020642424242424244,
      "loss": 1.4442,
      "theoretical_loss": 3.3549365850129598,
      "tokens_seen": 2630615040
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002064040404040404,
      "loss": 1.1017,
      "theoretical_loss": 3.3549299859480834,
      "tokens_seen": 2630680576
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002063838383838384,
      "loss": 1.2537,
      "theoretical_loss": 3.354923387093632,
      "tokens_seen": 2630746112
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020636363636363636,
      "loss": 1.2753,
      "theoretical_loss": 3.354916788449594,
      "tokens_seen": 2630811648
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020634343434343437,
      "loss": 1.4803,
      "theoretical_loss": 3.354910190015957,
      "tokens_seen": 2630877184
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020632323232323233,
      "loss": 1.3297,
      "theoretical_loss": 3.354903591792709,
      "tokens_seen": 2630942720
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002063030303030303,
      "loss": 1.4878,
      "theoretical_loss": 3.3548969937798376,
      "tokens_seen": 2631008256
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002062828282828283,
      "loss": 1.2412,
      "theoretical_loss": 3.354890395977332,
      "tokens_seen": 2631073792
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020626262626262626,
      "loss": 1.2563,
      "theoretical_loss": 3.3548837983851794,
      "tokens_seen": 2631139328
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020624242424242427,
      "loss": 1.5206,
      "theoretical_loss": 3.3548772010033687,
      "tokens_seen": 2631204864
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020622222222222222,
      "loss": 1.5557,
      "theoretical_loss": 3.3548706038318867,
      "tokens_seen": 2631270400
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002062020202020202,
      "loss": 1.3512,
      "theoretical_loss": 3.3548640068707223,
      "tokens_seen": 2631335936
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002061818181818182,
      "loss": 1.3213,
      "theoretical_loss": 3.3548574101198634,
      "tokens_seen": 2631401472
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020616161616161615,
      "loss": 1.2218,
      "theoretical_loss": 3.354850813579298,
      "tokens_seen": 2631467008
    },
    {
      "epoch": 0.59,
      "objective/train/aligned_token_count": 0.5670821070671082,
      "objective/train/avg_token_score": 0.005377893336117268,
      "objective/train/docs_used": 1481256,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8459515571594238,
      "objective/train/original_loss": 3.8112692832946777,
      "objective/train/theoretical_loss": 3.3548442172490147,
      "objective/train/tokens_used": 990391776,
      "objective/train/unlikelihood_loss": 0.054917339235544205,
      "theoretical_loss": 3.3548442172490147,
      "tokens_seen": 2631532544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020614141414141416,
      "loss": 1.588,
      "theoretical_loss": 3.3548442172490147,
      "tokens_seen": 2631532544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020612121212121212,
      "loss": 1.3215,
      "theoretical_loss": 3.3548376211290005,
      "tokens_seen": 2631598080
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002061010101010101,
      "loss": 1.5161,
      "theoretical_loss": 3.3548310252192444,
      "tokens_seen": 2631663616
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020608080808080808,
      "loss": 1.4152,
      "theoretical_loss": 3.354824429519734,
      "tokens_seen": 2631729152
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020606060606060607,
      "loss": 1.4153,
      "theoretical_loss": 3.354817834030457,
      "tokens_seen": 2631794688
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020604040404040405,
      "loss": 1.2901,
      "theoretical_loss": 3.3548112387514024,
      "tokens_seen": 2631860224
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020602020202020204,
      "loss": 1.4594,
      "theoretical_loss": 3.3548046436825576,
      "tokens_seen": 2631925760
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000206,
      "loss": 1.2848,
      "theoretical_loss": 3.354798048823911,
      "tokens_seen": 2631991296
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020597979797979798,
      "loss": 1.3681,
      "theoretical_loss": 3.35479145417545,
      "tokens_seen": 2632056832
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020595959595959596,
      "loss": 1.2751,
      "theoretical_loss": 3.3547848597371637,
      "tokens_seen": 2632122368
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020593939393939394,
      "loss": 1.1498,
      "theoretical_loss": 3.3547782655090397,
      "tokens_seen": 2632187904
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020591919191919193,
      "loss": 1.31,
      "theoretical_loss": 3.354771671491066,
      "tokens_seen": 2632253440
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020589898989898989,
      "loss": 1.151,
      "theoretical_loss": 3.3547650776832305,
      "tokens_seen": 2632318976
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002058787878787879,
      "loss": 1.4229,
      "theoretical_loss": 3.3547584840855214,
      "tokens_seen": 2632384512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020585858585858585,
      "loss": 1.2767,
      "theoretical_loss": 3.3547518906979272,
      "tokens_seen": 2632450048
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020583838383838386,
      "loss": 1.157,
      "theoretical_loss": 3.3547452975204353,
      "tokens_seen": 2632515584
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020581818181818182,
      "loss": 1.2531,
      "theoretical_loss": 3.354738704553034,
      "tokens_seen": 2632581120
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020579797979797978,
      "loss": 1.5091,
      "theoretical_loss": 3.3547321117957116,
      "tokens_seen": 2632646656
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002057777777777778,
      "loss": 1.2399,
      "theoretical_loss": 3.354725519248456,
      "tokens_seen": 2632712192
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020575757575757575,
      "loss": 1.3927,
      "theoretical_loss": 3.3547189269112554,
      "tokens_seen": 2632777728
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020573737373737376,
      "loss": 1.467,
      "theoretical_loss": 3.354712334784098,
      "tokens_seen": 2632843264
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020571717171717171,
      "loss": 1.4112,
      "theoretical_loss": 3.354705742866971,
      "tokens_seen": 2632908800
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002056969696969697,
      "loss": 1.1768,
      "theoretical_loss": 3.354699151159864,
      "tokens_seen": 2632974336
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020567676767676768,
      "loss": 1.2527,
      "theoretical_loss": 3.3546925596627637,
      "tokens_seen": 2633039872
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020565656565656567,
      "loss": 1.262,
      "theoretical_loss": 3.3546859683756587,
      "tokens_seen": 2633105408
    },
    {
      "epoch": 0.59,
      "objective/train/aligned_token_count": 0.3105449676513672,
      "objective/train/avg_token_score": 0.011045201681554317,
      "objective/train/docs_used": 1482418,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0906933546066284,
      "objective/train/original_loss": 4.825595378875732,
      "objective/train/theoretical_loss": 3.3546793772985373,
      "objective/train/tokens_used": 992030176,
      "objective/train/unlikelihood_loss": 0.04736040160059929,
      "theoretical_loss": 3.3546793772985373,
      "tokens_seen": 2633170944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020563636363636365,
      "loss": 1.3566,
      "theoretical_loss": 3.3546793772985373,
      "tokens_seen": 2633170944
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002056161616161616,
      "loss": 1.4001,
      "theoretical_loss": 3.3546727864313874,
      "tokens_seen": 2633236480
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020559595959595962,
      "loss": 1.2746,
      "theoretical_loss": 3.354666195774197,
      "tokens_seen": 2633302016
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020557575757575757,
      "loss": 1.5432,
      "theoretical_loss": 3.354659605326954,
      "tokens_seen": 2633367552
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020555555555555556,
      "loss": 1.4564,
      "theoretical_loss": 3.354653015089647,
      "tokens_seen": 2633433088
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020553535353535354,
      "loss": 1.6502,
      "theoretical_loss": 3.354646425062264,
      "tokens_seen": 2633498624
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020551515151515153,
      "loss": 1.4682,
      "theoretical_loss": 3.3546398352447926,
      "tokens_seen": 2633564160
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002054949494949495,
      "loss": 1.51,
      "theoretical_loss": 3.3546332456372214,
      "tokens_seen": 2633629696
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002054747474747475,
      "loss": 1.4719,
      "theoretical_loss": 3.3546266562395384,
      "tokens_seen": 2633695232
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020545454545454545,
      "loss": 1.5888,
      "theoretical_loss": 3.3546200670517314,
      "tokens_seen": 2633760768
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020543434343434344,
      "loss": 1.3312,
      "theoretical_loss": 3.354613478073789,
      "tokens_seen": 2633826304
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020541414141414142,
      "loss": 1.3378,
      "theoretical_loss": 3.354606889305699,
      "tokens_seen": 2633891840
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002053939393939394,
      "loss": 1.1668,
      "theoretical_loss": 3.3546003007474487,
      "tokens_seen": 2633957376
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002053737373737374,
      "loss": 1.6481,
      "theoretical_loss": 3.3545937123990273,
      "tokens_seen": 2634022912
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020535353535353534,
      "loss": 1.3373,
      "theoretical_loss": 3.354587124260423,
      "tokens_seen": 2634088448
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020533333333333336,
      "loss": 1.5786,
      "theoretical_loss": 3.354580536331623,
      "tokens_seen": 2634153984
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002053131313131313,
      "loss": 1.358,
      "theoretical_loss": 3.3545739486126163,
      "tokens_seen": 2634219520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002052929292929293,
      "loss": 1.2426,
      "theoretical_loss": 3.3545673611033906,
      "tokens_seen": 2634285056
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020527272727272728,
      "loss": 1.498,
      "theoretical_loss": 3.3545607738039336,
      "tokens_seen": 2634350592
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020525252525252524,
      "loss": 1.0362,
      "theoretical_loss": 3.3545541867142337,
      "tokens_seen": 2634416128
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020523232323232325,
      "loss": 1.3936,
      "theoretical_loss": 3.3545475998342793,
      "tokens_seen": 2634481664
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002052121212121212,
      "loss": 1.3611,
      "theoretical_loss": 3.3545410131640585,
      "tokens_seen": 2634547200
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020519191919191922,
      "loss": 1.2467,
      "theoretical_loss": 3.354534426703559,
      "tokens_seen": 2634612736
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020517171717171717,
      "loss": 1.5955,
      "theoretical_loss": 3.3545278404527687,
      "tokens_seen": 2634678272
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020515151515151516,
      "loss": 1.5266,
      "theoretical_loss": 3.3545212544116767,
      "tokens_seen": 2634743808
    },
    {
      "epoch": 0.59,
      "objective/train/aligned_token_count": 0.4293743968009949,
      "objective/train/avg_token_score": 0.01933373138308525,
      "objective/train/docs_used": 1483155,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4439805746078491,
      "objective/train/original_loss": 4.442872047424316,
      "objective/train/theoretical_loss": 3.3545146685802703,
      "objective/train/tokens_used": 993668576,
      "objective/train/unlikelihood_loss": 0.0548425130546093,
      "theoretical_loss": 3.3545146685802703,
      "tokens_seen": 2634809344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020513131313131314,
      "loss": 1.4686,
      "theoretical_loss": 3.3545146685802703,
      "tokens_seen": 2634809344
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002051111111111111,
      "loss": 1.311,
      "theoretical_loss": 3.354508082958538,
      "tokens_seen": 2634874880
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002050909090909091,
      "loss": 1.5608,
      "theoretical_loss": 3.354501497546467,
      "tokens_seen": 2634940416
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020507070707070707,
      "loss": 1.4932,
      "theoretical_loss": 3.354494912344047,
      "tokens_seen": 2635005952
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020505050505050505,
      "loss": 1.1804,
      "theoretical_loss": 3.3544883273512647,
      "tokens_seen": 2635071488
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020503030303030303,
      "loss": 1.2223,
      "theoretical_loss": 3.354481742568109,
      "tokens_seen": 2635137024
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020501010101010102,
      "loss": 1.4905,
      "theoretical_loss": 3.3544751579945675,
      "tokens_seen": 2635202560
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.000204989898989899,
      "loss": 1.3415,
      "theoretical_loss": 3.354468573630629,
      "tokens_seen": 2635268096
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020496969696969699,
      "loss": 1.479,
      "theoretical_loss": 3.354461989476281,
      "tokens_seen": 2635333632
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020494949494949494,
      "loss": 1.3738,
      "theoretical_loss": 3.3544554055315117,
      "tokens_seen": 2635399168
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020492929292929293,
      "loss": 1.1714,
      "theoretical_loss": 3.3544488217963093,
      "tokens_seen": 2635464704
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002049090909090909,
      "loss": 1.3918,
      "theoretical_loss": 3.3544422382706625,
      "tokens_seen": 2635530240
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002048888888888889,
      "loss": 1.4268,
      "theoretical_loss": 3.354435654954558,
      "tokens_seen": 2635595776
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020486868686868688,
      "loss": 1.3078,
      "theoretical_loss": 3.3544290718479854,
      "tokens_seen": 2635661312
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020484848484848483,
      "loss": 1.3288,
      "theoretical_loss": 3.354422488950932,
      "tokens_seen": 2635726848
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020482828282828285,
      "loss": 1.4016,
      "theoretical_loss": 3.3544159062633865,
      "tokens_seen": 2635792384
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002048080808080808,
      "loss": 1.3434,
      "theoretical_loss": 3.3544093237853363,
      "tokens_seen": 2635857920
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020478787878787881,
      "loss": 1.2618,
      "theoretical_loss": 3.35440274151677,
      "tokens_seen": 2635923456
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020476767676767677,
      "loss": 1.4329,
      "theoretical_loss": 3.3543961594576754,
      "tokens_seen": 2635988992
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020474747474747473,
      "loss": 1.3711,
      "theoretical_loss": 3.3543895776080412,
      "tokens_seen": 2636054528
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020472727272727274,
      "loss": 1.4806,
      "theoretical_loss": 3.354382995967855,
      "tokens_seen": 2636120064
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002047070707070707,
      "loss": 1.2406,
      "theoretical_loss": 3.354376414537105,
      "tokens_seen": 2636185600
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002046868686868687,
      "loss": 1.2343,
      "theoretical_loss": 3.354369833315779,
      "tokens_seen": 2636251136
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020466666666666666,
      "loss": 1.3467,
      "theoretical_loss": 3.3543632523038664,
      "tokens_seen": 2636316672
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020464646464646467,
      "loss": 1.2764,
      "theoretical_loss": 3.354356671501354,
      "tokens_seen": 2636382208
    },
    {
      "epoch": 0.59,
      "objective/train/aligned_token_count": 0.4084188640117645,
      "objective/train/avg_token_score": 0.007876947522163391,
      "objective/train/docs_used": 1483578,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.365606427192688,
      "objective/train/original_loss": 4.529594421386719,
      "objective/train/theoretical_loss": 3.3543500909082304,
      "objective/train/tokens_used": 995306976,
      "objective/train/unlikelihood_loss": 0.052933886647224426,
      "theoretical_loss": 3.3543500909082304,
      "tokens_seen": 2636447744
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00020462626262626263,
      "loss": 1.4475,
      "theoretical_loss": 3.3543500909082304,
      "tokens_seen": 2636447744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002046060606060606,
      "loss": 1.3852,
      "theoretical_loss": 3.354343510524484,
      "tokens_seen": 2636513280
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002045858585858586,
      "loss": 1.3862,
      "theoretical_loss": 3.3543369303501027,
      "tokens_seen": 2636578816
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020456565656565656,
      "loss": 1.4775,
      "theoretical_loss": 3.3543303503850748,
      "tokens_seen": 2636644352
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020454545454545457,
      "loss": 1.3421,
      "theoretical_loss": 3.354323770629388,
      "tokens_seen": 2636709888
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020452525252525252,
      "loss": 1.5999,
      "theoretical_loss": 3.3543171910830303,
      "tokens_seen": 2636775424
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002045050505050505,
      "loss": 1.4763,
      "theoretical_loss": 3.3543106117459907,
      "tokens_seen": 2636840960
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002044848484848485,
      "loss": 1.2972,
      "theoretical_loss": 3.3543040326182565,
      "tokens_seen": 2636906496
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020446464646464648,
      "loss": 1.5101,
      "theoretical_loss": 3.3542974536998162,
      "tokens_seen": 2636972032
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020444444444444446,
      "loss": 1.2526,
      "theoretical_loss": 3.3542908749906584,
      "tokens_seen": 2637037568
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020442424242424242,
      "loss": 1.3939,
      "theoretical_loss": 3.3542842964907704,
      "tokens_seen": 2637103104
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002044040404040404,
      "loss": 1.481,
      "theoretical_loss": 3.354277718200141,
      "tokens_seen": 2637168640
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020438383838383838,
      "loss": 1.3188,
      "theoretical_loss": 3.354271140118758,
      "tokens_seen": 2637234176
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020436363636363637,
      "loss": 1.4618,
      "theoretical_loss": 3.3542645622466094,
      "tokens_seen": 2637299712
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020434343434343435,
      "loss": 1.115,
      "theoretical_loss": 3.354257984583684,
      "tokens_seen": 2637365248
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020432323232323234,
      "loss": 1.2822,
      "theoretical_loss": 3.354251407129969,
      "tokens_seen": 2637430784
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002043030303030303,
      "loss": 1.2815,
      "theoretical_loss": 3.3542448298854532,
      "tokens_seen": 2637496320
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002042828282828283,
      "loss": 1.1246,
      "theoretical_loss": 3.354238252850125,
      "tokens_seen": 2637561856
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020426262626262626,
      "loss": 1.3588,
      "theoretical_loss": 3.3542316760239714,
      "tokens_seen": 2637627392
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020424242424242425,
      "loss": 1.1798,
      "theoretical_loss": 3.3542250994069818,
      "tokens_seen": 2637692928
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020422222222222223,
      "loss": 1.2456,
      "theoretical_loss": 3.354218522999144,
      "tokens_seen": 2637758464
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020420202020202019,
      "loss": 1.359,
      "theoretical_loss": 3.354211946800446,
      "tokens_seen": 2637824000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002041818181818182,
      "loss": 1.4676,
      "theoretical_loss": 3.3542053708108757,
      "tokens_seen": 2637889536
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020416161616161615,
      "loss": 1.114,
      "theoretical_loss": 3.3541987950304213,
      "tokens_seen": 2637955072
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020414141414141416,
      "loss": 1.4396,
      "theoretical_loss": 3.3541922194590716,
      "tokens_seen": 2638020608
    },
    {
      "epoch": 0.6,
      "objective/train/aligned_token_count": 0.47727271914482117,
      "objective/train/avg_token_score": 0.010965626686811447,
      "objective/train/docs_used": 1484926,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4964243173599243,
      "objective/train/original_loss": 4.793129920959473,
      "objective/train/theoretical_loss": 3.3541856440968143,
      "objective/train/tokens_used": 996945376,
      "objective/train/unlikelihood_loss": 0.04090041667222977,
      "theoretical_loss": 3.3541856440968143,
      "tokens_seen": 2638086144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020412121212121212,
      "loss": 1.3307,
      "theoretical_loss": 3.3541856440968143,
      "tokens_seen": 2638086144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002041010101010101,
      "loss": 1.4124,
      "theoretical_loss": 3.3541790689436377,
      "tokens_seen": 2638151680
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002040808080808081,
      "loss": 1.3624,
      "theoretical_loss": 3.3541724939995294,
      "tokens_seen": 2638217216
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020406060606060605,
      "loss": 1.5096,
      "theoretical_loss": 3.3541659192644784,
      "tokens_seen": 2638282752
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020404040404040406,
      "loss": 1.3408,
      "theoretical_loss": 3.354159344738472,
      "tokens_seen": 2638348288
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020402020202020201,
      "loss": 1.411,
      "theoretical_loss": 3.354152770421499,
      "tokens_seen": 2638413824
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000204,
      "loss": 1.2414,
      "theoretical_loss": 3.3541461963135477,
      "tokens_seen": 2638479360
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020397979797979798,
      "loss": 1.4321,
      "theoretical_loss": 3.354139622414606,
      "tokens_seen": 2638544896
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020395959595959597,
      "loss": 1.2656,
      "theoretical_loss": 3.3541330487246617,
      "tokens_seen": 2638610432
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020393939393939395,
      "loss": 1.1689,
      "theoretical_loss": 3.354126475243703,
      "tokens_seen": 2638675968
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020391919191919193,
      "loss": 1.3044,
      "theoretical_loss": 3.354119901971719,
      "tokens_seen": 2638741504
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002038989898989899,
      "loss": 1.4651,
      "theoretical_loss": 3.3541133289086966,
      "tokens_seen": 2638807040
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020387878787878788,
      "loss": 1.3027,
      "theoretical_loss": 3.354106756054625,
      "tokens_seen": 2638872576
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020385858585858586,
      "loss": 1.511,
      "theoretical_loss": 3.354100183409492,
      "tokens_seen": 2638938112
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020383838383838384,
      "loss": 1.3883,
      "theoretical_loss": 3.3540936109732855,
      "tokens_seen": 2639003648
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020381818181818183,
      "loss": 1.3979,
      "theoretical_loss": 3.354087038745994,
      "tokens_seen": 2639069184
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002037979797979798,
      "loss": 1.4926,
      "theoretical_loss": 3.3540804667276056,
      "tokens_seen": 2639134720
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002037777777777778,
      "loss": 1.4507,
      "theoretical_loss": 3.354073894918108,
      "tokens_seen": 2639200256
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020375757575757575,
      "loss": 1.419,
      "theoretical_loss": 3.3540673233174902,
      "tokens_seen": 2639265792
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020373737373737374,
      "loss": 1.5157,
      "theoretical_loss": 3.35406075192574,
      "tokens_seen": 2639331328
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020371717171717172,
      "loss": 1.0873,
      "theoretical_loss": 3.3540541807428452,
      "tokens_seen": 2639396864
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002036969696969697,
      "loss": 0.9508,
      "theoretical_loss": 3.354047609768795,
      "tokens_seen": 2639462400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002036767676767677,
      "loss": 1.4253,
      "theoretical_loss": 3.3540410390035764,
      "tokens_seen": 2639527936
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020365656565656564,
      "loss": 1.373,
      "theoretical_loss": 3.354034468447178,
      "tokens_seen": 2639593472
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020363636363636366,
      "loss": 1.4295,
      "theoretical_loss": 3.354027898099588,
      "tokens_seen": 2639659008
    },
    {
      "epoch": 0.6,
      "objective/train/aligned_token_count": 0.3816593289375305,
      "objective/train/avg_token_score": 0.015349997207522392,
      "objective/train/docs_used": 1485572,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2490715980529785,
      "objective/train/original_loss": 4.949139595031738,
      "objective/train/theoretical_loss": 3.354021327960795,
      "objective/train/tokens_used": 998583776,
      "objective/train/unlikelihood_loss": 0.05009213462471962,
      "theoretical_loss": 3.354021327960795,
      "tokens_seen": 2639724544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002036161616161616,
      "loss": 1.3466,
      "theoretical_loss": 3.354021327960795,
      "tokens_seen": 2639724544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020359595959595962,
      "loss": 1.338,
      "theoretical_loss": 3.354014758030787,
      "tokens_seen": 2639790080
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020357575757575758,
      "loss": 1.1746,
      "theoretical_loss": 3.3540081883095514,
      "tokens_seen": 2639855616
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020355555555555554,
      "loss": 1.1299,
      "theoretical_loss": 3.3540016187970774,
      "tokens_seen": 2639921152
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020353535353535355,
      "loss": 1.2035,
      "theoretical_loss": 3.353995049493353,
      "tokens_seen": 2639986688
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002035151515151515,
      "loss": 1.3571,
      "theoretical_loss": 3.3539884803983657,
      "tokens_seen": 2640052224
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020349494949494952,
      "loss": 1.4896,
      "theoretical_loss": 3.353981911512104,
      "tokens_seen": 2640117760
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020347474747474747,
      "loss": 1.5448,
      "theoretical_loss": 3.353975342834557,
      "tokens_seen": 2640183296
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020345454545454546,
      "loss": 1.1024,
      "theoretical_loss": 3.353968774365711,
      "tokens_seen": 2640248832
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020343434343434344,
      "loss": 1.3399,
      "theoretical_loss": 3.3539622061055563,
      "tokens_seen": 2640314368
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020341414141414142,
      "loss": 1.2978,
      "theoretical_loss": 3.35395563805408,
      "tokens_seen": 2640379904
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002033939393939394,
      "loss": 1.4374,
      "theoretical_loss": 3.3539490702112698,
      "tokens_seen": 2640445440
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020337373737373737,
      "loss": 1.1234,
      "theoretical_loss": 3.353942502577115,
      "tokens_seen": 2640510976
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020335353535353535,
      "loss": 1.1891,
      "theoretical_loss": 3.353935935151603,
      "tokens_seen": 2640576512
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020333333333333333,
      "loss": 1.3986,
      "theoretical_loss": 3.3539293679347226,
      "tokens_seen": 2640642048
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020331313131313132,
      "loss": 1.3591,
      "theoretical_loss": 3.353922800926461,
      "tokens_seen": 2640707584
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002032929292929293,
      "loss": 1.2573,
      "theoretical_loss": 3.3539162341268076,
      "tokens_seen": 2640773120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020327272727272729,
      "loss": 0.9312,
      "theoretical_loss": 3.35390966753575,
      "tokens_seen": 2640838656
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020325252525252524,
      "loss": 1.2477,
      "theoretical_loss": 3.3539031011532767,
      "tokens_seen": 2640904192
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020323232323232325,
      "loss": 1.2016,
      "theoretical_loss": 3.3538965349793752,
      "tokens_seen": 2640969728
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002032121212121212,
      "loss": 1.4889,
      "theoretical_loss": 3.3538899690140345,
      "tokens_seen": 2641035264
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002031919191919192,
      "loss": 1.2291,
      "theoretical_loss": 3.3538834032572424,
      "tokens_seen": 2641100800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020317171717171718,
      "loss": 1.3029,
      "theoretical_loss": 3.3538768377089867,
      "tokens_seen": 2641166336
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020315151515151514,
      "loss": 1.3734,
      "theoretical_loss": 3.3538702723692566,
      "tokens_seen": 2641231872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020313131313131315,
      "loss": 1.2375,
      "theoretical_loss": 3.3538637072380393,
      "tokens_seen": 2641297408
    },
    {
      "epoch": 0.6,
      "objective/train/aligned_token_count": 0.32007575035095215,
      "objective/train/avg_token_score": 0.02881203219294548,
      "objective/train/docs_used": 1487130,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0979609489440918,
      "objective/train/original_loss": 5.87798547744751,
      "objective/train/theoretical_loss": 3.3538571423153236,
      "objective/train/tokens_used": 1000222176,
      "objective/train/unlikelihood_loss": 0.03302318975329399,
      "theoretical_loss": 3.3538571423153236,
      "tokens_seen": 2641362944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002031111111111111,
      "loss": 1.4255,
      "theoretical_loss": 3.3538571423153236,
      "tokens_seen": 2641362944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020309090909090911,
      "loss": 1.2142,
      "theoretical_loss": 3.353850577601098,
      "tokens_seen": 2641428480
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020307070707070707,
      "loss": 1.4897,
      "theoretical_loss": 3.3538440130953497,
      "tokens_seen": 2641494016
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020305050505050503,
      "loss": 1.3384,
      "theoretical_loss": 3.353837448798068,
      "tokens_seen": 2641559552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020303030303030304,
      "loss": 1.3549,
      "theoretical_loss": 3.3538308847092404,
      "tokens_seen": 2641625088
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000203010101010101,
      "loss": 1.2808,
      "theoretical_loss": 3.353824320828855,
      "tokens_seen": 2641690624
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000202989898989899,
      "loss": 1.1666,
      "theoretical_loss": 3.3538177571569006,
      "tokens_seen": 2641756160
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020296969696969696,
      "loss": 1.4997,
      "theoretical_loss": 3.3538111936933652,
      "tokens_seen": 2641821696
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020294949494949497,
      "loss": 1.3562,
      "theoretical_loss": 3.3538046304382365,
      "tokens_seen": 2641887232
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020292929292929293,
      "loss": 1.3575,
      "theoretical_loss": 3.3537980673915033,
      "tokens_seen": 2641952768
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020290909090909092,
      "loss": 1.4345,
      "theoretical_loss": 3.353791504553154,
      "tokens_seen": 2642018304
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002028888888888889,
      "loss": 1.1827,
      "theoretical_loss": 3.3537849419231764,
      "tokens_seen": 2642083840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020286868686868686,
      "loss": 1.4813,
      "theoretical_loss": 3.3537783795015588,
      "tokens_seen": 2642149376
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020284848484848487,
      "loss": 1.566,
      "theoretical_loss": 3.353771817288289,
      "tokens_seen": 2642214912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020282828282828282,
      "loss": 1.4678,
      "theoretical_loss": 3.3537652552833563,
      "tokens_seen": 2642280448
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002028080808080808,
      "loss": 1.5469,
      "theoretical_loss": 3.3537586934867476,
      "tokens_seen": 2642345984
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002027878787878788,
      "loss": 1.3667,
      "theoretical_loss": 3.3537521318984522,
      "tokens_seen": 2642411520
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020276767676767678,
      "loss": 1.4754,
      "theoretical_loss": 3.3537455705184573,
      "tokens_seen": 2642477056
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020274747474747476,
      "loss": 1.3219,
      "theoretical_loss": 3.3537390093467523,
      "tokens_seen": 2642542592
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020272727272727274,
      "loss": 1.3023,
      "theoretical_loss": 3.3537324483833246,
      "tokens_seen": 2642608128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002027070707070707,
      "loss": 1.3669,
      "theoretical_loss": 3.353725887628163,
      "tokens_seen": 2642673664
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020268686868686868,
      "loss": 1.167,
      "theoretical_loss": 3.3537193270812544,
      "tokens_seen": 2642739200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020266666666666667,
      "loss": 1.2353,
      "theoretical_loss": 3.3537127667425892,
      "tokens_seen": 2642804736
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020264646464646465,
      "loss": 1.2332,
      "theoretical_loss": 3.353706206612154,
      "tokens_seen": 2642870272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020262626262626264,
      "loss": 1.1943,
      "theoretical_loss": 3.353699646689937,
      "tokens_seen": 2642935808
    },
    {
      "epoch": 0.6,
      "objective/train/aligned_token_count": 0.3800097703933716,
      "objective/train/avg_token_score": 0.013477552682161331,
      "objective/train/docs_used": 1487881,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1496537923812866,
      "objective/train/original_loss": 4.358090400695801,
      "objective/train/theoretical_loss": 3.353693086975927,
      "objective/train/tokens_used": 1001860576,
      "objective/train/unlikelihood_loss": 0.06138288602232933,
      "theoretical_loss": 3.353693086975927,
      "tokens_seen": 2643001344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002026060606060606,
      "loss": 1.303,
      "theoretical_loss": 3.353693086975927,
      "tokens_seen": 2643001344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002025858585858586,
      "loss": 1.3085,
      "theoretical_loss": 3.3536865274701126,
      "tokens_seen": 2643066880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020256565656565656,
      "loss": 1.4577,
      "theoretical_loss": 3.3536799681724814,
      "tokens_seen": 2643132416
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020254545454545457,
      "loss": 1.1933,
      "theoretical_loss": 3.3536734090830214,
      "tokens_seen": 2643197952
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020252525252525253,
      "loss": 1.5612,
      "theoretical_loss": 3.3536668502017215,
      "tokens_seen": 2643263488
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020250505050505049,
      "loss": 1.3009,
      "theoretical_loss": 3.35366029152857,
      "tokens_seen": 2643329024
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002024848484848485,
      "loss": 1.381,
      "theoretical_loss": 3.353653733063554,
      "tokens_seen": 2643394560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020246464646464645,
      "loss": 1.4232,
      "theoretical_loss": 3.353647174806663,
      "tokens_seen": 2643460096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020244444444444447,
      "loss": 1.3914,
      "theoretical_loss": 3.3536406167578843,
      "tokens_seen": 2643525632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020242424242424242,
      "loss": 1.4903,
      "theoretical_loss": 3.353634058917207,
      "tokens_seen": 2643591168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002024040404040404,
      "loss": 1.7199,
      "theoretical_loss": 3.353627501284619,
      "tokens_seen": 2643656704
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002023838383838384,
      "loss": 1.2575,
      "theoretical_loss": 3.353620943860108,
      "tokens_seen": 2643722240
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020236363636363637,
      "loss": 1.3975,
      "theoretical_loss": 3.353614386643663,
      "tokens_seen": 2643787776
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020234343434343436,
      "loss": 1.2169,
      "theoretical_loss": 3.353607829635272,
      "tokens_seen": 2643853312
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020232323232323231,
      "loss": 1.4352,
      "theoretical_loss": 3.3536012728349234,
      "tokens_seen": 2643918848
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002023030303030303,
      "loss": 1.2518,
      "theoretical_loss": 3.3535947162426045,
      "tokens_seen": 2643984384
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020228282828282828,
      "loss": 1.35,
      "theoretical_loss": 3.353588159858305,
      "tokens_seen": 2644049920
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020226262626262627,
      "loss": 1.1307,
      "theoretical_loss": 3.353581603682012,
      "tokens_seen": 2644115456
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020224242424242425,
      "loss": 1.4211,
      "theoretical_loss": 3.353575047713714,
      "tokens_seen": 2644180992
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020222222222222223,
      "loss": 1.4068,
      "theoretical_loss": 3.3535684919534,
      "tokens_seen": 2644246528
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002022020202020202,
      "loss": 1.6353,
      "theoretical_loss": 3.353561936401057,
      "tokens_seen": 2644312064
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020218181818181818,
      "loss": 1.2623,
      "theoretical_loss": 3.3535553810566743,
      "tokens_seen": 2644377600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020216161616161616,
      "loss": 1.4069,
      "theoretical_loss": 3.3535488259202397,
      "tokens_seen": 2644443136
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020214141414141414,
      "loss": 1.1306,
      "theoretical_loss": 3.3535422709917415,
      "tokens_seen": 2644508672
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020212121212121213,
      "loss": 1.4691,
      "theoretical_loss": 3.3535357162711676,
      "tokens_seen": 2644574208
    },
    {
      "debugging/Self-BLEU-5": 0.44043301378291366,
      "debugging/distinct-1-grams": 0.8028592116073892,
      "debugging/distinct-2-grams": 0.9709771731359748,
      "debugging/entropy-1-grams": 5.991297910662675,
      "debugging/entropy-2-grams": 6.773013872721587,
      "debugging/length": 450.2142857142857,
      "debugging/num_segments": 14,
      "epoch": 0.6,
      "objective/train/aligned_token_count": 0.4979838728904724,
      "objective/train/avg_token_score": 0.003441630629822612,
      "objective/train/docs_used": 1488919,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4609434604644775,
      "objective/train/original_loss": 3.914118766784668,
      "objective/train/theoretical_loss": 3.353529161758507,
      "objective/train/tokens_used": 1003498976,
      "objective/train/unlikelihood_loss": 0.07803807407617569,
      "theoretical_loss": 3.353529161758507,
      "tokens_seen": 2644639744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002021010101010101,
      "loss": 1.3054,
      "theoretical_loss": 3.353529161758507,
      "tokens_seen": 2644639744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002020808080808081,
      "loss": 1.2219,
      "theoretical_loss": 3.3535226074537476,
      "tokens_seen": 2644705280
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020206060606060605,
      "loss": 1.3998,
      "theoretical_loss": 3.353516053356877,
      "tokens_seen": 2644770816
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020204040404040406,
      "loss": 1.4236,
      "theoretical_loss": 3.3535094994678847,
      "tokens_seen": 2644836352
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020202020202020202,
      "loss": 1.2944,
      "theoretical_loss": 3.353502945786758,
      "tokens_seen": 2644901888
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000202,
      "loss": 1.5161,
      "theoretical_loss": 3.353496392313486,
      "tokens_seen": 2644967424
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000201979797979798,
      "loss": 1.5322,
      "theoretical_loss": 3.3534898390480556,
      "tokens_seen": 2645032960
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020195959595959594,
      "loss": 1.3949,
      "theoretical_loss": 3.353483285990457,
      "tokens_seen": 2645098496
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020193939393939396,
      "loss": 1.4088,
      "theoretical_loss": 3.3534767331406763,
      "tokens_seen": 2645164032
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002019191919191919,
      "loss": 1.5874,
      "theoretical_loss": 3.3534701804987033,
      "tokens_seen": 2645229568
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020189898989898992,
      "loss": 1.2493,
      "theoretical_loss": 3.3534636280645254,
      "tokens_seen": 2645295104
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020187878787878788,
      "loss": 1.6246,
      "theoretical_loss": 3.353457075838132,
      "tokens_seen": 2645360640
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020185858585858586,
      "loss": 1.4005,
      "theoretical_loss": 3.35345052381951,
      "tokens_seen": 2645426176
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020183838383838385,
      "loss": 1.4624,
      "theoretical_loss": 3.353443972008648,
      "tokens_seen": 2645491712
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002018181818181818,
      "loss": 1.2584,
      "theoretical_loss": 3.353437420405535,
      "tokens_seen": 2645557248
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020179797979797982,
      "loss": 1.6251,
      "theoretical_loss": 3.3534308690101584,
      "tokens_seen": 2645622784
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020177777777777777,
      "loss": 1.3404,
      "theoretical_loss": 3.3534243178225074,
      "tokens_seen": 2645688320
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020175757575757576,
      "loss": 1.3556,
      "theoretical_loss": 3.3534177668425698,
      "tokens_seen": 2645753856
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020173737373737374,
      "loss": 1.5036,
      "theoretical_loss": 3.353411216070333,
      "tokens_seen": 2645819392
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020171717171717173,
      "loss": 1.5591,
      "theoretical_loss": 3.3534046655057868,
      "tokens_seen": 2645884928
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002016969696969697,
      "loss": 1.5313,
      "theoretical_loss": 3.3533981151489187,
      "tokens_seen": 2645950464
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002016767676767677,
      "loss": 1.6612,
      "theoretical_loss": 3.3533915649997166,
      "tokens_seen": 2646016000
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020165656565656565,
      "loss": 1.4434,
      "theoretical_loss": 3.3533850150581697,
      "tokens_seen": 2646081536
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020163636363636363,
      "loss": 1.3919,
      "theoretical_loss": 3.3533784653242655,
      "tokens_seen": 2646147072
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020161616161616162,
      "loss": 1.3354,
      "theoretical_loss": 3.353371915797992,
      "tokens_seen": 2646212608
    },
    {
      "epoch": 0.6,
      "objective/train/aligned_token_count": 0.29820382595062256,
      "objective/train/avg_token_score": 0.007026410195976496,
      "objective/train/docs_used": 1489355,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9937134385108948,
      "objective/train/original_loss": 4.702162265777588,
      "objective/train/theoretical_loss": 3.3533653664793386,
      "objective/train/tokens_used": 1005137376,
      "objective/train/unlikelihood_loss": 0.06563170254230499,
      "theoretical_loss": 3.3533653664793386,
      "tokens_seen": 2646278144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002015959595959596,
      "loss": 1.2753,
      "theoretical_loss": 3.3533653664793386,
      "tokens_seen": 2646278144
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020157575757575759,
      "loss": 1.2285,
      "theoretical_loss": 3.353358817368293,
      "tokens_seen": 2646343680
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020155555555555554,
      "loss": 1.3373,
      "theoretical_loss": 3.3533522684648434,
      "tokens_seen": 2646409216
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020153535353535355,
      "loss": 1.596,
      "theoretical_loss": 3.353345719768978,
      "tokens_seen": 2646474752
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002015151515151515,
      "loss": 1.2329,
      "theoretical_loss": 3.3533391712806857,
      "tokens_seen": 2646540288
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020149494949494952,
      "loss": 1.33,
      "theoretical_loss": 3.353332622999954,
      "tokens_seen": 2646605824
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020147474747474748,
      "loss": 1.4246,
      "theoretical_loss": 3.3533260749267715,
      "tokens_seen": 2646671360
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020145454545454544,
      "loss": 1.4265,
      "theoretical_loss": 3.3533195270611262,
      "tokens_seen": 2646736896
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020143434343434345,
      "loss": 1.1806,
      "theoretical_loss": 3.353312979403007,
      "tokens_seen": 2646802432
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002014141414141414,
      "loss": 1.4028,
      "theoretical_loss": 3.3533064319524017,
      "tokens_seen": 2646867968
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020139393939393941,
      "loss": 1.3308,
      "theoretical_loss": 3.353299884709299,
      "tokens_seen": 2646933504
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020137373737373737,
      "loss": 1.3952,
      "theoretical_loss": 3.3532933376736866,
      "tokens_seen": 2646999040
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020135353535353536,
      "loss": 1.3755,
      "theoretical_loss": 3.3532867908455533,
      "tokens_seen": 2647064576
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020133333333333334,
      "loss": 1.5982,
      "theoretical_loss": 3.353280244224887,
      "tokens_seen": 2647130112
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002013131313131313,
      "loss": 1.2681,
      "theoretical_loss": 3.353273697811676,
      "tokens_seen": 2647195648
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002012929292929293,
      "loss": 1.4256,
      "theoretical_loss": 3.353267151605909,
      "tokens_seen": 2647261184
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020127272727272726,
      "loss": 1.4363,
      "theoretical_loss": 3.3532606056075744,
      "tokens_seen": 2647326720
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020125252525252528,
      "loss": 1.4536,
      "theoretical_loss": 3.35325405981666,
      "tokens_seen": 2647392256
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020123232323232323,
      "loss": 1.5244,
      "theoretical_loss": 3.3532475142331535,
      "tokens_seen": 2647457792
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020121212121212122,
      "loss": 1.313,
      "theoretical_loss": 3.353240968857045,
      "tokens_seen": 2647523328
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002011919191919192,
      "loss": 1.4155,
      "theoretical_loss": 3.353234423688321,
      "tokens_seen": 2647588864
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020117171717171718,
      "loss": 1.3218,
      "theoretical_loss": 3.3532278787269707,
      "tokens_seen": 2647654400
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020115151515151517,
      "loss": 1.2197,
      "theoretical_loss": 3.353221333972982,
      "tokens_seen": 2647719936
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020113131313131312,
      "loss": 1.3166,
      "theoretical_loss": 3.353214789426344,
      "tokens_seen": 2647785472
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002011111111111111,
      "loss": 1.5364,
      "theoretical_loss": 3.3532082450870444,
      "tokens_seen": 2647851008
    },
    {
      "epoch": 0.6,
      "objective/train/aligned_token_count": 0.4590664803981781,
      "objective/train/avg_token_score": 0.018356328830122948,
      "objective/train/docs_used": 1489706,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.340805172920227,
      "objective/train/original_loss": 4.358427047729492,
      "objective/train/theoretical_loss": 3.353201700955071,
      "objective/train/tokens_used": 1006775776,
      "objective/train/unlikelihood_loss": 0.050564903765916824,
      "theoretical_loss": 3.353201700955071,
      "tokens_seen": 2647916544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002010909090909091,
      "loss": 1.3496,
      "theoretical_loss": 3.353201700955071,
      "tokens_seen": 2647916544
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020107070707070708,
      "loss": 1.3144,
      "theoretical_loss": 3.353195157030413,
      "tokens_seen": 2647982080
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020105050505050506,
      "loss": 1.2787,
      "theoretical_loss": 3.3531886133130584,
      "tokens_seen": 2648047616
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020103030303030304,
      "loss": 1.1294,
      "theoretical_loss": 3.3531820698029953,
      "tokens_seen": 2648113152
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000201010101010101,
      "loss": 1.2267,
      "theoretical_loss": 3.353175526500212,
      "tokens_seen": 2648178688
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000200989898989899,
      "loss": 1.3933,
      "theoretical_loss": 3.353168983404697,
      "tokens_seen": 2648244224
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020096969696969697,
      "loss": 1.3847,
      "theoretical_loss": 3.3531624405164386,
      "tokens_seen": 2648309760
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020094949494949495,
      "loss": 1.2689,
      "theoretical_loss": 3.353155897835425,
      "tokens_seen": 2648375296
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020092929292929294,
      "loss": 1.4263,
      "theoretical_loss": 3.3531493553616447,
      "tokens_seen": 2648440832
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002009090909090909,
      "loss": 1.2877,
      "theoretical_loss": 3.3531428130950856,
      "tokens_seen": 2648506368
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002008888888888889,
      "loss": 1.2822,
      "theoretical_loss": 3.353136271035736,
      "tokens_seen": 2648571904
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020086868686868686,
      "loss": 1.2816,
      "theoretical_loss": 3.353129729183585,
      "tokens_seen": 2648637440
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020084848484848487,
      "loss": 1.5119,
      "theoretical_loss": 3.3531231875386207,
      "tokens_seen": 2648702976
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020082828282828283,
      "loss": 1.1565,
      "theoretical_loss": 3.3531166461008306,
      "tokens_seen": 2648768512
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020080808080808081,
      "loss": 1.3274,
      "theoretical_loss": 3.3531101048702037,
      "tokens_seen": 2648834048
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002007878787878788,
      "loss": 1.5963,
      "theoretical_loss": 3.353103563846728,
      "tokens_seen": 2648899584
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020076767676767675,
      "loss": 1.1537,
      "theoretical_loss": 3.353097023030392,
      "tokens_seen": 2648965120
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020074747474747477,
      "loss": 1.3523,
      "theoretical_loss": 3.353090482421184,
      "tokens_seen": 2649030656
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020072727272727272,
      "loss": 1.2407,
      "theoretical_loss": 3.3530839420190923,
      "tokens_seen": 2649096192
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002007070707070707,
      "loss": 1.3398,
      "theoretical_loss": 3.353077401824105,
      "tokens_seen": 2649161728
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002006868686868687,
      "loss": 1.2263,
      "theoretical_loss": 3.353070861836211,
      "tokens_seen": 2649227264
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020066666666666667,
      "loss": 1.2683,
      "theoretical_loss": 3.3530643220553977,
      "tokens_seen": 2649292800
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020064646464646466,
      "loss": 1.3756,
      "theoretical_loss": 3.3530577824816543,
      "tokens_seen": 2649358336
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020062626262626264,
      "loss": 1.4209,
      "theoretical_loss": 3.353051243114969,
      "tokens_seen": 2649423872
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002006060606060606,
      "loss": 1.2851,
      "theoretical_loss": 3.3530447039553293,
      "tokens_seen": 2649489408
    },
    {
      "epoch": 0.6,
      "objective/train/aligned_token_count": 0.3911290466785431,
      "objective/train/avg_token_score": 0.022906459867954254,
      "objective/train/docs_used": 1491205,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1416980028152466,
      "objective/train/original_loss": 5.172658443450928,
      "objective/train/theoretical_loss": 3.3530381650027246,
      "objective/train/tokens_used": 1008414176,
      "objective/train/unlikelihood_loss": 0.03421119600534439,
      "theoretical_loss": 3.3530381650027246,
      "tokens_seen": 2649554944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020058585858585858,
      "loss": 1.4555,
      "theoretical_loss": 3.3530381650027246,
      "tokens_seen": 2649554944
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020056565656565657,
      "loss": 1.3752,
      "theoretical_loss": 3.3530316262571427,
      "tokens_seen": 2649620480
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020054545454545455,
      "loss": 1.3557,
      "theoretical_loss": 3.3530250877185717,
      "tokens_seen": 2649686016
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020052525252525254,
      "loss": 1.3561,
      "theoretical_loss": 3.3530185493870004,
      "tokens_seen": 2649751552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002005050505050505,
      "loss": 1.552,
      "theoretical_loss": 3.3530120112624173,
      "tokens_seen": 2649817088
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002004848484848485,
      "loss": 1.3612,
      "theoretical_loss": 3.3530054733448096,
      "tokens_seen": 2649882624
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020046464646464646,
      "loss": 1.5926,
      "theoretical_loss": 3.352998935634167,
      "tokens_seen": 2649948160
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020044444444444444,
      "loss": 1.3131,
      "theoretical_loss": 3.352992398130477,
      "tokens_seen": 2650013696
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020042424242424243,
      "loss": 1.2442,
      "theoretical_loss": 3.352985860833728,
      "tokens_seen": 2650079232
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002004040404040404,
      "loss": 1.2727,
      "theoretical_loss": 3.3529793237439085,
      "tokens_seen": 2650144768
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002003838383838384,
      "loss": 1.3681,
      "theoretical_loss": 3.352972786861007,
      "tokens_seen": 2650210304
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020036363636363635,
      "loss": 1.2331,
      "theoretical_loss": 3.3529662501850117,
      "tokens_seen": 2650275840
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020034343434343436,
      "loss": 1.2881,
      "theoretical_loss": 3.3529597137159106,
      "tokens_seen": 2650341376
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020032323232323232,
      "loss": 1.1103,
      "theoretical_loss": 3.352953177453693,
      "tokens_seen": 2650406912
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020030303030303033,
      "loss": 1.2999,
      "theoretical_loss": 3.3529466413983458,
      "tokens_seen": 2650472448
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002002828282828283,
      "loss": 1.4282,
      "theoretical_loss": 3.352940105549858,
      "tokens_seen": 2650537984
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020026262626262625,
      "loss": 1.3311,
      "theoretical_loss": 3.3529335699082186,
      "tokens_seen": 2650603520
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020024242424242426,
      "loss": 1.3756,
      "theoretical_loss": 3.3529270344734154,
      "tokens_seen": 2650669056
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002002222222222222,
      "loss": 1.458,
      "theoretical_loss": 3.3529204992454362,
      "tokens_seen": 2650734592
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020020202020202022,
      "loss": 1.2947,
      "theoretical_loss": 3.3529139642242702,
      "tokens_seen": 2650800128
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020018181818181818,
      "loss": 1.4033,
      "theoretical_loss": 3.352907429409905,
      "tokens_seen": 2650865664
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020016161616161616,
      "loss": 1.3923,
      "theoretical_loss": 3.35290089480233,
      "tokens_seen": 2650931200
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020014141414141415,
      "loss": 1.6062,
      "theoretical_loss": 3.352894360401532,
      "tokens_seen": 2650996736
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020012121212121213,
      "loss": 1.5141,
      "theoretical_loss": 3.352887826207501,
      "tokens_seen": 2651062272
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020010101010101012,
      "loss": 1.4132,
      "theoretical_loss": 3.3528812922202245,
      "tokens_seen": 2651127808
    },
    {
      "epoch": 0.6,
      "objective/train/aligned_token_count": 0.4533846378326416,
      "objective/train/avg_token_score": 0.01504614856094122,
      "objective/train/docs_used": 1491801,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6054449081420898,
      "objective/train/original_loss": 4.346124649047852,
      "objective/train/theoretical_loss": 3.3528747584396905,
      "objective/train/tokens_used": 1010052576,
      "objective/train/unlikelihood_loss": 0.05576375871896744,
      "theoretical_loss": 3.3528747584396905,
      "tokens_seen": 2651193344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020008080808080807,
      "loss": 1.241,
      "theoretical_loss": 3.3528747584396905,
      "tokens_seen": 2651193344
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020006060606060606,
      "loss": 1.5442,
      "theoretical_loss": 3.352868224865888,
      "tokens_seen": 2651258880
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020004040404040404,
      "loss": 1.3292,
      "theoretical_loss": 3.352861691498805,
      "tokens_seen": 2651324416
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00020002020202020203,
      "loss": 1.3422,
      "theoretical_loss": 3.3528551583384303,
      "tokens_seen": 2651389952
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 1.2044,
      "theoretical_loss": 3.352848625384752,
      "tokens_seen": 2651455488
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.000199979797979798,
      "loss": 1.0564,
      "theoretical_loss": 3.352842092637758,
      "tokens_seen": 2651521024
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019995959595959595,
      "loss": 1.4517,
      "theoretical_loss": 3.352835560097437,
      "tokens_seen": 2651586560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019993939393939396,
      "loss": 1.4145,
      "theoretical_loss": 3.352829027763778,
      "tokens_seen": 2651652096
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019991919191919192,
      "loss": 1.3146,
      "theoretical_loss": 3.352822495636768,
      "tokens_seen": 2651717632
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001998989898989899,
      "loss": 1.2064,
      "theoretical_loss": 3.352815963716396,
      "tokens_seen": 2651783168
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019987878787878789,
      "loss": 1.2051,
      "theoretical_loss": 3.3528094320026507,
      "tokens_seen": 2651848704
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019985858585858584,
      "loss": 1.3374,
      "theoretical_loss": 3.3528029004955204,
      "tokens_seen": 2651914240
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019983838383838385,
      "loss": 1.3772,
      "theoretical_loss": 3.352796369194993,
      "tokens_seen": 2651979776
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001998181818181818,
      "loss": 1.5226,
      "theoretical_loss": 3.3527898381010575,
      "tokens_seen": 2652045312
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019979797979797982,
      "loss": 1.4098,
      "theoretical_loss": 3.352783307213701,
      "tokens_seen": 2652110848
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019977777777777778,
      "loss": 1.3013,
      "theoretical_loss": 3.352776776532914,
      "tokens_seen": 2652176384
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019975757575757574,
      "loss": 1.4465,
      "theoretical_loss": 3.352770246058683,
      "tokens_seen": 2652241920
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019973737373737375,
      "loss": 1.4207,
      "theoretical_loss": 3.3527637157909966,
      "tokens_seen": 2652307456
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001997171717171717,
      "loss": 1.4362,
      "theoretical_loss": 3.352757185729844,
      "tokens_seen": 2652372992
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019969696969696971,
      "loss": 1.3088,
      "theoretical_loss": 3.3527506558752127,
      "tokens_seen": 2652438528
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019967676767676767,
      "loss": 1.3446,
      "theoretical_loss": 3.352744126227092,
      "tokens_seen": 2652504064
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019965656565656566,
      "loss": 1.2971,
      "theoretical_loss": 3.3527375967854693,
      "tokens_seen": 2652569600
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019963636363636364,
      "loss": 1.2554,
      "theoretical_loss": 3.3527310675503337,
      "tokens_seen": 2652635136
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019961616161616162,
      "loss": 1.189,
      "theoretical_loss": 3.352724538521673,
      "tokens_seen": 2652700672
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0001995959595959596,
      "loss": 1.2868,
      "theoretical_loss": 3.352718009699476,
      "tokens_seen": 2652766208
    },
    {
      "epoch": 0.6,
      "objective/train/aligned_token_count": 0.4159335196018219,
      "objective/train/avg_token_score": 0.02335064485669136,
      "objective/train/docs_used": 1493158,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3423142433166504,
      "objective/train/original_loss": 3.9414682388305664,
      "objective/train/theoretical_loss": 3.352711481083731,
      "objective/train/tokens_used": 1011690976,
      "objective/train/unlikelihood_loss": 0.0677536204457283,
      "theoretical_loss": 3.352711481083731,
      "tokens_seen": 2652831744
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.00019957575757575756,
      "loss": 1.2459,
      "theoretical_loss": 3.352711481083731,
      "tokens_seen": 2652831744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019955555555555558,
      "loss": 1.4926,
      "theoretical_loss": 3.3527049526744257,
      "tokens_seen": 2652897280
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019953535353535353,
      "loss": 1.2913,
      "theoretical_loss": 3.3526984244715496,
      "tokens_seen": 2652962816
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019951515151515152,
      "loss": 1.096,
      "theoretical_loss": 3.3526918964750907,
      "tokens_seen": 2653028352
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001994949494949495,
      "loss": 1.1434,
      "theoretical_loss": 3.352685368685037,
      "tokens_seen": 2653093888
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019947474747474748,
      "loss": 1.4521,
      "theoretical_loss": 3.3526788411013766,
      "tokens_seen": 2653159424
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019945454545454547,
      "loss": 1.3701,
      "theoretical_loss": 3.352672313724099,
      "tokens_seen": 2653224960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019943434343434345,
      "loss": 1.4585,
      "theoretical_loss": 3.3526657865531915,
      "tokens_seen": 2653290496
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001994141414141414,
      "loss": 1.2294,
      "theoretical_loss": 3.352659259588643,
      "tokens_seen": 2653356032
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001993939393939394,
      "loss": 1.2391,
      "theoretical_loss": 3.352652732830442,
      "tokens_seen": 2653421568
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019937373737373738,
      "loss": 1.2583,
      "theoretical_loss": 3.3526462062785765,
      "tokens_seen": 2653487104
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019935353535353536,
      "loss": 1.356,
      "theoretical_loss": 3.352639679933035,
      "tokens_seen": 2653552640
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019933333333333334,
      "loss": 1.4682,
      "theoretical_loss": 3.352633153793806,
      "tokens_seen": 2653618176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001993131313131313,
      "loss": 1.4003,
      "theoretical_loss": 3.352626627860878,
      "tokens_seen": 2653683712
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001992929292929293,
      "loss": 1.1616,
      "theoretical_loss": 3.3526201021342388,
      "tokens_seen": 2653749248
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019927272727272727,
      "loss": 1.2119,
      "theoretical_loss": 3.3526135766138774,
      "tokens_seen": 2653814784
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019925252525252528,
      "loss": 1.3761,
      "theoretical_loss": 3.352607051299782,
      "tokens_seen": 2653880320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019923232323232324,
      "loss": 1.2639,
      "theoretical_loss": 3.352600526191941,
      "tokens_seen": 2653945856
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001992121212121212,
      "loss": 1.3854,
      "theoretical_loss": 3.352594001290343,
      "tokens_seen": 2654011392
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001991919191919192,
      "loss": 1.2649,
      "theoretical_loss": 3.3525874765949752,
      "tokens_seen": 2654076928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019917171717171716,
      "loss": 1.4305,
      "theoretical_loss": 3.3525809521058276,
      "tokens_seen": 2654142464
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019915151515151517,
      "loss": 1.4164,
      "theoretical_loss": 3.352574427822888,
      "tokens_seen": 2654208000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019913131313131313,
      "loss": 1.2837,
      "theoretical_loss": 3.352567903746144,
      "tokens_seen": 2654273536
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019911111111111111,
      "loss": 1.4187,
      "theoretical_loss": 3.352561379875586,
      "tokens_seen": 2654339072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001990909090909091,
      "loss": 1.2894,
      "theoretical_loss": 3.3525548562112,
      "tokens_seen": 2654404608
    },
    {
      "epoch": 0.61,
      "objective/train/aligned_token_count": 0.32526880502700806,
      "objective/train/avg_token_score": 0.07225336879491806,
      "objective/train/docs_used": 1493841,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9618147611618042,
      "objective/train/original_loss": 4.679859638214111,
      "objective/train/theoretical_loss": 3.3525483327529755,
      "objective/train/tokens_used": 1013329376,
      "objective/train/unlikelihood_loss": 0.07530628889799118,
      "theoretical_loss": 3.3525483327529755,
      "tokens_seen": 2654470144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019907070707070708,
      "loss": 1.403,
      "theoretical_loss": 3.3525483327529755,
      "tokens_seen": 2654470144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019905050505050507,
      "loss": 1.2498,
      "theoretical_loss": 3.3525418095009014,
      "tokens_seen": 2654535680
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019903030303030302,
      "loss": 1.3364,
      "theoretical_loss": 3.352535286454965,
      "tokens_seen": 2654601216
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000199010101010101,
      "loss": 1.3282,
      "theoretical_loss": 3.352528763615156,
      "tokens_seen": 2654666752
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000198989898989899,
      "loss": 1.5521,
      "theoretical_loss": 3.3525222409814615,
      "tokens_seen": 2654732288
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019896969696969697,
      "loss": 1.2838,
      "theoretical_loss": 3.3525157185538705,
      "tokens_seen": 2654797824
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019894949494949496,
      "loss": 1.4068,
      "theoretical_loss": 3.3525091963323717,
      "tokens_seen": 2654863360
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019892929292929294,
      "loss": 1.0854,
      "theoretical_loss": 3.352502674316953,
      "tokens_seen": 2654928896
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001989090909090909,
      "loss": 1.3579,
      "theoretical_loss": 3.352496152507603,
      "tokens_seen": 2654994432
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019888888888888888,
      "loss": 1.398,
      "theoretical_loss": 3.35248963090431,
      "tokens_seen": 2655059968
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019886868686868687,
      "loss": 1.2336,
      "theoretical_loss": 3.3524831095070624,
      "tokens_seen": 2655125504
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019884848484848485,
      "loss": 1.5299,
      "theoretical_loss": 3.3524765883158487,
      "tokens_seen": 2655191040
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019882828282828284,
      "loss": 1.3368,
      "theoretical_loss": 3.3524700673306573,
      "tokens_seen": 2655256576
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001988080808080808,
      "loss": 1.2302,
      "theoretical_loss": 3.3524635465514767,
      "tokens_seen": 2655322112
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001987878787878788,
      "loss": 1.1968,
      "theoretical_loss": 3.352457025978295,
      "tokens_seen": 2655387648
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019876767676767676,
      "loss": 1.2189,
      "theoretical_loss": 3.3524505056111007,
      "tokens_seen": 2655453184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019874747474747477,
      "loss": 1.1636,
      "theoretical_loss": 3.3524439854498826,
      "tokens_seen": 2655518720
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019872727272727273,
      "loss": 1.3451,
      "theoretical_loss": 3.352437465494629,
      "tokens_seen": 2655584256
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019870707070707068,
      "loss": 1.3645,
      "theoretical_loss": 3.352430945745328,
      "tokens_seen": 2655649792
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001986868686868687,
      "loss": 1.4773,
      "theoretical_loss": 3.352424426201968,
      "tokens_seen": 2655715328
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019866666666666665,
      "loss": 1.3068,
      "theoretical_loss": 3.3524179068645377,
      "tokens_seen": 2655780864
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019864646464646466,
      "loss": 1.2259,
      "theoretical_loss": 3.3524113877330253,
      "tokens_seen": 2655846400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019862626262626262,
      "loss": 1.3465,
      "theoretical_loss": 3.3524048688074193,
      "tokens_seen": 2655911936
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019860606060606063,
      "loss": 1.4528,
      "theoretical_loss": 3.352398350087708,
      "tokens_seen": 2655977472
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001985858585858586,
      "loss": 1.372,
      "theoretical_loss": 3.35239183157388,
      "tokens_seen": 2656043008
    },
    {
      "epoch": 0.61,
      "objective/train/aligned_token_count": 0.32551318407058716,
      "objective/train/avg_token_score": 0.014923589304089546,
      "objective/train/docs_used": 1495048,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0850740671157837,
      "objective/train/original_loss": 4.398007869720459,
      "objective/train/theoretical_loss": 3.352385313265924,
      "objective/train/tokens_used": 1014967776,
      "objective/train/unlikelihood_loss": 0.06202024221420288,
      "theoretical_loss": 3.352385313265924,
      "tokens_seen": 2656108544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019856565656565657,
      "loss": 1.3066,
      "theoretical_loss": 3.352385313265924,
      "tokens_seen": 2656108544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019854545454545456,
      "loss": 1.339,
      "theoretical_loss": 3.3523787951638275,
      "tokens_seen": 2656174080
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001985252525252525,
      "loss": 1.4425,
      "theoretical_loss": 3.35237227726758,
      "tokens_seen": 2656239616
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019850505050505052,
      "loss": 1.4923,
      "theoretical_loss": 3.352365759577169,
      "tokens_seen": 2656305152
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019848484848484848,
      "loss": 1.3137,
      "theoretical_loss": 3.352359242092583,
      "tokens_seen": 2656370688
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019846464646464647,
      "loss": 1.3345,
      "theoretical_loss": 3.3523527248138114,
      "tokens_seen": 2656436224
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019844444444444445,
      "loss": 1.3258,
      "theoretical_loss": 3.3523462077408417,
      "tokens_seen": 2656501760
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019842424242424243,
      "loss": 1.3943,
      "theoretical_loss": 3.352339690873663,
      "tokens_seen": 2656567296
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019840404040404042,
      "loss": 1.3717,
      "theoretical_loss": 3.352333174212263,
      "tokens_seen": 2656632832
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001983838383838384,
      "loss": 1.5515,
      "theoretical_loss": 3.35232665775663,
      "tokens_seen": 2656698368
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019836363636363636,
      "loss": 1.4439,
      "theoretical_loss": 3.3523201415067536,
      "tokens_seen": 2656763904
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019834343434343434,
      "loss": 1.3815,
      "theoretical_loss": 3.352313625462621,
      "tokens_seen": 2656829440
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019832323232323233,
      "loss": 1.4307,
      "theoretical_loss": 3.3523071096242214,
      "tokens_seen": 2656894976
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001983030303030303,
      "loss": 1.4253,
      "theoretical_loss": 3.352300593991543,
      "tokens_seen": 2656960512
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001982828282828283,
      "loss": 1.3121,
      "theoretical_loss": 3.352294078564574,
      "tokens_seen": 2657026048
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019826262626262625,
      "loss": 1.4152,
      "theoretical_loss": 3.352287563343303,
      "tokens_seen": 2657091584
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019824242424242426,
      "loss": 1.3349,
      "theoretical_loss": 3.352281048327719,
      "tokens_seen": 2657157120
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019822222222222222,
      "loss": 1.522,
      "theoretical_loss": 3.352274533517809,
      "tokens_seen": 2657222656
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019820202020202023,
      "loss": 1.2481,
      "theoretical_loss": 3.3522680189135627,
      "tokens_seen": 2657288192
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001981818181818182,
      "loss": 1.2869,
      "theoretical_loss": 3.3522615045149684,
      "tokens_seen": 2657353728
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019816161616161614,
      "loss": 1.2787,
      "theoretical_loss": 3.352254990322014,
      "tokens_seen": 2657419264
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019814141414141415,
      "loss": 1.237,
      "theoretical_loss": 3.3522484763346885,
      "tokens_seen": 2657484800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001981212121212121,
      "loss": 1.6149,
      "theoretical_loss": 3.35224196255298,
      "tokens_seen": 2657550336
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019810101010101012,
      "loss": 1.1922,
      "theoretical_loss": 3.352235448976877,
      "tokens_seen": 2657615872
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019808080808080808,
      "loss": 1.5716,
      "theoretical_loss": 3.3522289356063677,
      "tokens_seen": 2657681408
    },
    {
      "epoch": 0.61,
      "objective/train/aligned_token_count": 0.35551074147224426,
      "objective/train/avg_token_score": 0.02391725592315197,
      "objective/train/docs_used": 1495638,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1210973262786865,
      "objective/train/original_loss": 4.727067947387695,
      "objective/train/theoretical_loss": 3.3522224224414408,
      "objective/train/tokens_used": 1016606176,
      "objective/train/unlikelihood_loss": 0.05063033849000931,
      "theoretical_loss": 3.3522224224414408,
      "tokens_seen": 2657746944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019806060606060606,
      "loss": 1.3434,
      "theoretical_loss": 3.3522224224414408,
      "tokens_seen": 2657746944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019804040404040405,
      "loss": 1.2934,
      "theoretical_loss": 3.352215909482085,
      "tokens_seen": 2657812480
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000198020202020202,
      "loss": 1.1613,
      "theoretical_loss": 3.352209396728288,
      "tokens_seen": 2657878016
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019800000000000002,
      "loss": 1.267,
      "theoretical_loss": 3.3522028841800395,
      "tokens_seen": 2657943552
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019797979797979797,
      "loss": 1.5543,
      "theoretical_loss": 3.3521963718373264,
      "tokens_seen": 2658009088
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019795959595959596,
      "loss": 1.256,
      "theoretical_loss": 3.3521898597001387,
      "tokens_seen": 2658074624
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019793939393939394,
      "loss": 1.3131,
      "theoretical_loss": 3.3521833477684635,
      "tokens_seen": 2658140160
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019791919191919192,
      "loss": 1.3977,
      "theoretical_loss": 3.3521768360422897,
      "tokens_seen": 2658205696
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001978989898989899,
      "loss": 1.2327,
      "theoretical_loss": 3.352170324521606,
      "tokens_seen": 2658271232
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001978787878787879,
      "loss": 1.2777,
      "theoretical_loss": 3.352163813206401,
      "tokens_seen": 2658336768
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019785858585858585,
      "loss": 1.6722,
      "theoretical_loss": 3.3521573020966624,
      "tokens_seen": 2658402304
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019783838383838383,
      "loss": 1.5599,
      "theoretical_loss": 3.3521507911923796,
      "tokens_seen": 2658467840
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019781818181818182,
      "loss": 1.5112,
      "theoretical_loss": 3.3521442804935404,
      "tokens_seen": 2658533376
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001977979797979798,
      "loss": 1.2427,
      "theoretical_loss": 3.352137770000133,
      "tokens_seen": 2658598912
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019777777777777778,
      "loss": 1.1728,
      "theoretical_loss": 3.352131259712147,
      "tokens_seen": 2658664448
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019775757575757577,
      "loss": 1.3322,
      "theoretical_loss": 3.3521247496295694,
      "tokens_seen": 2658729984
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019773737373737375,
      "loss": 1.395,
      "theoretical_loss": 3.3521182397523894,
      "tokens_seen": 2658795520
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001977171717171717,
      "loss": 1.3658,
      "theoretical_loss": 3.352111730080596,
      "tokens_seen": 2658861056
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019769696969696972,
      "loss": 1.6082,
      "theoretical_loss": 3.3521052206141766,
      "tokens_seen": 2658926592
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019767676767676768,
      "loss": 1.4557,
      "theoretical_loss": 3.35209871135312,
      "tokens_seen": 2658992128
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019765656565656566,
      "loss": 1.2748,
      "theoretical_loss": 3.3520922022974156,
      "tokens_seen": 2659057664
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019763636363636365,
      "loss": 1.2838,
      "theoretical_loss": 3.3520856934470507,
      "tokens_seen": 2659123200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001976161616161616,
      "loss": 1.6863,
      "theoretical_loss": 3.352079184802014,
      "tokens_seen": 2659188736
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001975959595959596,
      "loss": 1.3185,
      "theoretical_loss": 3.352072676362294,
      "tokens_seen": 2659254272
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019757575757575757,
      "loss": 1.4559,
      "theoretical_loss": 3.3520661681278794,
      "tokens_seen": 2659319808
    },
    {
      "epoch": 0.61,
      "objective/train/aligned_token_count": 0.42375367879867554,
      "objective/train/avg_token_score": 0.007491173688322306,
      "objective/train/docs_used": 1497122,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2350248098373413,
      "objective/train/original_loss": 4.087658882141113,
      "objective/train/theoretical_loss": 3.3520596600987584,
      "objective/train/tokens_used": 1018244576,
      "objective/train/unlikelihood_loss": 0.05739041045308113,
      "theoretical_loss": 3.3520596600987584,
      "tokens_seen": 2659385344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019755555555555558,
      "loss": 1.4293,
      "theoretical_loss": 3.3520596600987584,
      "tokens_seen": 2659385344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019753535353535354,
      "loss": 1.2412,
      "theoretical_loss": 3.35205315227492,
      "tokens_seen": 2659450880
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019751515151515152,
      "loss": 1.6077,
      "theoretical_loss": 3.352046644656352,
      "tokens_seen": 2659516416
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001974949494949495,
      "loss": 1.3077,
      "theoretical_loss": 3.352040137243043,
      "tokens_seen": 2659581952
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019747474747474746,
      "loss": 1.4503,
      "theoretical_loss": 3.3520336300349816,
      "tokens_seen": 2659647488
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019745454545454547,
      "loss": 1.2452,
      "theoretical_loss": 3.3520271230321566,
      "tokens_seen": 2659713024
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019743434343434343,
      "loss": 1.5024,
      "theoretical_loss": 3.3520206162345554,
      "tokens_seen": 2659778560
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019741414141414141,
      "loss": 1.3175,
      "theoretical_loss": 3.352014109642168,
      "tokens_seen": 2659844096
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001973939393939394,
      "loss": 1.4991,
      "theoretical_loss": 3.3520076032549815,
      "tokens_seen": 2659909632
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019737373737373738,
      "loss": 1.3053,
      "theoretical_loss": 3.352001097072985,
      "tokens_seen": 2659975168
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019735353535353537,
      "loss": 1.3457,
      "theoretical_loss": 3.3519945910961675,
      "tokens_seen": 2660040704
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019733333333333335,
      "loss": 1.3447,
      "theoretical_loss": 3.3519880853245163,
      "tokens_seen": 2660106240
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001973131313131313,
      "loss": 1.3131,
      "theoretical_loss": 3.3519815797580206,
      "tokens_seen": 2660171776
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001972929292929293,
      "loss": 1.6132,
      "theoretical_loss": 3.3519750743966688,
      "tokens_seen": 2660237312
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019727272727272728,
      "loss": 1.5592,
      "theoretical_loss": 3.3519685692404493,
      "tokens_seen": 2660302848
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019725252525252526,
      "loss": 1.3037,
      "theoretical_loss": 3.3519620642893506,
      "tokens_seen": 2660368384
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019723232323232324,
      "loss": 1.4062,
      "theoretical_loss": 3.351955559543361,
      "tokens_seen": 2660433920
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001972121212121212,
      "loss": 1.5989,
      "theoretical_loss": 3.3519490550024695,
      "tokens_seen": 2660499456
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001971919191919192,
      "loss": 1.1182,
      "theoretical_loss": 3.3519425506666645,
      "tokens_seen": 2660564992
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019717171717171717,
      "loss": 1.513,
      "theoretical_loss": 3.3519360465359336,
      "tokens_seen": 2660630528
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019715151515151515,
      "loss": 1.4674,
      "theoretical_loss": 3.351929542610266,
      "tokens_seen": 2660696064
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019713131313131314,
      "loss": 1.4299,
      "theoretical_loss": 3.3519230388896504,
      "tokens_seen": 2660761600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001971111111111111,
      "loss": 1.5147,
      "theoretical_loss": 3.351916535374075,
      "tokens_seen": 2660827136
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001970909090909091,
      "loss": 1.3864,
      "theoretical_loss": 3.351910032063528,
      "tokens_seen": 2660892672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019707070707070706,
      "loss": 1.3559,
      "theoretical_loss": 3.351903528957998,
      "tokens_seen": 2660958208
    },
    {
      "epoch": 0.61,
      "objective/train/aligned_token_count": 0.30364125967025757,
      "objective/train/avg_token_score": 0.029923921450972557,
      "objective/train/docs_used": 1497834,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0159965753555298,
      "objective/train/original_loss": 5.4581451416015625,
      "objective/train/theoretical_loss": 3.3518970260574736,
      "objective/train/tokens_used": 1019882976,
      "objective/train/unlikelihood_loss": 0.02088344469666481,
      "theoretical_loss": 3.3518970260574736,
      "tokens_seen": 2661023744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019705050505050507,
      "loss": 1.3551,
      "theoretical_loss": 3.3518970260574736,
      "tokens_seen": 2661023744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019703030303030303,
      "loss": 1.459,
      "theoretical_loss": 3.351890523361944,
      "tokens_seen": 2661089280
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000197010101010101,
      "loss": 1.5369,
      "theoretical_loss": 3.3518840208713967,
      "tokens_seen": 2661154816
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000196989898989899,
      "loss": 1.453,
      "theoretical_loss": 3.35187751858582,
      "tokens_seen": 2661220352
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019696969696969695,
      "loss": 1.1912,
      "theoretical_loss": 3.3518710165052035,
      "tokens_seen": 2661285888
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019694949494949496,
      "loss": 1.4529,
      "theoretical_loss": 3.351864514629535,
      "tokens_seen": 2661351424
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019692929292929292,
      "loss": 1.3451,
      "theoretical_loss": 3.351858012958803,
      "tokens_seen": 2661416960
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019690909090909093,
      "loss": 1.281,
      "theoretical_loss": 3.351851511492996,
      "tokens_seen": 2661482496
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001968888888888889,
      "loss": 1.1243,
      "theoretical_loss": 3.3518450102321027,
      "tokens_seen": 2661548032
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019686868686868687,
      "loss": 1.4269,
      "theoretical_loss": 3.3518385091761114,
      "tokens_seen": 2661613568
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019684848484848486,
      "loss": 1.3171,
      "theoretical_loss": 3.3518320083250104,
      "tokens_seen": 2661679104
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019682828282828284,
      "loss": 1.2119,
      "theoretical_loss": 3.3518255076787886,
      "tokens_seen": 2661744640
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019680808080808082,
      "loss": 1.2874,
      "theoretical_loss": 3.3518190072374345,
      "tokens_seen": 2661810176
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019678787878787878,
      "loss": 1.4864,
      "theoretical_loss": 3.3518125070009366,
      "tokens_seen": 2661875712
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019676767676767677,
      "loss": 1.4496,
      "theoretical_loss": 3.351806006969283,
      "tokens_seen": 2661941248
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019674747474747475,
      "loss": 1.1961,
      "theoretical_loss": 3.351799507142463,
      "tokens_seen": 2662006784
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019672727272727273,
      "loss": 1.4395,
      "theoretical_loss": 3.351793007520464,
      "tokens_seen": 2662072320
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019670707070707072,
      "loss": 1.1861,
      "theoretical_loss": 3.3517865081032747,
      "tokens_seen": 2662137856
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001966868686868687,
      "loss": 1.4657,
      "theoretical_loss": 3.351780008890885,
      "tokens_seen": 2662203392
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019666666666666666,
      "loss": 1.351,
      "theoretical_loss": 3.351773509883282,
      "tokens_seen": 2662268928
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019664646464646467,
      "loss": 1.2518,
      "theoretical_loss": 3.351767011080454,
      "tokens_seen": 2662334464
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019662626262626263,
      "loss": 1.1438,
      "theoretical_loss": 3.3517605124823904,
      "tokens_seen": 2662400000
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001966060606060606,
      "loss": 1.4185,
      "theoretical_loss": 3.35175401408908,
      "tokens_seen": 2662465536
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001965858585858586,
      "loss": 1.2958,
      "theoretical_loss": 3.3517475159005103,
      "tokens_seen": 2662531072
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019656565656565655,
      "loss": 1.4583,
      "theoretical_loss": 3.35174101791667,
      "tokens_seen": 2662596608
    },
    {
      "epoch": 0.61,
      "objective/train/aligned_token_count": 0.4177052676677704,
      "objective/train/avg_token_score": 0.011851279065012932,
      "objective/train/docs_used": 1499033,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3157339096069336,
      "objective/train/original_loss": 3.7377519607543945,
      "objective/train/theoretical_loss": 3.351734520137548,
      "objective/train/tokens_used": 1021521376,
      "objective/train/unlikelihood_loss": 0.08110038936138153,
      "theoretical_loss": 3.351734520137548,
      "tokens_seen": 2662662144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019654545454545456,
      "loss": 1.376,
      "theoretical_loss": 3.351734520137548,
      "tokens_seen": 2662662144
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019652525252525252,
      "loss": 1.2347,
      "theoretical_loss": 3.3517280225631327,
      "tokens_seen": 2662727680
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019650505050505053,
      "loss": 1.2999,
      "theoretical_loss": 3.3517215251934123,
      "tokens_seen": 2662793216
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001964848484848485,
      "loss": 1.4907,
      "theoretical_loss": 3.351715028028376,
      "tokens_seen": 2662858752
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019646464646464644,
      "loss": 1.3177,
      "theoretical_loss": 3.3517085310680117,
      "tokens_seen": 2662924288
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019644444444444445,
      "loss": 1.2321,
      "theoretical_loss": 3.351702034312308,
      "tokens_seen": 2662989824
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001964242424242424,
      "loss": 1.2027,
      "theoretical_loss": 3.3516955377612536,
      "tokens_seen": 2663055360
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019640404040404042,
      "loss": 1.3741,
      "theoretical_loss": 3.351689041414837,
      "tokens_seen": 2663120896
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019638383838383838,
      "loss": 1.3575,
      "theoretical_loss": 3.3516825452730465,
      "tokens_seen": 2663186432
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019636363636363636,
      "loss": 1.4228,
      "theoretical_loss": 3.351676049335871,
      "tokens_seen": 2663251968
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019634343434343435,
      "loss": 1.1792,
      "theoretical_loss": 3.3516695536032985,
      "tokens_seen": 2663317504
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019632323232323233,
      "loss": 1.4604,
      "theoretical_loss": 3.351663058075318,
      "tokens_seen": 2663383040
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019630303030303032,
      "loss": 1.5346,
      "theoretical_loss": 3.3516565627519177,
      "tokens_seen": 2663448576
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019628282828282827,
      "loss": 1.3814,
      "theoretical_loss": 3.3516500676330865,
      "tokens_seen": 2663514112
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019626262626262626,
      "loss": 1.5084,
      "theoretical_loss": 3.3516435727188125,
      "tokens_seen": 2663579648
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019624242424242424,
      "loss": 1.3449,
      "theoretical_loss": 3.3516370780090847,
      "tokens_seen": 2663645184
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019622222222222222,
      "loss": 1.3693,
      "theoretical_loss": 3.3516305835038906,
      "tokens_seen": 2663710720
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001962020202020202,
      "loss": 1.4276,
      "theoretical_loss": 3.35162408920322,
      "tokens_seen": 2663776256
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001961818181818182,
      "loss": 1.3559,
      "theoretical_loss": 3.351617595107061,
      "tokens_seen": 2663841792
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019616161616161615,
      "loss": 1.3908,
      "theoretical_loss": 3.351611101215402,
      "tokens_seen": 2663907328
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019614141414141416,
      "loss": 1.4,
      "theoretical_loss": 3.3516046075282313,
      "tokens_seen": 2663972864
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019612121212121212,
      "loss": 1.2966,
      "theoretical_loss": 3.3515981140455375,
      "tokens_seen": 2664038400
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001961010101010101,
      "loss": 1.2596,
      "theoretical_loss": 3.35159162076731,
      "tokens_seen": 2664103936
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019608080808080808,
      "loss": 1.25,
      "theoretical_loss": 3.3515851276935362,
      "tokens_seen": 2664169472
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019606060606060607,
      "loss": 1.2331,
      "theoretical_loss": 3.351578634824205,
      "tokens_seen": 2664235008
    },
    {
      "epoch": 0.61,
      "objective/train/aligned_token_count": 0.353128045797348,
      "objective/train/avg_token_score": 0.03188995271921158,
      "objective/train/docs_used": 1499572,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0189653635025024,
      "objective/train/original_loss": 4.6517229080200195,
      "objective/train/theoretical_loss": 3.351572142159305,
      "objective/train/tokens_used": 1023159776,
      "objective/train/unlikelihood_loss": 0.04986146464943886,
      "theoretical_loss": 3.351572142159305,
      "tokens_seen": 2664300544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019604040404040405,
      "loss": 1.3044,
      "theoretical_loss": 3.351572142159305,
      "tokens_seen": 2664300544
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000196020202020202,
      "loss": 1.4979,
      "theoretical_loss": 3.351565649698825,
      "tokens_seen": 2664366080
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019600000000000002,
      "loss": 1.35,
      "theoretical_loss": 3.351559157442753,
      "tokens_seen": 2664431616
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019597979797979798,
      "loss": 1.4597,
      "theoretical_loss": 3.351552665391078,
      "tokens_seen": 2664497152
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000195959595959596,
      "loss": 1.5107,
      "theoretical_loss": 3.351546173543788,
      "tokens_seen": 2664562688
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019593939393939395,
      "loss": 1.5328,
      "theoretical_loss": 3.3515396819008725,
      "tokens_seen": 2664628224
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001959191919191919,
      "loss": 1.3207,
      "theoretical_loss": 3.351533190462319,
      "tokens_seen": 2664693760
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001958989898989899,
      "loss": 1.459,
      "theoretical_loss": 3.351526699228117,
      "tokens_seen": 2664759296
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019587878787878787,
      "loss": 1.324,
      "theoretical_loss": 3.351520208198254,
      "tokens_seen": 2664824832
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019585858585858588,
      "loss": 1.4229,
      "theoretical_loss": 3.351513717372719,
      "tokens_seen": 2664890368
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019583838383838384,
      "loss": 1.4196,
      "theoretical_loss": 3.351507226751501,
      "tokens_seen": 2664955904
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019581818181818182,
      "loss": 1.4638,
      "theoretical_loss": 3.3515007363345877,
      "tokens_seen": 2665021440
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001957979797979798,
      "loss": 1.2814,
      "theoretical_loss": 3.3514942461219683,
      "tokens_seen": 2665086976
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001957777777777778,
      "loss": 1.43,
      "theoretical_loss": 3.351487756113631,
      "tokens_seen": 2665152512
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019575757575757577,
      "loss": 1.4307,
      "theoretical_loss": 3.3514812663095648,
      "tokens_seen": 2665218048
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019573737373737373,
      "loss": 1.4634,
      "theoretical_loss": 3.3514747767097575,
      "tokens_seen": 2665283584
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019571717171717171,
      "loss": 1.5704,
      "theoretical_loss": 3.3514682873141988,
      "tokens_seen": 2665349120
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001956969696969697,
      "loss": 1.2056,
      "theoretical_loss": 3.351461798122876,
      "tokens_seen": 2665414656
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019567676767676768,
      "loss": 1.4683,
      "theoretical_loss": 3.351455309135778,
      "tokens_seen": 2665480192
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019565656565656567,
      "loss": 1.3966,
      "theoretical_loss": 3.351448820352894,
      "tokens_seen": 2665545728
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019563636363636365,
      "loss": 1.2821,
      "theoretical_loss": 3.3514423317742117,
      "tokens_seen": 2665611264
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001956161616161616,
      "loss": 1.1198,
      "theoretical_loss": 3.3514358433997202,
      "tokens_seen": 2665676800
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001955959595959596,
      "loss": 1.1784,
      "theoretical_loss": 3.351429355229408,
      "tokens_seen": 2665742336
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019557575757575758,
      "loss": 1.4678,
      "theoretical_loss": 3.3514228672632633,
      "tokens_seen": 2665807872
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019555555555555556,
      "loss": 1.1976,
      "theoretical_loss": 3.3514163795012752,
      "tokens_seen": 2665873408
    },
    {
      "epoch": 0.61,
      "objective/train/aligned_token_count": 0.22922776639461517,
      "objective/train/avg_token_score": 0.018859460949897766,
      "objective/train/docs_used": 1500816,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.7384244799613953,
      "objective/train/original_loss": 4.877727031707764,
      "objective/train/theoretical_loss": 3.351409891943432,
      "objective/train/tokens_used": 1024798176,
      "objective/train/unlikelihood_loss": 0.04011557996273041,
      "theoretical_loss": 3.351409891943432,
      "tokens_seen": 2665938944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019553535353535354,
      "loss": 1.1928,
      "theoretical_loss": 3.351409891943432,
      "tokens_seen": 2665938944
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001955151515151515,
      "loss": 1.5273,
      "theoretical_loss": 3.3514034045897216,
      "tokens_seen": 2666004480
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001954949494949495,
      "loss": 1.2079,
      "theoretical_loss": 3.351396917440134,
      "tokens_seen": 2666070016
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019547474747474747,
      "loss": 1.2384,
      "theoretical_loss": 3.3513904304946562,
      "tokens_seen": 2666135552
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019545454545454548,
      "loss": 1.3233,
      "theoretical_loss": 3.351383943753278,
      "tokens_seen": 2666201088
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019543434343434344,
      "loss": 1.3573,
      "theoretical_loss": 3.351377457215987,
      "tokens_seen": 2666266624
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001954141414141414,
      "loss": 1.6283,
      "theoretical_loss": 3.351370970882773,
      "tokens_seen": 2666332160
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001953939393939394,
      "loss": 1.2702,
      "theoretical_loss": 3.3513644847536233,
      "tokens_seen": 2666397696
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019537373737373736,
      "loss": 1.4154,
      "theoretical_loss": 3.3513579988285267,
      "tokens_seen": 2666463232
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019535353535353537,
      "loss": 1.5106,
      "theoretical_loss": 3.3513515131074723,
      "tokens_seen": 2666528768
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019533333333333333,
      "loss": 1.383,
      "theoretical_loss": 3.3513450275904484,
      "tokens_seen": 2666594304
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001953131313131313,
      "loss": 1.5987,
      "theoretical_loss": 3.3513385422774435,
      "tokens_seen": 2666659840
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001952929292929293,
      "loss": 1.5387,
      "theoretical_loss": 3.3513320571684466,
      "tokens_seen": 2666725376
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019527272727272728,
      "loss": 1.2949,
      "theoretical_loss": 3.3513255722634456,
      "tokens_seen": 2666790912
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019525252525252526,
      "loss": 1.3473,
      "theoretical_loss": 3.351319087562429,
      "tokens_seen": 2666856448
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019523232323232322,
      "loss": 1.6141,
      "theoretical_loss": 3.351312603065386,
      "tokens_seen": 2666921984
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019521212121212123,
      "loss": 1.5793,
      "theoretical_loss": 3.351306118772305,
      "tokens_seen": 2666987520
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001951919191919192,
      "loss": 1.5745,
      "theoretical_loss": 3.3512996346831745,
      "tokens_seen": 2667053056
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019517171717171717,
      "loss": 1.403,
      "theoretical_loss": 3.351293150797983,
      "tokens_seen": 2667118592
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019515151515151516,
      "loss": 1.2199,
      "theoretical_loss": 3.351286667116719,
      "tokens_seen": 2667184128
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019513131313131314,
      "loss": 1.2129,
      "theoretical_loss": 3.3512801836393713,
      "tokens_seen": 2667249664
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019511111111111113,
      "loss": 1.3733,
      "theoretical_loss": 3.3512737003659288,
      "tokens_seen": 2667315200
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001950909090909091,
      "loss": 1.2798,
      "theoretical_loss": 3.351267217296379,
      "tokens_seen": 2667380736
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019507070707070707,
      "loss": 1.2316,
      "theoretical_loss": 3.3512607344307113,
      "tokens_seen": 2667446272
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019505050505050505,
      "loss": 1.1807,
      "theoretical_loss": 3.351254251768914,
      "tokens_seen": 2667511808
    },
    {
      "epoch": 0.61,
      "objective/train/aligned_token_count": 0.4104349911212921,
      "objective/train/avg_token_score": 0.016366634517908096,
      "objective/train/docs_used": 1501537,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.317261815071106,
      "objective/train/original_loss": 4.390142917633057,
      "objective/train/theoretical_loss": 3.351247769310976,
      "objective/train/tokens_used": 1026436576,
      "objective/train/unlikelihood_loss": 0.04805602878332138,
      "theoretical_loss": 3.351247769310976,
      "tokens_seen": 2667577344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019503030303030303,
      "loss": 1.3404,
      "theoretical_loss": 3.351247769310976,
      "tokens_seen": 2667577344
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019501010101010102,
      "loss": 1.3312,
      "theoretical_loss": 3.3512412870568857,
      "tokens_seen": 2667642880
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.000194989898989899,
      "loss": 1.4689,
      "theoretical_loss": 3.351234805006632,
      "tokens_seen": 2667708416
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019496969696969696,
      "loss": 1.4963,
      "theoretical_loss": 3.3512283231602025,
      "tokens_seen": 2667773952
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019494949494949497,
      "loss": 1.2474,
      "theoretical_loss": 3.3512218415175865,
      "tokens_seen": 2667839488
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019492929292929293,
      "loss": 1.3488,
      "theoretical_loss": 3.351215360078773,
      "tokens_seen": 2667905024
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019490909090909094,
      "loss": 1.3565,
      "theoretical_loss": 3.3512088788437495,
      "tokens_seen": 2667970560
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001948888888888889,
      "loss": 1.2401,
      "theoretical_loss": 3.3512023978125054,
      "tokens_seen": 2668036096
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019486868686868685,
      "loss": 1.3676,
      "theoretical_loss": 3.351195916985029,
      "tokens_seen": 2668101632
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019484848484848486,
      "loss": 1.2613,
      "theoretical_loss": 3.351189436361309,
      "tokens_seen": 2668167168
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019482828282828282,
      "loss": 1.1748,
      "theoretical_loss": 3.351182955941334,
      "tokens_seen": 2668232704
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019480808080808083,
      "loss": 1.4457,
      "theoretical_loss": 3.351176475725093,
      "tokens_seen": 2668298240
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001947878787878788,
      "loss": 1.3374,
      "theoretical_loss": 3.3511699957125733,
      "tokens_seen": 2668363776
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019476767676767677,
      "loss": 1.2553,
      "theoretical_loss": 3.3511635159037647,
      "tokens_seen": 2668429312
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019474747474747476,
      "loss": 0.9898,
      "theoretical_loss": 3.3511570362986554,
      "tokens_seen": 2668494848
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001947272727272727,
      "loss": 1.2747,
      "theoretical_loss": 3.3511505568972337,
      "tokens_seen": 2668560384
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019470707070707072,
      "loss": 1.5521,
      "theoretical_loss": 3.351144077699489,
      "tokens_seen": 2668625920
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019468686868686868,
      "loss": 1.4227,
      "theoretical_loss": 3.3511375987054093,
      "tokens_seen": 2668691456
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019466666666666666,
      "loss": 1.4088,
      "theoretical_loss": 3.3511311199149834,
      "tokens_seen": 2668756992
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019464646464646465,
      "loss": 1.3836,
      "theoretical_loss": 3.3511246413281994,
      "tokens_seen": 2668822528
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019462626262626263,
      "loss": 1.3357,
      "theoretical_loss": 3.3511181629450464,
      "tokens_seen": 2668888064
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019460606060606062,
      "loss": 1.1736,
      "theoretical_loss": 3.351111684765513,
      "tokens_seen": 2668953600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0001945858585858586,
      "loss": 1.3433,
      "theoretical_loss": 3.3511052067895877,
      "tokens_seen": 2669019136
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019456565656565656,
      "loss": 1.3618,
      "theoretical_loss": 3.351098729017259,
      "tokens_seen": 2669084672
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019454545454545454,
      "loss": 1.5368,
      "theoretical_loss": 3.3510922514485157,
      "tokens_seen": 2669150208
    },
    {
      "epoch": 0.61,
      "objective/train/aligned_token_count": 0.3608871102333069,
      "objective/train/avg_token_score": 0.023790957406163216,
      "objective/train/docs_used": 1503056,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2031161785125732,
      "objective/train/original_loss": 4.928018569946289,
      "objective/train/theoretical_loss": 3.351085774083346,
      "objective/train/tokens_used": 1028074976,
      "objective/train/unlikelihood_loss": 0.04655122384428978,
      "theoretical_loss": 3.351085774083346,
      "tokens_seen": 2669215744
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00019452525252525252,
      "loss": 1.4552,
      "theoretical_loss": 3.351085774083346,
      "tokens_seen": 2669215744
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001945050505050505,
      "loss": 1.5429,
      "theoretical_loss": 3.351079296921739,
      "tokens_seen": 2669281280
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001944848484848485,
      "loss": 1.5497,
      "theoretical_loss": 3.3510728199636834,
      "tokens_seen": 2669346816
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019446464646464645,
      "loss": 1.1726,
      "theoretical_loss": 3.351066343209167,
      "tokens_seen": 2669412352
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019444444444444446,
      "loss": 1.3738,
      "theoretical_loss": 3.3510598666581792,
      "tokens_seen": 2669477888
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019442424242424242,
      "loss": 1.2526,
      "theoretical_loss": 3.3510533903107085,
      "tokens_seen": 2669543424
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019440404040404043,
      "loss": 1.3464,
      "theoretical_loss": 3.3510469141667434,
      "tokens_seen": 2669608960
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019438383838383839,
      "loss": 1.3446,
      "theoretical_loss": 3.351040438226272,
      "tokens_seen": 2669674496
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019436363636363637,
      "loss": 1.3041,
      "theoretical_loss": 3.3510339624892835,
      "tokens_seen": 2669740032
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019434343434343435,
      "loss": 1.3364,
      "theoretical_loss": 3.3510274869557666,
      "tokens_seen": 2669805568
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001943232323232323,
      "loss": 1.5013,
      "theoretical_loss": 3.3510210116257095,
      "tokens_seen": 2669871104
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019430303030303032,
      "loss": 1.4984,
      "theoretical_loss": 3.351014536499101,
      "tokens_seen": 2669936640
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019428282828282828,
      "loss": 1.3332,
      "theoretical_loss": 3.3510080615759295,
      "tokens_seen": 2670002176
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001942626262626263,
      "loss": 1.1197,
      "theoretical_loss": 3.351001586856184,
      "tokens_seen": 2670067712
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019424242424242425,
      "loss": 1.2679,
      "theoretical_loss": 3.350995112339853,
      "tokens_seen": 2670133248
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019422222222222223,
      "loss": 1.092,
      "theoretical_loss": 3.3509886380269247,
      "tokens_seen": 2670198784
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019420202020202021,
      "loss": 1.2429,
      "theoretical_loss": 3.3509821639173887,
      "tokens_seen": 2670264320
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019418181818181817,
      "loss": 1.1345,
      "theoretical_loss": 3.3509756900112326,
      "tokens_seen": 2670329856
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019416161616161618,
      "loss": 1.2312,
      "theoretical_loss": 3.3509692163084455,
      "tokens_seen": 2670395392
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019414141414141414,
      "loss": 1.1492,
      "theoretical_loss": 3.350962742809016,
      "tokens_seen": 2670460928
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019412121212121212,
      "loss": 1.4894,
      "theoretical_loss": 3.3509562695129325,
      "tokens_seen": 2670526464
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001941010101010101,
      "loss": 1.551,
      "theoretical_loss": 3.350949796420184,
      "tokens_seen": 2670592000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001940808080808081,
      "loss": 1.2106,
      "theoretical_loss": 3.3509433235307586,
      "tokens_seen": 2670657536
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019406060606060607,
      "loss": 1.2727,
      "theoretical_loss": 3.3509368508446453,
      "tokens_seen": 2670723072
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019404040404040406,
      "loss": 1.2901,
      "theoretical_loss": 3.350930378361833,
      "tokens_seen": 2670788608
    },
    {
      "epoch": 0.62,
      "objective/train/aligned_token_count": 0.4226539731025696,
      "objective/train/avg_token_score": 0.00859240535646677,
      "objective/train/docs_used": 1503662,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2434051036834717,
      "objective/train/original_loss": 4.143040180206299,
      "objective/train/theoretical_loss": 3.3509239060823095,
      "objective/train/tokens_used": 1029713376,
      "objective/train/unlikelihood_loss": 0.058222685009241104,
      "theoretical_loss": 3.3509239060823095,
      "tokens_seen": 2670854144
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019402020202020202,
      "loss": 1.4397,
      "theoretical_loss": 3.3509239060823095,
      "tokens_seen": 2670854144
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000194,
      "loss": 1.4156,
      "theoretical_loss": 3.3509174340060643,
      "tokens_seen": 2670919680
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019397979797979798,
      "loss": 1.23,
      "theoretical_loss": 3.3509109621330855,
      "tokens_seen": 2670985216
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019395959595959597,
      "loss": 1.4047,
      "theoretical_loss": 3.350904490463362,
      "tokens_seen": 2671050752
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019393939393939395,
      "loss": 1.3174,
      "theoretical_loss": 3.3508980189968818,
      "tokens_seen": 2671116288
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001939191919191919,
      "loss": 1.3716,
      "theoretical_loss": 3.3508915477336343,
      "tokens_seen": 2671181824
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019389898989898992,
      "loss": 1.4195,
      "theoretical_loss": 3.350885076673608,
      "tokens_seen": 2671247360
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019387878787878788,
      "loss": 1.2076,
      "theoretical_loss": 3.350878605816791,
      "tokens_seen": 2671312896
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019385858585858586,
      "loss": 1.2259,
      "theoretical_loss": 3.350872135163173,
      "tokens_seen": 2671378432
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019383838383838384,
      "loss": 1.4415,
      "theoretical_loss": 3.3508656647127415,
      "tokens_seen": 2671443968
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001938181818181818,
      "loss": 1.3035,
      "theoretical_loss": 3.3508591944654853,
      "tokens_seen": 2671509504
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001937979797979798,
      "loss": 1.3799,
      "theoretical_loss": 3.3508527244213937,
      "tokens_seen": 2671575040
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019377777777777777,
      "loss": 1.3652,
      "theoretical_loss": 3.350846254580455,
      "tokens_seen": 2671640576
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019375757575757578,
      "loss": 1.6323,
      "theoretical_loss": 3.3508397849426577,
      "tokens_seen": 2671706112
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019373737373737374,
      "loss": 1.4007,
      "theoretical_loss": 3.3508333155079906,
      "tokens_seen": 2671771648
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019371717171717172,
      "loss": 1.374,
      "theoretical_loss": 3.3508268462764423,
      "tokens_seen": 2671837184
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001936969696969697,
      "loss": 1.1384,
      "theoretical_loss": 3.3508203772480014,
      "tokens_seen": 2671902720
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019367676767676766,
      "loss": 1.2559,
      "theoretical_loss": 3.3508139084226563,
      "tokens_seen": 2671968256
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019365656565656567,
      "loss": 1.4753,
      "theoretical_loss": 3.350807439800396,
      "tokens_seen": 2672033792
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019363636363636363,
      "loss": 1.4767,
      "theoretical_loss": 3.3508009713812092,
      "tokens_seen": 2672099328
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001936161616161616,
      "loss": 1.264,
      "theoretical_loss": 3.3507945031650843,
      "tokens_seen": 2672164864
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001935959595959596,
      "loss": 1.4898,
      "theoretical_loss": 3.3507880351520103,
      "tokens_seen": 2672230400
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019357575757575758,
      "loss": 1.2806,
      "theoretical_loss": 3.3507815673419756,
      "tokens_seen": 2672295936
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019355555555555557,
      "loss": 1.345,
      "theoretical_loss": 3.3507750997349683,
      "tokens_seen": 2672361472
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019353535353535355,
      "loss": 1.3031,
      "theoretical_loss": 3.350768632330978,
      "tokens_seen": 2672427008
    },
    {
      "epoch": 0.62,
      "objective/train/aligned_token_count": 0.36644673347473145,
      "objective/train/avg_token_score": 0.023111937567591667,
      "objective/train/docs_used": 1505160,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.139733076095581,
      "objective/train/original_loss": 4.362312316894531,
      "objective/train/theoretical_loss": 3.350762165129993,
      "objective/train/tokens_used": 1031351776,
      "objective/train/unlikelihood_loss": 0.06176440045237541,
      "theoretical_loss": 3.350762165129993,
      "tokens_seen": 2672492544
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019351515151515153,
      "loss": 1.2803,
      "theoretical_loss": 3.350762165129993,
      "tokens_seen": 2672492544
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001934949494949495,
      "loss": 1.3552,
      "theoretical_loss": 3.3507556981320015,
      "tokens_seen": 2672558080
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019347474747474747,
      "loss": 1.5288,
      "theoretical_loss": 3.3507492313369927,
      "tokens_seen": 2672623616
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019345454545454546,
      "loss": 1.2331,
      "theoretical_loss": 3.3507427647449552,
      "tokens_seen": 2672689152
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019343434343434344,
      "loss": 1.2554,
      "theoretical_loss": 3.3507362983558773,
      "tokens_seen": 2672754688
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019341414141414143,
      "loss": 1.2677,
      "theoretical_loss": 3.350729832169748,
      "tokens_seen": 2672820224
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001933939393939394,
      "loss": 1.3359,
      "theoretical_loss": 3.350723366186556,
      "tokens_seen": 2672885760
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019337373737373737,
      "loss": 1.0023,
      "theoretical_loss": 3.3507169004062893,
      "tokens_seen": 2672951296
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019335353535353538,
      "loss": 1.4549,
      "theoretical_loss": 3.3507104348289376,
      "tokens_seen": 2673016832
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019333333333333333,
      "loss": 1.4354,
      "theoretical_loss": 3.350703969454489,
      "tokens_seen": 2673082368
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019331313131313132,
      "loss": 1.3798,
      "theoretical_loss": 3.350697504282932,
      "tokens_seen": 2673147904
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001932929292929293,
      "loss": 1.5247,
      "theoretical_loss": 3.3506910393142553,
      "tokens_seen": 2673213440
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019327272727272726,
      "loss": 1.5601,
      "theoretical_loss": 3.3506845745484477,
      "tokens_seen": 2673278976
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019325252525252527,
      "loss": 1.4069,
      "theoretical_loss": 3.3506781099854983,
      "tokens_seen": 2673344512
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019323232323232323,
      "loss": 1.3612,
      "theoretical_loss": 3.3506716456253947,
      "tokens_seen": 2673410048
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019321212121212124,
      "loss": 1.3823,
      "theoretical_loss": 3.3506651814681265,
      "tokens_seen": 2673475584
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001931919191919192,
      "loss": 1.2857,
      "theoretical_loss": 3.350658717513682,
      "tokens_seen": 2673541120
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019317171717171715,
      "loss": 1.4113,
      "theoretical_loss": 3.35065225376205,
      "tokens_seen": 2673606656
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019315151515151516,
      "loss": 1.4927,
      "theoretical_loss": 3.3506457902132185,
      "tokens_seen": 2673672192
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019313131313131312,
      "loss": 1.2804,
      "theoretical_loss": 3.3506393268671775,
      "tokens_seen": 2673737728
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019311111111111113,
      "loss": 1.3814,
      "theoretical_loss": 3.3506328637239147,
      "tokens_seen": 2673803264
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001930909090909091,
      "loss": 1.4891,
      "theoretical_loss": 3.3506264007834186,
      "tokens_seen": 2673868800
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019307070707070707,
      "loss": 1.5678,
      "theoretical_loss": 3.3506199380456785,
      "tokens_seen": 2673934336
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019305050505050506,
      "loss": 1.4234,
      "theoretical_loss": 3.350613475510683,
      "tokens_seen": 2673999872
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019303030303030304,
      "loss": 1.286,
      "theoretical_loss": 3.3506070131784202,
      "tokens_seen": 2674065408
    },
    {
      "epoch": 0.62,
      "objective/train/aligned_token_count": 0.507331371307373,
      "objective/train/avg_token_score": 0.010844988748431206,
      "objective/train/docs_used": 1505818,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6578856706619263,
      "objective/train/original_loss": 3.4896798133850098,
      "objective/train/theoretical_loss": 3.3506005510488794,
      "objective/train/tokens_used": 1032990176,
      "objective/train/unlikelihood_loss": 0.08209670335054398,
      "theoretical_loss": 3.3506005510488794,
      "tokens_seen": 2674130944
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019301010101010102,
      "loss": 1.4347,
      "theoretical_loss": 3.3506005510488794,
      "tokens_seen": 2674130944
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019298989898989898,
      "loss": 1.0281,
      "theoretical_loss": 3.350594089122049,
      "tokens_seen": 2674196480
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019296969696969696,
      "loss": 1.3033,
      "theoretical_loss": 3.3505876273979176,
      "tokens_seen": 2674262016
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019294949494949495,
      "loss": 1.4453,
      "theoretical_loss": 3.350581165876474,
      "tokens_seen": 2674327552
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019292929292929293,
      "loss": 1.3046,
      "theoretical_loss": 3.350574704557707,
      "tokens_seen": 2674393088
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019290909090909092,
      "loss": 1.2362,
      "theoretical_loss": 3.3505682434416046,
      "tokens_seen": 2674458624
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001928888888888889,
      "loss": 1.4839,
      "theoretical_loss": 3.3505617825281564,
      "tokens_seen": 2674524160
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019286868686868686,
      "loss": 1.649,
      "theoretical_loss": 3.3505553218173505,
      "tokens_seen": 2674589696
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019284848484848487,
      "loss": 1.5372,
      "theoretical_loss": 3.350548861309176,
      "tokens_seen": 2674655232
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019282828282828283,
      "loss": 1.2472,
      "theoretical_loss": 3.3505424010036213,
      "tokens_seen": 2674720768
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001928080808080808,
      "loss": 1.3649,
      "theoretical_loss": 3.350535940900675,
      "tokens_seen": 2674786304
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001927878787878788,
      "loss": 1.3186,
      "theoretical_loss": 3.350529481000326,
      "tokens_seen": 2674851840
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019276767676767675,
      "loss": 1.3018,
      "theoretical_loss": 3.3505230213025627,
      "tokens_seen": 2674917376
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019274747474747476,
      "loss": 1.2125,
      "theoretical_loss": 3.3505165618073742,
      "tokens_seen": 2674982912
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019272727272727272,
      "loss": 1.3061,
      "theoretical_loss": 3.350510102514749,
      "tokens_seen": 2675048448
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019270707070707073,
      "loss": 1.4823,
      "theoretical_loss": 3.3505036434246755,
      "tokens_seen": 2675113984
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019268686868686869,
      "loss": 1.1613,
      "theoretical_loss": 3.3504971845371427,
      "tokens_seen": 2675179520
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001926666666666667,
      "loss": 1.4579,
      "theoretical_loss": 3.350490725852139,
      "tokens_seen": 2675245056
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019264646464646465,
      "loss": 1.1617,
      "theoretical_loss": 3.3504842673696538,
      "tokens_seen": 2675310592
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001926262626262626,
      "loss": 1.4988,
      "theoretical_loss": 3.350477809089675,
      "tokens_seen": 2675376128
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019260606060606062,
      "loss": 1.5052,
      "theoretical_loss": 3.3504713510121915,
      "tokens_seen": 2675441664
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019258585858585858,
      "loss": 1.4358,
      "theoretical_loss": 3.3504648931371923,
      "tokens_seen": 2675507200
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001925656565656566,
      "loss": 1.4311,
      "theoretical_loss": 3.350458435464666,
      "tokens_seen": 2675572736
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019254545454545455,
      "loss": 1.3512,
      "theoretical_loss": 3.3504519779946005,
      "tokens_seen": 2675638272
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019252525252525253,
      "loss": 1.4274,
      "theoretical_loss": 3.3504455207269856,
      "tokens_seen": 2675703808
    },
    {
      "epoch": 0.62,
      "objective/train/aligned_token_count": 0.41055718064308167,
      "objective/train/avg_token_score": 0.01150602102279663,
      "objective/train/docs_used": 1507167,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.271159291267395,
      "objective/train/original_loss": 4.6576337814331055,
      "objective/train/theoretical_loss": 3.3504390636618093,
      "objective/train/tokens_used": 1034628576,
      "objective/train/unlikelihood_loss": 0.03692620247602463,
      "theoretical_loss": 3.3504390636618093,
      "tokens_seen": 2675769344
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019250505050505051,
      "loss": 1.395,
      "theoretical_loss": 3.3504390636618093,
      "tokens_seen": 2675769344
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001924848484848485,
      "loss": 1.0912,
      "theoretical_loss": 3.350432606799061,
      "tokens_seen": 2675834880
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019246464646464648,
      "loss": 1.3583,
      "theoretical_loss": 3.3504261501387282,
      "tokens_seen": 2675900416
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019244444444444444,
      "loss": 1.4751,
      "theoretical_loss": 3.350419693680801,
      "tokens_seen": 2675965952
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019242424242424242,
      "loss": 1.477,
      "theoretical_loss": 3.350413237425267,
      "tokens_seen": 2676031488
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001924040404040404,
      "loss": 1.5703,
      "theoretical_loss": 3.350406781372116,
      "tokens_seen": 2676097024
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001923838383838384,
      "loss": 1.484,
      "theoretical_loss": 3.3504003255213353,
      "tokens_seen": 2676162560
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019236363636363637,
      "loss": 1.4811,
      "theoretical_loss": 3.3503938698729145,
      "tokens_seen": 2676228096
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019234343434343436,
      "loss": 1.4763,
      "theoretical_loss": 3.350387414426842,
      "tokens_seen": 2676293632
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019232323232323232,
      "loss": 1.3158,
      "theoretical_loss": 3.350380959183107,
      "tokens_seen": 2676359168
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001923030303030303,
      "loss": 1.4673,
      "theoretical_loss": 3.3503745041416977,
      "tokens_seen": 2676424704
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019228282828282828,
      "loss": 1.414,
      "theoretical_loss": 3.350368049302603,
      "tokens_seen": 2676490240
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019226262626262627,
      "loss": 1.2193,
      "theoretical_loss": 3.350361594665811,
      "tokens_seen": 2676555776
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019224242424242425,
      "loss": 1.3007,
      "theoretical_loss": 3.350355140231312,
      "tokens_seen": 2676621312
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001922222222222222,
      "loss": 1.2928,
      "theoretical_loss": 3.3503486859990925,
      "tokens_seen": 2676686848
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019220202020202022,
      "loss": 1.354,
      "theoretical_loss": 3.350342231969143,
      "tokens_seen": 2676752384
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019218181818181818,
      "loss": 1.0806,
      "theoretical_loss": 3.3503357781414516,
      "tokens_seen": 2676817920
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001921616161616162,
      "loss": 1.3161,
      "theoretical_loss": 3.350329324516007,
      "tokens_seen": 2676883456
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019214141414141414,
      "loss": 1.4682,
      "theoretical_loss": 3.3503228710927977,
      "tokens_seen": 2676948992
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001921212121212121,
      "loss": 1.2768,
      "theoretical_loss": 3.350316417871813,
      "tokens_seen": 2677014528
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001921010101010101,
      "loss": 1.4669,
      "theoretical_loss": 3.350309964853041,
      "tokens_seen": 2677080064
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019208080808080807,
      "loss": 1.2579,
      "theoretical_loss": 3.3503035120364704,
      "tokens_seen": 2677145600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019206060606060608,
      "loss": 1.2791,
      "theoretical_loss": 3.35029705942209,
      "tokens_seen": 2677211136
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019204040404040404,
      "loss": 1.4942,
      "theoretical_loss": 3.350290607009889,
      "tokens_seen": 2677276672
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019202020202020202,
      "loss": 1.3143,
      "theoretical_loss": 3.350284154799856,
      "tokens_seen": 2677342208
    },
    {
      "epoch": 0.62,
      "objective/train/aligned_token_count": 0.45472875237464905,
      "objective/train/avg_token_score": 0.012021453119814396,
      "objective/train/docs_used": 1507786,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3495817184448242,
      "objective/train/original_loss": 4.130172252655029,
      "objective/train/theoretical_loss": 3.350277702791979,
      "objective/train/tokens_used": 1036266976,
      "objective/train/unlikelihood_loss": 0.06219317018985748,
      "theoretical_loss": 3.350277702791979,
      "tokens_seen": 2677407744
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000192,
      "loss": 1.1598,
      "theoretical_loss": 3.350277702791979,
      "tokens_seen": 2677407744
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.000191979797979798,
      "loss": 1.4779,
      "theoretical_loss": 3.3502712509862476,
      "tokens_seen": 2677473280
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019195959595959597,
      "loss": 1.4681,
      "theoretical_loss": 3.35026479938265,
      "tokens_seen": 2677538816
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019193939393939393,
      "loss": 1.427,
      "theoretical_loss": 3.3502583479811747,
      "tokens_seen": 2677604352
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019191919191919191,
      "loss": 1.4736,
      "theoretical_loss": 3.3502518967818116,
      "tokens_seen": 2677669888
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001918989898989899,
      "loss": 1.3006,
      "theoretical_loss": 3.350245445784548,
      "tokens_seen": 2677735424
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019187878787878788,
      "loss": 1.4305,
      "theoretical_loss": 3.3502389949893736,
      "tokens_seen": 2677800960
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019185858585858587,
      "loss": 1.4321,
      "theoretical_loss": 3.3502325443962766,
      "tokens_seen": 2677866496
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019183838383838385,
      "loss": 1.1557,
      "theoretical_loss": 3.3502260940052455,
      "tokens_seen": 2677932032
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001918181818181818,
      "loss": 1.3619,
      "theoretical_loss": 3.3502196438162697,
      "tokens_seen": 2677997568
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019179797979797982,
      "loss": 1.4405,
      "theoretical_loss": 3.3502131938293376,
      "tokens_seen": 2678063104
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019177777777777777,
      "loss": 1.1385,
      "theoretical_loss": 3.3502067440444376,
      "tokens_seen": 2678128640
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019175757575757576,
      "loss": 1.4977,
      "theoretical_loss": 3.350200294461559,
      "tokens_seen": 2678194176
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019173737373737374,
      "loss": 1.4073,
      "theoretical_loss": 3.3501938450806907,
      "tokens_seen": 2678259712
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019171717171717173,
      "loss": 1.3796,
      "theoretical_loss": 3.3501873959018207,
      "tokens_seen": 2678325248
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001916969696969697,
      "loss": 1.3474,
      "theoretical_loss": 3.3501809469249384,
      "tokens_seen": 2678390784
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019167676767676767,
      "loss": 1.1991,
      "theoretical_loss": 3.350174498150032,
      "tokens_seen": 2678456320
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019165656565656568,
      "loss": 1.3827,
      "theoretical_loss": 3.35016804957709,
      "tokens_seen": 2678521856
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019163636363636363,
      "loss": 1.148,
      "theoretical_loss": 3.350161601206102,
      "tokens_seen": 2678587392
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019161616161616165,
      "loss": 1.3739,
      "theoretical_loss": 3.3501551530370564,
      "tokens_seen": 2678652928
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001915959595959596,
      "loss": 1.2735,
      "theoretical_loss": 3.3501487050699414,
      "tokens_seen": 2678718464
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019157575757575756,
      "loss": 1.3973,
      "theoretical_loss": 3.3501422573047464,
      "tokens_seen": 2678784000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019155555555555557,
      "loss": 1.3313,
      "theoretical_loss": 3.35013580974146,
      "tokens_seen": 2678849536
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019153535353535353,
      "loss": 1.2152,
      "theoretical_loss": 3.3501293623800708,
      "tokens_seen": 2678915072
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019151515151515154,
      "loss": 1.4805,
      "theoretical_loss": 3.3501229152205676,
      "tokens_seen": 2678980608
    },
    {
      "epoch": 0.62,
      "objective/train/aligned_token_count": 0.4699413478374481,
      "objective/train/avg_token_score": 0.004387288354337215,
      "objective/train/docs_used": 1509162,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4834542274475098,
      "objective/train/original_loss": 3.727614641189575,
      "objective/train/theoretical_loss": 3.350116468262939,
      "objective/train/tokens_used": 1037905376,
      "objective/train/unlikelihood_loss": 0.06014406681060791,
      "theoretical_loss": 3.350116468262939,
      "tokens_seen": 2679046144
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001914949494949495,
      "loss": 1.4252,
      "theoretical_loss": 3.350116468262939,
      "tokens_seen": 2679046144
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019147474747474748,
      "loss": 1.3489,
      "theoretical_loss": 3.350110021507174,
      "tokens_seen": 2679111680
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019145454545454546,
      "loss": 1.4857,
      "theoretical_loss": 3.3501035749532613,
      "tokens_seen": 2679177216
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019143434343434342,
      "loss": 1.4312,
      "theoretical_loss": 3.350097128601189,
      "tokens_seen": 2679242752
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019141414141414143,
      "loss": 1.3967,
      "theoretical_loss": 3.350090682450947,
      "tokens_seen": 2679308288
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001913939393939394,
      "loss": 1.2991,
      "theoretical_loss": 3.350084236502523,
      "tokens_seen": 2679373824
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019137373737373737,
      "loss": 1.4341,
      "theoretical_loss": 3.3500777907559067,
      "tokens_seen": 2679439360
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019135353535353536,
      "loss": 1.415,
      "theoretical_loss": 3.350071345211086,
      "tokens_seen": 2679504896
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019133333333333334,
      "loss": 1.4207,
      "theoretical_loss": 3.35006489986805,
      "tokens_seen": 2679570432
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019131313131313132,
      "loss": 1.1473,
      "theoretical_loss": 3.3500584547267875,
      "tokens_seen": 2679635968
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001912929292929293,
      "loss": 1.508,
      "theoretical_loss": 3.3500520097872872,
      "tokens_seen": 2679701504
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019127272727272726,
      "loss": 1.3258,
      "theoretical_loss": 3.3500455650495375,
      "tokens_seen": 2679767040
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019125252525252525,
      "loss": 1.5069,
      "theoretical_loss": 3.350039120513528,
      "tokens_seen": 2679832576
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019123232323232323,
      "loss": 1.6786,
      "theoretical_loss": 3.3500326761792465,
      "tokens_seen": 2679898112
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019121212121212122,
      "loss": 1.5812,
      "theoretical_loss": 3.350026232046682,
      "tokens_seen": 2679963648
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001911919191919192,
      "loss": 1.437,
      "theoretical_loss": 3.3500197881158242,
      "tokens_seen": 2680029184
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019117171717171716,
      "loss": 1.4935,
      "theoretical_loss": 3.35001334438666,
      "tokens_seen": 2680094720
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019115151515151517,
      "loss": 1.3211,
      "theoretical_loss": 3.35000690085918,
      "tokens_seen": 2680160256
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019113131313131313,
      "loss": 1.3174,
      "theoretical_loss": 3.3500004575333726,
      "tokens_seen": 2680225792
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019111111111111114,
      "loss": 1.6961,
      "theoretical_loss": 3.3499940144092255,
      "tokens_seen": 2680291328
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001910909090909091,
      "loss": 1.325,
      "theoretical_loss": 3.349987571486728,
      "tokens_seen": 2680356864
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019107070707070705,
      "loss": 1.4296,
      "theoretical_loss": 3.349981128765869,
      "tokens_seen": 2680422400
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019105050505050506,
      "loss": 1.3574,
      "theoretical_loss": 3.3499746862466377,
      "tokens_seen": 2680487936
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019103030303030302,
      "loss": 1.3934,
      "theoretical_loss": 3.3499682439290224,
      "tokens_seen": 2680553472
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019101010101010103,
      "loss": 1.2469,
      "theoretical_loss": 3.349961801813011,
      "tokens_seen": 2680619008
    },
    {
      "epoch": 0.62,
      "objective/train/aligned_token_count": 0.3221529722213745,
      "objective/train/avg_token_score": 0.011578681878745556,
      "objective/train/docs_used": 1509790,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.052809476852417,
      "objective/train/original_loss": 4.378558158874512,
      "objective/train/theoretical_loss": 3.349955359898594,
      "objective/train/tokens_used": 1039543776,
      "objective/train/unlikelihood_loss": 0.06155547499656677,
      "theoretical_loss": 3.349955359898594,
      "tokens_seen": 2680684544
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019098989898989899,
      "loss": 1.3951,
      "theoretical_loss": 3.349955359898594,
      "tokens_seen": 2680684544
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019096969696969697,
      "loss": 1.2549,
      "theoretical_loss": 3.349948918185759,
      "tokens_seen": 2680750080
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019094949494949495,
      "loss": 1.3099,
      "theoretical_loss": 3.349942476674495,
      "tokens_seen": 2680815616
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019092929292929294,
      "loss": 1.5128,
      "theoretical_loss": 3.3499360353647907,
      "tokens_seen": 2680881152
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019090909090909092,
      "loss": 1.4392,
      "theoretical_loss": 3.349929594256635,
      "tokens_seen": 2680946688
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019088888888888888,
      "loss": 1.1826,
      "theoretical_loss": 3.3499231533500167,
      "tokens_seen": 2681012224
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001908686868686869,
      "loss": 1.5455,
      "theoretical_loss": 3.3499167126449243,
      "tokens_seen": 2681077760
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019084848484848485,
      "loss": 1.5843,
      "theoretical_loss": 3.349910272141347,
      "tokens_seen": 2681143296
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019082828282828283,
      "loss": 1.3469,
      "theoretical_loss": 3.3499038318392733,
      "tokens_seen": 2681208832
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019080808080808081,
      "loss": 1.2539,
      "theoretical_loss": 3.3498973917386925,
      "tokens_seen": 2681274368
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001907878787878788,
      "loss": 1.5069,
      "theoretical_loss": 3.3498909518395923,
      "tokens_seen": 2681339904
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019076767676767678,
      "loss": 1.3765,
      "theoretical_loss": 3.349884512141962,
      "tokens_seen": 2681405440
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019074747474747474,
      "loss": 1.5237,
      "theoretical_loss": 3.3498780726457906,
      "tokens_seen": 2681470976
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019072727272727272,
      "loss": 1.3868,
      "theoretical_loss": 3.349871633351067,
      "tokens_seen": 2681536512
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001907070707070707,
      "loss": 1.277,
      "theoretical_loss": 3.349865194257779,
      "tokens_seen": 2681602048
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001906868686868687,
      "loss": 1.1166,
      "theoretical_loss": 3.3498587553659167,
      "tokens_seen": 2681667584
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019066666666666668,
      "loss": 1.4861,
      "theoretical_loss": 3.349852316675468,
      "tokens_seen": 2681733120
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019064646464646466,
      "loss": 1.5294,
      "theoretical_loss": 3.349845878186422,
      "tokens_seen": 2681798656
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019062626262626262,
      "loss": 1.4901,
      "theoretical_loss": 3.3498394398987674,
      "tokens_seen": 2681864192
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019060606060606063,
      "loss": 1.5934,
      "theoretical_loss": 3.3498330018124927,
      "tokens_seen": 2681929728
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019058585858585858,
      "loss": 1.3598,
      "theoretical_loss": 3.349826563927587,
      "tokens_seen": 2681995264
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019056565656565657,
      "loss": 1.313,
      "theoretical_loss": 3.3498201262440395,
      "tokens_seen": 2682060800
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019054545454545455,
      "loss": 1.4145,
      "theoretical_loss": 3.349813688761838,
      "tokens_seen": 2682126336
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001905252525252525,
      "loss": 1.3129,
      "theoretical_loss": 3.349807251480972,
      "tokens_seen": 2682191872
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019050505050505052,
      "loss": 1.4742,
      "theoretical_loss": 3.34980081440143,
      "tokens_seen": 2682257408
    },
    {
      "epoch": 0.62,
      "objective/train/aligned_token_count": 0.41403958201408386,
      "objective/train/avg_token_score": 0.029405483976006508,
      "objective/train/docs_used": 1510817,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2975271940231323,
      "objective/train/original_loss": 4.779730796813965,
      "objective/train/theoretical_loss": 3.349794377523201,
      "objective/train/tokens_used": 1041182176,
      "objective/train/unlikelihood_loss": 0.047045063227415085,
      "theoretical_loss": 3.349794377523201,
      "tokens_seen": 2682322944
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019048484848484848,
      "loss": 1.6724,
      "theoretical_loss": 3.349794377523201,
      "tokens_seen": 2682322944
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001904646464646465,
      "loss": 1.5166,
      "theoretical_loss": 3.3497879408462734,
      "tokens_seen": 2682388480
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019044444444444444,
      "loss": 1.3146,
      "theoretical_loss": 3.3497815043706365,
      "tokens_seen": 2682454016
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019042424242424243,
      "loss": 1.4403,
      "theoretical_loss": 3.3497750680962786,
      "tokens_seen": 2682519552
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001904040404040404,
      "loss": 1.37,
      "theoretical_loss": 3.349768632023189,
      "tokens_seen": 2682585088
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019038383838383837,
      "loss": 1.5882,
      "theoretical_loss": 3.3497621961513557,
      "tokens_seen": 2682650624
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019036363636363638,
      "loss": 1.2595,
      "theoretical_loss": 3.3497557604807686,
      "tokens_seen": 2682716160
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019034343434343434,
      "loss": 1.2528,
      "theoretical_loss": 3.3497493250114156,
      "tokens_seen": 2682781696
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019032323232323232,
      "loss": 1.3953,
      "theoretical_loss": 3.3497428897432857,
      "tokens_seen": 2682847232
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001903030303030303,
      "loss": 1.5141,
      "theoretical_loss": 3.3497364546763677,
      "tokens_seen": 2682912768
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001902828282828283,
      "loss": 1.4439,
      "theoretical_loss": 3.349730019810651,
      "tokens_seen": 2682978304
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019026262626262627,
      "loss": 1.3091,
      "theoretical_loss": 3.3497235851461236,
      "tokens_seen": 2683043840
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019024242424242426,
      "loss": 1.4472,
      "theoretical_loss": 3.3497171506827743,
      "tokens_seen": 2683109376
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019022222222222221,
      "loss": 1.2734,
      "theoretical_loss": 3.349710716420592,
      "tokens_seen": 2683174912
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001902020202020202,
      "loss": 1.3079,
      "theoretical_loss": 3.349704282359566,
      "tokens_seen": 2683240448
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019018181818181818,
      "loss": 1.4649,
      "theoretical_loss": 3.3496978484996847,
      "tokens_seen": 2683305984
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019016161616161617,
      "loss": 1.4296,
      "theoretical_loss": 3.349691414840937,
      "tokens_seen": 2683371520
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019014141414141415,
      "loss": 1.4897,
      "theoretical_loss": 3.349684981383312,
      "tokens_seen": 2683437056
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001901212121212121,
      "loss": 1.4795,
      "theoretical_loss": 3.349678548126797,
      "tokens_seen": 2683502592
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019010101010101012,
      "loss": 1.4632,
      "theoretical_loss": 3.349672115071383,
      "tokens_seen": 2683568128
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019008080808080807,
      "loss": 1.093,
      "theoretical_loss": 3.349665682217058,
      "tokens_seen": 2683633664
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019006060606060609,
      "loss": 1.325,
      "theoretical_loss": 3.3496592495638096,
      "tokens_seen": 2683699200
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019004040404040404,
      "loss": 1.4232,
      "theoretical_loss": 3.349652817111628,
      "tokens_seen": 2683764736
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019002020202020203,
      "loss": 1.4536,
      "theoretical_loss": 3.3496463848605016,
      "tokens_seen": 2683830272
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00019,
      "loss": 1.2542,
      "theoretical_loss": 3.349639952810419,
      "tokens_seen": 2683895808
    },
    {
      "epoch": 0.62,
      "objective/train/aligned_token_count": 0.440493643283844,
      "objective/train/avg_token_score": 0.012817805632948875,
      "objective/train/docs_used": 1511447,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4204206466674805,
      "objective/train/original_loss": 4.105740547180176,
      "objective/train/theoretical_loss": 3.3496335209613695,
      "objective/train/tokens_used": 1042820576,
      "objective/train/unlikelihood_loss": 0.08086664229631424,
      "theoretical_loss": 3.3496335209613695,
      "tokens_seen": 2683961344
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018997979797979797,
      "loss": 1.383,
      "theoretical_loss": 3.3496335209613695,
      "tokens_seen": 2683961344
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018995959595959598,
      "loss": 1.2428,
      "theoretical_loss": 3.349627089313341,
      "tokens_seen": 2684026880
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018993939393939394,
      "loss": 1.5524,
      "theoretical_loss": 3.3496206578663235,
      "tokens_seen": 2684092416
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018991919191919195,
      "loss": 1.2513,
      "theoretical_loss": 3.349614226620305,
      "tokens_seen": 2684157952
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001898989898989899,
      "loss": 1.4298,
      "theoretical_loss": 3.3496077955752743,
      "tokens_seen": 2684223488
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018987878787878786,
      "loss": 1.0104,
      "theoretical_loss": 3.3496013647312206,
      "tokens_seen": 2684289024
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018985858585858587,
      "loss": 1.1832,
      "theoretical_loss": 3.349594934088133,
      "tokens_seen": 2684354560
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018983838383838383,
      "loss": 1.4947,
      "theoretical_loss": 3.349588503645999,
      "tokens_seen": 2684420096
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018981818181818184,
      "loss": 1.2059,
      "theoretical_loss": 3.3495820734048087,
      "tokens_seen": 2684485632
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001897979797979798,
      "loss": 1.4153,
      "theoretical_loss": 3.3495756433645507,
      "tokens_seen": 2684551168
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018977777777777778,
      "loss": 1.3485,
      "theoretical_loss": 3.3495692135252133,
      "tokens_seen": 2684616704
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018975757575757576,
      "loss": 1.5269,
      "theoretical_loss": 3.3495627838867854,
      "tokens_seen": 2684682240
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018973737373737375,
      "loss": 1.1254,
      "theoretical_loss": 3.3495563544492564,
      "tokens_seen": 2684747776
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018971717171717173,
      "loss": 1.2984,
      "theoretical_loss": 3.3495499252126146,
      "tokens_seen": 2684813312
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001896969696969697,
      "loss": 1.4649,
      "theoretical_loss": 3.3495434961768495,
      "tokens_seen": 2684878848
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018967676767676767,
      "loss": 1.3631,
      "theoretical_loss": 3.3495370673419487,
      "tokens_seen": 2684944384
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018965656565656566,
      "loss": 1.2872,
      "theoretical_loss": 3.3495306387079022,
      "tokens_seen": 2685009920
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018963636363636364,
      "loss": 1.4343,
      "theoretical_loss": 3.349524210274698,
      "tokens_seen": 2685075456
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018961616161616162,
      "loss": 1.3297,
      "theoretical_loss": 3.3495177820423256,
      "tokens_seen": 2685140992
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001895959595959596,
      "loss": 1.5431,
      "theoretical_loss": 3.3495113540107733,
      "tokens_seen": 2685206528
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018957575757575757,
      "loss": 1.5247,
      "theoretical_loss": 3.34950492618003,
      "tokens_seen": 2685272064
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018955555555555558,
      "loss": 1.4754,
      "theoretical_loss": 3.3494984985500844,
      "tokens_seen": 2685337600
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018953535353535353,
      "loss": 1.3131,
      "theoretical_loss": 3.349492071120926,
      "tokens_seen": 2685403136
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018951515151515152,
      "loss": 1.481,
      "theoretical_loss": 3.3494856438925433,
      "tokens_seen": 2685468672
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0001894949494949495,
      "loss": 1.4869,
      "theoretical_loss": 3.349479216864925,
      "tokens_seen": 2685534208
    },
    {
      "epoch": 0.62,
      "objective/train/aligned_token_count": 0.5303030014038086,
      "objective/train/avg_token_score": 0.010919919237494469,
      "objective/train/docs_used": 1512730,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6341955661773682,
      "objective/train/original_loss": 3.877470016479492,
      "objective/train/theoretical_loss": 3.3494727900380594,
      "objective/train/tokens_used": 1044458976,
      "objective/train/unlikelihood_loss": 0.06329912692308426,
      "theoretical_loss": 3.3494727900380594,
      "tokens_seen": 2685599744
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.00018947474747474746,
      "loss": 1.4358,
      "theoretical_loss": 3.3494727900380594,
      "tokens_seen": 2685599744
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018945454545454547,
      "loss": 1.141,
      "theoretical_loss": 3.3494663634119366,
      "tokens_seen": 2685665280
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018943434343434343,
      "loss": 1.5621,
      "theoretical_loss": 3.3494599369865443,
      "tokens_seen": 2685730816
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018941414141414144,
      "loss": 1.5601,
      "theoretical_loss": 3.349453510761872,
      "tokens_seen": 2685796352
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001893939393939394,
      "loss": 1.4494,
      "theoretical_loss": 3.349447084737908,
      "tokens_seen": 2685861888
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018937373737373738,
      "loss": 1.5752,
      "theoretical_loss": 3.3494406589146415,
      "tokens_seen": 2685927424
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018935353535353536,
      "loss": 1.1495,
      "theoretical_loss": 3.3494342332920612,
      "tokens_seen": 2685992960
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018933333333333332,
      "loss": 1.6487,
      "theoretical_loss": 3.349427807870156,
      "tokens_seen": 2686058496
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018931313131313133,
      "loss": 1.5957,
      "theoretical_loss": 3.3494213826489148,
      "tokens_seen": 2686124032
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001892929292929293,
      "loss": 1.164,
      "theoretical_loss": 3.3494149576283263,
      "tokens_seen": 2686189568
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018927272727272727,
      "loss": 1.359,
      "theoretical_loss": 3.3494085328083796,
      "tokens_seen": 2686255104
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018925252525252525,
      "loss": 1.169,
      "theoretical_loss": 3.3494021081890635,
      "tokens_seen": 2686320640
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018923232323232324,
      "loss": 1.0794,
      "theoretical_loss": 3.349395683770366,
      "tokens_seen": 2686386176
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018921212121212122,
      "loss": 1.4105,
      "theoretical_loss": 3.349389259552277,
      "tokens_seen": 2686451712
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001891919191919192,
      "loss": 1.3493,
      "theoretical_loss": 3.349382835534785,
      "tokens_seen": 2686517248
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001891717171717172,
      "loss": 1.1174,
      "theoretical_loss": 3.349376411717879,
      "tokens_seen": 2686582784
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018915151515151515,
      "loss": 1.2719,
      "theoretical_loss": 3.3493699881015475,
      "tokens_seen": 2686648320
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018913131313131313,
      "loss": 1.1652,
      "theoretical_loss": 3.3493635646857793,
      "tokens_seen": 2686713856
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018911111111111112,
      "loss": 1.3142,
      "theoretical_loss": 3.3493571414705636,
      "tokens_seen": 2686779392
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001890909090909091,
      "loss": 1.5773,
      "theoretical_loss": 3.349350718455889,
      "tokens_seen": 2686844928
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018907070707070708,
      "loss": 1.465,
      "theoretical_loss": 3.3493442956417443,
      "tokens_seen": 2686910464
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018905050505050507,
      "loss": 1.3979,
      "theoretical_loss": 3.349337873028119,
      "tokens_seen": 2686976000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018903030303030302,
      "loss": 1.4307,
      "theoretical_loss": 3.349331450615001,
      "tokens_seen": 2687041536
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000189010101010101,
      "loss": 1.3295,
      "theoretical_loss": 3.3493250284023794,
      "tokens_seen": 2687107072
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000188989898989899,
      "loss": 1.5736,
      "theoretical_loss": 3.349318606390244,
      "tokens_seen": 2687172608
    },
    {
      "epoch": 0.63,
      "objective/train/aligned_token_count": 0.32526880502700806,
      "objective/train/avg_token_score": 0.021030981093645096,
      "objective/train/docs_used": 1513440,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1697689294815063,
      "objective/train/original_loss": 5.072680950164795,
      "objective/train/theoretical_loss": 3.349312184578582,
      "objective/train/tokens_used": 1046097376,
      "objective/train/unlikelihood_loss": 0.036334652453660965,
      "theoretical_loss": 3.349312184578582,
      "tokens_seen": 2687238144
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018896969696969698,
      "loss": 1.2347,
      "theoretical_loss": 3.349312184578582,
      "tokens_seen": 2687238144
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018894949494949496,
      "loss": 1.4723,
      "theoretical_loss": 3.3493057629673837,
      "tokens_seen": 2687303680
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018892929292929292,
      "loss": 1.3571,
      "theoretical_loss": 3.3492993415566374,
      "tokens_seen": 2687369216
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018890909090909093,
      "loss": 1.4911,
      "theoretical_loss": 3.3492929203463317,
      "tokens_seen": 2687434752
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018888888888888888,
      "loss": 1.4776,
      "theoretical_loss": 3.349286499336456,
      "tokens_seen": 2687500288
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001888686868686869,
      "loss": 1.1326,
      "theoretical_loss": 3.3492800785269985,
      "tokens_seen": 2687565824
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018884848484848485,
      "loss": 1.2889,
      "theoretical_loss": 3.3492736579179487,
      "tokens_seen": 2687631360
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001888282828282828,
      "loss": 1.3438,
      "theoretical_loss": 3.349267237509295,
      "tokens_seen": 2687696896
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018880808080808082,
      "loss": 1.4007,
      "theoretical_loss": 3.3492608173010265,
      "tokens_seen": 2687762432
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018878787878787878,
      "loss": 1.5419,
      "theoretical_loss": 3.349254397293132,
      "tokens_seen": 2687827968
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001887676767676768,
      "loss": 1.4402,
      "theoretical_loss": 3.3492479774856005,
      "tokens_seen": 2687893504
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018874747474747475,
      "loss": 1.3981,
      "theoretical_loss": 3.3492415578784205,
      "tokens_seen": 2687959040
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018872727272727273,
      "loss": 1.2385,
      "theoretical_loss": 3.349235138471581,
      "tokens_seen": 2688024576
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001887070707070707,
      "loss": 1.4308,
      "theoretical_loss": 3.3492287192650716,
      "tokens_seen": 2688090112
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001886868686868687,
      "loss": 1.4517,
      "theoretical_loss": 3.3492223002588797,
      "tokens_seen": 2688155648
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001886868686868687,
      "loss": 1.3181,
      "theoretical_loss": 3.349215881452995,
      "tokens_seen": 2688221184
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018866666666666668,
      "loss": 1.2702,
      "theoretical_loss": 3.3492094628474067,
      "tokens_seen": 2688286720
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018864646464646464,
      "loss": 1.6276,
      "theoretical_loss": 3.3492030444421035,
      "tokens_seen": 2688352256
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018862626262626262,
      "loss": 1.3265,
      "theoretical_loss": 3.349196626237074,
      "tokens_seen": 2688417792
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001886060606060606,
      "loss": 1.3698,
      "theoretical_loss": 3.349190208232307,
      "tokens_seen": 2688483328
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001885858585858586,
      "loss": 1.3775,
      "theoretical_loss": 3.3491837904277912,
      "tokens_seen": 2688548864
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018856565656565657,
      "loss": 1.4397,
      "theoretical_loss": 3.349177372823516,
      "tokens_seen": 2688614400
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018854545454545456,
      "loss": 1.061,
      "theoretical_loss": 3.34917095541947,
      "tokens_seen": 2688679936
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018852525252525251,
      "loss": 1.256,
      "theoretical_loss": 3.3491645382156428,
      "tokens_seen": 2688745472
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018850505050505053,
      "loss": 1.4919,
      "theoretical_loss": 3.349158121212022,
      "tokens_seen": 2688811008
    },
    {
      "epoch": 0.63,
      "objective/train/aligned_token_count": 0.2733382284641266,
      "objective/train/avg_token_score": 0.022246329113841057,
      "objective/train/docs_used": 1514443,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8296374678611755,
      "objective/train/original_loss": 4.8193559646606445,
      "objective/train/theoretical_loss": 3.349151704408597,
      "objective/train/tokens_used": 1047735776,
      "objective/train/unlikelihood_loss": 0.05123618617653847,
      "theoretical_loss": 3.349151704408597,
      "tokens_seen": 2688876544
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018848484848484848,
      "loss": 1.1997,
      "theoretical_loss": 3.349151704408597,
      "tokens_seen": 2688876544
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018846464646464647,
      "loss": 1.4158,
      "theoretical_loss": 3.349145287805357,
      "tokens_seen": 2688942080
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018844444444444445,
      "loss": 1.3023,
      "theoretical_loss": 3.3491388714022907,
      "tokens_seen": 2689007616
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001884242424242424,
      "loss": 1.4217,
      "theoretical_loss": 3.3491324551993866,
      "tokens_seen": 2689073152
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018840404040404042,
      "loss": 1.1856,
      "theoretical_loss": 3.349126039196634,
      "tokens_seen": 2689138688
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018838383838383838,
      "loss": 1.4123,
      "theoretical_loss": 3.3491196233940217,
      "tokens_seen": 2689204224
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018836363636363639,
      "loss": 1.2833,
      "theoretical_loss": 3.349113207791538,
      "tokens_seen": 2689269760
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018834343434343434,
      "loss": 1.3166,
      "theoretical_loss": 3.3491067923891733,
      "tokens_seen": 2689335296
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018832323232323235,
      "loss": 1.1055,
      "theoretical_loss": 3.349100377186915,
      "tokens_seen": 2689400832
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001883030303030303,
      "loss": 1.4116,
      "theoretical_loss": 3.3490939621847526,
      "tokens_seen": 2689466368
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018828282828282827,
      "loss": 1.4614,
      "theoretical_loss": 3.3490875473826747,
      "tokens_seen": 2689531904
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018826262626262628,
      "loss": 1.3217,
      "theoretical_loss": 3.3490811327806704,
      "tokens_seen": 2689597440
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018824242424242424,
      "loss": 1.4419,
      "theoretical_loss": 3.349074718378729,
      "tokens_seen": 2689662976
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018822222222222225,
      "loss": 1.2902,
      "theoretical_loss": 3.3490683041768383,
      "tokens_seen": 2689728512
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001882020202020202,
      "loss": 1.188,
      "theoretical_loss": 3.349061890174988,
      "tokens_seen": 2689794048
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001881818181818182,
      "loss": 1.2128,
      "theoretical_loss": 3.3490554763731666,
      "tokens_seen": 2689859584
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018816161616161617,
      "loss": 1.4206,
      "theoretical_loss": 3.3490490627713636,
      "tokens_seen": 2689925120
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018814141414141413,
      "loss": 1.3353,
      "theoretical_loss": 3.3490426493695673,
      "tokens_seen": 2689990656
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018812121212121214,
      "loss": 1.5001,
      "theoretical_loss": 3.349036236167767,
      "tokens_seen": 2690056192
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001881010101010101,
      "loss": 1.4035,
      "theoretical_loss": 3.349029823165951,
      "tokens_seen": 2690121728
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018808080808080808,
      "loss": 1.4581,
      "theoretical_loss": 3.3490234103641088,
      "tokens_seen": 2690187264
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018806060606060606,
      "loss": 1.5373,
      "theoretical_loss": 3.349016997762229,
      "tokens_seen": 2690252800
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018804040404040405,
      "loss": 1.1968,
      "theoretical_loss": 3.3490105853603005,
      "tokens_seen": 2690318336
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018802020202020203,
      "loss": 1.2373,
      "theoretical_loss": 3.349004173158312,
      "tokens_seen": 2690383872
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018800000000000002,
      "loss": 1.4498,
      "theoretical_loss": 3.348997761156253,
      "tokens_seen": 2690449408
    },
    {
      "epoch": 0.63,
      "objective/train/aligned_token_count": 0.45204055309295654,
      "objective/train/avg_token_score": 0.009940666146576405,
      "objective/train/docs_used": 1514886,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2701870203018188,
      "objective/train/original_loss": 3.9016923904418945,
      "objective/train/theoretical_loss": 3.3489913493541117,
      "objective/train/tokens_used": 1049374176,
      "objective/train/unlikelihood_loss": 0.09443721920251846,
      "theoretical_loss": 3.3489913493541117,
      "tokens_seen": 2690514944
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018797979797979797,
      "loss": 1.385,
      "theoretical_loss": 3.3489913493541117,
      "tokens_seen": 2690514944
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018795959595959596,
      "loss": 1.4128,
      "theoretical_loss": 3.3489849377518777,
      "tokens_seen": 2690580480
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018793939393939394,
      "loss": 1.2784,
      "theoretical_loss": 3.3489785263495393,
      "tokens_seen": 2690646016
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018791919191919192,
      "loss": 1.25,
      "theoretical_loss": 3.3489721151470855,
      "tokens_seen": 2690711552
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001878989898989899,
      "loss": 1.3592,
      "theoretical_loss": 3.3489657041445056,
      "tokens_seen": 2690777088
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018787878787878787,
      "loss": 1.5732,
      "theoretical_loss": 3.348959293341788,
      "tokens_seen": 2690842624
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018785858585858588,
      "loss": 1.3847,
      "theoretical_loss": 3.3489528827389217,
      "tokens_seen": 2690908160
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018783838383838383,
      "loss": 1.4602,
      "theoretical_loss": 3.348946472335896,
      "tokens_seen": 2690973696
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018781818181818184,
      "loss": 1.1388,
      "theoretical_loss": 3.3489400621326997,
      "tokens_seen": 2691039232
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001877979797979798,
      "loss": 1.4264,
      "theoretical_loss": 3.348933652129321,
      "tokens_seen": 2691104768
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018777777777777776,
      "loss": 1.4522,
      "theoretical_loss": 3.3489272423257495,
      "tokens_seen": 2691170304
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018775757575757577,
      "loss": 1.1099,
      "theoretical_loss": 3.348920832721974,
      "tokens_seen": 2691235840
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018773737373737373,
      "loss": 1.4591,
      "theoretical_loss": 3.3489144233179835,
      "tokens_seen": 2691301376
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018771717171717174,
      "loss": 1.5092,
      "theoretical_loss": 3.348908014113767,
      "tokens_seen": 2691366912
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001876969696969697,
      "loss": 1.3581,
      "theoretical_loss": 3.3489016051093126,
      "tokens_seen": 2691432448
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018767676767676768,
      "loss": 1.5784,
      "theoretical_loss": 3.34889519630461,
      "tokens_seen": 2691497984
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018765656565656566,
      "loss": 1.3875,
      "theoretical_loss": 3.3488887876996483,
      "tokens_seen": 2691563520
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018763636363636365,
      "loss": 1.3007,
      "theoretical_loss": 3.348882379294415,
      "tokens_seen": 2691629056
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018761616161616163,
      "loss": 1.2708,
      "theoretical_loss": 3.348875971088901,
      "tokens_seen": 2691694592
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001875959595959596,
      "loss": 1.4893,
      "theoretical_loss": 3.3488695630830936,
      "tokens_seen": 2691760128
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018757575757575757,
      "loss": 1.3918,
      "theoretical_loss": 3.348863155276983,
      "tokens_seen": 2691825664
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018755555555555555,
      "loss": 1.5221,
      "theoretical_loss": 3.348856747670557,
      "tokens_seen": 2691891200
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018753535353535354,
      "loss": 1.335,
      "theoretical_loss": 3.3488503402638052,
      "tokens_seen": 2691956736
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018751515151515152,
      "loss": 1.2936,
      "theoretical_loss": 3.348843933056716,
      "tokens_seen": 2692022272
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001874949494949495,
      "loss": 1.3671,
      "theoretical_loss": 3.3488375260492784,
      "tokens_seen": 2692087808
    },
    {
      "epoch": 0.63,
      "objective/train/aligned_token_count": 0.4923020601272583,
      "objective/train/avg_token_score": 0.029123054817318916,
      "objective/train/docs_used": 1515566,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.57011079788208,
      "objective/train/original_loss": 3.9811103343963623,
      "objective/train/theoretical_loss": 3.348831119241482,
      "objective/train/tokens_used": 1051012576,
      "objective/train/unlikelihood_loss": 0.06739974021911621,
      "theoretical_loss": 3.348831119241482,
      "tokens_seen": 2692153344
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001874747474747475,
      "loss": 1.5827,
      "theoretical_loss": 3.348831119241482,
      "tokens_seen": 2692153344
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018745454545454545,
      "loss": 1.0419,
      "theoretical_loss": 3.348824712633315,
      "tokens_seen": 2692218880
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018743434343434343,
      "loss": 1.4866,
      "theoretical_loss": 3.348818306224767,
      "tokens_seen": 2692284416
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018741414141414142,
      "loss": 1.2558,
      "theoretical_loss": 3.348811900015826,
      "tokens_seen": 2692349952
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001873939393939394,
      "loss": 1.3017,
      "theoretical_loss": 3.3488054940064815,
      "tokens_seen": 2692415488
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018737373737373738,
      "loss": 1.3205,
      "theoretical_loss": 3.3487990881967225,
      "tokens_seen": 2692481024
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018735353535353537,
      "loss": 1.4279,
      "theoretical_loss": 3.3487926825865375,
      "tokens_seen": 2692546560
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018733333333333332,
      "loss": 1.4949,
      "theoretical_loss": 3.348786277175916,
      "tokens_seen": 2692612096
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018731313131313134,
      "loss": 1.433,
      "theoretical_loss": 3.3487798719648465,
      "tokens_seen": 2692677632
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001872929292929293,
      "loss": 1.3064,
      "theoretical_loss": 3.348773466953318,
      "tokens_seen": 2692743168
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018727272727272728,
      "loss": 1.359,
      "theoretical_loss": 3.3487670621413193,
      "tokens_seen": 2692808704
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018725252525252526,
      "loss": 1.2017,
      "theoretical_loss": 3.34876065752884,
      "tokens_seen": 2692874240
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018723232323232322,
      "loss": 1.2094,
      "theoretical_loss": 3.348754253115868,
      "tokens_seen": 2692939776
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018721212121212123,
      "loss": 1.5746,
      "theoretical_loss": 3.3487478489023927,
      "tokens_seen": 2693005312
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018719191919191918,
      "loss": 1.4491,
      "theoretical_loss": 3.348741444888403,
      "tokens_seen": 2693070848
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001871717171717172,
      "loss": 1.1872,
      "theoretical_loss": 3.3487350410738883,
      "tokens_seen": 2693136384
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018715151515151515,
      "loss": 1.3195,
      "theoretical_loss": 3.348728637458837,
      "tokens_seen": 2693201920
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018713131313131314,
      "loss": 1.1841,
      "theoretical_loss": 3.3487222340432385,
      "tokens_seen": 2693267456
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018711111111111112,
      "loss": 1.5931,
      "theoretical_loss": 3.3487158308270812,
      "tokens_seen": 2693332992
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018709090909090908,
      "loss": 1.3745,
      "theoretical_loss": 3.3487094278103537,
      "tokens_seen": 2693398528
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001870707070707071,
      "loss": 1.3089,
      "theoretical_loss": 3.3487030249930463,
      "tokens_seen": 2693464064
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018705050505050505,
      "loss": 1.4127,
      "theoretical_loss": 3.348696622375147,
      "tokens_seen": 2693529600
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018703030303030303,
      "loss": 1.3329,
      "theoretical_loss": 3.3486902199566444,
      "tokens_seen": 2693595136
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000187010101010101,
      "loss": 1.4002,
      "theoretical_loss": 3.348683817737528,
      "tokens_seen": 2693660672
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000186989898989899,
      "loss": 1.2743,
      "theoretical_loss": 3.3486774157177868,
      "tokens_seen": 2693726208
    },
    {
      "epoch": 0.63,
      "objective/train/aligned_token_count": 0.3918010890483856,
      "objective/train/avg_token_score": 0.013431337662041187,
      "objective/train/docs_used": 1516840,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2430692911148071,
      "objective/train/original_loss": 4.125279903411865,
      "objective/train/theoretical_loss": 3.34867101389741,
      "objective/train/tokens_used": 1052650976,
      "objective/train/unlikelihood_loss": 0.07370205223560333,
      "theoretical_loss": 3.34867101389741,
      "tokens_seen": 2693791744
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018696969696969698,
      "loss": 1.4171,
      "theoretical_loss": 3.34867101389741,
      "tokens_seen": 2693791744
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018694949494949497,
      "loss": 1.43,
      "theoretical_loss": 3.3486646122763855,
      "tokens_seen": 2693857280
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018692929292929292,
      "loss": 1.2482,
      "theoretical_loss": 3.348658210854703,
      "tokens_seen": 2693922816
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001869090909090909,
      "loss": 1.483,
      "theoretical_loss": 3.348651809632351,
      "tokens_seen": 2693988352
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001868888888888889,
      "loss": 1.3873,
      "theoretical_loss": 3.348645408609319,
      "tokens_seen": 2694053888
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018686868686868687,
      "loss": 1.3616,
      "theoretical_loss": 3.348639007785596,
      "tokens_seen": 2694119424
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018684848484848486,
      "loss": 1.2742,
      "theoretical_loss": 3.3486326071611705,
      "tokens_seen": 2694184960
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018682828282828281,
      "loss": 1.4152,
      "theoretical_loss": 3.3486262067360317,
      "tokens_seen": 2694250496
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018680808080808083,
      "loss": 1.4016,
      "theoretical_loss": 3.348619806510168,
      "tokens_seen": 2694316032
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018678787878787878,
      "loss": 1.4676,
      "theoretical_loss": 3.3486134064835693,
      "tokens_seen": 2694381568
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001867676767676768,
      "loss": 1.5262,
      "theoretical_loss": 3.348607006656224,
      "tokens_seen": 2694447104
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018674747474747475,
      "loss": 1.53,
      "theoretical_loss": 3.3486006070281205,
      "tokens_seen": 2694512640
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001867272727272727,
      "loss": 1.2309,
      "theoretical_loss": 3.3485942075992488,
      "tokens_seen": 2694578176
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018670707070707072,
      "loss": 1.3337,
      "theoretical_loss": 3.3485878083695972,
      "tokens_seen": 2694643712
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018668686868686868,
      "loss": 1.3564,
      "theoretical_loss": 3.3485814093391553,
      "tokens_seen": 2694709248
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001866666666666667,
      "loss": 1.3556,
      "theoretical_loss": 3.3485750105079113,
      "tokens_seen": 2694774784
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018664646464646464,
      "loss": 1.2643,
      "theoretical_loss": 3.348568611875854,
      "tokens_seen": 2694840320
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018662626262626265,
      "loss": 1.4992,
      "theoretical_loss": 3.3485622134429733,
      "tokens_seen": 2694905856
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001866060606060606,
      "loss": 1.3887,
      "theoretical_loss": 3.3485558152092576,
      "tokens_seen": 2694971392
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018658585858585857,
      "loss": 1.2495,
      "theoretical_loss": 3.348549417174696,
      "tokens_seen": 2695036928
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018656565656565658,
      "loss": 1.2536,
      "theoretical_loss": 3.348543019339277,
      "tokens_seen": 2695102464
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018654545454545454,
      "loss": 1.2034,
      "theoretical_loss": 3.3485366217029906,
      "tokens_seen": 2695168000
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018652525252525255,
      "loss": 1.4355,
      "theoretical_loss": 3.348530224265825,
      "tokens_seen": 2695233536
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001865050505050505,
      "loss": 1.3397,
      "theoretical_loss": 3.3485238270277686,
      "tokens_seen": 2695299072
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001864848484848485,
      "loss": 1.6281,
      "theoretical_loss": 3.3485174299888114,
      "tokens_seen": 2695364608
    },
    {
      "epoch": 0.63,
      "objective/train/aligned_token_count": 0.5631109476089478,
      "objective/train/avg_token_score": 0.015946222469210625,
      "objective/train/docs_used": 1517438,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8454149961471558,
      "objective/train/original_loss": 4.184688568115234,
      "objective/train/theoretical_loss": 3.348511033148942,
      "objective/train/tokens_used": 1054289376,
      "objective/train/unlikelihood_loss": 0.043036095798015594,
      "theoretical_loss": 3.348511033148942,
      "tokens_seen": 2695430144
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018646464646464647,
      "loss": 1.5109,
      "theoretical_loss": 3.348511033148942,
      "tokens_seen": 2695430144
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018644444444444446,
      "loss": 1.3117,
      "theoretical_loss": 3.3485046365081494,
      "tokens_seen": 2695495680
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018642424242424244,
      "loss": 1.1664,
      "theoretical_loss": 3.3484982400664225,
      "tokens_seen": 2695561216
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001864040404040404,
      "loss": 1.1731,
      "theoretical_loss": 3.34849184382375,
      "tokens_seen": 2695626752
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018638383838383838,
      "loss": 1.0806,
      "theoretical_loss": 3.3484854477801216,
      "tokens_seen": 2695692288
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018636363636363636,
      "loss": 1.6081,
      "theoretical_loss": 3.3484790519355254,
      "tokens_seen": 2695757824
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018634343434343435,
      "loss": 1.501,
      "theoretical_loss": 3.3484726562899514,
      "tokens_seen": 2695823360
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018632323232323233,
      "loss": 1.3704,
      "theoretical_loss": 3.348466260843387,
      "tokens_seen": 2695888896
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018630303030303032,
      "loss": 1.2485,
      "theoretical_loss": 3.3484598655958226,
      "tokens_seen": 2695954432
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018628282828282827,
      "loss": 1.1774,
      "theoretical_loss": 3.348453470547247,
      "tokens_seen": 2696019968
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018626262626262628,
      "loss": 1.2936,
      "theoretical_loss": 3.3484470756976488,
      "tokens_seen": 2696085504
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018624242424242424,
      "loss": 1.2027,
      "theoretical_loss": 3.3484406810470166,
      "tokens_seen": 2696151040
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018622222222222223,
      "loss": 1.259,
      "theoretical_loss": 3.34843428659534,
      "tokens_seen": 2696216576
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001862020202020202,
      "loss": 1.4428,
      "theoretical_loss": 3.348427892342608,
      "tokens_seen": 2696282112
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018618181818181817,
      "loss": 1.3335,
      "theoretical_loss": 3.348421498288809,
      "tokens_seen": 2696347648
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018616161616161618,
      "loss": 1.2112,
      "theoretical_loss": 3.3484151044339328,
      "tokens_seen": 2696413184
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018614141414141413,
      "loss": 1.6416,
      "theoretical_loss": 3.348408710777967,
      "tokens_seen": 2696478720
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018612121212121214,
      "loss": 1.5387,
      "theoretical_loss": 3.3484023173209025,
      "tokens_seen": 2696544256
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001861010101010101,
      "loss": 1.1611,
      "theoretical_loss": 3.348395924062727,
      "tokens_seen": 2696609792
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018608080808080809,
      "loss": 1.4948,
      "theoretical_loss": 3.3483895310034297,
      "tokens_seen": 2696675328
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018606060606060607,
      "loss": 1.3124,
      "theoretical_loss": 3.3483831381429994,
      "tokens_seen": 2696740864
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018604040404040403,
      "loss": 1.5638,
      "theoretical_loss": 3.3483767454814255,
      "tokens_seen": 2696806400
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018602020202020204,
      "loss": 1.2062,
      "theoretical_loss": 3.3483703530186966,
      "tokens_seen": 2696871936
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000186,
      "loss": 1.2954,
      "theoretical_loss": 3.348363960754802,
      "tokens_seen": 2696937472
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018597979797979798,
      "loss": 1.4639,
      "theoretical_loss": 3.348357568689731,
      "tokens_seen": 2697003008
    },
    {
      "epoch": 0.63,
      "objective/train/aligned_token_count": 0.4538734257221222,
      "objective/train/avg_token_score": 0.007988063618540764,
      "objective/train/docs_used": 1518738,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5575000047683716,
      "objective/train/original_loss": 4.582637786865234,
      "objective/train/theoretical_loss": 3.3483511768234715,
      "objective/train/tokens_used": 1055927776,
      "objective/train/unlikelihood_loss": 0.048297274857759476,
      "theoretical_loss": 3.3483511768234715,
      "tokens_seen": 2697068544
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018595959595959596,
      "loss": 1.3106,
      "theoretical_loss": 3.3483511768234715,
      "tokens_seen": 2697068544
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018593939393939395,
      "loss": 1.3305,
      "theoretical_loss": 3.348344785156013,
      "tokens_seen": 2697134080
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018591919191919193,
      "loss": 1.3476,
      "theoretical_loss": 3.348338393687345,
      "tokens_seen": 2697199616
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018589898989898991,
      "loss": 1.3075,
      "theoretical_loss": 3.3483320024174565,
      "tokens_seen": 2697265152
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018587878787878787,
      "loss": 1.3159,
      "theoretical_loss": 3.3483256113463353,
      "tokens_seen": 2697330688
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018585858585858586,
      "loss": 1.5388,
      "theoretical_loss": 3.3483192204739716,
      "tokens_seen": 2697396224
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018583838383838384,
      "loss": 1.273,
      "theoretical_loss": 3.3483128298003537,
      "tokens_seen": 2697461760
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018581818181818182,
      "loss": 1.1688,
      "theoretical_loss": 3.348306439325471,
      "tokens_seen": 2697527296
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001857979797979798,
      "loss": 1.5131,
      "theoretical_loss": 3.3483000490493127,
      "tokens_seen": 2697592832
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018577777777777776,
      "loss": 1.6178,
      "theoretical_loss": 3.3482936589718673,
      "tokens_seen": 2697658368
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018575757575757577,
      "loss": 1.3837,
      "theoretical_loss": 3.348287269093124,
      "tokens_seen": 2697723904
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018573737373737373,
      "loss": 1.3638,
      "theoretical_loss": 3.348280879413071,
      "tokens_seen": 2697789440
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018571717171717172,
      "loss": 1.4791,
      "theoretical_loss": 3.3482744899316987,
      "tokens_seen": 2697854976
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001856969696969697,
      "loss": 1.5501,
      "theoretical_loss": 3.3482681006489954,
      "tokens_seen": 2697920512
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018567676767676768,
      "loss": 1.3933,
      "theoretical_loss": 3.34826171156495,
      "tokens_seen": 2697986048
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018565656565656567,
      "loss": 1.2343,
      "theoretical_loss": 3.3482553226795515,
      "tokens_seen": 2698051584
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018563636363636362,
      "loss": 1.4323,
      "theoretical_loss": 3.348248933992789,
      "tokens_seen": 2698117120
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018561616161616164,
      "loss": 1.4722,
      "theoretical_loss": 3.3482425455046516,
      "tokens_seen": 2698182656
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001855959595959596,
      "loss": 1.4115,
      "theoretical_loss": 3.3482361572151285,
      "tokens_seen": 2698248192
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001855757575757576,
      "loss": 1.4189,
      "theoretical_loss": 3.3482297691242078,
      "tokens_seen": 2698313728
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018555555555555556,
      "loss": 1.2513,
      "theoretical_loss": 3.34822338123188,
      "tokens_seen": 2698379264
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018553535353535352,
      "loss": 1.269,
      "theoretical_loss": 3.3482169935381325,
      "tokens_seen": 2698444800
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018551515151515153,
      "loss": 1.3218,
      "theoretical_loss": 3.3482106060429553,
      "tokens_seen": 2698510336
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018549494949494949,
      "loss": 1.4001,
      "theoretical_loss": 3.348204218746337,
      "tokens_seen": 2698575872
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001854747474747475,
      "loss": 1.3883,
      "theoretical_loss": 3.348197831648267,
      "tokens_seen": 2698641408
    },
    {
      "epoch": 0.63,
      "objective/train/aligned_token_count": 0.49163001775741577,
      "objective/train/avg_token_score": 0.013117627240717411,
      "objective/train/docs_used": 1519063,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4972667694091797,
      "objective/train/original_loss": 3.9530982971191406,
      "objective/train/theoretical_loss": 3.3481914447487338,
      "objective/train/tokens_used": 1057566176,
      "objective/train/unlikelihood_loss": 0.07717724144458771,
      "theoretical_loss": 3.3481914447487338,
      "tokens_seen": 2698706944
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018545454545454545,
      "loss": 1.3892,
      "theoretical_loss": 3.3481914447487338,
      "tokens_seen": 2698706944
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018543434343434344,
      "loss": 1.3545,
      "theoretical_loss": 3.3481850580477266,
      "tokens_seen": 2698772480
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018541414141414142,
      "loss": 1.4775,
      "theoretical_loss": 3.3481786715452344,
      "tokens_seen": 2698838016
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001853939393939394,
      "loss": 1.1082,
      "theoretical_loss": 3.3481722852412465,
      "tokens_seen": 2698903552
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001853737373737374,
      "loss": 1.4821,
      "theoretical_loss": 3.348165899135752,
      "tokens_seen": 2698969088
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018535353535353535,
      "loss": 1.4279,
      "theoretical_loss": 3.348159513228739,
      "tokens_seen": 2699034624
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018533333333333333,
      "loss": 1.2109,
      "theoretical_loss": 3.348153127520197,
      "tokens_seen": 2699100160
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018531313131313131,
      "loss": 1.2698,
      "theoretical_loss": 3.348146742010116,
      "tokens_seen": 2699165696
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001852929292929293,
      "loss": 1.1016,
      "theoretical_loss": 3.348140356698483,
      "tokens_seen": 2699231232
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018527272727272728,
      "loss": 1.3972,
      "theoretical_loss": 3.348133971585289,
      "tokens_seen": 2699296768
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018525252525252527,
      "loss": 1.4511,
      "theoretical_loss": 3.348127586670522,
      "tokens_seen": 2699362304
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018523232323232322,
      "loss": 1.3583,
      "theoretical_loss": 3.348121201954171,
      "tokens_seen": 2699427840
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018521212121212123,
      "loss": 1.2887,
      "theoretical_loss": 3.348114817436225,
      "tokens_seen": 2699493376
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001851919191919192,
      "loss": 1.3195,
      "theoretical_loss": 3.3481084331166735,
      "tokens_seen": 2699558912
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018517171717171717,
      "loss": 1.3206,
      "theoretical_loss": 3.3481020489955053,
      "tokens_seen": 2699624448
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018515151515151516,
      "loss": 1.3569,
      "theoretical_loss": 3.3480956650727096,
      "tokens_seen": 2699689984
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018513131313131312,
      "loss": 1.443,
      "theoretical_loss": 3.3480892813482743,
      "tokens_seen": 2699755520
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018511111111111113,
      "loss": 1.4259,
      "theoretical_loss": 3.34808289782219,
      "tokens_seen": 2699821056
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018509090909090908,
      "loss": 1.3482,
      "theoretical_loss": 3.348076514494445,
      "tokens_seen": 2699886592
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001850707070707071,
      "loss": 1.376,
      "theoretical_loss": 3.348070131365028,
      "tokens_seen": 2699952128
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018505050505050505,
      "loss": 1.5123,
      "theoretical_loss": 3.348063748433929,
      "tokens_seen": 2700017664
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000185030303030303,
      "loss": 1.3183,
      "theoretical_loss": 3.3480573657011354,
      "tokens_seen": 2700083200
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018501010101010102,
      "loss": 1.2413,
      "theoretical_loss": 3.348050983166638,
      "tokens_seen": 2700148736
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018498989898989898,
      "loss": 1.3727,
      "theoretical_loss": 3.3480446008304248,
      "tokens_seen": 2700214272
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.000184969696969697,
      "loss": 1.3747,
      "theoretical_loss": 3.348038218692485,
      "tokens_seen": 2700279808
    },
    {
      "epoch": 0.63,
      "objective/train/aligned_token_count": 0.310056209564209,
      "objective/train/avg_token_score": 0.011696191504597664,
      "objective/train/docs_used": 1519691,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9081757068634033,
      "objective/train/original_loss": 4.300036430358887,
      "objective/train/theoretical_loss": 3.3480318367528077,
      "objective/train/tokens_used": 1059204576,
      "objective/train/unlikelihood_loss": 0.06119450926780701,
      "theoretical_loss": 3.3480318367528077,
      "tokens_seen": 2700345344
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018494949494949494,
      "loss": 1.177,
      "theoretical_loss": 3.3480318367528077,
      "tokens_seen": 2700345344
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018492929292929293,
      "loss": 1.5176,
      "theoretical_loss": 3.3480254550113817,
      "tokens_seen": 2700410880
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001849090909090909,
      "loss": 1.577,
      "theoretical_loss": 3.3480190734681967,
      "tokens_seen": 2700476416
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001848888888888889,
      "loss": 1.1891,
      "theoretical_loss": 3.348012692123241,
      "tokens_seen": 2700541952
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018486868686868688,
      "loss": 1.4227,
      "theoretical_loss": 3.348006310976504,
      "tokens_seen": 2700607488
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018484848484848484,
      "loss": 1.5888,
      "theoretical_loss": 3.347999930027975,
      "tokens_seen": 2700673024
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018482828282828285,
      "loss": 1.2909,
      "theoretical_loss": 3.3479935492776423,
      "tokens_seen": 2700738560
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001848080808080808,
      "loss": 1.6326,
      "theoretical_loss": 3.347987168725495,
      "tokens_seen": 2700804096
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001847878787878788,
      "loss": 1.5232,
      "theoretical_loss": 3.3479807883715234,
      "tokens_seen": 2700869632
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018476767676767677,
      "loss": 1.3051,
      "theoretical_loss": 3.347974408215715,
      "tokens_seen": 2700935168
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018474747474747476,
      "loss": 1.2339,
      "theoretical_loss": 3.3479680282580593,
      "tokens_seen": 2701000704
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018472727272727274,
      "loss": 1.2321,
      "theoretical_loss": 3.3479616484985457,
      "tokens_seen": 2701066240
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018470707070707072,
      "loss": 1.2389,
      "theoretical_loss": 3.347955268937163,
      "tokens_seen": 2701131776
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018468686868686868,
      "loss": 1.1309,
      "theoretical_loss": 3.3479488895739,
      "tokens_seen": 2701197312
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018466666666666666,
      "loss": 1.4745,
      "theoretical_loss": 3.347942510408746,
      "tokens_seen": 2701262848
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018464646464646465,
      "loss": 1.2391,
      "theoretical_loss": 3.3479361314416907,
      "tokens_seen": 2701328384
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018462626262626263,
      "loss": 1.278,
      "theoretical_loss": 3.347929752672722,
      "tokens_seen": 2701393920
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018460606060606062,
      "loss": 1.3891,
      "theoretical_loss": 3.3479233741018293,
      "tokens_seen": 2701459456
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018458585858585857,
      "loss": 1.5163,
      "theoretical_loss": 3.347916995729002,
      "tokens_seen": 2701524992
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018456565656565658,
      "loss": 1.2388,
      "theoretical_loss": 3.347910617554229,
      "tokens_seen": 2701590528
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018454545454545454,
      "loss": 1.3981,
      "theoretical_loss": 3.3479042395774985,
      "tokens_seen": 2701656064
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018452525252525255,
      "loss": 1.1227,
      "theoretical_loss": 3.347897861798801,
      "tokens_seen": 2701721600
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0001845050505050505,
      "loss": 1.3306,
      "theoretical_loss": 3.347891484218125,
      "tokens_seen": 2701787136
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018448484848484847,
      "loss": 1.5682,
      "theoretical_loss": 3.3478851068354585,
      "tokens_seen": 2701852672
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00018446464646464648,
      "loss": 1.3277,
      "theoretical_loss": 3.3478787296507924,
      "tokens_seen": 2701918208
    },
    {
      "epoch": 0.63,
      "objective/train/aligned_token_count": 0.39002931118011475,
      "objective/train/avg_token_score": 0.01600056327879429,
      "objective/train/docs_used": 1520269,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.325368046760559,
      "objective/train/original_loss": 4.517643928527832,
      "objective/train/theoretical_loss": 3.347872352664114,
      "objective/train/tokens_used": 1060842976,
      "objective/train/unlikelihood_loss": 0.06534574180841446,
      "theoretical_loss": 3.347872352664114,
      "tokens_seen": 2701983744
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018444444444444443,
      "loss": 1.3504,
      "theoretical_loss": 3.347872352664114,
      "tokens_seen": 2701983744
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018442424242424245,
      "loss": 1.2348,
      "theoretical_loss": 3.347865975875414,
      "tokens_seen": 2702049280
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001844040404040404,
      "loss": 1.5489,
      "theoretical_loss": 3.34785959928468,
      "tokens_seen": 2702114816
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018438383838383839,
      "loss": 1.2698,
      "theoretical_loss": 3.3478532228919016,
      "tokens_seen": 2702180352
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018436363636363637,
      "loss": 1.5959,
      "theoretical_loss": 3.347846846697068,
      "tokens_seen": 2702245888
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018434343434343435,
      "loss": 1.3631,
      "theoretical_loss": 3.3478404707001683,
      "tokens_seen": 2702311424
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018432323232323234,
      "loss": 1.5257,
      "theoretical_loss": 3.3478340949011915,
      "tokens_seen": 2702376960
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001843030303030303,
      "loss": 1.5492,
      "theoretical_loss": 3.3478277193001267,
      "tokens_seen": 2702442496
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018428282828282828,
      "loss": 1.0737,
      "theoretical_loss": 3.347821343896962,
      "tokens_seen": 2702508032
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018426262626262626,
      "loss": 1.2543,
      "theoretical_loss": 3.3478149686916883,
      "tokens_seen": 2702573568
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018424242424242425,
      "loss": 1.3264,
      "theoretical_loss": 3.3478085936842934,
      "tokens_seen": 2702639104
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018422222222222223,
      "loss": 1.3274,
      "theoretical_loss": 3.347802218874766,
      "tokens_seen": 2702704640
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018420202020202021,
      "loss": 1.5218,
      "theoretical_loss": 3.3477958442630964,
      "tokens_seen": 2702770176
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018418181818181817,
      "loss": 1.167,
      "theoretical_loss": 3.347789469849273,
      "tokens_seen": 2702835712
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018416161616161616,
      "loss": 1.1769,
      "theoretical_loss": 3.3477830956332846,
      "tokens_seen": 2702901248
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018414141414141414,
      "loss": 1.4332,
      "theoretical_loss": 3.3477767216151206,
      "tokens_seen": 2702966784
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018412121212121212,
      "loss": 1.2981,
      "theoretical_loss": 3.34777034779477,
      "tokens_seen": 2703032320
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001841010101010101,
      "loss": 1.2559,
      "theoretical_loss": 3.347763974172222,
      "tokens_seen": 2703097856
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018408080808080806,
      "loss": 1.1283,
      "theoretical_loss": 3.3477576007474656,
      "tokens_seen": 2703163392
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018406060606060608,
      "loss": 1.2547,
      "theoretical_loss": 3.3477512275204897,
      "tokens_seen": 2703228928
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018404040404040403,
      "loss": 1.2548,
      "theoretical_loss": 3.3477448544912836,
      "tokens_seen": 2703294464
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018402020202020204,
      "loss": 1.3688,
      "theoretical_loss": 3.347738481659836,
      "tokens_seen": 2703360000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000184,
      "loss": 1.2463,
      "theoretical_loss": 3.3477321090261363,
      "tokens_seen": 2703425536
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018397979797979798,
      "loss": 1.2738,
      "theoretical_loss": 3.347725736590174,
      "tokens_seen": 2703491072
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018395959595959597,
      "loss": 1.1768,
      "theoretical_loss": 3.347719364351937,
      "tokens_seen": 2703556608
    },
    {
      "epoch": 0.64,
      "objective/train/aligned_token_count": 0.2509775161743164,
      "objective/train/avg_token_score": 0.02934659831225872,
      "objective/train/docs_used": 1521568,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8301218152046204,
      "objective/train/original_loss": 5.349640846252441,
      "objective/train/theoretical_loss": 3.3477129923114153,
      "objective/train/tokens_used": 1062481376,
      "objective/train/unlikelihood_loss": 0.06215095892548561,
      "theoretical_loss": 3.3477129923114153,
      "tokens_seen": 2703622144
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018393939393939392,
      "loss": 1.1431,
      "theoretical_loss": 3.3477129923114153,
      "tokens_seen": 2703622144
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018391919191919194,
      "loss": 1.2441,
      "theoretical_loss": 3.347706620468598,
      "tokens_seen": 2703687680
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001838989898989899,
      "loss": 1.5488,
      "theoretical_loss": 3.3477002488234735,
      "tokens_seen": 2703753216
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001838787878787879,
      "loss": 1.1507,
      "theoretical_loss": 3.347693877376031,
      "tokens_seen": 2703818752
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018385858585858586,
      "loss": 1.4766,
      "theoretical_loss": 3.3476875061262605,
      "tokens_seen": 2703884288
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018383838383838384,
      "loss": 1.5484,
      "theoretical_loss": 3.3476811350741498,
      "tokens_seen": 2703949824
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018381818181818183,
      "loss": 1.2563,
      "theoretical_loss": 3.347674764219689,
      "tokens_seen": 2704015360
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018379797979797979,
      "loss": 1.2542,
      "theoretical_loss": 3.3476683935628664,
      "tokens_seen": 2704080896
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001837777777777778,
      "loss": 1.3255,
      "theoretical_loss": 3.3476620231036716,
      "tokens_seen": 2704146432
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018375757575757575,
      "loss": 1.0614,
      "theoretical_loss": 3.3476556528420938,
      "tokens_seen": 2704211968
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018373737373737374,
      "loss": 1.474,
      "theoretical_loss": 3.3476492827781215,
      "tokens_seen": 2704277504
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018371717171717172,
      "loss": 1.3191,
      "theoretical_loss": 3.3476429129117444,
      "tokens_seen": 2704343040
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001836969696969697,
      "loss": 1.2428,
      "theoretical_loss": 3.347636543242951,
      "tokens_seen": 2704408576
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001836767676767677,
      "loss": 1.284,
      "theoretical_loss": 3.347630173771731,
      "tokens_seen": 2704474112
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018365656565656567,
      "loss": 1.5234,
      "theoretical_loss": 3.3476238044980726,
      "tokens_seen": 2704539648
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018363636363636363,
      "loss": 1.4283,
      "theoretical_loss": 3.3476174354219657,
      "tokens_seen": 2704605184
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018361616161616161,
      "loss": 1.5448,
      "theoretical_loss": 3.347611066543399,
      "tokens_seen": 2704670720
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001835959595959596,
      "loss": 1.1179,
      "theoretical_loss": 3.347604697862362,
      "tokens_seen": 2704736256
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018357575757575758,
      "loss": 1.4118,
      "theoretical_loss": 3.347598329378843,
      "tokens_seen": 2704801792
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018355555555555557,
      "loss": 1.2892,
      "theoretical_loss": 3.347591961092832,
      "tokens_seen": 2704867328
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018353535353535352,
      "loss": 1.2164,
      "theoretical_loss": 3.3475855930043172,
      "tokens_seen": 2704932864
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018351515151515153,
      "loss": 1.3788,
      "theoretical_loss": 3.3475792251132885,
      "tokens_seen": 2704998400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001834949494949495,
      "loss": 1.3237,
      "theoretical_loss": 3.3475728574197348,
      "tokens_seen": 2705063936
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001834747474747475,
      "loss": 1.0409,
      "theoretical_loss": 3.347566489923645,
      "tokens_seen": 2705129472
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018345454545454546,
      "loss": 1.3473,
      "theoretical_loss": 3.347560122625008,
      "tokens_seen": 2705195008
    },
    {
      "epoch": 0.64,
      "objective/train/aligned_token_count": 0.41923263669013977,
      "objective/train/avg_token_score": 0.011428485624492168,
      "objective/train/docs_used": 1522374,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2564101219177246,
      "objective/train/original_loss": 3.5934181213378906,
      "objective/train/theoretical_loss": 3.3475537555238133,
      "objective/train/tokens_used": 1064119776,
      "objective/train/unlikelihood_loss": 0.07359922677278519,
      "theoretical_loss": 3.3475537555238133,
      "tokens_seen": 2705260544
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018343434343434342,
      "loss": 1.2171,
      "theoretical_loss": 3.3475537555238133,
      "tokens_seen": 2705260544
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018341414141414143,
      "loss": 1.4705,
      "theoretical_loss": 3.3475473886200495,
      "tokens_seen": 2705326080
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018339393939393938,
      "loss": 1.3634,
      "theoretical_loss": 3.347541021913706,
      "tokens_seen": 2705391616
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001833737373737374,
      "loss": 1.3771,
      "theoretical_loss": 3.3475346554047722,
      "tokens_seen": 2705457152
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018335353535353535,
      "loss": 1.3714,
      "theoretical_loss": 3.347528289093237,
      "tokens_seen": 2705522688
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018333333333333334,
      "loss": 1.4868,
      "theoretical_loss": 3.347521922979089,
      "tokens_seen": 2705588224
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018331313131313132,
      "loss": 1.5615,
      "theoretical_loss": 3.347515557062318,
      "tokens_seen": 2705653760
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018329292929292928,
      "loss": 1.182,
      "theoretical_loss": 3.3475091913429127,
      "tokens_seen": 2705719296
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001832727272727273,
      "loss": 1.2188,
      "theoretical_loss": 3.347502825820863,
      "tokens_seen": 2705784832
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018325252525252524,
      "loss": 1.2567,
      "theoretical_loss": 3.3474964604961563,
      "tokens_seen": 2705850368
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018323232323232323,
      "loss": 1.4506,
      "theoretical_loss": 3.3474900953687827,
      "tokens_seen": 2705915904
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001832121212121212,
      "loss": 1.4039,
      "theoretical_loss": 3.3474837304387317,
      "tokens_seen": 2705981440
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001831919191919192,
      "loss": 1.5292,
      "theoretical_loss": 3.347477365705992,
      "tokens_seen": 2706046976
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018317171717171718,
      "loss": 1.2975,
      "theoretical_loss": 3.3474710011705526,
      "tokens_seen": 2706112512
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018315151515151516,
      "loss": 1.3313,
      "theoretical_loss": 3.3474646368324033,
      "tokens_seen": 2706178048
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018313131313131315,
      "loss": 1.4695,
      "theoretical_loss": 3.347458272691532,
      "tokens_seen": 2706243584
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001831111111111111,
      "loss": 1.4079,
      "theoretical_loss": 3.3474519087479284,
      "tokens_seen": 2706309120
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001830909090909091,
      "loss": 1.3089,
      "theoretical_loss": 3.347445545001582,
      "tokens_seen": 2706374656
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018307070707070707,
      "loss": 1.2825,
      "theoretical_loss": 3.347439181452481,
      "tokens_seen": 2706440192
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018305050505050506,
      "loss": 1.2685,
      "theoretical_loss": 3.3474328181006157,
      "tokens_seen": 2706505728
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018303030303030304,
      "loss": 1.3746,
      "theoretical_loss": 3.3474264549459742,
      "tokens_seen": 2706571264
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018301010101010102,
      "loss": 1.3402,
      "theoretical_loss": 3.347420091988546,
      "tokens_seen": 2706636800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018298989898989898,
      "loss": 1.3627,
      "theoretical_loss": 3.347413729228321,
      "tokens_seen": 2706702336
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000182969696969697,
      "loss": 1.3207,
      "theoretical_loss": 3.3474073666652866,
      "tokens_seen": 2706767872
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018294949494949495,
      "loss": 1.4254,
      "theoretical_loss": 3.3474010042994333,
      "tokens_seen": 2706833408
    },
    {
      "epoch": 0.64,
      "objective/train/aligned_token_count": 0.45546188950538635,
      "objective/train/avg_token_score": 0.025383848696947098,
      "objective/train/docs_used": 1523555,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3006609678268433,
      "objective/train/original_loss": 4.263200759887695,
      "objective/train/theoretical_loss": 3.347394642130749,
      "objective/train/tokens_used": 1065758176,
      "objective/train/unlikelihood_loss": 0.042930908501148224,
      "theoretical_loss": 3.347394642130749,
      "tokens_seen": 2706898944
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018292929292929293,
      "loss": 1.2484,
      "theoretical_loss": 3.347394642130749,
      "tokens_seen": 2706898944
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018290909090909092,
      "loss": 1.502,
      "theoretical_loss": 3.3473882801592243,
      "tokens_seen": 2706964480
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018288888888888887,
      "loss": 1.2744,
      "theoretical_loss": 3.3473819183848477,
      "tokens_seen": 2707030016
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018286868686868689,
      "loss": 1.484,
      "theoretical_loss": 3.3473755568076076,
      "tokens_seen": 2707095552
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018284848484848484,
      "loss": 1.1044,
      "theoretical_loss": 3.3473691954274942,
      "tokens_seen": 2707161088
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018282828282828285,
      "loss": 1.3974,
      "theoretical_loss": 3.3473628342444957,
      "tokens_seen": 2707226624
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001828080808080808,
      "loss": 1.2055,
      "theoretical_loss": 3.347356473258602,
      "tokens_seen": 2707292160
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001827878787878788,
      "loss": 1.5096,
      "theoretical_loss": 3.347350112469802,
      "tokens_seen": 2707357696
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018276767676767678,
      "loss": 1.1181,
      "theoretical_loss": 3.347343751878084,
      "tokens_seen": 2707423232
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018274747474747473,
      "loss": 1.6094,
      "theoretical_loss": 3.3473373914834386,
      "tokens_seen": 2707488768
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018272727272727275,
      "loss": 1.2696,
      "theoretical_loss": 3.3473310312858535,
      "tokens_seen": 2707554304
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001827070707070707,
      "loss": 1.3617,
      "theoretical_loss": 3.3473246712853193,
      "tokens_seen": 2707619840
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001826868686868687,
      "loss": 1.3157,
      "theoretical_loss": 3.347318311481824,
      "tokens_seen": 2707685376
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018266666666666667,
      "loss": 1.358,
      "theoretical_loss": 3.3473119518753567,
      "tokens_seen": 2707750912
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018264646464646465,
      "loss": 1.1904,
      "theoretical_loss": 3.347305592465907,
      "tokens_seen": 2707816448
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018262626262626264,
      "loss": 1.3125,
      "theoretical_loss": 3.3472992332534637,
      "tokens_seen": 2707881984
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018260606060606062,
      "loss": 1.2993,
      "theoretical_loss": 3.347292874238016,
      "tokens_seen": 2707947520
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018258585858585858,
      "loss": 1.6713,
      "theoretical_loss": 3.3472865154195537,
      "tokens_seen": 2708013056
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018256565656565656,
      "loss": 1.5167,
      "theoretical_loss": 3.3472801567980652,
      "tokens_seen": 2708078592
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018254545454545455,
      "loss": 1.2505,
      "theoretical_loss": 3.34727379837354,
      "tokens_seen": 2708144128
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018252525252525253,
      "loss": 1.3944,
      "theoretical_loss": 3.3472674401459663,
      "tokens_seen": 2708209664
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018250505050505052,
      "loss": 1.2206,
      "theoretical_loss": 3.3472610821153346,
      "tokens_seen": 2708275200
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018248484848484847,
      "loss": 1.1937,
      "theoretical_loss": 3.3472547242816333,
      "tokens_seen": 2708340736
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018246464646464648,
      "loss": 1.3628,
      "theoretical_loss": 3.347248366644852,
      "tokens_seen": 2708406272
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018244444444444444,
      "loss": 1.5501,
      "theoretical_loss": 3.3472420092049786,
      "tokens_seen": 2708471808
    },
    {
      "epoch": 0.64,
      "objective/train/aligned_token_count": 0.3534335196018219,
      "objective/train/avg_token_score": 0.02238118276000023,
      "objective/train/docs_used": 1524121,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1129951477050781,
      "objective/train/original_loss": 4.4213457107543945,
      "objective/train/theoretical_loss": 3.347235651962004,
      "objective/train/tokens_used": 1067396576,
      "objective/train/unlikelihood_loss": 0.06573837995529175,
      "theoretical_loss": 3.347235651962004,
      "tokens_seen": 2708537344
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018242424242424242,
      "loss": 1.2166,
      "theoretical_loss": 3.347235651962004,
      "tokens_seen": 2708537344
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001824040404040404,
      "loss": 1.3339,
      "theoretical_loss": 3.3472292949159157,
      "tokens_seen": 2708602880
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018238383838383836,
      "loss": 1.4059,
      "theoretical_loss": 3.347222938066704,
      "tokens_seen": 2708668416
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018236363636363638,
      "loss": 1.1798,
      "theoretical_loss": 3.3472165814143575,
      "tokens_seen": 2708733952
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018234343434343433,
      "loss": 1.2124,
      "theoretical_loss": 3.3472102249588653,
      "tokens_seen": 2708799488
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018232323232323234,
      "loss": 1.2814,
      "theoretical_loss": 3.347203868700217,
      "tokens_seen": 2708865024
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001823030303030303,
      "loss": 1.1416,
      "theoretical_loss": 3.3471975126384015,
      "tokens_seen": 2708930560
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001822828282828283,
      "loss": 1.2844,
      "theoretical_loss": 3.347191156773408,
      "tokens_seen": 2708996096
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018226262626262627,
      "loss": 1.38,
      "theoretical_loss": 3.347184801105225,
      "tokens_seen": 2709061632
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018224242424242423,
      "loss": 1.4442,
      "theoretical_loss": 3.3471784456338427,
      "tokens_seen": 2709127168
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018222222222222224,
      "loss": 1.3995,
      "theoretical_loss": 3.3471720903592495,
      "tokens_seen": 2709192704
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001822020202020202,
      "loss": 1.3205,
      "theoretical_loss": 3.3471657352814352,
      "tokens_seen": 2709258240
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001821818181818182,
      "loss": 1.3927,
      "theoretical_loss": 3.347159380400388,
      "tokens_seen": 2709323776
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018216161616161616,
      "loss": 1.2845,
      "theoretical_loss": 3.3471530257160977,
      "tokens_seen": 2709389312
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018214141414141415,
      "loss": 1.5644,
      "theoretical_loss": 3.347146671228553,
      "tokens_seen": 2709454848
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018212121212121213,
      "loss": 1.2501,
      "theoretical_loss": 3.347140316937744,
      "tokens_seen": 2709520384
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001821010101010101,
      "loss": 1.2903,
      "theoretical_loss": 3.347133962843659,
      "tokens_seen": 2709585920
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001820808080808081,
      "loss": 1.3415,
      "theoretical_loss": 3.3471276089462876,
      "tokens_seen": 2709651456
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018206060606060605,
      "loss": 1.2406,
      "theoretical_loss": 3.3471212552456184,
      "tokens_seen": 2709716992
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018204040404040404,
      "loss": 1.4192,
      "theoretical_loss": 3.347114901741641,
      "tokens_seen": 2709782528
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018202020202020202,
      "loss": 1.2541,
      "theoretical_loss": 3.347108548434344,
      "tokens_seen": 2709848064
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000182,
      "loss": 1.1603,
      "theoretical_loss": 3.3471021953237177,
      "tokens_seen": 2709913600
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000181979797979798,
      "loss": 1.2542,
      "theoretical_loss": 3.34709584240975,
      "tokens_seen": 2709979136
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018195959595959597,
      "loss": 1.4076,
      "theoretical_loss": 3.347089489692431,
      "tokens_seen": 2710044672
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018193939393939393,
      "loss": 1.5613,
      "theoretical_loss": 3.347083137171749,
      "tokens_seen": 2710110208
    },
    {
      "epoch": 0.64,
      "objective/train/aligned_token_count": 0.24792277812957764,
      "objective/train/avg_token_score": 0.012780007906258106,
      "objective/train/docs_used": 1525702,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8622182607650757,
      "objective/train/original_loss": 5.701241493225098,
      "objective/train/theoretical_loss": 3.347076784847694,
      "objective/train/tokens_used": 1069034976,
      "objective/train/unlikelihood_loss": 0.04030083492398262,
      "theoretical_loss": 3.347076784847694,
      "tokens_seen": 2710175744
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018191919191919194,
      "loss": 1.4114,
      "theoretical_loss": 3.347076784847694,
      "tokens_seen": 2710175744
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001818989898989899,
      "loss": 1.174,
      "theoretical_loss": 3.3470704327202547,
      "tokens_seen": 2710241280
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018187878787878788,
      "loss": 1.1185,
      "theoretical_loss": 3.3470640807894205,
      "tokens_seen": 2710306816
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018185858585858587,
      "loss": 1.4551,
      "theoretical_loss": 3.34705772905518,
      "tokens_seen": 2710372352
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018183838383838382,
      "loss": 1.2167,
      "theoretical_loss": 3.347051377517523,
      "tokens_seen": 2710437888
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018181818181818183,
      "loss": 1.2771,
      "theoretical_loss": 3.3470450261764384,
      "tokens_seen": 2710503424
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001817979797979798,
      "loss": 1.5323,
      "theoretical_loss": 3.3470386750319157,
      "tokens_seen": 2710568960
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001817777777777778,
      "loss": 1.3532,
      "theoretical_loss": 3.3470323240839432,
      "tokens_seen": 2710634496
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018175757575757576,
      "loss": 1.4138,
      "theoretical_loss": 3.347025973332511,
      "tokens_seen": 2710700032
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018173737373737372,
      "loss": 1.6048,
      "theoretical_loss": 3.3470196227776077,
      "tokens_seen": 2710765568
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018171717171717173,
      "loss": 1.4522,
      "theoretical_loss": 3.3470132724192228,
      "tokens_seen": 2710831104
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018169696969696968,
      "loss": 1.2549,
      "theoretical_loss": 3.3470069222573455,
      "tokens_seen": 2710896640
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001816767676767677,
      "loss": 1.2111,
      "theoretical_loss": 3.3470005722919645,
      "tokens_seen": 2710962176
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018165656565656565,
      "loss": 1.3403,
      "theoretical_loss": 3.346994222523069,
      "tokens_seen": 2711027712
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018163636363636364,
      "loss": 1.3154,
      "theoretical_loss": 3.346987872950649,
      "tokens_seen": 2711093248
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018161616161616162,
      "loss": 1.4739,
      "theoretical_loss": 3.3469815235746925,
      "tokens_seen": 2711158784
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001815959595959596,
      "loss": 1.5379,
      "theoretical_loss": 3.34697517439519,
      "tokens_seen": 2711224320
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001815757575757576,
      "loss": 1.4534,
      "theoretical_loss": 3.3469688254121293,
      "tokens_seen": 2711289856
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018155555555555554,
      "loss": 1.4741,
      "theoretical_loss": 3.3469624766255004,
      "tokens_seen": 2711355392
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018153535353535353,
      "loss": 1.4875,
      "theoretical_loss": 3.346956128035292,
      "tokens_seen": 2711420928
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001815151515151515,
      "loss": 1.0798,
      "theoretical_loss": 3.3469497796414944,
      "tokens_seen": 2711486464
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001814949494949495,
      "loss": 1.2007,
      "theoretical_loss": 3.3469434314440956,
      "tokens_seen": 2711552000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018147474747474748,
      "loss": 1.6392,
      "theoretical_loss": 3.346937083443085,
      "tokens_seen": 2711617536
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018145454545454546,
      "loss": 1.3404,
      "theoretical_loss": 3.346930735638452,
      "tokens_seen": 2711683072
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018143434343434345,
      "loss": 1.4254,
      "theoretical_loss": 3.3469243880301853,
      "tokens_seen": 2711748608
    },
    {
      "epoch": 0.64,
      "objective/train/aligned_token_count": 0.32801806926727295,
      "objective/train/avg_token_score": 0.024438899010419846,
      "objective/train/docs_used": 1526439,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0598777532577515,
      "objective/train/original_loss": 4.360488414764404,
      "objective/train/theoretical_loss": 3.3469180406182746,
      "objective/train/tokens_used": 1070673376,
      "objective/train/unlikelihood_loss": 0.06372426450252533,
      "theoretical_loss": 3.3469180406182746,
      "tokens_seen": 2711814144
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018141414141414143,
      "loss": 1.4836,
      "theoretical_loss": 3.3469180406182746,
      "tokens_seen": 2711814144
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001813939393939394,
      "loss": 1.265,
      "theoretical_loss": 3.3469116934027094,
      "tokens_seen": 2711879680
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018137373737373737,
      "loss": 1.4132,
      "theoretical_loss": 3.346905346383478,
      "tokens_seen": 2711945216
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018135353535353536,
      "loss": 1.3565,
      "theoretical_loss": 3.3468989995605702,
      "tokens_seen": 2712010752
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018133333333333334,
      "loss": 1.285,
      "theoretical_loss": 3.346892652933975,
      "tokens_seen": 2712076288
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018131313131313132,
      "loss": 1.3542,
      "theoretical_loss": 3.3468863065036816,
      "tokens_seen": 2712141824
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018129292929292928,
      "loss": 1.2502,
      "theoretical_loss": 3.346879960269679,
      "tokens_seen": 2712207360
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001812727272727273,
      "loss": 1.3651,
      "theoretical_loss": 3.3468736142319564,
      "tokens_seen": 2712272896
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018125252525252525,
      "loss": 1.3428,
      "theoretical_loss": 3.346867268390503,
      "tokens_seen": 2712338432
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018123232323232326,
      "loss": 1.4555,
      "theoretical_loss": 3.346860922745309,
      "tokens_seen": 2712403968
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018121212121212122,
      "loss": 1.3304,
      "theoretical_loss": 3.3468545772963623,
      "tokens_seen": 2712469504
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018119191919191917,
      "loss": 1.423,
      "theoretical_loss": 3.3468482320436523,
      "tokens_seen": 2712535040
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018117171717171719,
      "loss": 1.3872,
      "theoretical_loss": 3.3468418869871686,
      "tokens_seen": 2712600576
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018115151515151514,
      "loss": 1.1861,
      "theoretical_loss": 3.3468355421269,
      "tokens_seen": 2712666112
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018113131313131315,
      "loss": 1.4095,
      "theoretical_loss": 3.346829197462836,
      "tokens_seen": 2712731648
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001811111111111111,
      "loss": 1.5423,
      "theoretical_loss": 3.3468228529949653,
      "tokens_seen": 2712797184
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001810909090909091,
      "loss": 1.3876,
      "theoretical_loss": 3.3468165087232777,
      "tokens_seen": 2712862720
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018107070707070708,
      "loss": 1.5401,
      "theoretical_loss": 3.3468101646477626,
      "tokens_seen": 2712928256
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018105050505050506,
      "loss": 1.4211,
      "theoretical_loss": 3.3468038207684083,
      "tokens_seen": 2712993792
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018103030303030305,
      "loss": 1.3705,
      "theoretical_loss": 3.3467974770852043,
      "tokens_seen": 2713059328
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000181010101010101,
      "loss": 1.3935,
      "theoretical_loss": 3.34679113359814,
      "tokens_seen": 2713124864
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000180989898989899,
      "loss": 1.3474,
      "theoretical_loss": 3.346784790307205,
      "tokens_seen": 2713190400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018096969696969697,
      "loss": 1.3571,
      "theoretical_loss": 3.3467784472123876,
      "tokens_seen": 2713255936
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018094949494949495,
      "loss": 1.4277,
      "theoretical_loss": 3.3467721043136773,
      "tokens_seen": 2713321472
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018092929292929294,
      "loss": 1.3723,
      "theoretical_loss": 3.346765761611064,
      "tokens_seen": 2713387008
    },
    {
      "epoch": 0.64,
      "objective/train/aligned_token_count": 0.3475073277950287,
      "objective/train/avg_token_score": 0.040864743292331696,
      "objective/train/docs_used": 1527722,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1181749105453491,
      "objective/train/original_loss": 5.045957565307617,
      "objective/train/theoretical_loss": 3.346759419104536,
      "objective/train/tokens_used": 1072311776,
      "objective/train/unlikelihood_loss": 0.0420418381690979,
      "theoretical_loss": 3.346759419104536,
      "tokens_seen": 2713452544
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018090909090909092,
      "loss": 1.3673,
      "theoretical_loss": 3.346759419104536,
      "tokens_seen": 2713452544
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018088888888888888,
      "loss": 1.5371,
      "theoretical_loss": 3.3467530767940827,
      "tokens_seen": 2713518080
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018086868686868686,
      "loss": 1.3391,
      "theoretical_loss": 3.346746734679694,
      "tokens_seen": 2713583616
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018084848484848485,
      "loss": 1.3223,
      "theoretical_loss": 3.346740392761358,
      "tokens_seen": 2713649152
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018082828282828283,
      "loss": 1.3599,
      "theoretical_loss": 3.3467340510390646,
      "tokens_seen": 2713714688
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018080808080808082,
      "loss": 1.3166,
      "theoretical_loss": 3.346727709512803,
      "tokens_seen": 2713780224
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018078787878787877,
      "loss": 1.3474,
      "theoretical_loss": 3.3467213681825623,
      "tokens_seen": 2713845760
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018076767676767678,
      "loss": 1.5393,
      "theoretical_loss": 3.3467150270483312,
      "tokens_seen": 2713911296
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018074747474747474,
      "loss": 1.2385,
      "theoretical_loss": 3.3467086861101,
      "tokens_seen": 2713976832
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018072727272727275,
      "loss": 1.3036,
      "theoretical_loss": 3.346702345367857,
      "tokens_seen": 2714042368
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001807070707070707,
      "loss": 1.3354,
      "theoretical_loss": 3.3466960048215917,
      "tokens_seen": 2714107904
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018068686868686867,
      "loss": 1.3355,
      "theoretical_loss": 3.3466896644712936,
      "tokens_seen": 2714173440
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018066666666666668,
      "loss": 1.565,
      "theoretical_loss": 3.3466833243169516,
      "tokens_seen": 2714238976
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018064646464646463,
      "loss": 1.312,
      "theoretical_loss": 3.3466769843585547,
      "tokens_seen": 2714304512
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018062626262626264,
      "loss": 1.59,
      "theoretical_loss": 3.346670644596092,
      "tokens_seen": 2714370048
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001806060606060606,
      "loss": 1.3035,
      "theoretical_loss": 3.3466643050295537,
      "tokens_seen": 2714435584
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001805858585858586,
      "loss": 1.4579,
      "theoretical_loss": 3.346657965658928,
      "tokens_seen": 2714501120
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018056565656565657,
      "loss": 1.319,
      "theoretical_loss": 3.3466516264842046,
      "tokens_seen": 2714566656
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018054545454545455,
      "loss": 1.362,
      "theoretical_loss": 3.346645287505373,
      "tokens_seen": 2714632192
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018052525252525254,
      "loss": 1.4272,
      "theoretical_loss": 3.3466389487224215,
      "tokens_seen": 2714697728
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001805050505050505,
      "loss": 1.6075,
      "theoretical_loss": 3.34663261013534,
      "tokens_seen": 2714763264
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001804848484848485,
      "loss": 1.3248,
      "theoretical_loss": 3.3466262717441175,
      "tokens_seen": 2714828800
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018046464646464646,
      "loss": 1.5162,
      "theoretical_loss": 3.346619933548743,
      "tokens_seen": 2714894336
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018044444444444445,
      "loss": 1.3405,
      "theoretical_loss": 3.346613595549207,
      "tokens_seen": 2714959872
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018042424242424243,
      "loss": 1.4374,
      "theoretical_loss": 3.3466072577454966,
      "tokens_seen": 2715025408
    },
    {
      "epoch": 0.64,
      "objective/train/aligned_token_count": 0.4741568863391876,
      "objective/train/avg_token_score": 0.010101846419274807,
      "objective/train/docs_used": 1528322,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6042120456695557,
      "objective/train/original_loss": 4.234947204589844,
      "objective/train/theoretical_loss": 3.346600920137603,
      "objective/train/tokens_used": 1073950176,
      "objective/train/unlikelihood_loss": 0.04748011752963066,
      "theoretical_loss": 3.346600920137603,
      "tokens_seen": 2715090944
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001804040404040404,
      "loss": 1.4297,
      "theoretical_loss": 3.346600920137603,
      "tokens_seen": 2715090944
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001803838383838384,
      "loss": 1.6236,
      "theoretical_loss": 3.346594582725514,
      "tokens_seen": 2715156480
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018036363636363638,
      "loss": 1.5581,
      "theoretical_loss": 3.3465882455092197,
      "tokens_seen": 2715222016
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018034343434343434,
      "loss": 1.2004,
      "theoretical_loss": 3.346581908488709,
      "tokens_seen": 2715287552
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018032323232323232,
      "loss": 1.5535,
      "theoretical_loss": 3.3465755716639713,
      "tokens_seen": 2715353088
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001803030303030303,
      "loss": 1.305,
      "theoretical_loss": 3.3465692350349956,
      "tokens_seen": 2715418624
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001802828282828283,
      "loss": 1.5925,
      "theoretical_loss": 3.3465628986017713,
      "tokens_seen": 2715484160
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018026262626262627,
      "loss": 1.0743,
      "theoretical_loss": 3.3465565623642872,
      "tokens_seen": 2715549696
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018024242424242423,
      "loss": 1.2635,
      "theoretical_loss": 3.3465502263225333,
      "tokens_seen": 2715615232
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018022222222222224,
      "loss": 1.5623,
      "theoretical_loss": 3.346543890476498,
      "tokens_seen": 2715680768
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001802020202020202,
      "loss": 1.4707,
      "theoretical_loss": 3.346537554826171,
      "tokens_seen": 2715746304
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001801818181818182,
      "loss": 1.2256,
      "theoretical_loss": 3.3465312193715415,
      "tokens_seen": 2715811840
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018016161616161617,
      "loss": 1.246,
      "theoretical_loss": 3.3465248841125987,
      "tokens_seen": 2715877376
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018014141414141412,
      "loss": 1.323,
      "theoretical_loss": 3.346518549049332,
      "tokens_seen": 2715942912
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018012121212121213,
      "loss": 1.5507,
      "theoretical_loss": 3.3465122141817303,
      "tokens_seen": 2716008448
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001801010101010101,
      "loss": 1.1997,
      "theoretical_loss": 3.3465058795097833,
      "tokens_seen": 2716073984
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001800808080808081,
      "loss": 1.1439,
      "theoretical_loss": 3.3464995450334794,
      "tokens_seen": 2716139520
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018006060606060606,
      "loss": 1.2405,
      "theoretical_loss": 3.3464932107528087,
      "tokens_seen": 2716205056
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018004040404040404,
      "loss": 1.3241,
      "theoretical_loss": 3.3464868766677602,
      "tokens_seen": 2716270592
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00018002020202020203,
      "loss": 1.4814,
      "theoretical_loss": 3.346480542778323,
      "tokens_seen": 2716336128
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017999999999999998,
      "loss": 1.3656,
      "theoretical_loss": 3.3464742090844863,
      "tokens_seen": 2716401664
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000179979797979798,
      "loss": 1.4343,
      "theoretical_loss": 3.346467875586239,
      "tokens_seen": 2716467200
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017995959595959595,
      "loss": 1.1505,
      "theoretical_loss": 3.3464615422835715,
      "tokens_seen": 2716532736
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017993939393939394,
      "loss": 1.2615,
      "theoretical_loss": 3.346455209176472,
      "tokens_seen": 2716598272
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017991919191919192,
      "loss": 1.3119,
      "theoretical_loss": 3.34644887626493,
      "tokens_seen": 2716663808
    },
    {
      "epoch": 0.64,
      "objective/train/aligned_token_count": 0.2854349911212921,
      "objective/train/avg_token_score": 0.01961679197847843,
      "objective/train/docs_used": 1529677,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8881481289863586,
      "objective/train/original_loss": 4.9914937019348145,
      "objective/train/theoretical_loss": 3.346442543548935,
      "objective/train/tokens_used": 1075588576,
      "objective/train/unlikelihood_loss": 0.0534181110560894,
      "theoretical_loss": 3.346442543548935,
      "tokens_seen": 2716729344
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001798989898989899,
      "loss": 1.1748,
      "theoretical_loss": 3.346442543548935,
      "tokens_seen": 2716729344
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001798787878787879,
      "loss": 1.5484,
      "theoretical_loss": 3.346436211028476,
      "tokens_seen": 2716794880
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017985858585858587,
      "loss": 1.5798,
      "theoretical_loss": 3.346429878703542,
      "tokens_seen": 2716860416
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017983838383838383,
      "loss": 1.3668,
      "theoretical_loss": 3.346423546574123,
      "tokens_seen": 2716925952
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001798181818181818,
      "loss": 1.2053,
      "theoretical_loss": 3.3464172146402076,
      "tokens_seen": 2716991488
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001797979797979798,
      "loss": 1.2352,
      "theoretical_loss": 3.346410882901785,
      "tokens_seen": 2717057024
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017977777777777778,
      "loss": 1.2733,
      "theoretical_loss": 3.346404551358845,
      "tokens_seen": 2717122560
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017975757575757576,
      "loss": 1.4311,
      "theoretical_loss": 3.3463982200113764,
      "tokens_seen": 2717188096
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017973737373737375,
      "loss": 1.2709,
      "theoretical_loss": 3.3463918888593684,
      "tokens_seen": 2717253632
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017971717171717173,
      "loss": 1.3543,
      "theoretical_loss": 3.346385557902811,
      "tokens_seen": 2717319168
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001796969696969697,
      "loss": 1.6125,
      "theoretical_loss": 3.3463792271416923,
      "tokens_seen": 2717384704
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001796767676767677,
      "loss": 1.314,
      "theoretical_loss": 3.3463728965760025,
      "tokens_seen": 2717450240
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017965656565656566,
      "loss": 1.4822,
      "theoretical_loss": 3.34636656620573,
      "tokens_seen": 2717515776
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017963636363636364,
      "loss": 1.4892,
      "theoretical_loss": 3.3463602360308653,
      "tokens_seen": 2717581312
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017961616161616163,
      "loss": 1.3496,
      "theoretical_loss": 3.3463539060513963,
      "tokens_seen": 2717646848
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017959595959595958,
      "loss": 1.7039,
      "theoretical_loss": 3.346347576267313,
      "tokens_seen": 2717712384
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001795757575757576,
      "loss": 1.4575,
      "theoretical_loss": 3.346341246678605,
      "tokens_seen": 2717777920
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017955555555555555,
      "loss": 1.4378,
      "theoretical_loss": 3.346334917285261,
      "tokens_seen": 2717843456
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017953535353535356,
      "loss": 1.4636,
      "theoretical_loss": 3.3463285880872697,
      "tokens_seen": 2717908992
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017951515151515152,
      "loss": 1.391,
      "theoretical_loss": 3.346322259084621,
      "tokens_seen": 2717974528
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001794949494949495,
      "loss": 1.4312,
      "theoretical_loss": 3.3463159302773047,
      "tokens_seen": 2718040064
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017947474747474749,
      "loss": 1.1851,
      "theoretical_loss": 3.3463096016653093,
      "tokens_seen": 2718105600
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017945454545454544,
      "loss": 1.3369,
      "theoretical_loss": 3.3463032732486244,
      "tokens_seen": 2718171136
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00017943434343434345,
      "loss": 1.4794,
      "theoretical_loss": 3.3462969450272393,
      "tokens_seen": 2718236672
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0001794141414141414,
      "loss": 1.2989,
      "theoretical_loss": 3.346290617001143,
      "tokens_seen": 2718302208
    },
    {
      "epoch": 0.64,
      "objective/train/aligned_token_count": 0.5992180109024048,
      "objective/train/avg_token_score": 0.04368416592478752,
      "objective/train/docs_used": 1530384,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8810855150222778,
      "objective/train/original_loss": 4.055726051330566,
      "objective/train/theoretical_loss": 3.3462842891703244,
      "objective/train/tokens_used": 1077226976,
      "objective/train/unlikelihood_loss": 0.062255896627902985,
      "theoretical_loss": 3.3462842891703244,
      "tokens_seen": 2718367744
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001793939393939394,
      "loss": 1.5774,
      "theoretical_loss": 3.3462842891703244,
      "tokens_seen": 2718367744
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017937373737373738,
      "loss": 1.4215,
      "theoretical_loss": 3.3462779615347737,
      "tokens_seen": 2718433280
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017935353535353536,
      "loss": 1.4765,
      "theoretical_loss": 3.34627163409448,
      "tokens_seen": 2718498816
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017933333333333335,
      "loss": 1.3733,
      "theoretical_loss": 3.3462653068494315,
      "tokens_seen": 2718564352
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001793131313131313,
      "loss": 1.3719,
      "theoretical_loss": 3.346258979799619,
      "tokens_seen": 2718629888
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001792929292929293,
      "loss": 1.3414,
      "theoretical_loss": 3.3462526529450303,
      "tokens_seen": 2718695424
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017927272727272727,
      "loss": 1.5134,
      "theoretical_loss": 3.346246326285656,
      "tokens_seen": 2718760960
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017925252525252526,
      "loss": 1.2312,
      "theoretical_loss": 3.3462399998214845,
      "tokens_seen": 2718826496
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017923232323232324,
      "loss": 1.2117,
      "theoretical_loss": 3.3462336735525056,
      "tokens_seen": 2718892032
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017921212121212122,
      "loss": 1.415,
      "theoretical_loss": 3.346227347478708,
      "tokens_seen": 2718957568
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017919191919191918,
      "loss": 1.3907,
      "theoretical_loss": 3.3462210216000816,
      "tokens_seen": 2719023104
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001791717171717172,
      "loss": 1.3395,
      "theoretical_loss": 3.346214695916615,
      "tokens_seen": 2719088640
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017915151515151515,
      "loss": 1.2748,
      "theoretical_loss": 3.3462083704282977,
      "tokens_seen": 2719154176
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017913131313131313,
      "loss": 1.555,
      "theoretical_loss": 3.3462020451351195,
      "tokens_seen": 2719219712
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017911111111111112,
      "loss": 1.3393,
      "theoretical_loss": 3.3461957200370693,
      "tokens_seen": 2719285248
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017909090909090907,
      "loss": 1.0866,
      "theoretical_loss": 3.346189395134136,
      "tokens_seen": 2719350784
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017907070707070708,
      "loss": 1.4287,
      "theoretical_loss": 3.3461830704263096,
      "tokens_seen": 2719416320
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017905050505050504,
      "loss": 1.2095,
      "theoretical_loss": 3.3461767459135787,
      "tokens_seen": 2719481856
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017903030303030305,
      "loss": 1.2844,
      "theoretical_loss": 3.346170421595933,
      "tokens_seen": 2719547392
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000179010101010101,
      "loss": 1.3897,
      "theoretical_loss": 3.346164097473362,
      "tokens_seen": 2719612928
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000178989898989899,
      "loss": 1.5116,
      "theoretical_loss": 3.3461577735458548,
      "tokens_seen": 2719678464
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017896969696969698,
      "loss": 1.3025,
      "theoretical_loss": 3.3461514498133997,
      "tokens_seen": 2719744000
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017894949494949493,
      "loss": 1.3767,
      "theoretical_loss": 3.346145126275988,
      "tokens_seen": 2719809536
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017892929292929294,
      "loss": 1.5236,
      "theoretical_loss": 3.3461388029336065,
      "tokens_seen": 2719875072
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001789090909090909,
      "loss": 1.1506,
      "theoretical_loss": 3.3461324797862466,
      "tokens_seen": 2719940608
    },
    {
      "epoch": 0.65,
      "objective/train/aligned_token_count": 0.41526147723197937,
      "objective/train/avg_token_score": 0.00773067818954587,
      "objective/train/docs_used": 1531036,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2557531595230103,
      "objective/train/original_loss": 4.072418212890625,
      "objective/train/theoretical_loss": 3.346126156833897,
      "objective/train/tokens_used": 1078865376,
      "objective/train/unlikelihood_loss": 0.08055424690246582,
      "theoretical_loss": 3.346126156833897,
      "tokens_seen": 2720006144
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017888888888888889,
      "loss": 1.3546,
      "theoretical_loss": 3.346126156833897,
      "tokens_seen": 2720006144
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017886868686868687,
      "loss": 1.3315,
      "theoretical_loss": 3.3461198340765463,
      "tokens_seen": 2720071680
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017884848484848485,
      "loss": 1.4495,
      "theoretical_loss": 3.3461135115141842,
      "tokens_seen": 2720137216
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017882828282828284,
      "loss": 1.4473,
      "theoretical_loss": 3.3461071891468004,
      "tokens_seen": 2720202752
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017880808080808082,
      "loss": 1.4507,
      "theoretical_loss": 3.3461008669743837,
      "tokens_seen": 2720268288
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001787878787878788,
      "loss": 1.3204,
      "theoretical_loss": 3.3460945449969235,
      "tokens_seen": 2720333824
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017876767676767676,
      "loss": 1.4504,
      "theoretical_loss": 3.346088223214409,
      "tokens_seen": 2720399360
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017874747474747475,
      "loss": 1.3642,
      "theoretical_loss": 3.34608190162683,
      "tokens_seen": 2720464896
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017872727272727273,
      "loss": 1.3966,
      "theoretical_loss": 3.3460755802341753,
      "tokens_seen": 2720530432
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017870707070707071,
      "loss": 1.4953,
      "theoretical_loss": 3.346069259036434,
      "tokens_seen": 2720595968
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001786868686868687,
      "loss": 1.5027,
      "theoretical_loss": 3.346062938033596,
      "tokens_seen": 2720661504
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017866666666666668,
      "loss": 1.2443,
      "theoretical_loss": 3.34605661722565,
      "tokens_seen": 2720727040
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017864646464646464,
      "loss": 1.3989,
      "theoretical_loss": 3.346050296612586,
      "tokens_seen": 2720792576
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017862626262626265,
      "loss": 1.3895,
      "theoretical_loss": 3.3460439761943928,
      "tokens_seen": 2720858112
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001786060606060606,
      "loss": 1.3675,
      "theoretical_loss": 3.3460376559710596,
      "tokens_seen": 2720923648
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001785858585858586,
      "loss": 1.4324,
      "theoretical_loss": 3.346031335942576,
      "tokens_seen": 2720989184
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017856565656565657,
      "loss": 1.4431,
      "theoretical_loss": 3.346025016108931,
      "tokens_seen": 2721054720
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017854545454545453,
      "loss": 1.3426,
      "theoretical_loss": 3.3460186964701144,
      "tokens_seen": 2721120256
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017852525252525254,
      "loss": 1.3557,
      "theoretical_loss": 3.3460123770261148,
      "tokens_seen": 2721185792
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001785050505050505,
      "loss": 1.2746,
      "theoretical_loss": 3.3460060577769224,
      "tokens_seen": 2721251328
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001784848484848485,
      "loss": 1.2434,
      "theoretical_loss": 3.3459997387225258,
      "tokens_seen": 2721316864
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017846464646464647,
      "loss": 1.2422,
      "theoretical_loss": 3.345993419862914,
      "tokens_seen": 2721382400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017844444444444442,
      "loss": 1.4216,
      "theoretical_loss": 3.3459871011980775,
      "tokens_seen": 2721447936
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017842424242424244,
      "loss": 1.4252,
      "theoretical_loss": 3.3459807827280046,
      "tokens_seen": 2721513472
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001784040404040404,
      "loss": 1.3636,
      "theoretical_loss": 3.345974464452685,
      "tokens_seen": 2721579008
    },
    {
      "epoch": 0.65,
      "objective/train/aligned_token_count": 0.42662513256073,
      "objective/train/avg_token_score": 0.01378691103309393,
      "objective/train/docs_used": 1532195,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.415401816368103,
      "objective/train/original_loss": 4.73554801940918,
      "objective/train/theoretical_loss": 3.3459681463721083,
      "objective/train/tokens_used": 1080503776,
      "objective/train/unlikelihood_loss": 0.04374285414814949,
      "theoretical_loss": 3.3459681463721083,
      "tokens_seen": 2721644544
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001783838383838384,
      "loss": 1.4891,
      "theoretical_loss": 3.3459681463721083,
      "tokens_seen": 2721644544
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017836363636363636,
      "loss": 0.9695,
      "theoretical_loss": 3.3459618284862627,
      "tokens_seen": 2721710080
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017834343434343434,
      "loss": 1.2317,
      "theoretical_loss": 3.3459555107951386,
      "tokens_seen": 2721775616
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017832323232323233,
      "loss": 1.2741,
      "theoretical_loss": 3.345949193298725,
      "tokens_seen": 2721841152
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001783030303030303,
      "loss": 1.5108,
      "theoretical_loss": 3.3459428759970113,
      "tokens_seen": 2721906688
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001782828282828283,
      "loss": 1.3452,
      "theoretical_loss": 3.3459365588899863,
      "tokens_seen": 2721972224
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017826262626262625,
      "loss": 1.251,
      "theoretical_loss": 3.34593024197764,
      "tokens_seen": 2722037760
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017824242424242424,
      "loss": 1.3142,
      "theoretical_loss": 3.3459239252599615,
      "tokens_seen": 2722103296
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017822222222222222,
      "loss": 1.3307,
      "theoretical_loss": 3.34591760873694,
      "tokens_seen": 2722168832
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001782020202020202,
      "loss": 1.3315,
      "theoretical_loss": 3.3459112924085646,
      "tokens_seen": 2722234368
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001781818181818182,
      "loss": 1.0049,
      "theoretical_loss": 3.345904976274825,
      "tokens_seen": 2722299904
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017816161616161617,
      "loss": 1.3487,
      "theoretical_loss": 3.34589866033571,
      "tokens_seen": 2722365440
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017814141414141413,
      "loss": 1.4146,
      "theoretical_loss": 3.34589234459121,
      "tokens_seen": 2722430976
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017812121212121214,
      "loss": 1.2284,
      "theoretical_loss": 3.345886029041313,
      "tokens_seen": 2722496512
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001781010101010101,
      "loss": 1.4162,
      "theoretical_loss": 3.345879713686009,
      "tokens_seen": 2722562048
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017808080808080808,
      "loss": 1.2139,
      "theoretical_loss": 3.3458733985252875,
      "tokens_seen": 2722627584
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017806060606060607,
      "loss": 1.486,
      "theoretical_loss": 3.3458670835591375,
      "tokens_seen": 2722693120
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017804040404040402,
      "loss": 1.4104,
      "theoretical_loss": 3.3458607687875483,
      "tokens_seen": 2722758656
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017802020202020203,
      "loss": 1.2329,
      "theoretical_loss": 3.3458544542105093,
      "tokens_seen": 2722824192
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000178,
      "loss": 1.2943,
      "theoretical_loss": 3.3458481398280098,
      "tokens_seen": 2722889728
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000177979797979798,
      "loss": 1.2861,
      "theoretical_loss": 3.345841825640039,
      "tokens_seen": 2722955264
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017795959595959596,
      "loss": 1.2259,
      "theoretical_loss": 3.345835511646587,
      "tokens_seen": 2723020800
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017793939393939397,
      "loss": 1.2392,
      "theoretical_loss": 3.345829197847642,
      "tokens_seen": 2723086336
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017791919191919193,
      "loss": 1.4033,
      "theoretical_loss": 3.345822884243194,
      "tokens_seen": 2723151872
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017789898989898988,
      "loss": 1.3348,
      "theoretical_loss": 3.3458165708332324,
      "tokens_seen": 2723217408
    },
    {
      "epoch": 0.65,
      "objective/train/aligned_token_count": 0.3048631548881531,
      "objective/train/avg_token_score": 0.007430032826960087,
      "objective/train/docs_used": 1532891,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8858935236930847,
      "objective/train/original_loss": 4.314034461975098,
      "objective/train/theoretical_loss": 3.345810257617746,
      "objective/train/tokens_used": 1082142176,
      "objective/train/unlikelihood_loss": 0.05527576804161072,
      "theoretical_loss": 3.345810257617746,
      "tokens_seen": 2723282944
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001778787878787879,
      "loss": 1.1395,
      "theoretical_loss": 3.345810257617746,
      "tokens_seen": 2723282944
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017785858585858585,
      "loss": 1.2608,
      "theoretical_loss": 3.3458039445967245,
      "tokens_seen": 2723348480
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017783838383838386,
      "loss": 1.3934,
      "theoretical_loss": 3.3457976317701568,
      "tokens_seen": 2723414016
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017781818181818182,
      "loss": 1.3625,
      "theoretical_loss": 3.3457913191380326,
      "tokens_seen": 2723479552
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001777979797979798,
      "loss": 1.2202,
      "theoretical_loss": 3.345785006700342,
      "tokens_seen": 2723545088
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017777777777777779,
      "loss": 1.4218,
      "theoretical_loss": 3.345778694457073,
      "tokens_seen": 2723610624
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017775757575757577,
      "loss": 1.2429,
      "theoretical_loss": 3.3457723824082155,
      "tokens_seen": 2723676160
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017773737373737375,
      "loss": 1.3827,
      "theoretical_loss": 3.3457660705537586,
      "tokens_seen": 2723741696
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001777171717171717,
      "loss": 1.2782,
      "theoretical_loss": 3.3457597588936925,
      "tokens_seen": 2723807232
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001776969696969697,
      "loss": 1.2036,
      "theoretical_loss": 3.3457534474280055,
      "tokens_seen": 2723872768
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017767676767676768,
      "loss": 1.0773,
      "theoretical_loss": 3.3457471361566875,
      "tokens_seen": 2723938304
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017765656565656566,
      "loss": 1.2471,
      "theoretical_loss": 3.3457408250797274,
      "tokens_seen": 2724003840
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017763636363636365,
      "loss": 1.4716,
      "theoretical_loss": 3.345734514197115,
      "tokens_seen": 2724069376
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017761616161616163,
      "loss": 1.3822,
      "theoretical_loss": 3.3457282035088394,
      "tokens_seen": 2724134912
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001775959595959596,
      "loss": 1.5064,
      "theoretical_loss": 3.3457218930148898,
      "tokens_seen": 2724200448
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017757575757575757,
      "loss": 1.3987,
      "theoretical_loss": 3.345715582715256,
      "tokens_seen": 2724265984
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017755555555555556,
      "loss": 1.5493,
      "theoretical_loss": 3.345709272609927,
      "tokens_seen": 2724331520
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017753535353535354,
      "loss": 1.3561,
      "theoretical_loss": 3.3457029626988923,
      "tokens_seen": 2724397056
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017751515151515152,
      "loss": 1.4031,
      "theoretical_loss": 3.345696652982141,
      "tokens_seen": 2724462592
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017749494949494948,
      "loss": 1.3398,
      "theoretical_loss": 3.3456903434596628,
      "tokens_seen": 2724528128
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001774747474747475,
      "loss": 1.4075,
      "theoretical_loss": 3.3456840341314464,
      "tokens_seen": 2724593664
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017745454545454545,
      "loss": 1.5008,
      "theoretical_loss": 3.345677724997482,
      "tokens_seen": 2724659200
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017743434343434346,
      "loss": 1.274,
      "theoretical_loss": 3.3456714160577583,
      "tokens_seen": 2724724736
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017741414141414142,
      "loss": 1.2144,
      "theoretical_loss": 3.345665107312265,
      "tokens_seen": 2724790272
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017739393939393937,
      "loss": 1.1843,
      "theoretical_loss": 3.3456587987609914,
      "tokens_seen": 2724855808
    },
    {
      "epoch": 0.65,
      "objective/train/aligned_token_count": 0.5315859913825989,
      "objective/train/avg_token_score": 0.007586340419948101,
      "objective/train/docs_used": 1534156,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5279138088226318,
      "objective/train/original_loss": 3.7258191108703613,
      "objective/train/theoretical_loss": 3.3456524904039266,
      "objective/train/tokens_used": 1083780576,
      "objective/train/unlikelihood_loss": 0.0683053508400917,
      "theoretical_loss": 3.3456524904039266,
      "tokens_seen": 2724921344
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017737373737373738,
      "loss": 1.3873,
      "theoretical_loss": 3.3456524904039266,
      "tokens_seen": 2724921344
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017735353535353534,
      "loss": 1.3559,
      "theoretical_loss": 3.3456461822410604,
      "tokens_seen": 2724986880
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017733333333333335,
      "loss": 1.2491,
      "theoretical_loss": 3.345639874272382,
      "tokens_seen": 2725052416
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001773131313131313,
      "loss": 1.4056,
      "theoretical_loss": 3.3456335664978805,
      "tokens_seen": 2725117952
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001772929292929293,
      "loss": 1.332,
      "theoretical_loss": 3.345627258917545,
      "tokens_seen": 2725183488
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017727272727272728,
      "loss": 1.3464,
      "theoretical_loss": 3.345620951531366,
      "tokens_seen": 2725249024
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017725252525252526,
      "loss": 1.1668,
      "theoretical_loss": 3.3456146443393315,
      "tokens_seen": 2725314560
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017723232323232324,
      "loss": 1.3798,
      "theoretical_loss": 3.3456083373414316,
      "tokens_seen": 2725380096
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001772121212121212,
      "loss": 1.6852,
      "theoretical_loss": 3.3456020305376555,
      "tokens_seen": 2725445632
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017719191919191919,
      "loss": 1.5405,
      "theoretical_loss": 3.3455957239279925,
      "tokens_seen": 2725511168
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017717171717171717,
      "loss": 1.4168,
      "theoretical_loss": 3.345589417512432,
      "tokens_seen": 2725576704
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017715151515151515,
      "loss": 1.3752,
      "theoretical_loss": 3.345583111290964,
      "tokens_seen": 2725642240
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017713131313131314,
      "loss": 1.3863,
      "theoretical_loss": 3.3455768052635766,
      "tokens_seen": 2725707776
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017711111111111112,
      "loss": 1.4323,
      "theoretical_loss": 3.34557049943026,
      "tokens_seen": 2725773312
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001770909090909091,
      "loss": 1.2802,
      "theoretical_loss": 3.3455641937910037,
      "tokens_seen": 2725838848
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001770707070707071,
      "loss": 1.3624,
      "theoretical_loss": 3.3455578883457964,
      "tokens_seen": 2725904384
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017705050505050505,
      "loss": 1.4386,
      "theoretical_loss": 3.3455515830946276,
      "tokens_seen": 2725969920
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017703030303030303,
      "loss": 1.4205,
      "theoretical_loss": 3.3455452780374872,
      "tokens_seen": 2726035456
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017701010101010101,
      "loss": 1.2774,
      "theoretical_loss": 3.345538973174364,
      "tokens_seen": 2726100992
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000176989898989899,
      "loss": 1.2043,
      "theoretical_loss": 3.345532668505248,
      "tokens_seen": 2726166528
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017696969696969698,
      "loss": 1.3523,
      "theoretical_loss": 3.345526364030128,
      "tokens_seen": 2726232064
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017694949494949494,
      "loss": 1.5592,
      "theoretical_loss": 3.345520059748993,
      "tokens_seen": 2726297600
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017692929292929295,
      "loss": 1.6018,
      "theoretical_loss": 3.3455137556618335,
      "tokens_seen": 2726363136
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001769090909090909,
      "loss": 1.4069,
      "theoretical_loss": 3.345507451768638,
      "tokens_seen": 2726428672
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017688888888888892,
      "loss": 1.3893,
      "theoretical_loss": 3.3455011480693964,
      "tokens_seen": 2726494208
    },
    {
      "epoch": 0.65,
      "objective/train/aligned_token_count": 0.5120356678962708,
      "objective/train/avg_token_score": 0.013210717588663101,
      "objective/train/docs_used": 1534846,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6189095973968506,
      "objective/train/original_loss": 3.688598394393921,
      "objective/train/theoretical_loss": 3.3454948445640973,
      "objective/train/tokens_used": 1085418976,
      "objective/train/unlikelihood_loss": 0.06267130374908447,
      "theoretical_loss": 3.3454948445640973,
      "tokens_seen": 2726559744
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017686868686868687,
      "loss": 1.532,
      "theoretical_loss": 3.3454948445640973,
      "tokens_seen": 2726559744
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017684848484848483,
      "loss": 1.0971,
      "theoretical_loss": 3.345488541252731,
      "tokens_seen": 2726625280
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017682828282828284,
      "loss": 1.3414,
      "theoretical_loss": 3.3454822381352862,
      "tokens_seen": 2726690816
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001768080808080808,
      "loss": 1.239,
      "theoretical_loss": 3.345475935211753,
      "tokens_seen": 2726756352
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001767878787878788,
      "loss": 1.3563,
      "theoretical_loss": 3.3454696324821196,
      "tokens_seen": 2726821888
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017676767676767677,
      "loss": 1.226,
      "theoretical_loss": 3.3454633299463765,
      "tokens_seen": 2726887424
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017674747474747475,
      "loss": 1.2845,
      "theoretical_loss": 3.345457027604513,
      "tokens_seen": 2726952960
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017672727272727274,
      "loss": 1.421,
      "theoretical_loss": 3.3454507254565176,
      "tokens_seen": 2727018496
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001767070707070707,
      "loss": 1.3516,
      "theoretical_loss": 3.3454444235023804,
      "tokens_seen": 2727084032
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001766868686868687,
      "loss": 1.2418,
      "theoretical_loss": 3.3454381217420903,
      "tokens_seen": 2727149568
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017666666666666666,
      "loss": 1.2833,
      "theoretical_loss": 3.3454318201756372,
      "tokens_seen": 2727215104
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017664646464646464,
      "loss": 1.3449,
      "theoretical_loss": 3.34542551880301,
      "tokens_seen": 2727280640
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017662626262626263,
      "loss": 1.1014,
      "theoretical_loss": 3.345419217624199,
      "tokens_seen": 2727346176
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001766060606060606,
      "loss": 1.3741,
      "theoretical_loss": 3.345412916639192,
      "tokens_seen": 2727411712
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001765858585858586,
      "loss": 1.4643,
      "theoretical_loss": 3.34540661584798,
      "tokens_seen": 2727477248
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017656565656565658,
      "loss": 1.2585,
      "theoretical_loss": 3.345400315250551,
      "tokens_seen": 2727542784
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017654545454545454,
      "loss": 1.2258,
      "theoretical_loss": 3.3453940148468955,
      "tokens_seen": 2727608320
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017652525252525252,
      "loss": 1.3744,
      "theoretical_loss": 3.3453877146370026,
      "tokens_seen": 2727673856
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001765050505050505,
      "loss": 1.4414,
      "theoretical_loss": 3.345381414620861,
      "tokens_seen": 2727739392
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001764848484848485,
      "loss": 1.3282,
      "theoretical_loss": 3.345375114798461,
      "tokens_seen": 2727804928
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017646464646464647,
      "loss": 1.2336,
      "theoretical_loss": 3.345368815169792,
      "tokens_seen": 2727870464
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017644444444444443,
      "loss": 1.3731,
      "theoretical_loss": 3.345362515734842,
      "tokens_seen": 2727936000
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017642424242424244,
      "loss": 1.4204,
      "theoretical_loss": 3.345356216493602,
      "tokens_seen": 2728001536
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001764040404040404,
      "loss": 1.1734,
      "theoretical_loss": 3.3453499174460606,
      "tokens_seen": 2728067072
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001763838383838384,
      "loss": 1.3641,
      "theoretical_loss": 3.3453436185922074,
      "tokens_seen": 2728132608
    },
    {
      "epoch": 0.65,
      "objective/train/aligned_token_count": 0.40120968222618103,
      "objective/train/avg_token_score": 0.013250082731246948,
      "objective/train/docs_used": 1535777,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3175878524780273,
      "objective/train/original_loss": 4.49712610244751,
      "objective/train/theoretical_loss": 3.3453373199320318,
      "objective/train/tokens_used": 1087057376,
      "objective/train/unlikelihood_loss": 0.06207995116710663,
      "theoretical_loss": 3.3453373199320318,
      "tokens_seen": 2728198144
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017636363636363637,
      "loss": 1.3078,
      "theoretical_loss": 3.3453373199320318,
      "tokens_seen": 2728198144
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017634343434343432,
      "loss": 1.3086,
      "theoretical_loss": 3.345331021465523,
      "tokens_seen": 2728263680
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017632323232323233,
      "loss": 1.4927,
      "theoretical_loss": 3.3453247231926704,
      "tokens_seen": 2728329216
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001763030303030303,
      "loss": 1.3655,
      "theoretical_loss": 3.345318425113464,
      "tokens_seen": 2728394752
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001762828282828283,
      "loss": 1.5495,
      "theoretical_loss": 3.345312127227892,
      "tokens_seen": 2728460288
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017626262626262626,
      "loss": 1.5247,
      "theoretical_loss": 3.3453058295359455,
      "tokens_seen": 2728525824
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017624242424242427,
      "loss": 1.2768,
      "theoretical_loss": 3.345299532037612,
      "tokens_seen": 2728591360
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017622222222222223,
      "loss": 1.275,
      "theoretical_loss": 3.3452932347328823,
      "tokens_seen": 2728656896
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001762020202020202,
      "loss": 1.4005,
      "theoretical_loss": 3.3452869376217453,
      "tokens_seen": 2728722432
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001761818181818182,
      "loss": 1.2209,
      "theoretical_loss": 3.3452806407041904,
      "tokens_seen": 2728787968
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017616161616161615,
      "loss": 1.1645,
      "theoretical_loss": 3.345274343980207,
      "tokens_seen": 2728853504
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017614141414141416,
      "loss": 1.2647,
      "theoretical_loss": 3.3452680474497845,
      "tokens_seen": 2728919040
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017612121212121212,
      "loss": 1.2516,
      "theoretical_loss": 3.345261751112912,
      "tokens_seen": 2728984576
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001761010101010101,
      "loss": 1.2482,
      "theoretical_loss": 3.3452554549695797,
      "tokens_seen": 2729050112
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001760808080808081,
      "loss": 1.2261,
      "theoretical_loss": 3.345249159019776,
      "tokens_seen": 2729115648
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017606060606060607,
      "loss": 1.2287,
      "theoretical_loss": 3.3452428632634916,
      "tokens_seen": 2729181184
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017604040404040405,
      "loss": 1.3733,
      "theoretical_loss": 3.3452365677007143,
      "tokens_seen": 2729246720
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000176020202020202,
      "loss": 1.5077,
      "theoretical_loss": 3.3452302723314347,
      "tokens_seen": 2729312256
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000176,
      "loss": 1.2225,
      "theoretical_loss": 3.345223977155642,
      "tokens_seen": 2729377792
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017597979797979798,
      "loss": 1.4041,
      "theoretical_loss": 3.3452176821733253,
      "tokens_seen": 2729443328
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017595959595959596,
      "loss": 1.2701,
      "theoretical_loss": 3.345211387384474,
      "tokens_seen": 2729508864
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017593939393939395,
      "loss": 1.4149,
      "theoretical_loss": 3.3452050927890777,
      "tokens_seen": 2729574400
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017591919191919193,
      "loss": 1.1917,
      "theoretical_loss": 3.345198798387126,
      "tokens_seen": 2729639936
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001758989898989899,
      "loss": 1.2258,
      "theoretical_loss": 3.345192504178608,
      "tokens_seen": 2729705472
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001758787878787879,
      "loss": 1.2656,
      "theoretical_loss": 3.3451862101635133,
      "tokens_seen": 2729771008
    },
    {
      "epoch": 0.65,
      "objective/train/aligned_token_count": 0.5224828720092773,
      "objective/train/avg_token_score": 0.00965793989598751,
      "objective/train/docs_used": 1536637,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6528210639953613,
      "objective/train/original_loss": 3.8776376247406006,
      "objective/train/theoretical_loss": 3.345179916341831,
      "objective/train/tokens_used": 1088695776,
      "objective/train/unlikelihood_loss": 0.061210375279188156,
      "theoretical_loss": 3.345179916341831,
      "tokens_seen": 2729836544
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017585858585858586,
      "loss": 1.5655,
      "theoretical_loss": 3.345179916341831,
      "tokens_seen": 2729836544
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017583838383838384,
      "loss": 1.4202,
      "theoretical_loss": 3.345173622713551,
      "tokens_seen": 2729902080
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017581818181818182,
      "loss": 1.2688,
      "theoretical_loss": 3.345167329278662,
      "tokens_seen": 2729967616
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017579797979797978,
      "loss": 1.3188,
      "theoretical_loss": 3.345161036037154,
      "tokens_seen": 2730033152
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001757777777777778,
      "loss": 1.0389,
      "theoretical_loss": 3.3451547429890165,
      "tokens_seen": 2730098688
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017575757575757575,
      "loss": 1.4079,
      "theoretical_loss": 3.3451484501342383,
      "tokens_seen": 2730164224
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017573737373737376,
      "loss": 1.4507,
      "theoretical_loss": 3.34514215747281,
      "tokens_seen": 2730229760
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017571717171717172,
      "loss": 1.509,
      "theoretical_loss": 3.3451358650047194,
      "tokens_seen": 2730295296
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001756969696969697,
      "loss": 1.2549,
      "theoretical_loss": 3.3451295727299573,
      "tokens_seen": 2730360832
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017567676767676768,
      "loss": 1.054,
      "theoretical_loss": 3.345123280648512,
      "tokens_seen": 2730426368
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017565656565656564,
      "loss": 1.6175,
      "theoretical_loss": 3.345116988760374,
      "tokens_seen": 2730491904
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017563636363636365,
      "loss": 1.3173,
      "theoretical_loss": 3.345110697065532,
      "tokens_seen": 2730557440
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001756161616161616,
      "loss": 1.4103,
      "theoretical_loss": 3.3451044055639754,
      "tokens_seen": 2730622976
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001755959595959596,
      "loss": 1.3102,
      "theoretical_loss": 3.345098114255694,
      "tokens_seen": 2730688512
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017557575757575758,
      "loss": 1.3661,
      "theoretical_loss": 3.3450918231406774,
      "tokens_seen": 2730754048
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017555555555555556,
      "loss": 1.39,
      "theoretical_loss": 3.3450855322189144,
      "tokens_seen": 2730819584
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017553535353535355,
      "loss": 1.4136,
      "theoretical_loss": 3.345079241490395,
      "tokens_seen": 2730885120
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017551515151515153,
      "loss": 1.0704,
      "theoretical_loss": 3.345072950955108,
      "tokens_seen": 2730950656
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017549494949494949,
      "loss": 1.2706,
      "theoretical_loss": 3.345066660613043,
      "tokens_seen": 2731016192
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017547474747474747,
      "loss": 1.3137,
      "theoretical_loss": 3.34506037046419,
      "tokens_seen": 2731081728
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017545454545454545,
      "loss": 1.5217,
      "theoretical_loss": 3.345054080508538,
      "tokens_seen": 2731147264
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017543434343434344,
      "loss": 1.3348,
      "theoretical_loss": 3.3450477907460767,
      "tokens_seen": 2731212800
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017541414141414142,
      "loss": 1.2538,
      "theoretical_loss": 3.345041501176795,
      "tokens_seen": 2731278336
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001753939393939394,
      "loss": 1.3019,
      "theoretical_loss": 3.3450352118006825,
      "tokens_seen": 2731343872
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001753737373737374,
      "loss": 1.4036,
      "theoretical_loss": 3.345028922617729,
      "tokens_seen": 2731409408
    },
    {
      "epoch": 0.65,
      "objective/train/aligned_token_count": 0.4312683343887329,
      "objective/train/avg_token_score": 0.009266785345971584,
      "objective/train/docs_used": 1538108,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4779798984527588,
      "objective/train/original_loss": 4.360762119293213,
      "objective/train/theoretical_loss": 3.3450226336279236,
      "objective/train/tokens_used": 1090334176,
      "objective/train/unlikelihood_loss": 0.05111461132764816,
      "theoretical_loss": 3.3450226336279236,
      "tokens_seen": 2731474944
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017535353535353535,
      "loss": 1.3456,
      "theoretical_loss": 3.3450226336279236,
      "tokens_seen": 2731474944
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017533333333333336,
      "loss": 1.4166,
      "theoretical_loss": 3.3450163448312558,
      "tokens_seen": 2731540480
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017531313131313131,
      "loss": 1.1914,
      "theoretical_loss": 3.345010056227715,
      "tokens_seen": 2731606016
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001752929292929293,
      "loss": 1.2865,
      "theoretical_loss": 3.345003767817291,
      "tokens_seen": 2731671552
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017527272727272728,
      "loss": 1.2223,
      "theoretical_loss": 3.3449974795999724,
      "tokens_seen": 2731737088
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017525252525252524,
      "loss": 1.5351,
      "theoretical_loss": 3.3449911915757498,
      "tokens_seen": 2731802624
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017523232323232325,
      "loss": 1.2006,
      "theoretical_loss": 3.3449849037446118,
      "tokens_seen": 2731868160
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001752121212121212,
      "loss": 1.2723,
      "theoretical_loss": 3.344978616106548,
      "tokens_seen": 2731933696
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017519191919191922,
      "loss": 1.6793,
      "theoretical_loss": 3.344972328661548,
      "tokens_seen": 2731999232
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017517171717171718,
      "loss": 1.2952,
      "theoretical_loss": 3.344966041409601,
      "tokens_seen": 2732064768
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017515151515151513,
      "loss": 1.3022,
      "theoretical_loss": 3.3449597543506964,
      "tokens_seen": 2732130304
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017513131313131314,
      "loss": 1.3454,
      "theoretical_loss": 3.344953467484824,
      "tokens_seen": 2732195840
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001751111111111111,
      "loss": 1.4,
      "theoretical_loss": 3.3449471808119733,
      "tokens_seen": 2732261376
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001750909090909091,
      "loss": 1.1784,
      "theoretical_loss": 3.344940894332133,
      "tokens_seen": 2732326912
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017507070707070707,
      "loss": 1.4556,
      "theoretical_loss": 3.3449346080452935,
      "tokens_seen": 2732392448
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017505050505050505,
      "loss": 1.4365,
      "theoretical_loss": 3.3449283219514436,
      "tokens_seen": 2732457984
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017503030303030304,
      "loss": 1.3628,
      "theoretical_loss": 3.344922036050573,
      "tokens_seen": 2732523520
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017501010101010102,
      "loss": 1.5031,
      "theoretical_loss": 3.344915750342671,
      "tokens_seen": 2732589056
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.000174989898989899,
      "loss": 1.317,
      "theoretical_loss": 3.3449094648277273,
      "tokens_seen": 2732654592
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017496969696969696,
      "loss": 1.1814,
      "theoretical_loss": 3.344903179505731,
      "tokens_seen": 2732720128
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017494949494949494,
      "loss": 1.4239,
      "theoretical_loss": 3.344896894376672,
      "tokens_seen": 2732785664
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017492929292929293,
      "loss": 1.3499,
      "theoretical_loss": 3.3448906094405393,
      "tokens_seen": 2732851200
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001749090909090909,
      "loss": 1.5416,
      "theoretical_loss": 3.3448843246973228,
      "tokens_seen": 2732916736
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001748888888888889,
      "loss": 1.3149,
      "theoretical_loss": 3.3448780401470115,
      "tokens_seen": 2732982272
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017486868686868688,
      "loss": 1.2689,
      "theoretical_loss": 3.344871755789595,
      "tokens_seen": 2733047808
    },
    {
      "epoch": 0.65,
      "objective/train/aligned_token_count": 0.2663123309612274,
      "objective/train/avg_token_score": 0.023486346006393433,
      "objective/train/docs_used": 1538743,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.7882217168807983,
      "objective/train/original_loss": 5.28279972076416,
      "objective/train/theoretical_loss": 3.344865471625063,
      "objective/train/tokens_used": 1091972576,
      "objective/train/unlikelihood_loss": 0.04963473603129387,
      "theoretical_loss": 3.344865471625063,
      "tokens_seen": 2733113344
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017484848484848484,
      "loss": 1.2079,
      "theoretical_loss": 3.344865471625063,
      "tokens_seen": 2733113344
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017482828282828285,
      "loss": 1.4817,
      "theoretical_loss": 3.3448591876534044,
      "tokens_seen": 2733178880
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001748080808080808,
      "loss": 1.4044,
      "theoretical_loss": 3.3448529038746093,
      "tokens_seen": 2733244416
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001747878787878788,
      "loss": 1.4438,
      "theoretical_loss": 3.3448466202886666,
      "tokens_seen": 2733309952
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017476767676767677,
      "loss": 1.4137,
      "theoretical_loss": 3.344840336895566,
      "tokens_seen": 2733375488
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017474747474747473,
      "loss": 1.1319,
      "theoretical_loss": 3.3448340536952976,
      "tokens_seen": 2733441024
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017472727272727274,
      "loss": 1.3778,
      "theoretical_loss": 3.3448277706878495,
      "tokens_seen": 2733506560
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001747070707070707,
      "loss": 1.3098,
      "theoretical_loss": 3.3448214878732125,
      "tokens_seen": 2733572096
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001746868686868687,
      "loss": 1.3128,
      "theoretical_loss": 3.344815205251375,
      "tokens_seen": 2733637632
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017466666666666667,
      "loss": 1.273,
      "theoretical_loss": 3.3448089228223274,
      "tokens_seen": 2733703168
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017464646464646465,
      "loss": 1.3722,
      "theoretical_loss": 3.344802640586058,
      "tokens_seen": 2733768704
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017462626262626263,
      "loss": 1.4833,
      "theoretical_loss": 3.3447963585425575,
      "tokens_seen": 2733834240
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001746060606060606,
      "loss": 1.2691,
      "theoretical_loss": 3.344790076691815,
      "tokens_seen": 2733899776
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001745858585858586,
      "loss": 1.1567,
      "theoretical_loss": 3.3447837950338193,
      "tokens_seen": 2733965312
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017456565656565656,
      "loss": 1.1282,
      "theoretical_loss": 3.3447775135685607,
      "tokens_seen": 2734030848
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017454545454545457,
      "loss": 1.176,
      "theoretical_loss": 3.344771232296028,
      "tokens_seen": 2734096384
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017452525252525253,
      "loss": 1.5378,
      "theoretical_loss": 3.3447649512162108,
      "tokens_seen": 2734161920
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001745050505050505,
      "loss": 1.4532,
      "theoretical_loss": 3.3447586703290995,
      "tokens_seen": 2734227456
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001744848484848485,
      "loss": 1.1513,
      "theoretical_loss": 3.344752389634682,
      "tokens_seen": 2734292992
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017446464646464648,
      "loss": 1.4319,
      "theoretical_loss": 3.344746109132949,
      "tokens_seen": 2734358528
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017444444444444446,
      "loss": 1.4652,
      "theoretical_loss": 3.3447398288238896,
      "tokens_seen": 2734424064
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017442424242424242,
      "loss": 1.2067,
      "theoretical_loss": 3.344733548707493,
      "tokens_seen": 2734489600
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001744040404040404,
      "loss": 1.2497,
      "theoretical_loss": 3.3447272687837493,
      "tokens_seen": 2734555136
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0001743838383838384,
      "loss": 1.4817,
      "theoretical_loss": 3.344720989052647,
      "tokens_seen": 2734620672
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.00017436363636363637,
      "loss": 1.1781,
      "theoretical_loss": 3.344714709514177,
      "tokens_seen": 2734686208
    },
    {
      "epoch": 0.65,
      "objective/train/aligned_token_count": 0.2500610947608948,
      "objective/train/avg_token_score": 0.021418025717139244,
      "objective/train/docs_used": 1539437,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.7693347334861755,
      "objective/train/original_loss": 5.000955104827881,
      "objective/train/theoretical_loss": 3.344708430168327,
      "objective/train/tokens_used": 1093610976,
      "objective/train/unlikelihood_loss": 0.05918896943330765,
      "theoretical_loss": 3.344708430168327,
      "tokens_seen": 2734751744
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017434343434343435,
      "loss": 1.1693,
      "theoretical_loss": 3.344708430168327,
      "tokens_seen": 2734751744
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017432323232323234,
      "loss": 1.3033,
      "theoretical_loss": 3.344702151015088,
      "tokens_seen": 2734817280
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001743030303030303,
      "loss": 1.3602,
      "theoretical_loss": 3.3446958720544484,
      "tokens_seen": 2734882816
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017428282828282828,
      "loss": 1.2805,
      "theoretical_loss": 3.3446895932863985,
      "tokens_seen": 2734948352
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017426262626262626,
      "loss": 1.2513,
      "theoretical_loss": 3.3446833147109274,
      "tokens_seen": 2735013888
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017424242424242425,
      "loss": 1.1231,
      "theoretical_loss": 3.3446770363280245,
      "tokens_seen": 2735079424
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017422222222222223,
      "loss": 1.5091,
      "theoretical_loss": 3.3446707581376796,
      "tokens_seen": 2735144960
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001742020202020202,
      "loss": 1.3222,
      "theoretical_loss": 3.344664480139882,
      "tokens_seen": 2735210496
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001741818181818182,
      "loss": 1.1803,
      "theoretical_loss": 3.344658202334621,
      "tokens_seen": 2735276032
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017416161616161616,
      "loss": 1.4028,
      "theoretical_loss": 3.344651924721886,
      "tokens_seen": 2735341568
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017414141414141417,
      "loss": 1.3862,
      "theoretical_loss": 3.3446456473016672,
      "tokens_seen": 2735407104
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017412121212121212,
      "loss": 1.5503,
      "theoretical_loss": 3.344639370073953,
      "tokens_seen": 2735472640
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017410101010101008,
      "loss": 1.4445,
      "theoretical_loss": 3.344633093038734,
      "tokens_seen": 2735538176
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001740808080808081,
      "loss": 1.2943,
      "theoretical_loss": 3.344626816195999,
      "tokens_seen": 2735603712
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017406060606060605,
      "loss": 1.4337,
      "theoretical_loss": 3.344620539545738,
      "tokens_seen": 2735669248
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017404040404040406,
      "loss": 1.266,
      "theoretical_loss": 3.3446142630879394,
      "tokens_seen": 2735734784
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017402020202020202,
      "loss": 1.355,
      "theoretical_loss": 3.344607986822594,
      "tokens_seen": 2735800320
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000174,
      "loss": 1.2749,
      "theoretical_loss": 3.3446017107496906,
      "tokens_seen": 2735865856
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017397979797979798,
      "loss": 1.1409,
      "theoretical_loss": 3.344595434869219,
      "tokens_seen": 2735931392
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017395959595959597,
      "loss": 1.4951,
      "theoretical_loss": 3.3445891591811683,
      "tokens_seen": 2735996928
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017393939393939395,
      "loss": 1.3205,
      "theoretical_loss": 3.3445828836855283,
      "tokens_seen": 2736062464
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001739191919191919,
      "loss": 1.3567,
      "theoretical_loss": 3.344576608382288,
      "tokens_seen": 2736128000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001738989898989899,
      "loss": 1.1669,
      "theoretical_loss": 3.3445703332714376,
      "tokens_seen": 2736193536
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017387878787878788,
      "loss": 1.2797,
      "theoretical_loss": 3.344564058352966,
      "tokens_seen": 2736259072
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017385858585858586,
      "loss": 1.138,
      "theoretical_loss": 3.3445577836268634,
      "tokens_seen": 2736324608
    },
    {
      "epoch": 0.66,
      "objective/train/aligned_token_count": 0.34335288405418396,
      "objective/train/avg_token_score": 0.018947679549455643,
      "objective/train/docs_used": 1540841,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.168357491493225,
      "objective/train/original_loss": 5.046326160430908,
      "objective/train/theoretical_loss": 3.344551509093119,
      "objective/train/tokens_used": 1095249376,
      "objective/train/unlikelihood_loss": 0.06833254545927048,
      "theoretical_loss": 3.344551509093119,
      "tokens_seen": 2736390144
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017383838383838385,
      "loss": 1.3125,
      "theoretical_loss": 3.344551509093119,
      "tokens_seen": 2736390144
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017381818181818183,
      "loss": 1.4773,
      "theoretical_loss": 3.3445452347517217,
      "tokens_seen": 2736455680
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017379797979797979,
      "loss": 1.3238,
      "theoretical_loss": 3.3445389606026614,
      "tokens_seen": 2736521216
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001737777777777778,
      "loss": 1.3949,
      "theoretical_loss": 3.3445326866459277,
      "tokens_seen": 2736586752
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017375757575757575,
      "loss": 1.3565,
      "theoretical_loss": 3.3445264128815104,
      "tokens_seen": 2736652288
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017373737373737374,
      "loss": 1.3967,
      "theoretical_loss": 3.3445201393093984,
      "tokens_seen": 2736717824
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017371717171717172,
      "loss": 1.5245,
      "theoretical_loss": 3.344513865929582,
      "tokens_seen": 2736783360
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001736969696969697,
      "loss": 1.4644,
      "theoretical_loss": 3.3445075927420493,
      "tokens_seen": 2736848896
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001736767676767677,
      "loss": 1.1938,
      "theoretical_loss": 3.344501319746791,
      "tokens_seen": 2736914432
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017365656565656565,
      "loss": 1.27,
      "theoretical_loss": 3.3444950469437966,
      "tokens_seen": 2736979968
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017363636363636366,
      "loss": 1.2762,
      "theoretical_loss": 3.3444887743330547,
      "tokens_seen": 2737045504
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017361616161616161,
      "loss": 1.3342,
      "theoretical_loss": 3.344482501914556,
      "tokens_seen": 2737111040
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017359595959595963,
      "loss": 1.1932,
      "theoretical_loss": 3.344476229688289,
      "tokens_seen": 2737176576
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017357575757575758,
      "loss": 1.477,
      "theoretical_loss": 3.3444699576542436,
      "tokens_seen": 2737242112
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017355555555555554,
      "loss": 1.3444,
      "theoretical_loss": 3.3444636858124093,
      "tokens_seen": 2737307648
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017353535353535355,
      "loss": 1.3756,
      "theoretical_loss": 3.3444574141627754,
      "tokens_seen": 2737373184
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001735151515151515,
      "loss": 1.4026,
      "theoretical_loss": 3.344451142705332,
      "tokens_seen": 2737438720
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017349494949494952,
      "loss": 1.1333,
      "theoretical_loss": 3.344444871440068,
      "tokens_seen": 2737504256
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017347474747474748,
      "loss": 1.0299,
      "theoretical_loss": 3.344438600366973,
      "tokens_seen": 2737569792
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017345454545454546,
      "loss": 1.2118,
      "theoretical_loss": 3.3444323294860365,
      "tokens_seen": 2737635328
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017343434343434344,
      "loss": 1.3028,
      "theoretical_loss": 3.3444260587972487,
      "tokens_seen": 2737700864
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001734141414141414,
      "loss": 1.1173,
      "theoretical_loss": 3.3444197883005984,
      "tokens_seen": 2737766400
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001733939393939394,
      "loss": 1.2948,
      "theoretical_loss": 3.344413517996075,
      "tokens_seen": 2737831936
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017337373737373737,
      "loss": 1.6238,
      "theoretical_loss": 3.3444072478836686,
      "tokens_seen": 2737897472
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017335353535353535,
      "loss": 1.4683,
      "theoretical_loss": 3.3444009779633683,
      "tokens_seen": 2737963008
    },
    {
      "epoch": 0.66,
      "objective/train/aligned_token_count": 0.3283235430717468,
      "objective/train/avg_token_score": 0.016833538189530373,
      "objective/train/docs_used": 1541498,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0295722484588623,
      "objective/train/original_loss": 5.159447193145752,
      "objective/train/theoretical_loss": 3.3443947082351633,
      "objective/train/tokens_used": 1096887776,
      "objective/train/unlikelihood_loss": 0.050457969307899475,
      "theoretical_loss": 3.3443947082351633,
      "tokens_seen": 2738028544
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017333333333333334,
      "loss": 1.3219,
      "theoretical_loss": 3.3443947082351633,
      "tokens_seen": 2738028544
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017331313131313132,
      "loss": 1.0798,
      "theoretical_loss": 3.344388438699044,
      "tokens_seen": 2738094080
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001732929292929293,
      "loss": 1.2847,
      "theoretical_loss": 3.3443821693549993,
      "tokens_seen": 2738159616
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001732727272727273,
      "loss": 1.3767,
      "theoretical_loss": 3.344375900203019,
      "tokens_seen": 2738225152
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017325252525252524,
      "loss": 1.4173,
      "theoretical_loss": 3.3443696312430924,
      "tokens_seen": 2738290688
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017323232323232323,
      "loss": 1.2933,
      "theoretical_loss": 3.3443633624752094,
      "tokens_seen": 2738356224
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001732121212121212,
      "loss": 1.391,
      "theoretical_loss": 3.344357093899359,
      "tokens_seen": 2738421760
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001731919191919192,
      "loss": 1.3911,
      "theoretical_loss": 3.344350825515531,
      "tokens_seen": 2738487296
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017317171717171718,
      "loss": 1.2736,
      "theoretical_loss": 3.3443445573237147,
      "tokens_seen": 2738552832
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017315151515151514,
      "loss": 1.3594,
      "theoretical_loss": 3.3443382893238995,
      "tokens_seen": 2738618368
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017313131313131315,
      "loss": 1.4173,
      "theoretical_loss": 3.3443320215160757,
      "tokens_seen": 2738683904
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001731111111111111,
      "loss": 1.432,
      "theoretical_loss": 3.3443257539002325,
      "tokens_seen": 2738749440
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017309090909090912,
      "loss": 1.3174,
      "theoretical_loss": 3.3443194864763592,
      "tokens_seen": 2738814976
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017307070707070707,
      "loss": 1.4213,
      "theoretical_loss": 3.3443132192444454,
      "tokens_seen": 2738880512
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017305050505050503,
      "loss": 1.2744,
      "theoretical_loss": 3.34430695220448,
      "tokens_seen": 2738946048
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017303030303030304,
      "loss": 1.2022,
      "theoretical_loss": 3.344300685356454,
      "tokens_seen": 2739011584
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000173010101010101,
      "loss": 1.3409,
      "theoretical_loss": 3.344294418700356,
      "tokens_seen": 2739077120
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000172989898989899,
      "loss": 1.0979,
      "theoretical_loss": 3.3442881522361754,
      "tokens_seen": 2739142656
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017296969696969697,
      "loss": 1.3542,
      "theoretical_loss": 3.344281885963902,
      "tokens_seen": 2739208192
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017294949494949495,
      "loss": 1.4126,
      "theoretical_loss": 3.344275619883525,
      "tokens_seen": 2739273728
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017292929292929293,
      "loss": 1.3941,
      "theoretical_loss": 3.3442693539950343,
      "tokens_seen": 2739339264
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017290909090909092,
      "loss": 1.6379,
      "theoretical_loss": 3.34426308829842,
      "tokens_seen": 2739404800
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001728888888888889,
      "loss": 1.4632,
      "theoretical_loss": 3.34425682279367,
      "tokens_seen": 2739470336
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017286868686868686,
      "loss": 1.4087,
      "theoretical_loss": 3.3442505574807755,
      "tokens_seen": 2739535872
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017284848484848484,
      "loss": 1.182,
      "theoretical_loss": 3.344244292359725,
      "tokens_seen": 2739601408
    },
    {
      "epoch": 0.66,
      "objective/train/aligned_token_count": 0.49144673347473145,
      "objective/train/avg_token_score": 0.003644437761977315,
      "objective/train/docs_used": 1542857,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.592158555984497,
      "objective/train/original_loss": 4.220889568328857,
      "objective/train/theoretical_loss": 3.3442380274305084,
      "objective/train/tokens_used": 1098526176,
      "objective/train/unlikelihood_loss": 0.05682485178112984,
      "theoretical_loss": 3.3442380274305084,
      "tokens_seen": 2739666944
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017282828282828283,
      "loss": 1.5415,
      "theoretical_loss": 3.3442380274305084,
      "tokens_seen": 2739666944
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001728080808080808,
      "loss": 1.3786,
      "theoretical_loss": 3.3442317626931155,
      "tokens_seen": 2739732480
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001727878787878788,
      "loss": 1.2996,
      "theoretical_loss": 3.3442254981475354,
      "tokens_seen": 2739798016
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017276767676767678,
      "loss": 1.2712,
      "theoretical_loss": 3.344219233793758,
      "tokens_seen": 2739863552
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017274747474747476,
      "loss": 1.4534,
      "theoretical_loss": 3.344212969631773,
      "tokens_seen": 2739929088
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017272727272727272,
      "loss": 1.3964,
      "theoretical_loss": 3.3442067056615685,
      "tokens_seen": 2739994624
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001727070707070707,
      "loss": 1.2673,
      "theoretical_loss": 3.3442004418831357,
      "tokens_seen": 2740060160
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001726868686868687,
      "loss": 1.2809,
      "theoretical_loss": 3.3441941782964637,
      "tokens_seen": 2740125696
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017266666666666667,
      "loss": 1.3049,
      "theoretical_loss": 3.344187914901542,
      "tokens_seen": 2740191232
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017264646464646466,
      "loss": 1.1777,
      "theoretical_loss": 3.3441816516983596,
      "tokens_seen": 2740256768
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017262626262626264,
      "loss": 1.3853,
      "theoretical_loss": 3.3441753886869066,
      "tokens_seen": 2740322304
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001726060606060606,
      "loss": 1.265,
      "theoretical_loss": 3.3441691258671726,
      "tokens_seen": 2740387840
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001725858585858586,
      "loss": 1.0854,
      "theoretical_loss": 3.3441628632391476,
      "tokens_seen": 2740453376
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017256565656565656,
      "loss": 1.3479,
      "theoretical_loss": 3.34415660080282,
      "tokens_seen": 2740518912
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017254545454545455,
      "loss": 1.1195,
      "theoretical_loss": 3.3441503385581797,
      "tokens_seen": 2740584448
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017252525252525253,
      "loss": 1.2389,
      "theoretical_loss": 3.3441440765052164,
      "tokens_seen": 2740649984
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001725050505050505,
      "loss": 1.3357,
      "theoretical_loss": 3.34413781464392,
      "tokens_seen": 2740715520
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001724848484848485,
      "loss": 1.3342,
      "theoretical_loss": 3.3441315529742797,
      "tokens_seen": 2740781056
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017246464646464646,
      "loss": 1.2671,
      "theoretical_loss": 3.344125291496285,
      "tokens_seen": 2740846592
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017244444444444447,
      "loss": 1.2214,
      "theoretical_loss": 3.3441190302099253,
      "tokens_seen": 2740912128
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017242424242424242,
      "loss": 1.4867,
      "theoretical_loss": 3.3441127691151906,
      "tokens_seen": 2740977664
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001724040404040404,
      "loss": 1.0995,
      "theoretical_loss": 3.3441065082120707,
      "tokens_seen": 2741043200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001723838383838384,
      "loss": 1.4176,
      "theoretical_loss": 3.344100247500554,
      "tokens_seen": 2741108736
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017236363636363635,
      "loss": 1.259,
      "theoretical_loss": 3.344093986980631,
      "tokens_seen": 2741174272
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017234343434343436,
      "loss": 1.215,
      "theoretical_loss": 3.3440877266522913,
      "tokens_seen": 2741239808
    },
    {
      "epoch": 0.66,
      "objective/train/aligned_token_count": 0.44532012939453125,
      "objective/train/avg_token_score": 0.046160511672496796,
      "objective/train/docs_used": 1543628,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.32597815990448,
      "objective/train/original_loss": 4.830068111419678,
      "objective/train/theoretical_loss": 3.344081466515524,
      "objective/train/tokens_used": 1100164576,
      "objective/train/unlikelihood_loss": 0.03986622020602226,
      "theoretical_loss": 3.344081466515524,
      "tokens_seen": 2741305344
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017232323232323232,
      "loss": 1.1208,
      "theoretical_loss": 3.344081466515524,
      "tokens_seen": 2741305344
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001723030303030303,
      "loss": 1.3317,
      "theoretical_loss": 3.3440752065703183,
      "tokens_seen": 2741370880
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017228282828282829,
      "loss": 1.157,
      "theoretical_loss": 3.344068946816665,
      "tokens_seen": 2741436416
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017226262626262627,
      "loss": 1.4164,
      "theoretical_loss": 3.344062687254553,
      "tokens_seen": 2741501952
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017224242424242425,
      "loss": 1.4054,
      "theoretical_loss": 3.344056427883971,
      "tokens_seen": 2741567488
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017222222222222224,
      "loss": 1.074,
      "theoretical_loss": 3.3440501687049102,
      "tokens_seen": 2741633024
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001722020202020202,
      "loss": 1.4685,
      "theoretical_loss": 3.3440439097173593,
      "tokens_seen": 2741698560
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017218181818181818,
      "loss": 1.5456,
      "theoretical_loss": 3.3440376509213072,
      "tokens_seen": 2741764096
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017216161616161616,
      "loss": 1.2226,
      "theoretical_loss": 3.344031392316745,
      "tokens_seen": 2741829632
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017214141414141415,
      "loss": 1.2962,
      "theoretical_loss": 3.344025133903661,
      "tokens_seen": 2741895168
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017212121212121213,
      "loss": 1.4087,
      "theoretical_loss": 3.3440188756820453,
      "tokens_seen": 2741960704
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001721010101010101,
      "loss": 1.4768,
      "theoretical_loss": 3.344012617651887,
      "tokens_seen": 2742026240
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001720808080808081,
      "loss": 1.3699,
      "theoretical_loss": 3.3440063598131764,
      "tokens_seen": 2742091776
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017206060606060605,
      "loss": 1.2662,
      "theoretical_loss": 3.3440001021659027,
      "tokens_seen": 2742157312
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017204040404040407,
      "loss": 1.422,
      "theoretical_loss": 3.3439938447100555,
      "tokens_seen": 2742222848
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017202020202020202,
      "loss": 1.2541,
      "theoretical_loss": 3.343987587445624,
      "tokens_seen": 2742288384
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017199999999999998,
      "loss": 1.39,
      "theoretical_loss": 3.3439813303725985,
      "tokens_seen": 2742353920
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000171979797979798,
      "loss": 1.4197,
      "theoretical_loss": 3.343975073490968,
      "tokens_seen": 2742419456
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017195959595959595,
      "loss": 1.496,
      "theoretical_loss": 3.343968816800722,
      "tokens_seen": 2742484992
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017193939393939396,
      "loss": 1.3709,
      "theoretical_loss": 3.3439625603018506,
      "tokens_seen": 2742550528
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017191919191919192,
      "loss": 1.3865,
      "theoretical_loss": 3.343956303994343,
      "tokens_seen": 2742616064
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017189898989898993,
      "loss": 1.233,
      "theoretical_loss": 3.343950047878189,
      "tokens_seen": 2742681600
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017187878787878788,
      "loss": 1.3198,
      "theoretical_loss": 3.343943791953378,
      "tokens_seen": 2742747136
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017185858585858584,
      "loss": 1.4573,
      "theoretical_loss": 3.3439375362198995,
      "tokens_seen": 2742812672
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017183838383838385,
      "loss": 1.2744,
      "theoretical_loss": 3.3439312806777433,
      "tokens_seen": 2742878208
    },
    {
      "epoch": 0.66,
      "objective/train/aligned_token_count": 0.49786168336868286,
      "objective/train/avg_token_score": 0.007107509765774012,
      "objective/train/docs_used": 1544577,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5329992771148682,
      "objective/train/original_loss": 3.6849119663238525,
      "objective/train/theoretical_loss": 3.343925025326899,
      "objective/train/tokens_used": 1101802976,
      "objective/train/unlikelihood_loss": 0.10865108668804169,
      "theoretical_loss": 3.343925025326899,
      "tokens_seen": 2742943744
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001718181818181818,
      "loss": 1.6408,
      "theoretical_loss": 3.343925025326899,
      "tokens_seen": 2742943744
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017179797979797982,
      "loss": 1.4112,
      "theoretical_loss": 3.3439187701673556,
      "tokens_seen": 2743009280
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017177777777777778,
      "loss": 1.2393,
      "theoretical_loss": 3.3439125151991034,
      "tokens_seen": 2743074816
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017175757575757576,
      "loss": 1.3805,
      "theoretical_loss": 3.3439062604221315,
      "tokens_seen": 2743140352
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017173737373737374,
      "loss": 1.1941,
      "theoretical_loss": 3.34390000583643,
      "tokens_seen": 2743205888
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017171717171717173,
      "loss": 1.167,
      "theoretical_loss": 3.343893751441988,
      "tokens_seen": 2743271424
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001716969696969697,
      "loss": 1.4421,
      "theoretical_loss": 3.3438874972387955,
      "tokens_seen": 2743336960
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017167676767676767,
      "loss": 1.3558,
      "theoretical_loss": 3.3438812432268414,
      "tokens_seen": 2743402496
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017165656565656565,
      "loss": 1.1815,
      "theoretical_loss": 3.343874989406116,
      "tokens_seen": 2743468032
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017163636363636364,
      "loss": 1.5668,
      "theoretical_loss": 3.3438687357766086,
      "tokens_seen": 2743533568
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017161616161616162,
      "loss": 1.4276,
      "theoretical_loss": 3.343862482338309,
      "tokens_seen": 2743599104
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001715959595959596,
      "loss": 1.426,
      "theoretical_loss": 3.343856229091206,
      "tokens_seen": 2743664640
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001715757575757576,
      "loss": 1.5473,
      "theoretical_loss": 3.3438499760352904,
      "tokens_seen": 2743730176
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017155555555555555,
      "loss": 1.3895,
      "theoretical_loss": 3.3438437231705507,
      "tokens_seen": 2743795712
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017153535353535356,
      "loss": 1.1787,
      "theoretical_loss": 3.343837470496977,
      "tokens_seen": 2743861248
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001715151515151515,
      "loss": 0.989,
      "theoretical_loss": 3.343831218014559,
      "tokens_seen": 2743926784
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001714949494949495,
      "loss": 1.3652,
      "theoretical_loss": 3.343824965723286,
      "tokens_seen": 2743992320
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017147474747474748,
      "loss": 1.3929,
      "theoretical_loss": 3.343818713623148,
      "tokens_seen": 2744057856
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017145454545454544,
      "loss": 1.532,
      "theoretical_loss": 3.343812461714134,
      "tokens_seen": 2744123392
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017143434343434345,
      "loss": 1.1566,
      "theoretical_loss": 3.343806209996234,
      "tokens_seen": 2744188928
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001714141414141414,
      "loss": 1.2713,
      "theoretical_loss": 3.3437999584694373,
      "tokens_seen": 2744254464
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017139393939393942,
      "loss": 1.4262,
      "theoretical_loss": 3.343793707133734,
      "tokens_seen": 2744320000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017137373737373737,
      "loss": 1.1575,
      "theoretical_loss": 3.343787455989113,
      "tokens_seen": 2744385536
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017135353535353536,
      "loss": 1.2395,
      "theoretical_loss": 3.343781205035565,
      "tokens_seen": 2744451072
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017133333333333334,
      "loss": 1.3523,
      "theoretical_loss": 3.343774954273078,
      "tokens_seen": 2744516608
    },
    {
      "epoch": 0.66,
      "objective/train/aligned_token_count": 0.3514784872531891,
      "objective/train/avg_token_score": 0.011434218846261501,
      "objective/train/docs_used": 1545179,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0267837047576904,
      "objective/train/original_loss": 4.107002258300781,
      "objective/train/theoretical_loss": 3.343768703701643,
      "objective/train/tokens_used": 1103441376,
      "objective/train/unlikelihood_loss": 0.06931591778993607,
      "theoretical_loss": 3.343768703701643,
      "tokens_seen": 2744582144
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001713131313131313,
      "loss": 1.2963,
      "theoretical_loss": 3.343768703701643,
      "tokens_seen": 2744582144
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001712929292929293,
      "loss": 1.2243,
      "theoretical_loss": 3.343762453321249,
      "tokens_seen": 2744647680
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017127272727272727,
      "loss": 1.3523,
      "theoretical_loss": 3.3437562031318855,
      "tokens_seen": 2744713216
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017125252525252525,
      "loss": 1.3495,
      "theoretical_loss": 3.3437499531335426,
      "tokens_seen": 2744778752
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017123232323232323,
      "loss": 1.3882,
      "theoretical_loss": 3.3437437033262096,
      "tokens_seen": 2744844288
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017121212121212122,
      "loss": 1.2869,
      "theoretical_loss": 3.343737453709876,
      "tokens_seen": 2744909824
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001711919191919192,
      "loss": 1.2349,
      "theoretical_loss": 3.3437312042845306,
      "tokens_seen": 2744975360
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017117171717171716,
      "loss": 1.2503,
      "theoretical_loss": 3.343724955050165,
      "tokens_seen": 2745040896
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017115151515151514,
      "loss": 1.3003,
      "theoretical_loss": 3.3437187060067672,
      "tokens_seen": 2745106432
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017113131313131313,
      "loss": 1.6251,
      "theoretical_loss": 3.3437124571543273,
      "tokens_seen": 2745171968
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001711111111111111,
      "loss": 1.1508,
      "theoretical_loss": 3.343706208492835,
      "tokens_seen": 2745237504
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001710909090909091,
      "loss": 1.2141,
      "theoretical_loss": 3.34369996002228,
      "tokens_seen": 2745303040
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017107070707070708,
      "loss": 1.3902,
      "theoretical_loss": 3.3436937117426515,
      "tokens_seen": 2745368576
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017105050505050506,
      "loss": 1.3654,
      "theoretical_loss": 3.3436874636539393,
      "tokens_seen": 2745434112
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017103030303030305,
      "loss": 1.1584,
      "theoretical_loss": 3.343681215756133,
      "tokens_seen": 2745499648
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000171010101010101,
      "loss": 1.4004,
      "theoretical_loss": 3.3436749680492226,
      "tokens_seen": 2745565184
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000170989898989899,
      "loss": 1.5464,
      "theoretical_loss": 3.3436687205331967,
      "tokens_seen": 2745630720
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017096969696969697,
      "loss": 1.4059,
      "theoretical_loss": 3.343662473208046,
      "tokens_seen": 2745696256
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017094949494949496,
      "loss": 1.1416,
      "theoretical_loss": 3.34365622607376,
      "tokens_seen": 2745761792
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017092929292929294,
      "loss": 1.3606,
      "theoretical_loss": 3.3436499791303276,
      "tokens_seen": 2745827328
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001709090909090909,
      "loss": 1.4312,
      "theoretical_loss": 3.343643732377739,
      "tokens_seen": 2745892864
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001708888888888889,
      "loss": 1.3631,
      "theoretical_loss": 3.343637485815983,
      "tokens_seen": 2745958400
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017086868686868686,
      "loss": 1.2226,
      "theoretical_loss": 3.3436312394450507,
      "tokens_seen": 2746023936
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017084848484848488,
      "loss": 1.2882,
      "theoretical_loss": 3.34362499326493,
      "tokens_seen": 2746089472
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017082828282828283,
      "loss": 1.3633,
      "theoretical_loss": 3.3436187472756123,
      "tokens_seen": 2746155008
    },
    {
      "epoch": 0.66,
      "objective/train/aligned_token_count": 0.36143696308135986,
      "objective/train/avg_token_score": 0.01408354938030243,
      "objective/train/docs_used": 1546393,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1121612787246704,
      "objective/train/original_loss": 4.8793206214904785,
      "objective/train/theoretical_loss": 3.343612501477086,
      "objective/train/tokens_used": 1105079776,
      "objective/train/unlikelihood_loss": 0.05387447774410248,
      "theoretical_loss": 3.343612501477086,
      "tokens_seen": 2746220544
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001708080808080808,
      "loss": 1.3355,
      "theoretical_loss": 3.343612501477086,
      "tokens_seen": 2746220544
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001707878787878788,
      "loss": 1.4031,
      "theoretical_loss": 3.3436062558693402,
      "tokens_seen": 2746286080
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017076767676767676,
      "loss": 1.424,
      "theoretical_loss": 3.3436000104523664,
      "tokens_seen": 2746351616
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017074747474747477,
      "loss": 1.2687,
      "theoretical_loss": 3.3435937652261525,
      "tokens_seen": 2746417152
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017072727272727273,
      "loss": 1.1261,
      "theoretical_loss": 3.343587520190689,
      "tokens_seen": 2746482688
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001707070707070707,
      "loss": 1.4714,
      "theoretical_loss": 3.343581275345965,
      "tokens_seen": 2746548224
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001706868686868687,
      "loss": 1.4258,
      "theoretical_loss": 3.343575030691971,
      "tokens_seen": 2746613760
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017066666666666668,
      "loss": 1.4297,
      "theoretical_loss": 3.343568786228696,
      "tokens_seen": 2746679296
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017064646464646466,
      "loss": 1.1011,
      "theoretical_loss": 3.3435625419561292,
      "tokens_seen": 2746744832
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017062626262626262,
      "loss": 1.5826,
      "theoretical_loss": 3.343556297874261,
      "tokens_seen": 2746810368
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001706060606060606,
      "loss": 1.5472,
      "theoretical_loss": 3.3435500539830807,
      "tokens_seen": 2746875904
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017058585858585859,
      "loss": 1.4461,
      "theoretical_loss": 3.3435438102825774,
      "tokens_seen": 2746941440
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017056565656565657,
      "loss": 1.2367,
      "theoretical_loss": 3.343537566772742,
      "tokens_seen": 2747006976
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017054545454545455,
      "loss": 1.2557,
      "theoretical_loss": 3.343531323453563,
      "tokens_seen": 2747072512
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017052525252525254,
      "loss": 1.4719,
      "theoretical_loss": 3.3435250803250307,
      "tokens_seen": 2747138048
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001705050505050505,
      "loss": 1.22,
      "theoretical_loss": 3.3435188373871343,
      "tokens_seen": 2747203584
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001704848484848485,
      "loss": 1.3294,
      "theoretical_loss": 3.3435125946398636,
      "tokens_seen": 2747269120
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017046464646464646,
      "loss": 1.3318,
      "theoretical_loss": 3.3435063520832085,
      "tokens_seen": 2747334656
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017044444444444445,
      "loss": 1.5571,
      "theoretical_loss": 3.3435001097171577,
      "tokens_seen": 2747400192
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017042424242424243,
      "loss": 1.4673,
      "theoretical_loss": 3.343493867541702,
      "tokens_seen": 2747465728
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001704040404040404,
      "loss": 1.4209,
      "theoretical_loss": 3.3434876255568304,
      "tokens_seen": 2747531264
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001703838383838384,
      "loss": 1.5188,
      "theoretical_loss": 3.3434813837625326,
      "tokens_seen": 2747596800
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017036363636363636,
      "loss": 1.5071,
      "theoretical_loss": 3.3434751421587987,
      "tokens_seen": 2747662336
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017034343434343437,
      "loss": 1.5865,
      "theoretical_loss": 3.3434689007456173,
      "tokens_seen": 2747727872
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017032323232323232,
      "loss": 1.1203,
      "theoretical_loss": 3.343462659522979,
      "tokens_seen": 2747793408
    },
    {
      "epoch": 0.66,
      "objective/train/aligned_token_count": 0.4153836667537689,
      "objective/train/avg_token_score": 0.008824159391224384,
      "objective/train/docs_used": 1547118,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2697007656097412,
      "objective/train/original_loss": 4.4204816818237305,
      "objective/train/theoretical_loss": 3.3434564184908733,
      "objective/train/tokens_used": 1106718176,
      "objective/train/unlikelihood_loss": 0.062088046222925186,
      "theoretical_loss": 3.3434564184908733,
      "tokens_seen": 2747858944
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017030303030303028,
      "loss": 1.4865,
      "theoretical_loss": 3.3434564184908733,
      "tokens_seen": 2747858944
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001702828282828283,
      "loss": 1.1233,
      "theoretical_loss": 3.3434501776492898,
      "tokens_seen": 2747924480
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017026262626262625,
      "loss": 1.5818,
      "theoretical_loss": 3.3434439369982174,
      "tokens_seen": 2747990016
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017024242424242426,
      "loss": 1.3709,
      "theoretical_loss": 3.343437696537647,
      "tokens_seen": 2748055552
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017022222222222222,
      "loss": 1.2446,
      "theoretical_loss": 3.3434314562675667,
      "tokens_seen": 2748121088
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017020202020202023,
      "loss": 1.2824,
      "theoretical_loss": 3.343425216187968,
      "tokens_seen": 2748186624
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017018181818181818,
      "loss": 1.3978,
      "theoretical_loss": 3.343418976298839,
      "tokens_seen": 2748252160
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017016161616161617,
      "loss": 1.4285,
      "theoretical_loss": 3.3434127366001696,
      "tokens_seen": 2748317696
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017014141414141415,
      "loss": 1.4503,
      "theoretical_loss": 3.34340649709195,
      "tokens_seen": 2748383232
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001701212121212121,
      "loss": 1.4549,
      "theoretical_loss": 3.34340025777417,
      "tokens_seen": 2748448768
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017010101010101012,
      "loss": 1.1759,
      "theoretical_loss": 3.3433940186468187,
      "tokens_seen": 2748514304
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017008080808080808,
      "loss": 1.3724,
      "theoretical_loss": 3.343387779709886,
      "tokens_seen": 2748579840
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017006060606060606,
      "loss": 1.4158,
      "theoretical_loss": 3.3433815409633607,
      "tokens_seen": 2748645376
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017004040404040404,
      "loss": 1.3661,
      "theoretical_loss": 3.3433753024072335,
      "tokens_seen": 2748710912
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017002020202020203,
      "loss": 1.2978,
      "theoretical_loss": 3.3433690640414944,
      "tokens_seen": 2748776448
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00017,
      "loss": 1.5369,
      "theoretical_loss": 3.343362825866132,
      "tokens_seen": 2748841984
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.000169979797979798,
      "loss": 1.3201,
      "theoretical_loss": 3.3433565878811358,
      "tokens_seen": 2748907520
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016995959595959595,
      "loss": 1.0132,
      "theoretical_loss": 3.3433503500864967,
      "tokens_seen": 2748973056
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016993939393939394,
      "loss": 1.1465,
      "theoretical_loss": 3.343344112482203,
      "tokens_seen": 2749038592
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016991919191919192,
      "loss": 1.6433,
      "theoretical_loss": 3.343337875068246,
      "tokens_seen": 2749104128
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001698989898989899,
      "loss": 1.1433,
      "theoretical_loss": 3.3433316378446136,
      "tokens_seen": 2749169664
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001698787878787879,
      "loss": 1.2422,
      "theoretical_loss": 3.3433254008112963,
      "tokens_seen": 2749235200
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016985858585858585,
      "loss": 1.3404,
      "theoretical_loss": 3.3433191639682835,
      "tokens_seen": 2749300736
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016983838383838386,
      "loss": 1.4965,
      "theoretical_loss": 3.3433129273155657,
      "tokens_seen": 2749366272
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016981818181818181,
      "loss": 1.2563,
      "theoretical_loss": 3.3433066908531313,
      "tokens_seen": 2749431808
    },
    {
      "epoch": 0.66,
      "objective/train/aligned_token_count": 0.5243157148361206,
      "objective/train/avg_token_score": 0.06135053187608719,
      "objective/train/docs_used": 1548324,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.728110432624817,
      "objective/train/original_loss": 4.531412124633789,
      "objective/train/theoretical_loss": 3.343300454580971,
      "objective/train/tokens_used": 1108356576,
      "objective/train/unlikelihood_loss": 0.048239532858133316,
      "theoretical_loss": 3.343300454580971,
      "tokens_seen": 2749497344
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016979797979797982,
      "loss": 1.4656,
      "theoretical_loss": 3.343300454580971,
      "tokens_seen": 2749497344
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016977777777777778,
      "loss": 1.2461,
      "theoretical_loss": 3.3432942184990733,
      "tokens_seen": 2749562880
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016975757575757574,
      "loss": 1.4659,
      "theoretical_loss": 3.343287982607429,
      "tokens_seen": 2749628416
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016973737373737375,
      "loss": 1.3565,
      "theoretical_loss": 3.3432817469060274,
      "tokens_seen": 2749693952
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001697171717171717,
      "loss": 1.3754,
      "theoretical_loss": 3.343275511394858,
      "tokens_seen": 2749759488
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016969696969696972,
      "loss": 1.2379,
      "theoretical_loss": 3.3432692760739102,
      "tokens_seen": 2749825024
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016967676767676767,
      "loss": 1.3754,
      "theoretical_loss": 3.3432630409431745,
      "tokens_seen": 2749890560
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016965656565656566,
      "loss": 1.275,
      "theoretical_loss": 3.34325680600264,
      "tokens_seen": 2749956096
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016963636363636364,
      "loss": 1.2824,
      "theoretical_loss": 3.3432505712522964,
      "tokens_seen": 2750021632
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016961616161616163,
      "loss": 1.1973,
      "theoretical_loss": 3.343244336692133,
      "tokens_seen": 2750087168
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001695959595959596,
      "loss": 1.3458,
      "theoretical_loss": 3.3432381023221405,
      "tokens_seen": 2750152704
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016957575757575757,
      "loss": 1.4622,
      "theoretical_loss": 3.3432318681423077,
      "tokens_seen": 2750218240
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016955555555555555,
      "loss": 1.3955,
      "theoretical_loss": 3.3432256341526245,
      "tokens_seen": 2750283776
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016953535353535353,
      "loss": 1.4689,
      "theoretical_loss": 3.343219400353081,
      "tokens_seen": 2750349312
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016951515151515152,
      "loss": 1.3693,
      "theoretical_loss": 3.3432131667436655,
      "tokens_seen": 2750414848
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001694949494949495,
      "loss": 1.1766,
      "theoretical_loss": 3.3432069333243692,
      "tokens_seen": 2750480384
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001694747474747475,
      "loss": 1.2803,
      "theoretical_loss": 3.3432007000951813,
      "tokens_seen": 2750545920
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016945454545454544,
      "loss": 1.3837,
      "theoretical_loss": 3.343194467056091,
      "tokens_seen": 2750611456
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016943434343434343,
      "loss": 1.3413,
      "theoretical_loss": 3.343188234207089,
      "tokens_seen": 2750676992
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001694141414141414,
      "loss": 1.4939,
      "theoretical_loss": 3.3431820015481635,
      "tokens_seen": 2750742528
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001693939393939394,
      "loss": 1.2853,
      "theoretical_loss": 3.3431757690793056,
      "tokens_seen": 2750808064
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016937373737373738,
      "loss": 1.2412,
      "theoretical_loss": 3.343169536800504,
      "tokens_seen": 2750873600
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016935353535353536,
      "loss": 1.4026,
      "theoretical_loss": 3.343163304711749,
      "tokens_seen": 2750939136
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00016933333333333335,
      "loss": 1.561,
      "theoretical_loss": 3.3431570728130295,
      "tokens_seen": 2751004672
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001693131313131313,
      "loss": 1.4495,
      "theoretical_loss": 3.3431508411043365,
      "tokens_seen": 2751070208
    },
    {
      "epoch": 0.66,
      "objective/train/aligned_token_count": 0.40915200114250183,
      "objective/train/avg_token_score": 0.026998503133654594,
      "objective/train/docs_used": 1548689,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3291044235229492,
      "objective/train/original_loss": 5.020063400268555,
      "objective/train/theoretical_loss": 3.3431446095856585,
      "objective/train/tokens_used": 1109994976,
      "objective/train/unlikelihood_loss": 0.053244397044181824,
      "theoretical_loss": 3.3431446095856585,
      "tokens_seen": 2751135744
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016929292929292932,
      "loss": 1.511,
      "theoretical_loss": 3.3431446095856585,
      "tokens_seen": 2751135744
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016927272727272727,
      "loss": 1.1949,
      "theoretical_loss": 3.3431383782569855,
      "tokens_seen": 2751201280
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016925252525252526,
      "loss": 1.2787,
      "theoretical_loss": 3.343132147118307,
      "tokens_seen": 2751266816
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016923232323232324,
      "loss": 1.5091,
      "theoretical_loss": 3.343125916169613,
      "tokens_seen": 2751332352
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001692121212121212,
      "loss": 1.4228,
      "theoretical_loss": 3.3431196854108935,
      "tokens_seen": 2751397888
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001691919191919192,
      "loss": 1.3033,
      "theoretical_loss": 3.3431134548421375,
      "tokens_seen": 2751463424
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016917171717171716,
      "loss": 1.3891,
      "theoretical_loss": 3.3431072244633353,
      "tokens_seen": 2751528960
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016915151515151518,
      "loss": 1.365,
      "theoretical_loss": 3.343100994274476,
      "tokens_seen": 2751594496
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016913131313131313,
      "loss": 1.3486,
      "theoretical_loss": 3.3430947642755493,
      "tokens_seen": 2751660032
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016911111111111112,
      "loss": 1.3158,
      "theoretical_loss": 3.3430885344665455,
      "tokens_seen": 2751725568
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001690909090909091,
      "loss": 1.2401,
      "theoretical_loss": 3.343082304847454,
      "tokens_seen": 2751791104
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016907070707070706,
      "loss": 1.2551,
      "theoretical_loss": 3.343076075418264,
      "tokens_seen": 2751856640
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016905050505050507,
      "loss": 1.5036,
      "theoretical_loss": 3.343069846178966,
      "tokens_seen": 2751922176
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016903030303030303,
      "loss": 1.4007,
      "theoretical_loss": 3.3430636171295487,
      "tokens_seen": 2751987712
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000169010101010101,
      "loss": 1.2948,
      "theoretical_loss": 3.3430573882700028,
      "tokens_seen": 2752053248
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000168989898989899,
      "loss": 1.4329,
      "theoretical_loss": 3.3430511596003174,
      "tokens_seen": 2752118784
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016896969696969698,
      "loss": 1.2439,
      "theoretical_loss": 3.3430449311204824,
      "tokens_seen": 2752184320
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016894949494949496,
      "loss": 1.3921,
      "theoretical_loss": 3.3430387028304875,
      "tokens_seen": 2752249856
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016892929292929295,
      "loss": 1.352,
      "theoretical_loss": 3.343032474730322,
      "tokens_seen": 2752315392
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001689090909090909,
      "loss": 1.0929,
      "theoretical_loss": 3.3430262468199765,
      "tokens_seen": 2752380928
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016888888888888889,
      "loss": 1.4767,
      "theoretical_loss": 3.34302001909944,
      "tokens_seen": 2752446464
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016886868686868687,
      "loss": 1.3879,
      "theoretical_loss": 3.343013791568702,
      "tokens_seen": 2752512000
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016884848484848485,
      "loss": 1.3426,
      "theoretical_loss": 3.3430075642277526,
      "tokens_seen": 2752577536
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016882828282828284,
      "loss": 1.0767,
      "theoretical_loss": 3.343001337076582,
      "tokens_seen": 2752643072
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001688080808080808,
      "loss": 1.2166,
      "theoretical_loss": 3.3429951101151785,
      "tokens_seen": 2752708608
    },
    {
      "epoch": 0.67,
      "objective/train/aligned_token_count": 0.538306474685669,
      "objective/train/avg_token_score": 0.020273258909583092,
      "objective/train/docs_used": 1549978,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6482270956039429,
      "objective/train/original_loss": 3.950676918029785,
      "objective/train/theoretical_loss": 3.342988883343533,
      "objective/train/tokens_used": 1111633376,
      "objective/train/unlikelihood_loss": 0.0629475936293602,
      "theoretical_loss": 3.342988883343533,
      "tokens_seen": 2752774144
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001687878787878788,
      "loss": 1.6066,
      "theoretical_loss": 3.342988883343533,
      "tokens_seen": 2752774144
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016876767676767676,
      "loss": 1.1881,
      "theoretical_loss": 3.342982656761635,
      "tokens_seen": 2752839680
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016874747474747477,
      "loss": 1.4827,
      "theoretical_loss": 3.3429764303694736,
      "tokens_seen": 2752905216
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016872727272727273,
      "loss": 1.4807,
      "theoretical_loss": 3.342970204167039,
      "tokens_seen": 2752970752
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001687070707070707,
      "loss": 1.3242,
      "theoretical_loss": 3.3429639781543212,
      "tokens_seen": 2753036288
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001686868686868687,
      "loss": 1.4293,
      "theoretical_loss": 3.342957752331309,
      "tokens_seen": 2753101824
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016866666666666666,
      "loss": 1.4668,
      "theoretical_loss": 3.342951526697993,
      "tokens_seen": 2753167360
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016864646464646467,
      "loss": 1.2958,
      "theoretical_loss": 3.342945301254362,
      "tokens_seen": 2753232896
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016862626262626262,
      "loss": 1.3303,
      "theoretical_loss": 3.342939076000407,
      "tokens_seen": 2753298432
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001686060606060606,
      "loss": 1.0824,
      "theoretical_loss": 3.3429328509361165,
      "tokens_seen": 2753363968
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001685858585858586,
      "loss": 1.4854,
      "theoretical_loss": 3.342926626061481,
      "tokens_seen": 2753429504
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016856565656565655,
      "loss": 1.2508,
      "theoretical_loss": 3.3429204013764897,
      "tokens_seen": 2753495040
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016854545454545456,
      "loss": 1.258,
      "theoretical_loss": 3.3429141768811323,
      "tokens_seen": 2753560576
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016852525252525252,
      "loss": 1.3341,
      "theoretical_loss": 3.342907952575399,
      "tokens_seen": 2753626112
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016850505050505053,
      "loss": 1.4246,
      "theoretical_loss": 3.3429017284592786,
      "tokens_seen": 2753691648
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016848484848484848,
      "loss": 1.3972,
      "theoretical_loss": 3.3428955045327617,
      "tokens_seen": 2753757184
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016846464646464647,
      "loss": 1.3978,
      "theoretical_loss": 3.3428892807958377,
      "tokens_seen": 2753822720
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016844444444444445,
      "loss": 1.3505,
      "theoretical_loss": 3.3428830572484967,
      "tokens_seen": 2753888256
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016842424242424244,
      "loss": 1.2362,
      "theoretical_loss": 3.3428768338907275,
      "tokens_seen": 2753953792
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016840404040404042,
      "loss": 1.2371,
      "theoretical_loss": 3.342870610722521,
      "tokens_seen": 2754019328
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016838383838383838,
      "loss": 1.2788,
      "theoretical_loss": 3.3428643877438655,
      "tokens_seen": 2754084864
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016836363636363636,
      "loss": 1.0688,
      "theoretical_loss": 3.3428581649547517,
      "tokens_seen": 2754150400
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016834343434343434,
      "loss": 1.447,
      "theoretical_loss": 3.3428519423551695,
      "tokens_seen": 2754215936
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016832323232323233,
      "loss": 1.4935,
      "theoretical_loss": 3.342845719945108,
      "tokens_seen": 2754281472
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001683030303030303,
      "loss": 1.4061,
      "theoretical_loss": 3.3428394977245572,
      "tokens_seen": 2754347008
    },
    {
      "epoch": 0.67,
      "objective/train/aligned_token_count": 0.44287633895874023,
      "objective/train/avg_token_score": 0.01175125315785408,
      "objective/train/docs_used": 1550558,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.456912636756897,
      "objective/train/original_loss": 4.182289123535156,
      "objective/train/theoretical_loss": 3.3428332756935064,
      "objective/train/tokens_used": 1113271776,
      "objective/train/unlikelihood_loss": 0.08193713426589966,
      "theoretical_loss": 3.3428332756935064,
      "tokens_seen": 2754412544
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001682828282828283,
      "loss": 1.3747,
      "theoretical_loss": 3.3428332756935064,
      "tokens_seen": 2754412544
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016826262626262625,
      "loss": 1.2606,
      "theoretical_loss": 3.342827053851946,
      "tokens_seen": 2754478080
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016824242424242426,
      "loss": 1.3687,
      "theoretical_loss": 3.3428208321998656,
      "tokens_seen": 2754543616
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016822222222222222,
      "loss": 1.3984,
      "theoretical_loss": 3.3428146107372543,
      "tokens_seen": 2754609152
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001682020202020202,
      "loss": 1.3299,
      "theoretical_loss": 3.3428083894641025,
      "tokens_seen": 2754674688
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001681818181818182,
      "loss": 1.6815,
      "theoretical_loss": 3.3428021683803992,
      "tokens_seen": 2754740224
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016816161616161615,
      "loss": 1.3859,
      "theoretical_loss": 3.342795947486135,
      "tokens_seen": 2754805760
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016814141414141416,
      "loss": 1.0978,
      "theoretical_loss": 3.3427897267812994,
      "tokens_seen": 2754871296
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016812121212121211,
      "loss": 1.2488,
      "theoretical_loss": 3.3427835062658815,
      "tokens_seen": 2754936832
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016810101010101013,
      "loss": 1.4195,
      "theoretical_loss": 3.3427772859398717,
      "tokens_seen": 2755002368
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016808080808080808,
      "loss": 1.5385,
      "theoretical_loss": 3.3427710658032597,
      "tokens_seen": 2755067904
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016806060606060607,
      "loss": 1.3038,
      "theoretical_loss": 3.342764845856035,
      "tokens_seen": 2755133440
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016804040404040405,
      "loss": 1.4419,
      "theoretical_loss": 3.342758626098187,
      "tokens_seen": 2755198976
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000168020202020202,
      "loss": 1.2705,
      "theoretical_loss": 3.342752406529706,
      "tokens_seen": 2755264512
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016800000000000002,
      "loss": 1.4588,
      "theoretical_loss": 3.3427461871505812,
      "tokens_seen": 2755330048
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016797979797979797,
      "loss": 1.3799,
      "theoretical_loss": 3.342739967960803,
      "tokens_seen": 2755395584
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016795959595959596,
      "loss": 1.3719,
      "theoretical_loss": 3.3427337489603604,
      "tokens_seen": 2755461120
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016793939393939394,
      "loss": 1.3835,
      "theoretical_loss": 3.3427275301492436,
      "tokens_seen": 2755526656
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016791919191919193,
      "loss": 1.4184,
      "theoretical_loss": 3.342721311527443,
      "tokens_seen": 2755592192
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001678989898989899,
      "loss": 1.2523,
      "theoretical_loss": 3.342715093094947,
      "tokens_seen": 2755657728
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016787878787878787,
      "loss": 1.1937,
      "theoretical_loss": 3.3427088748517457,
      "tokens_seen": 2755723264
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016785858585858585,
      "loss": 1.4352,
      "theoretical_loss": 3.342702656797829,
      "tokens_seen": 2755788800
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016783838383838384,
      "loss": 1.2245,
      "theoretical_loss": 3.3426964389331872,
      "tokens_seen": 2755854336
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016781818181818182,
      "loss": 1.3354,
      "theoretical_loss": 3.342690221257809,
      "tokens_seen": 2755919872
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001677979797979798,
      "loss": 1.2231,
      "theoretical_loss": 3.342684003771685,
      "tokens_seen": 2755985408
    },
    {
      "epoch": 0.67,
      "objective/train/aligned_token_count": 0.5309139490127563,
      "objective/train/avg_token_score": 0.014330440200865269,
      "objective/train/docs_used": 1551539,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6581816673278809,
      "objective/train/original_loss": 3.764711380004883,
      "objective/train/theoretical_loss": 3.3426777864748045,
      "objective/train/tokens_used": 1114910176,
      "objective/train/unlikelihood_loss": 0.06403524428606033,
      "theoretical_loss": 3.3426777864748045,
      "tokens_seen": 2756050944
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001677777777777778,
      "loss": 1.3715,
      "theoretical_loss": 3.3426777864748045,
      "tokens_seen": 2756050944
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016775757575757574,
      "loss": 1.4279,
      "theoretical_loss": 3.3426715693671576,
      "tokens_seen": 2756116480
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016773737373737376,
      "loss": 1.335,
      "theoretical_loss": 3.3426653524487335,
      "tokens_seen": 2756182016
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001677171717171717,
      "loss": 1.2933,
      "theoretical_loss": 3.342659135719522,
      "tokens_seen": 2756247552
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001676969696969697,
      "loss": 1.4468,
      "theoretical_loss": 3.342652919179513,
      "tokens_seen": 2756313088
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016767676767676768,
      "loss": 1.2089,
      "theoretical_loss": 3.342646702828697,
      "tokens_seen": 2756378624
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016765656565656566,
      "loss": 1.4213,
      "theoretical_loss": 3.342640486667063,
      "tokens_seen": 2756444160
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016763636363636365,
      "loss": 1.4431,
      "theoretical_loss": 3.3426342706946004,
      "tokens_seen": 2756509696
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001676161616161616,
      "loss": 1.4855,
      "theoretical_loss": 3.342628054911299,
      "tokens_seen": 2756575232
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016759595959595962,
      "loss": 1.2,
      "theoretical_loss": 3.3426218393171494,
      "tokens_seen": 2756640768
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016757575757575757,
      "loss": 0.9773,
      "theoretical_loss": 3.342615623912141,
      "tokens_seen": 2756706304
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016755555555555558,
      "loss": 1.3787,
      "theoretical_loss": 3.342609408696263,
      "tokens_seen": 2756771840
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016753535353535354,
      "loss": 1.3168,
      "theoretical_loss": 3.3426031936695058,
      "tokens_seen": 2756837376
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001675151515151515,
      "loss": 1.4462,
      "theoretical_loss": 3.3425969788318586,
      "tokens_seen": 2756902912
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001674949494949495,
      "loss": 1.0996,
      "theoretical_loss": 3.3425907641833117,
      "tokens_seen": 2756968448
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016747474747474747,
      "loss": 1.2348,
      "theoretical_loss": 3.3425845497238544,
      "tokens_seen": 2757033984
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016745454545454548,
      "loss": 1.3013,
      "theoretical_loss": 3.342578335453477,
      "tokens_seen": 2757099520
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016743434343434343,
      "loss": 1.2962,
      "theoretical_loss": 3.3425721213721684,
      "tokens_seen": 2757165056
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016741414141414142,
      "loss": 1.219,
      "theoretical_loss": 3.342565907479919,
      "tokens_seen": 2757230592
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001673939393939394,
      "loss": 1.3881,
      "theoretical_loss": 3.3425596937767184,
      "tokens_seen": 2757296128
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016737373737373739,
      "loss": 1.3943,
      "theoretical_loss": 3.3425534802625565,
      "tokens_seen": 2757361664
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016735353535353537,
      "loss": 1.2844,
      "theoretical_loss": 3.342547266937423,
      "tokens_seen": 2757427200
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016733333333333333,
      "loss": 1.252,
      "theoretical_loss": 3.342541053801307,
      "tokens_seen": 2757492736
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001673131313131313,
      "loss": 1.4585,
      "theoretical_loss": 3.3425348408541993,
      "tokens_seen": 2757558272
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001672929292929293,
      "loss": 1.3801,
      "theoretical_loss": 3.3425286280960895,
      "tokens_seen": 2757623808
    },
    {
      "epoch": 0.67,
      "objective/train/aligned_token_count": 0.3102394938468933,
      "objective/train/avg_token_score": 0.028522228822112083,
      "objective/train/docs_used": 1552106,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8915727138519287,
      "objective/train/original_loss": 4.591017723083496,
      "objective/train/theoretical_loss": 3.3425224155269664,
      "objective/train/tokens_used": 1116548576,
      "objective/train/unlikelihood_loss": 0.07514321804046631,
      "theoretical_loss": 3.3425224155269664,
      "tokens_seen": 2757689344
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016727272727272728,
      "loss": 1.2668,
      "theoretical_loss": 3.3425224155269664,
      "tokens_seen": 2757689344
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016725252525252526,
      "loss": 1.3787,
      "theoretical_loss": 3.3425162031468205,
      "tokens_seen": 2757754880
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016723232323232325,
      "loss": 1.5889,
      "theoretical_loss": 3.3425099909556417,
      "tokens_seen": 2757820416
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001672121212121212,
      "loss": 1.2479,
      "theoretical_loss": 3.3425037789534198,
      "tokens_seen": 2757885952
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016719191919191921,
      "loss": 1.2476,
      "theoretical_loss": 3.3424975671401436,
      "tokens_seen": 2757951488
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016717171717171717,
      "loss": 1.1898,
      "theoretical_loss": 3.342491355515804,
      "tokens_seen": 2758017024
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016715151515151515,
      "loss": 1.3531,
      "theoretical_loss": 3.34248514408039,
      "tokens_seen": 2758082560
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016713131313131314,
      "loss": 1.244,
      "theoretical_loss": 3.3424789328338917,
      "tokens_seen": 2758148096
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001671111111111111,
      "loss": 1.3704,
      "theoretical_loss": 3.342472721776299,
      "tokens_seen": 2758213632
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001670909090909091,
      "loss": 1.24,
      "theoretical_loss": 3.342466510907602,
      "tokens_seen": 2758279168
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016707070707070706,
      "loss": 1.4819,
      "theoretical_loss": 3.3424603002277893,
      "tokens_seen": 2758344704
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016705050505050507,
      "loss": 1.514,
      "theoretical_loss": 3.3424540897368518,
      "tokens_seen": 2758410240
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016703030303030303,
      "loss": 1.3238,
      "theoretical_loss": 3.3424478794347783,
      "tokens_seen": 2758475776
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000167010101010101,
      "loss": 1.508,
      "theoretical_loss": 3.3424416693215595,
      "tokens_seen": 2758541312
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000166989898989899,
      "loss": 1.3409,
      "theoretical_loss": 3.3424354593971843,
      "tokens_seen": 2758606848
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016696969696969696,
      "loss": 1.1715,
      "theoretical_loss": 3.3424292496616435,
      "tokens_seen": 2758672384
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016694949494949497,
      "loss": 1.5669,
      "theoretical_loss": 3.342423040114926,
      "tokens_seen": 2758737920
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016692929292929292,
      "loss": 1.2629,
      "theoretical_loss": 3.342416830757022,
      "tokens_seen": 2758803456
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001669090909090909,
      "loss": 1.3808,
      "theoretical_loss": 3.342410621587921,
      "tokens_seen": 2758868992
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001668888888888889,
      "loss": 1.1156,
      "theoretical_loss": 3.342404412607613,
      "tokens_seen": 2758934528
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016686868686868688,
      "loss": 1.2074,
      "theoretical_loss": 3.3423982038160878,
      "tokens_seen": 2759000064
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016684848484848486,
      "loss": 1.309,
      "theoretical_loss": 3.342391995213335,
      "tokens_seen": 2759065600
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016682828282828282,
      "loss": 1.4565,
      "theoretical_loss": 3.342385786799344,
      "tokens_seen": 2759131136
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016680808080808083,
      "loss": 1.3654,
      "theoretical_loss": 3.3423795785741057,
      "tokens_seen": 2759196672
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016678787878787878,
      "loss": 1.5522,
      "theoretical_loss": 3.342373370537609,
      "tokens_seen": 2759262208
    },
    {
      "epoch": 0.67,
      "objective/train/aligned_token_count": 0.2606915831565857,
      "objective/train/avg_token_score": 0.03105284832417965,
      "objective/train/docs_used": 1553410,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8049302697181702,
      "objective/train/original_loss": 5.530904769897461,
      "objective/train/theoretical_loss": 3.3423671626898437,
      "objective/train/tokens_used": 1118186976,
      "objective/train/unlikelihood_loss": 0.0534069761633873,
      "theoretical_loss": 3.3423671626898437,
      "tokens_seen": 2759327744
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016676767676767677,
      "loss": 1.3514,
      "theoretical_loss": 3.3423671626898437,
      "tokens_seen": 2759327744
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016674747474747475,
      "loss": 1.4207,
      "theoretical_loss": 3.3423609550308,
      "tokens_seen": 2759393280
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016672727272727274,
      "loss": 1.2665,
      "theoretical_loss": 3.342354747560467,
      "tokens_seen": 2759458816
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016670707070707072,
      "loss": 1.4728,
      "theoretical_loss": 3.3423485402788353,
      "tokens_seen": 2759524352
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001666868686868687,
      "loss": 1.5739,
      "theoretical_loss": 3.3423423331858944,
      "tokens_seen": 2759589888
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016666666666666666,
      "loss": 1.3544,
      "theoretical_loss": 3.342336126281634,
      "tokens_seen": 2759655424
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016664646464646465,
      "loss": 1.1547,
      "theoretical_loss": 3.342329919566044,
      "tokens_seen": 2759720960
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016662626262626263,
      "loss": 1.3341,
      "theoretical_loss": 3.3423237130391135,
      "tokens_seen": 2759786496
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001666060606060606,
      "loss": 1.4382,
      "theoretical_loss": 3.3423175067008333,
      "tokens_seen": 2759852032
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001665858585858586,
      "loss": 1.1588,
      "theoretical_loss": 3.3423113005511924,
      "tokens_seen": 2759917568
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016656565656565655,
      "loss": 1.109,
      "theoretical_loss": 3.3423050945901815,
      "tokens_seen": 2759983104
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016654545454545456,
      "loss": 1.4132,
      "theoretical_loss": 3.342298888817789,
      "tokens_seen": 2760048640
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016652525252525252,
      "loss": 1.3641,
      "theoretical_loss": 3.3422926832340063,
      "tokens_seen": 2760114176
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016650505050505053,
      "loss": 1.2014,
      "theoretical_loss": 3.342286477838822,
      "tokens_seen": 2760179712
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001664848484848485,
      "loss": 1.5848,
      "theoretical_loss": 3.3422802726322263,
      "tokens_seen": 2760245248
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016646464646464645,
      "loss": 1.4285,
      "theoretical_loss": 3.3422740676142086,
      "tokens_seen": 2760310784
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016644444444444446,
      "loss": 1.2089,
      "theoretical_loss": 3.3422678627847597,
      "tokens_seen": 2760376320
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016642424242424241,
      "loss": 1.5698,
      "theoretical_loss": 3.3422616581438684,
      "tokens_seen": 2760441856
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016640404040404043,
      "loss": 1.3569,
      "theoretical_loss": 3.342255453691525,
      "tokens_seen": 2760507392
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016638383838383838,
      "loss": 1.1644,
      "theoretical_loss": 3.342249249427719,
      "tokens_seen": 2760572928
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016636363636363637,
      "loss": 1.2717,
      "theoretical_loss": 3.3422430453524403,
      "tokens_seen": 2760638464
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016634343434343435,
      "loss": 1.465,
      "theoretical_loss": 3.342236841465679,
      "tokens_seen": 2760704000
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016632323232323233,
      "loss": 1.5082,
      "theoretical_loss": 3.3422306377674245,
      "tokens_seen": 2760769536
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016630303030303032,
      "loss": 1.3749,
      "theoretical_loss": 3.3422244342576666,
      "tokens_seen": 2760835072
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016628282828282828,
      "loss": 1.418,
      "theoretical_loss": 3.3422182309363953,
      "tokens_seen": 2760900608
    },
    {
      "epoch": 0.67,
      "objective/train/aligned_token_count": 0.558956503868103,
      "objective/train/avg_token_score": 0.005308112129569054,
      "objective/train/docs_used": 1554093,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7049980163574219,
      "objective/train/original_loss": 3.736757755279541,
      "objective/train/theoretical_loss": 3.3422120278036003,
      "objective/train/tokens_used": 1119825376,
      "objective/train/unlikelihood_loss": 0.06288882344961166,
      "theoretical_loss": 3.3422120278036003,
      "tokens_seen": 2760966144
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016626262626262626,
      "loss": 1.4673,
      "theoretical_loss": 3.3422120278036003,
      "tokens_seen": 2760966144
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016624242424242424,
      "loss": 1.4702,
      "theoretical_loss": 3.3422058248592714,
      "tokens_seen": 2761031680
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016622222222222223,
      "loss": 1.3191,
      "theoretical_loss": 3.3421996221033985,
      "tokens_seen": 2761097216
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001662020202020202,
      "loss": 1.149,
      "theoretical_loss": 3.3421934195359713,
      "tokens_seen": 2761162752
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001661818181818182,
      "loss": 1.3704,
      "theoretical_loss": 3.3421872171569795,
      "tokens_seen": 2761228288
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016616161616161615,
      "loss": 1.3711,
      "theoretical_loss": 3.3421810149664135,
      "tokens_seen": 2761293824
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016614141414141414,
      "loss": 1.3705,
      "theoretical_loss": 3.342174812964262,
      "tokens_seen": 2761359360
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016612121212121212,
      "loss": 1.5683,
      "theoretical_loss": 3.342168611150516,
      "tokens_seen": 2761424896
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001661010101010101,
      "loss": 1.4301,
      "theoretical_loss": 3.3421624095251645,
      "tokens_seen": 2761490432
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001660808080808081,
      "loss": 1.4185,
      "theoretical_loss": 3.3421562080881975,
      "tokens_seen": 2761555968
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016606060606060604,
      "loss": 1.3669,
      "theoretical_loss": 3.3421500068396046,
      "tokens_seen": 2761621504
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016604040404040406,
      "loss": 1.2371,
      "theoretical_loss": 3.3421438057793766,
      "tokens_seen": 2761687040
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000166020202020202,
      "loss": 1.3696,
      "theoretical_loss": 3.342137604907502,
      "tokens_seen": 2761752576
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016600000000000002,
      "loss": 1.3772,
      "theoretical_loss": 3.3421314042239714,
      "tokens_seen": 2761818112
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016597979797979798,
      "loss": 1.3671,
      "theoretical_loss": 3.3421252037287745,
      "tokens_seen": 2761883648
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016595959595959594,
      "loss": 1.1935,
      "theoretical_loss": 3.3421190034219004,
      "tokens_seen": 2761949184
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016593939393939395,
      "loss": 1.3093,
      "theoretical_loss": 3.34211280330334,
      "tokens_seen": 2762014720
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001659191919191919,
      "loss": 1.2401,
      "theoretical_loss": 3.342106603373083,
      "tokens_seen": 2762080256
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016589898989898992,
      "loss": 1.4966,
      "theoretical_loss": 3.342100403631118,
      "tokens_seen": 2762145792
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016587878787878787,
      "loss": 1.3246,
      "theoretical_loss": 3.342094204077436,
      "tokens_seen": 2762211328
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016585858585858588,
      "loss": 1.3785,
      "theoretical_loss": 3.3420880047120267,
      "tokens_seen": 2762276864
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016583838383838384,
      "loss": 1.2534,
      "theoretical_loss": 3.34208180553488,
      "tokens_seen": 2762342400
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016581818181818182,
      "loss": 1.5029,
      "theoretical_loss": 3.3420756065459845,
      "tokens_seen": 2762407936
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001657979797979798,
      "loss": 1.551,
      "theoretical_loss": 3.3420694077453317,
      "tokens_seen": 2762473472
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016577777777777777,
      "loss": 1.1553,
      "theoretical_loss": 3.34206320913291,
      "tokens_seen": 2762539008
    },
    {
      "epoch": 0.67,
      "objective/train/aligned_token_count": 0.354227751493454,
      "objective/train/avg_token_score": 0.040860917419195175,
      "objective/train/docs_used": 1555178,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.212241291999817,
      "objective/train/original_loss": 5.11087703704834,
      "objective/train/theoretical_loss": 3.3420570107087104,
      "objective/train/tokens_used": 1121463776,
      "objective/train/unlikelihood_loss": 0.04472345858812332,
      "theoretical_loss": 3.3420570107087104,
      "tokens_seen": 2762604544
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016575757575757578,
      "loss": 1.4218,
      "theoretical_loss": 3.3420570107087104,
      "tokens_seen": 2762604544
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016573737373737373,
      "loss": 1.3722,
      "theoretical_loss": 3.342050812472722,
      "tokens_seen": 2762670080
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016571717171717172,
      "loss": 1.4809,
      "theoretical_loss": 3.3420446144249345,
      "tokens_seen": 2762735616
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001656969696969697,
      "loss": 1.4511,
      "theoretical_loss": 3.3420384165653383,
      "tokens_seen": 2762801152
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016567676767676769,
      "loss": 1.4284,
      "theoretical_loss": 3.342032218893923,
      "tokens_seen": 2762866688
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016565656565656567,
      "loss": 1.4311,
      "theoretical_loss": 3.3420260214106783,
      "tokens_seen": 2762932224
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016563636363636365,
      "loss": 1.208,
      "theoretical_loss": 3.342019824115594,
      "tokens_seen": 2762997760
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001656161616161616,
      "loss": 1.44,
      "theoretical_loss": 3.3420136270086602,
      "tokens_seen": 2763063296
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001655959595959596,
      "loss": 1.3479,
      "theoretical_loss": 3.3420074300898666,
      "tokens_seen": 2763128832
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016557575757575758,
      "loss": 1.3016,
      "theoretical_loss": 3.342001233359203,
      "tokens_seen": 2763194368
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016555555555555556,
      "loss": 1.3774,
      "theoretical_loss": 3.341995036816659,
      "tokens_seen": 2763259904
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016553535353535355,
      "loss": 1.4019,
      "theoretical_loss": 3.3419888404622244,
      "tokens_seen": 2763325440
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001655151515151515,
      "loss": 1.5072,
      "theoretical_loss": 3.3419826442958893,
      "tokens_seen": 2763390976
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016549494949494951,
      "loss": 1.5358,
      "theoretical_loss": 3.3419764483176437,
      "tokens_seen": 2763456512
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016547474747474747,
      "loss": 1.1009,
      "theoretical_loss": 3.3419702525274775,
      "tokens_seen": 2763522048
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016545454545454548,
      "loss": 1.2007,
      "theoretical_loss": 3.34196405692538,
      "tokens_seen": 2763587584
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016543434343434344,
      "loss": 1.3049,
      "theoretical_loss": 3.341957861511341,
      "tokens_seen": 2763653120
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001654141414141414,
      "loss": 1.135,
      "theoretical_loss": 3.341951666285351,
      "tokens_seen": 2763718656
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001653939393939394,
      "loss": 1.3239,
      "theoretical_loss": 3.341945471247399,
      "tokens_seen": 2763784192
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016537373737373736,
      "loss": 1.4831,
      "theoretical_loss": 3.3419392763974756,
      "tokens_seen": 2763849728
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016535353535353537,
      "loss": 1.2257,
      "theoretical_loss": 3.34193308173557,
      "tokens_seen": 2763915264
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016533333333333333,
      "loss": 1.3403,
      "theoretical_loss": 3.3419268872616725,
      "tokens_seen": 2763980800
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016531313131313132,
      "loss": 1.5057,
      "theoretical_loss": 3.3419206929757728,
      "tokens_seen": 2764046336
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001652929292929293,
      "loss": 1.3853,
      "theoretical_loss": 3.341914498877861,
      "tokens_seen": 2764111872
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016527272727272726,
      "loss": 1.4786,
      "theoretical_loss": 3.341908304967926,
      "tokens_seen": 2764177408
    },
    {
      "epoch": 0.67,
      "objective/train/aligned_token_count": 0.4699413478374481,
      "objective/train/avg_token_score": 0.0038468639831990004,
      "objective/train/docs_used": 1555836,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3642975091934204,
      "objective/train/original_loss": 3.884120464324951,
      "objective/train/theoretical_loss": 3.341902111245959,
      "objective/train/tokens_used": 1123102176,
      "objective/train/unlikelihood_loss": 0.06755229830741882,
      "theoretical_loss": 3.341902111245959,
      "tokens_seen": 2764242944
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016525252525252527,
      "loss": 1.3168,
      "theoretical_loss": 3.341902111245959,
      "tokens_seen": 2764242944
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016523232323232322,
      "loss": 1.5124,
      "theoretical_loss": 3.3418959177119487,
      "tokens_seen": 2764308480
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001652121212121212,
      "loss": 1.2504,
      "theoretical_loss": 3.3418897243658856,
      "tokens_seen": 2764374016
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001651919191919192,
      "loss": 1.1492,
      "theoretical_loss": 3.341883531207759,
      "tokens_seen": 2764439552
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016517171717171718,
      "loss": 1.5587,
      "theoretical_loss": 3.341877338237559,
      "tokens_seen": 2764505088
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016515151515151516,
      "loss": 1.1261,
      "theoretical_loss": 3.341871145455276,
      "tokens_seen": 2764570624
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016513131313131314,
      "loss": 1.0852,
      "theoretical_loss": 3.341864952860899,
      "tokens_seen": 2764636160
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001651111111111111,
      "loss": 1.4629,
      "theoretical_loss": 3.3418587604544183,
      "tokens_seen": 2764701696
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016509090909090908,
      "loss": 1.3109,
      "theoretical_loss": 3.3418525682358236,
      "tokens_seen": 2764767232
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016507070707070707,
      "loss": 1.3637,
      "theoretical_loss": 3.3418463762051047,
      "tokens_seen": 2764832768
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016505050505050505,
      "loss": 1.3932,
      "theoretical_loss": 3.3418401843622516,
      "tokens_seen": 2764898304
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016503030303030304,
      "loss": 1.3859,
      "theoretical_loss": 3.3418339927072545,
      "tokens_seen": 2764963840
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016501010101010102,
      "loss": 1.3232,
      "theoretical_loss": 3.341827801240102,
      "tokens_seen": 2765029376
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.000164989898989899,
      "loss": 1.358,
      "theoretical_loss": 3.3418216099607854,
      "tokens_seen": 2765094912
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016496969696969696,
      "loss": 0.9006,
      "theoretical_loss": 3.3418154188692935,
      "tokens_seen": 2765160448
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016494949494949497,
      "loss": 1.3841,
      "theoretical_loss": 3.3418092279656166,
      "tokens_seen": 2765225984
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016492929292929293,
      "loss": 1.229,
      "theoretical_loss": 3.3418030372497447,
      "tokens_seen": 2765291520
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001649090909090909,
      "loss": 1.4826,
      "theoretical_loss": 3.3417968467216674,
      "tokens_seen": 2765357056
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001648888888888889,
      "loss": 1.438,
      "theoretical_loss": 3.3417906563813746,
      "tokens_seen": 2765422592
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016486868686868685,
      "loss": 1.4736,
      "theoretical_loss": 3.341784466228856,
      "tokens_seen": 2765488128
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016484848484848487,
      "loss": 1.349,
      "theoretical_loss": 3.341778276264102,
      "tokens_seen": 2765553664
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016482828282828282,
      "loss": 1.2965,
      "theoretical_loss": 3.341772086487102,
      "tokens_seen": 2765619200
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016480808080808083,
      "loss": 1.4409,
      "theoretical_loss": 3.341765896897846,
      "tokens_seen": 2765684736
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001647878787878788,
      "loss": 1.5203,
      "theoretical_loss": 3.3417597074963235,
      "tokens_seen": 2765750272
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016476767676767677,
      "loss": 1.4486,
      "theoretical_loss": 3.341753518282525,
      "tokens_seen": 2765815808
    },
    {
      "epoch": 0.67,
      "objective/train/aligned_token_count": 0.37066227197647095,
      "objective/train/avg_token_score": 0.010151724331080914,
      "objective/train/docs_used": 1556560,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.152061104774475,
      "objective/train/original_loss": 4.8078999519348145,
      "objective/train/theoretical_loss": 3.3417473292564397,
      "objective/train/tokens_used": 1124740576,
      "objective/train/unlikelihood_loss": 0.06019650399684906,
      "theoretical_loss": 3.3417473292564397,
      "tokens_seen": 2765881344
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016474747474747476,
      "loss": 1.1906,
      "theoretical_loss": 3.3417473292564397,
      "tokens_seen": 2765881344
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016472727272727271,
      "loss": 1.3852,
      "theoretical_loss": 3.3417411404180584,
      "tokens_seen": 2765946880
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016470707070707073,
      "loss": 1.4248,
      "theoretical_loss": 3.3417349517673696,
      "tokens_seen": 2766012416
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016468686868686868,
      "loss": 1.3611,
      "theoretical_loss": 3.3417287633043644,
      "tokens_seen": 2766077952
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016466666666666667,
      "loss": 1.1798,
      "theoretical_loss": 3.341722575029032,
      "tokens_seen": 2766143488
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016464646464646465,
      "loss": 1.5547,
      "theoretical_loss": 3.341716386941362,
      "tokens_seen": 2766209024
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016462626262626263,
      "loss": 1.6777,
      "theoretical_loss": 3.341710199041345,
      "tokens_seen": 2766274560
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016460606060606062,
      "loss": 1.4685,
      "theoretical_loss": 3.3417040113289707,
      "tokens_seen": 2766340096
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016458585858585858,
      "loss": 1.5038,
      "theoretical_loss": 3.3416978238042283,
      "tokens_seen": 2766405632
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016456565656565656,
      "loss": 1.6703,
      "theoretical_loss": 3.3416916364671088,
      "tokens_seen": 2766471168
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016454545454545454,
      "loss": 1.3101,
      "theoretical_loss": 3.341685449317601,
      "tokens_seen": 2766536704
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016452525252525253,
      "loss": 1.1985,
      "theoretical_loss": 3.3416792623556955,
      "tokens_seen": 2766602240
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001645050505050505,
      "loss": 1.4927,
      "theoretical_loss": 3.341673075581382,
      "tokens_seen": 2766667776
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001644848484848485,
      "loss": 1.4643,
      "theoretical_loss": 3.34166688899465,
      "tokens_seen": 2766733312
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016446464646464645,
      "loss": 1.5602,
      "theoretical_loss": 3.3416607025954894,
      "tokens_seen": 2766798848
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016444444444444446,
      "loss": 1.3814,
      "theoretical_loss": 3.341654516383891,
      "tokens_seen": 2766864384
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016442424242424242,
      "loss": 1.3068,
      "theoretical_loss": 3.341648330359843,
      "tokens_seen": 2766929920
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001644040404040404,
      "loss": 1.2479,
      "theoretical_loss": 3.341642144523337,
      "tokens_seen": 2766995456
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001643838383838384,
      "loss": 1.3185,
      "theoretical_loss": 3.341635958874362,
      "tokens_seen": 2767060992
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016436363636363634,
      "loss": 1.4529,
      "theoretical_loss": 3.3416297734129077,
      "tokens_seen": 2767126528
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016434343434343436,
      "loss": 1.2814,
      "theoretical_loss": 3.3416235881389644,
      "tokens_seen": 2767192064
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0001643232323232323,
      "loss": 1.2677,
      "theoretical_loss": 3.341617403052522,
      "tokens_seen": 2767257600
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016430303030303032,
      "loss": 1.3018,
      "theoretical_loss": 3.3416112181535698,
      "tokens_seen": 2767323136
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016428282828282828,
      "loss": 1.4275,
      "theoretical_loss": 3.3416050334420984,
      "tokens_seen": 2767388672
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016426262626262626,
      "loss": 1.246,
      "theoretical_loss": 3.341598848918097,
      "tokens_seen": 2767454208
    },
    {
      "epoch": 0.67,
      "objective/train/aligned_token_count": 0.46664223074913025,
      "objective/train/avg_token_score": 0.011671707965433598,
      "objective/train/docs_used": 1557767,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.49977707862854,
      "objective/train/original_loss": 4.0722761154174805,
      "objective/train/theoretical_loss": 3.3415926645815563,
      "objective/train/tokens_used": 1126378976,
      "objective/train/unlikelihood_loss": 0.10071424394845963,
      "theoretical_loss": 3.3415926645815563,
      "tokens_seen": 2767519744
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016424242424242425,
      "loss": 1.44,
      "theoretical_loss": 3.3415926645815563,
      "tokens_seen": 2767519744
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001642222222222222,
      "loss": 1.345,
      "theoretical_loss": 3.3415864804324653,
      "tokens_seen": 2767585280
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016420202020202022,
      "loss": 1.3349,
      "theoretical_loss": 3.3415802964708146,
      "tokens_seen": 2767650816
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016418181818181817,
      "loss": 1.1865,
      "theoretical_loss": 3.3415741126965934,
      "tokens_seen": 2767716352
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016416161616161618,
      "loss": 1.4461,
      "theoretical_loss": 3.341567929109792,
      "tokens_seen": 2767781888
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016414141414141414,
      "loss": 1.1328,
      "theoretical_loss": 3.3415617457104005,
      "tokens_seen": 2767847424
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016412121212121213,
      "loss": 1.3234,
      "theoretical_loss": 3.341555562498409,
      "tokens_seen": 2767912960
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001641010101010101,
      "loss": 1.2804,
      "theoretical_loss": 3.341549379473806,
      "tokens_seen": 2767978496
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001640808080808081,
      "loss": 1.4637,
      "theoretical_loss": 3.3415431966365827,
      "tokens_seen": 2768044032
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016406060606060608,
      "loss": 1.2673,
      "theoretical_loss": 3.3415370139867284,
      "tokens_seen": 2768109568
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016404040404040403,
      "loss": 1.416,
      "theoretical_loss": 3.341530831524233,
      "tokens_seen": 2768175104
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016402020202020202,
      "loss": 1.4731,
      "theoretical_loss": 3.341524649249087,
      "tokens_seen": 2768240640
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000164,
      "loss": 1.1592,
      "theoretical_loss": 3.34151846716128,
      "tokens_seen": 2768306176
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016397979797979799,
      "loss": 1.3718,
      "theoretical_loss": 3.3415122852608015,
      "tokens_seen": 2768371712
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016395959595959597,
      "loss": 1.4312,
      "theoretical_loss": 3.3415061035476414,
      "tokens_seen": 2768437248
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016393939393939395,
      "loss": 1.3397,
      "theoretical_loss": 3.34149992202179,
      "tokens_seen": 2768502784
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001639191919191919,
      "loss": 1.0975,
      "theoretical_loss": 3.3414937406832363,
      "tokens_seen": 2768568320
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016389898989898992,
      "loss": 1.5048,
      "theoretical_loss": 3.3414875595319717,
      "tokens_seen": 2768633856
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016387878787878788,
      "loss": 1.3454,
      "theoretical_loss": 3.341481378567985,
      "tokens_seen": 2768699392
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016385858585858586,
      "loss": 1.5478,
      "theoretical_loss": 3.3414751977912664,
      "tokens_seen": 2768764928
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016383838383838385,
      "loss": 1.453,
      "theoretical_loss": 3.341469017201806,
      "tokens_seen": 2768830464
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001638181818181818,
      "loss": 1.1783,
      "theoretical_loss": 3.341462836799593,
      "tokens_seen": 2768896000
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016379797979797981,
      "loss": 1.337,
      "theoretical_loss": 3.341456656584618,
      "tokens_seen": 2768961536
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016377777777777777,
      "loss": 1.3658,
      "theoretical_loss": 3.3414504765568704,
      "tokens_seen": 2769027072
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016375757575757578,
      "loss": 1.3598,
      "theoretical_loss": 3.341444296716341,
      "tokens_seen": 2769092608
    },
    {
      "epoch": 0.68,
      "objective/train/aligned_token_count": 0.3999877870082855,
      "objective/train/avg_token_score": 0.018196161836385727,
      "objective/train/docs_used": 1558514,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.216800570487976,
      "objective/train/original_loss": 4.837786674499512,
      "objective/train/theoretical_loss": 3.3414381170630185,
      "objective/train/tokens_used": 1128017376,
      "objective/train/unlikelihood_loss": 0.06602315604686737,
      "theoretical_loss": 3.3414381170630185,
      "tokens_seen": 2769158144
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016373737373737374,
      "loss": 1.2006,
      "theoretical_loss": 3.3414381170630185,
      "tokens_seen": 2769158144
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001637171717171717,
      "loss": 1.224,
      "theoretical_loss": 3.3414319375968935,
      "tokens_seen": 2769223680
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001636969696969697,
      "loss": 1.6148,
      "theoretical_loss": 3.341425758317955,
      "tokens_seen": 2769289216
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016367676767676766,
      "loss": 1.3152,
      "theoretical_loss": 3.3414195792261947,
      "tokens_seen": 2769354752
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016365656565656568,
      "loss": 1.3851,
      "theoretical_loss": 3.341413400321601,
      "tokens_seen": 2769420288
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016363636363636363,
      "loss": 1.2531,
      "theoretical_loss": 3.3414072216041646,
      "tokens_seen": 2769485824
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016361616161616162,
      "loss": 1.5093,
      "theoretical_loss": 3.3414010430738745,
      "tokens_seen": 2769551360
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001635959595959596,
      "loss": 1.5013,
      "theoretical_loss": 3.3413948647307214,
      "tokens_seen": 2769616896
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016357575757575758,
      "loss": 1.6407,
      "theoretical_loss": 3.341388686574695,
      "tokens_seen": 2769682432
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016355555555555557,
      "loss": 1.3961,
      "theoretical_loss": 3.3413825086057853,
      "tokens_seen": 2769747968
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016353535353535352,
      "loss": 1.3445,
      "theoretical_loss": 3.341376330823982,
      "tokens_seen": 2769813504
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001635151515151515,
      "loss": 1.3029,
      "theoretical_loss": 3.341370153229275,
      "tokens_seen": 2769879040
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001634949494949495,
      "loss": 1.2009,
      "theoretical_loss": 3.341363975821654,
      "tokens_seen": 2769944576
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016347474747474748,
      "loss": 1.1551,
      "theoretical_loss": 3.341357798601109,
      "tokens_seen": 2770010112
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016345454545454546,
      "loss": 1.2792,
      "theoretical_loss": 3.3413516215676307,
      "tokens_seen": 2770075648
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016343434343434344,
      "loss": 1.3368,
      "theoretical_loss": 3.3413454447212083,
      "tokens_seen": 2770141184
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001634141414141414,
      "loss": 1.6215,
      "theoretical_loss": 3.341339268061832,
      "tokens_seen": 2770206720
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001633939393939394,
      "loss": 1.1452,
      "theoretical_loss": 3.341333091589491,
      "tokens_seen": 2770272256
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016337373737373737,
      "loss": 1.3966,
      "theoretical_loss": 3.341326915304176,
      "tokens_seen": 2770337792
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016335353535353535,
      "loss": 1.3324,
      "theoretical_loss": 3.3413207392058766,
      "tokens_seen": 2770403328
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016333333333333334,
      "loss": 1.3164,
      "theoretical_loss": 3.3413145632945827,
      "tokens_seen": 2770468864
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016331313131313132,
      "loss": 1.2335,
      "theoretical_loss": 3.3413083875702845,
      "tokens_seen": 2770534400
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001632929292929293,
      "loss": 1.2232,
      "theoretical_loss": 3.3413022120329714,
      "tokens_seen": 2770599936
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016327272727272726,
      "loss": 1.1824,
      "theoretical_loss": 3.3412960366826336,
      "tokens_seen": 2770665472
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016325252525252527,
      "loss": 1.3719,
      "theoretical_loss": 3.3412898615192614,
      "tokens_seen": 2770731008
    },
    {
      "epoch": 0.68,
      "objective/train/aligned_token_count": 0.44477027654647827,
      "objective/train/avg_token_score": 0.0027839825488626957,
      "objective/train/docs_used": 1559839,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3901801109313965,
      "objective/train/original_loss": 3.8549821376800537,
      "objective/train/theoretical_loss": 3.341283686542844,
      "objective/train/tokens_used": 1129655776,
      "objective/train/unlikelihood_loss": 0.061734460294246674,
      "theoretical_loss": 3.341283686542844,
      "tokens_seen": 2770796544
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016323232323232323,
      "loss": 1.4517,
      "theoretical_loss": 3.341283686542844,
      "tokens_seen": 2770796544
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016321212121212124,
      "loss": 1.2419,
      "theoretical_loss": 3.341277511753372,
      "tokens_seen": 2770862080
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001631919191919192,
      "loss": 1.2722,
      "theoretical_loss": 3.3412713371508347,
      "tokens_seen": 2770927616
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016317171717171715,
      "loss": 1.2437,
      "theoretical_loss": 3.3412651627352226,
      "tokens_seen": 2770993152
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016315151515151517,
      "loss": 1.2659,
      "theoretical_loss": 3.341258988506525,
      "tokens_seen": 2771058688
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016313131313131312,
      "loss": 1.1481,
      "theoretical_loss": 3.341252814464732,
      "tokens_seen": 2771124224
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016311111111111113,
      "loss": 1.5667,
      "theoretical_loss": 3.341246640609834,
      "tokens_seen": 2771189760
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001630909090909091,
      "loss": 1.354,
      "theoretical_loss": 3.3412404669418203,
      "tokens_seen": 2771255296
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016307070707070707,
      "loss": 1.52,
      "theoretical_loss": 3.3412342934606816,
      "tokens_seen": 2771320832
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016305050505050506,
      "loss": 1.4182,
      "theoretical_loss": 3.3412281201664067,
      "tokens_seen": 2771386368
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016303030303030304,
      "loss": 1.3493,
      "theoretical_loss": 3.341221947058987,
      "tokens_seen": 2771451904
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016301010101010103,
      "loss": 1.5279,
      "theoretical_loss": 3.3412157741384108,
      "tokens_seen": 2771517440
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016298989898989898,
      "loss": 1.3084,
      "theoretical_loss": 3.341209601404669,
      "tokens_seen": 2771582976
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016296969696969697,
      "loss": 1.2601,
      "theoretical_loss": 3.3412034288577512,
      "tokens_seen": 2771648512
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016294949494949495,
      "loss": 1.2631,
      "theoretical_loss": 3.3411972564976478,
      "tokens_seen": 2771714048
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016292929292929294,
      "loss": 1.2656,
      "theoretical_loss": 3.341191084324348,
      "tokens_seen": 2771779584
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016290909090909092,
      "loss": 1.0345,
      "theoretical_loss": 3.3411849123378428,
      "tokens_seen": 2771845120
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001628888888888889,
      "loss": 1.3876,
      "theoretical_loss": 3.341178740538121,
      "tokens_seen": 2771910656
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016286868686868686,
      "loss": 1.0697,
      "theoretical_loss": 3.341172568925173,
      "tokens_seen": 2771976192
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016284848484848484,
      "loss": 1.1702,
      "theoretical_loss": 3.3411663974989887,
      "tokens_seen": 2772041728
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016282828282828283,
      "loss": 1.5029,
      "theoretical_loss": 3.341160226259558,
      "tokens_seen": 2772107264
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001628080808080808,
      "loss": 1.3332,
      "theoretical_loss": 3.341154055206871,
      "tokens_seen": 2772172800
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001627878787878788,
      "loss": 1.4306,
      "theoretical_loss": 3.3411478843409177,
      "tokens_seen": 2772238336
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016276767676767675,
      "loss": 1.2753,
      "theoretical_loss": 3.3411417136616874,
      "tokens_seen": 2772303872
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016274747474747476,
      "loss": 1.5548,
      "theoretical_loss": 3.341135543169171,
      "tokens_seen": 2772369408
    },
    {
      "epoch": 0.68,
      "objective/train/aligned_token_count": 0.2319159358739853,
      "objective/train/avg_token_score": 0.017299961298704147,
      "objective/train/docs_used": 1560640,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.6211682558059692,
      "objective/train/original_loss": 4.596543788909912,
      "objective/train/theoretical_loss": 3.3411293728633575,
      "objective/train/tokens_used": 1131294176,
      "objective/train/unlikelihood_loss": 0.0648721382021904,
      "theoretical_loss": 3.3411293728633575,
      "tokens_seen": 2772434944
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016272727272727272,
      "loss": 1.3135,
      "theoretical_loss": 3.3411293728633575,
      "tokens_seen": 2772434944
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016270707070707073,
      "loss": 1.4069,
      "theoretical_loss": 3.3411232027442375,
      "tokens_seen": 2772500480
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001626868686868687,
      "loss": 1.4257,
      "theoretical_loss": 3.341117032811801,
      "tokens_seen": 2772566016
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016266666666666665,
      "loss": 1.4257,
      "theoretical_loss": 3.341110863066037,
      "tokens_seen": 2772631552
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016264646464646466,
      "loss": 1.3411,
      "theoretical_loss": 3.3411046935069364,
      "tokens_seen": 2772697088
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001626262626262626,
      "loss": 1.4056,
      "theoretical_loss": 3.341098524134489,
      "tokens_seen": 2772762624
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016260606060606062,
      "loss": 1.2746,
      "theoretical_loss": 3.3410923549486844,
      "tokens_seen": 2772828160
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016258585858585858,
      "loss": 1.4389,
      "theoretical_loss": 3.3410861859495125,
      "tokens_seen": 2772893696
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016256565656565656,
      "loss": 1.2457,
      "theoretical_loss": 3.3410800171369637,
      "tokens_seen": 2772959232
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016254545454545455,
      "loss": 1.4134,
      "theoretical_loss": 3.3410738485110274,
      "tokens_seen": 2773024768
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016252525252525253,
      "loss": 1.3793,
      "theoretical_loss": 3.3410676800716943,
      "tokens_seen": 2773090304
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016250505050505052,
      "loss": 1.2465,
      "theoretical_loss": 3.3410615118189537,
      "tokens_seen": 2773155840
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016248484848484847,
      "loss": 1.2771,
      "theoretical_loss": 3.341055343752796,
      "tokens_seen": 2773221376
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016246464646464648,
      "loss": 1.4395,
      "theoretical_loss": 3.34104917587321,
      "tokens_seen": 2773286912
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016244444444444444,
      "loss": 1.481,
      "theoretical_loss": 3.341043008180187,
      "tokens_seen": 2773352448
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016242424242424243,
      "loss": 1.2821,
      "theoretical_loss": 3.3410368406737168,
      "tokens_seen": 2773417984
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001624040404040404,
      "loss": 1.4211,
      "theoretical_loss": 3.3410306733537887,
      "tokens_seen": 2773483520
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001623838383838384,
      "loss": 1.3858,
      "theoretical_loss": 3.3410245062203936,
      "tokens_seen": 2773549056
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016236363636363638,
      "loss": 1.5888,
      "theoretical_loss": 3.34101833927352,
      "tokens_seen": 2773614592
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016234343434343436,
      "loss": 1.0882,
      "theoretical_loss": 3.341012172513159,
      "tokens_seen": 2773680128
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016232323232323232,
      "loss": 1.6225,
      "theoretical_loss": 3.3410060059393003,
      "tokens_seen": 2773745664
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001623030303030303,
      "loss": 1.4296,
      "theoretical_loss": 3.340999839551934,
      "tokens_seen": 2773811200
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016228282828282829,
      "loss": 1.2377,
      "theoretical_loss": 3.3409936733510492,
      "tokens_seen": 2773876736
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016226262626262627,
      "loss": 1.3911,
      "theoretical_loss": 3.340987507336637,
      "tokens_seen": 2773942272
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016224242424242425,
      "loss": 1.2119,
      "theoretical_loss": 3.3409813415086864,
      "tokens_seen": 2774007808
    },
    {
      "epoch": 0.68,
      "objective/train/aligned_token_count": 0.39516130089759827,
      "objective/train/avg_token_score": 0.011921764351427555,
      "objective/train/docs_used": 1561903,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2056444883346558,
      "objective/train/original_loss": 4.241178035736084,
      "objective/train/theoretical_loss": 3.340975175867188,
      "objective/train/tokens_used": 1132932576,
      "objective/train/unlikelihood_loss": 0.0500645749270916,
      "theoretical_loss": 3.340975175867188,
      "tokens_seen": 2774073344
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001622222222222222,
      "loss": 1.3102,
      "theoretical_loss": 3.340975175867188,
      "tokens_seen": 2774073344
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016220202020202022,
      "loss": 1.3333,
      "theoretical_loss": 3.340969010412132,
      "tokens_seen": 2774138880
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016218181818181818,
      "loss": 1.5088,
      "theoretical_loss": 3.3409628451435074,
      "tokens_seen": 2774204416
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001621616161616162,
      "loss": 1.2883,
      "theoretical_loss": 3.340956680061305,
      "tokens_seen": 2774269952
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016214141414141415,
      "loss": 1.2354,
      "theoretical_loss": 3.340950515165514,
      "tokens_seen": 2774335488
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001621212121212121,
      "loss": 1.2396,
      "theoretical_loss": 3.340944350456125,
      "tokens_seen": 2774401024
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016210101010101011,
      "loss": 1.5133,
      "theoretical_loss": 3.340938185933128,
      "tokens_seen": 2774466560
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016208080808080807,
      "loss": 1.4944,
      "theoretical_loss": 3.3409320215965126,
      "tokens_seen": 2774532096
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016206060606060608,
      "loss": 1.457,
      "theoretical_loss": 3.340925857446269,
      "tokens_seen": 2774597632
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016204040404040404,
      "loss": 1.3883,
      "theoretical_loss": 3.3409196934823866,
      "tokens_seen": 2774663168
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016202020202020202,
      "loss": 1.0554,
      "theoretical_loss": 3.3409135297048564,
      "tokens_seen": 2774728704
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000162,
      "loss": 1.1526,
      "theoretical_loss": 3.340907366113667,
      "tokens_seen": 2774794240
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016197979797979796,
      "loss": 1.3733,
      "theoretical_loss": 3.34090120270881,
      "tokens_seen": 2774859776
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016195959595959598,
      "loss": 1.1144,
      "theoretical_loss": 3.340895039490274,
      "tokens_seen": 2774925312
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016193939393939393,
      "loss": 1.3084,
      "theoretical_loss": 3.3408888764580498,
      "tokens_seen": 2774990848
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016191919191919192,
      "loss": 1.2558,
      "theoretical_loss": 3.3408827136121264,
      "tokens_seen": 2775056384
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001618989898989899,
      "loss": 1.2412,
      "theoretical_loss": 3.340876550952495,
      "tokens_seen": 2775121920
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016187878787878788,
      "loss": 1.3495,
      "theoretical_loss": 3.340870388479145,
      "tokens_seen": 2775187456
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016185858585858587,
      "loss": 1.4499,
      "theoretical_loss": 3.3408642261920662,
      "tokens_seen": 2775252992
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016183838383838385,
      "loss": 1.2118,
      "theoretical_loss": 3.3408580640912486,
      "tokens_seen": 2775318528
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001618181818181818,
      "loss": 1.4708,
      "theoretical_loss": 3.3408519021766825,
      "tokens_seen": 2775384064
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001617979797979798,
      "loss": 1.1694,
      "theoretical_loss": 3.3408457404483576,
      "tokens_seen": 2775449600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016177777777777778,
      "loss": 1.3533,
      "theoretical_loss": 3.3408395789062637,
      "tokens_seen": 2775515136
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016175757575757576,
      "loss": 1.5047,
      "theoretical_loss": 3.3408334175503915,
      "tokens_seen": 2775580672
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016173737373737374,
      "loss": 1.3044,
      "theoretical_loss": 3.34082725638073,
      "tokens_seen": 2775646208
    },
    {
      "epoch": 0.68,
      "objective/train/aligned_token_count": 0.3003421425819397,
      "objective/train/avg_token_score": 0.01818271540105343,
      "objective/train/docs_used": 1562750,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9409250020980835,
      "objective/train/original_loss": 4.994516372680664,
      "objective/train/theoretical_loss": 3.3408210953972697,
      "objective/train/tokens_used": 1134570976,
      "objective/train/unlikelihood_loss": 0.06432691961526871,
      "theoretical_loss": 3.3408210953972697,
      "tokens_seen": 2775711744
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001617171717171717,
      "loss": 1.2092,
      "theoretical_loss": 3.3408210953972697,
      "tokens_seen": 2775711744
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001616969696969697,
      "loss": 1.4612,
      "theoretical_loss": 3.340814934600001,
      "tokens_seen": 2775777280
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016167676767676767,
      "loss": 1.2446,
      "theoretical_loss": 3.340808773988913,
      "tokens_seen": 2775842816
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016165656565656568,
      "loss": 1.3983,
      "theoretical_loss": 3.3408026135639965,
      "tokens_seen": 2775908352
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016163636363636364,
      "loss": 1.4358,
      "theoretical_loss": 3.3407964533252406,
      "tokens_seen": 2775973888
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016161616161616162,
      "loss": 1.2999,
      "theoretical_loss": 3.340790293272636,
      "tokens_seen": 2776039424
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001615959595959596,
      "loss": 1.1712,
      "theoretical_loss": 3.3407841334061725,
      "tokens_seen": 2776104960
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016157575757575756,
      "loss": 1.4784,
      "theoretical_loss": 3.34077797372584,
      "tokens_seen": 2776170496
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016155555555555557,
      "loss": 1.4005,
      "theoretical_loss": 3.3407718142316285,
      "tokens_seen": 2776236032
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016153535353535353,
      "loss": 1.2844,
      "theoretical_loss": 3.340765654923528,
      "tokens_seen": 2776301568
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016151515151515154,
      "loss": 1.1764,
      "theoretical_loss": 3.340759495801528,
      "tokens_seen": 2776367104
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001614949494949495,
      "loss": 1.3498,
      "theoretical_loss": 3.3407533368656197,
      "tokens_seen": 2776432640
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016147474747474748,
      "loss": 1.5721,
      "theoretical_loss": 3.340747178115792,
      "tokens_seen": 2776498176
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016145454545454547,
      "loss": 1.3639,
      "theoretical_loss": 3.340741019552035,
      "tokens_seen": 2776563712
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016143434343434342,
      "loss": 1.1244,
      "theoretical_loss": 3.3407348611743393,
      "tokens_seen": 2776629248
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016141414141414143,
      "loss": 1.4689,
      "theoretical_loss": 3.340728702982694,
      "tokens_seen": 2776694784
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001613939393939394,
      "loss": 1.3219,
      "theoretical_loss": 3.34072254497709,
      "tokens_seen": 2776760320
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016137373737373737,
      "loss": 1.4503,
      "theoretical_loss": 3.340716387157517,
      "tokens_seen": 2776825856
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016135353535353536,
      "loss": 1.4055,
      "theoretical_loss": 3.3407102295239643,
      "tokens_seen": 2776891392
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016133333333333334,
      "loss": 1.3199,
      "theoretical_loss": 3.340704072076423,
      "tokens_seen": 2776956928
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016131313131313133,
      "loss": 1.3237,
      "theoretical_loss": 3.340697914814882,
      "tokens_seen": 2777022464
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016129292929292928,
      "loss": 1.3311,
      "theoretical_loss": 3.3406917577393322,
      "tokens_seen": 2777088000
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016127272727272727,
      "loss": 1.4641,
      "theoretical_loss": 3.3406856008497634,
      "tokens_seen": 2777153536
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016125252525252525,
      "loss": 1.3634,
      "theoretical_loss": 3.3406794441461654,
      "tokens_seen": 2777219072
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016123232323232324,
      "loss": 1.2256,
      "theoretical_loss": 3.340673287628528,
      "tokens_seen": 2777284608
    },
    {
      "epoch": 0.68,
      "objective/train/aligned_token_count": 0.4546676576137543,
      "objective/train/avg_token_score": 0.0073591857217252254,
      "objective/train/docs_used": 1563516,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4633898735046387,
      "objective/train/original_loss": 4.481967449188232,
      "objective/train/theoretical_loss": 3.3406671312968417,
      "objective/train/tokens_used": 1136209376,
      "objective/train/unlikelihood_loss": 0.05388366058468819,
      "theoretical_loss": 3.3406671312968417,
      "tokens_seen": 2777350144
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016121212121212122,
      "loss": 1.2633,
      "theoretical_loss": 3.3406671312968417,
      "tokens_seen": 2777350144
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001611919191919192,
      "loss": 1.209,
      "theoretical_loss": 3.3406609751510956,
      "tokens_seen": 2777415680
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016117171717171716,
      "loss": 1.3474,
      "theoretical_loss": 3.340654819191281,
      "tokens_seen": 2777481216
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016115151515151517,
      "loss": 1.2282,
      "theoretical_loss": 3.3406486634173866,
      "tokens_seen": 2777546752
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016113131313131313,
      "loss": 1.4272,
      "theoretical_loss": 3.3406425078294033,
      "tokens_seen": 2777612288
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001611111111111111,
      "loss": 1.4686,
      "theoretical_loss": 3.3406363524273206,
      "tokens_seen": 2777677824
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001610909090909091,
      "loss": 1.2482,
      "theoretical_loss": 3.340630197211129,
      "tokens_seen": 2777743360
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016107070707070705,
      "loss": 1.4768,
      "theoretical_loss": 3.340624042180818,
      "tokens_seen": 2777808896
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016105050505050506,
      "loss": 1.3815,
      "theoretical_loss": 3.340617887336378,
      "tokens_seen": 2777874432
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016103030303030302,
      "loss": 1.3919,
      "theoretical_loss": 3.3406117326777984,
      "tokens_seen": 2777939968
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016101010101010103,
      "loss": 1.3125,
      "theoretical_loss": 3.34060557820507,
      "tokens_seen": 2778005504
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000160989898989899,
      "loss": 1.1679,
      "theoretical_loss": 3.3405994239181824,
      "tokens_seen": 2778071040
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016096969696969697,
      "loss": 1.5034,
      "theoretical_loss": 3.3405932698171257,
      "tokens_seen": 2778136576
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016094949494949496,
      "loss": 1.2367,
      "theoretical_loss": 3.3405871159018896,
      "tokens_seen": 2778202112
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001609292929292929,
      "loss": 1.257,
      "theoretical_loss": 3.3405809621724645,
      "tokens_seen": 2778267648
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016090909090909092,
      "loss": 1.3627,
      "theoretical_loss": 3.3405748086288396,
      "tokens_seen": 2778333184
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016088888888888888,
      "loss": 1.4809,
      "theoretical_loss": 3.340568655271006,
      "tokens_seen": 2778398720
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016086868686868687,
      "loss": 1.4314,
      "theoretical_loss": 3.340562502098954,
      "tokens_seen": 2778464256
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016084848484848485,
      "loss": 1.1755,
      "theoretical_loss": 3.3405563491126715,
      "tokens_seen": 2778529792
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016082828282828283,
      "loss": 1.4623,
      "theoretical_loss": 3.340550196312151,
      "tokens_seen": 2778595328
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016080808080808082,
      "loss": 1.5478,
      "theoretical_loss": 3.340544043697381,
      "tokens_seen": 2778660864
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001607878787878788,
      "loss": 1.3874,
      "theoretical_loss": 3.3405378912683514,
      "tokens_seen": 2778726400
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016076767676767679,
      "loss": 1.392,
      "theoretical_loss": 3.340531739025053,
      "tokens_seen": 2778791936
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016074747474747474,
      "loss": 1.3591,
      "theoretical_loss": 3.340525586967476,
      "tokens_seen": 2778857472
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016072727272727273,
      "loss": 1.2729,
      "theoretical_loss": 3.3405194350956093,
      "tokens_seen": 2778923008
    },
    {
      "epoch": 0.68,
      "objective/train/aligned_token_count": 0.45546188950538635,
      "objective/train/avg_token_score": 0.012778840959072113,
      "objective/train/docs_used": 1564998,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3331798315048218,
      "objective/train/original_loss": 4.611973762512207,
      "objective/train/theoretical_loss": 3.340513283409444,
      "objective/train/tokens_used": 1137847776,
      "objective/train/unlikelihood_loss": 0.040647611021995544,
      "theoretical_loss": 3.340513283409444,
      "tokens_seen": 2778988544
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001607070707070707,
      "loss": 1.2937,
      "theoretical_loss": 3.340513283409444,
      "tokens_seen": 2778988544
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001606868686868687,
      "loss": 1.3272,
      "theoretical_loss": 3.340507131908969,
      "tokens_seen": 2779054080
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016066666666666668,
      "loss": 1.365,
      "theoretical_loss": 3.3405009805941757,
      "tokens_seen": 2779119616
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016064646464646466,
      "loss": 1.1723,
      "theoretical_loss": 3.340494829465053,
      "tokens_seen": 2779185152
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016062626262626262,
      "loss": 1.3089,
      "theoretical_loss": 3.3404886785215915,
      "tokens_seen": 2779250688
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016060606060606063,
      "loss": 1.4143,
      "theoretical_loss": 3.340482527763781,
      "tokens_seen": 2779316224
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001605858585858586,
      "loss": 1.5666,
      "theoretical_loss": 3.3404763771916115,
      "tokens_seen": 2779381760
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016056565656565657,
      "loss": 1.4433,
      "theoretical_loss": 3.340470226805073,
      "tokens_seen": 2779447296
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016054545454545455,
      "loss": 1.3928,
      "theoretical_loss": 3.340464076604156,
      "tokens_seen": 2779512832
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001605252525252525,
      "loss": 1.3188,
      "theoretical_loss": 3.3404579265888494,
      "tokens_seen": 2779578368
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016050505050505052,
      "loss": 1.3907,
      "theoretical_loss": 3.3404517767591444,
      "tokens_seen": 2779643904
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016048484848484848,
      "loss": 1.3183,
      "theoretical_loss": 3.3404456271150305,
      "tokens_seen": 2779709440
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001604646464646465,
      "loss": 1.4024,
      "theoretical_loss": 3.3404394776564974,
      "tokens_seen": 2779774976
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016044444444444445,
      "loss": 1.4229,
      "theoretical_loss": 3.340433328383536,
      "tokens_seen": 2779840512
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001604242424242424,
      "loss": 1.338,
      "theoretical_loss": 3.3404271792961358,
      "tokens_seen": 2779906048
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016040404040404042,
      "loss": 1.2809,
      "theoretical_loss": 3.3404210303942863,
      "tokens_seen": 2779971584
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016038383838383837,
      "loss": 1.4037,
      "theoretical_loss": 3.3404148816779786,
      "tokens_seen": 2780037120
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016036363636363638,
      "loss": 1.2014,
      "theoretical_loss": 3.3404087331472025,
      "tokens_seen": 2780102656
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016034343434343434,
      "loss": 1.4521,
      "theoretical_loss": 3.340402584801947,
      "tokens_seen": 2780168192
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016032323232323232,
      "loss": 1.6276,
      "theoretical_loss": 3.3403964366422034,
      "tokens_seen": 2780233728
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001603030303030303,
      "loss": 1.203,
      "theoretical_loss": 3.340390288667961,
      "tokens_seen": 2780299264
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001602828282828283,
      "loss": 1.548,
      "theoretical_loss": 3.34038414087921,
      "tokens_seen": 2780364800
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016026262626262628,
      "loss": 1.5293,
      "theoretical_loss": 3.3403779932759403,
      "tokens_seen": 2780430336
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016024242424242423,
      "loss": 1.4763,
      "theoretical_loss": 3.3403718458581424,
      "tokens_seen": 2780495872
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016022222222222222,
      "loss": 1.4755,
      "theoretical_loss": 3.340365698625806,
      "tokens_seen": 2780561408
    },
    {
      "epoch": 0.68,
      "objective/train/aligned_token_count": 0.33125612139701843,
      "objective/train/avg_token_score": 0.004629293456673622,
      "objective/train/docs_used": 1565682,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1261192560195923,
      "objective/train/original_loss": 4.774572849273682,
      "objective/train/theoretical_loss": 3.340359551578921,
      "objective/train/tokens_used": 1139486176,
      "objective/train/unlikelihood_loss": 0.052373938262462616,
      "theoretical_loss": 3.340359551578921,
      "tokens_seen": 2780626944
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001602020202020202,
      "loss": 1.2069,
      "theoretical_loss": 3.340359551578921,
      "tokens_seen": 2780626944
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016018181818181818,
      "loss": 1.2133,
      "theoretical_loss": 3.3403534047174777,
      "tokens_seen": 2780692480
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016016161616161617,
      "loss": 1.3905,
      "theoretical_loss": 3.340347258041466,
      "tokens_seen": 2780758016
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016014141414141415,
      "loss": 1.308,
      "theoretical_loss": 3.3403411115508757,
      "tokens_seen": 2780823552
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001601212121212121,
      "loss": 1.3319,
      "theoretical_loss": 3.3403349652456975,
      "tokens_seen": 2780889088
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016010101010101012,
      "loss": 1.3992,
      "theoretical_loss": 3.340328819125921,
      "tokens_seen": 2780954624
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016008080808080808,
      "loss": 1.5345,
      "theoretical_loss": 3.340322673191536,
      "tokens_seen": 2781020160
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016006060606060606,
      "loss": 1.5074,
      "theoretical_loss": 3.340316527442533,
      "tokens_seen": 2781085696
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016004040404040405,
      "loss": 1.3953,
      "theoretical_loss": 3.340310381878902,
      "tokens_seen": 2781151232
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.000160020202020202,
      "loss": 1.2735,
      "theoretical_loss": 3.3403042365006326,
      "tokens_seen": 2781216768
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00016,
      "loss": 1.4892,
      "theoretical_loss": 3.3402980913077154,
      "tokens_seen": 2781282304
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015997979797979797,
      "loss": 1.2119,
      "theoretical_loss": 3.34029194630014,
      "tokens_seen": 2781347840
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015995959595959598,
      "loss": 1.4439,
      "theoretical_loss": 3.3402858014778967,
      "tokens_seen": 2781413376
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015993939393939394,
      "loss": 1.3682,
      "theoretical_loss": 3.3402796568409756,
      "tokens_seen": 2781478912
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015991919191919192,
      "loss": 1.4851,
      "theoretical_loss": 3.3402735123893663,
      "tokens_seen": 2781544448
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001598989898989899,
      "loss": 1.2678,
      "theoretical_loss": 3.3402673681230595,
      "tokens_seen": 2781609984
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015987878787878786,
      "loss": 1.4432,
      "theoretical_loss": 3.3402612240420444,
      "tokens_seen": 2781675520
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015985858585858587,
      "loss": 1.4332,
      "theoretical_loss": 3.340255080146312,
      "tokens_seen": 2781741056
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015983838383838383,
      "loss": 1.0317,
      "theoretical_loss": 3.340248936435852,
      "tokens_seen": 2781806592
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015981818181818184,
      "loss": 1.218,
      "theoretical_loss": 3.340242792910654,
      "tokens_seen": 2781872128
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001597979797979798,
      "loss": 1.3487,
      "theoretical_loss": 3.3402366495707083,
      "tokens_seen": 2781937664
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015977777777777778,
      "loss": 1.3769,
      "theoretical_loss": 3.340230506416005,
      "tokens_seen": 2782003200
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015975757575757577,
      "loss": 1.634,
      "theoretical_loss": 3.3402243634465347,
      "tokens_seen": 2782068736
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015973737373737372,
      "loss": 1.2219,
      "theoretical_loss": 3.340218220662287,
      "tokens_seen": 2782134272
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015971717171717173,
      "loss": 1.1976,
      "theoretical_loss": 3.340212078063251,
      "tokens_seen": 2782199808
    },
    {
      "epoch": 0.68,
      "objective/train/aligned_token_count": 0.48875856399536133,
      "objective/train/avg_token_score": 0.005170241929590702,
      "objective/train/docs_used": 1566838,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5390539169311523,
      "objective/train/original_loss": 3.9311375617980957,
      "objective/train/theoretical_loss": 3.3402059356494185,
      "objective/train/tokens_used": 1141124576,
      "objective/train/unlikelihood_loss": 0.0758662223815918,
      "theoretical_loss": 3.3402059356494185,
      "tokens_seen": 2782265344
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001596969696969697,
      "loss": 1.4809,
      "theoretical_loss": 3.3402059356494185,
      "tokens_seen": 2782265344
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015967676767676768,
      "loss": 1.3834,
      "theoretical_loss": 3.3401997934207786,
      "tokens_seen": 2782330880
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015965656565656566,
      "loss": 1.2438,
      "theoretical_loss": 3.340193651377321,
      "tokens_seen": 2782396416
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015963636363636364,
      "loss": 1.6484,
      "theoretical_loss": 3.340187509519037,
      "tokens_seen": 2782461952
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015961616161616163,
      "loss": 1.3949,
      "theoretical_loss": 3.3401813678459154,
      "tokens_seen": 2782527488
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001595959595959596,
      "loss": 1.1442,
      "theoretical_loss": 3.3401752263579465,
      "tokens_seen": 2782593024
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015957575757575757,
      "loss": 1.2392,
      "theoretical_loss": 3.3401690850551207,
      "tokens_seen": 2782658560
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015955555555555555,
      "loss": 1.4389,
      "theoretical_loss": 3.340162943937428,
      "tokens_seen": 2782724096
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015953535353535354,
      "loss": 1.3763,
      "theoretical_loss": 3.3401568030048585,
      "tokens_seen": 2782789632
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015951515151515152,
      "loss": 1.0824,
      "theoretical_loss": 3.340150662257402,
      "tokens_seen": 2782855168
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001594949494949495,
      "loss": 1.2597,
      "theoretical_loss": 3.3401445216950494,
      "tokens_seen": 2782920704
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015947474747474746,
      "loss": 1.6443,
      "theoretical_loss": 3.3401383813177894,
      "tokens_seen": 2782986240
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015945454545454547,
      "loss": 1.3269,
      "theoretical_loss": 3.340132241125613,
      "tokens_seen": 2783051776
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015943434343434343,
      "loss": 1.6301,
      "theoretical_loss": 3.3401261011185097,
      "tokens_seen": 2783117312
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015941414141414144,
      "loss": 1.5396,
      "theoretical_loss": 3.3401199612964705,
      "tokens_seen": 2783182848
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001593939393939394,
      "loss": 1.2769,
      "theoretical_loss": 3.340113821659484,
      "tokens_seen": 2783248384
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015937373737373735,
      "loss": 1.3833,
      "theoretical_loss": 3.3401076822075417,
      "tokens_seen": 2783313920
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015935353535353536,
      "loss": 1.4728,
      "theoretical_loss": 3.3401015429406327,
      "tokens_seen": 2783379456
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015933333333333332,
      "loss": 1.4339,
      "theoretical_loss": 3.340095403858748,
      "tokens_seen": 2783444992
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015931313131313133,
      "loss": 1.1497,
      "theoretical_loss": 3.3400892649618767,
      "tokens_seen": 2783510528
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0001592929292929293,
      "loss": 1.5182,
      "theoretical_loss": 3.3400831262500095,
      "tokens_seen": 2783576064
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015927272727272727,
      "loss": 1.2121,
      "theoretical_loss": 3.340076987723136,
      "tokens_seen": 2783641600
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015925252525252526,
      "loss": 1.5234,
      "theoretical_loss": 3.3400708493812465,
      "tokens_seen": 2783707136
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015923232323232324,
      "loss": 1.2678,
      "theoretical_loss": 3.3400647112243314,
      "tokens_seen": 2783772672
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00015921212121212122,
      "loss": 1.1948,
      "theoretical_loss": 3.3400585732523798,
      "tokens_seen": 2783838208
    },
    {
      "epoch": 0.68,
      "objective/train/aligned_token_count": 0.4261363744735718,
      "objective/train/avg_token_score": 0.010898923501372337,
      "objective/train/docs_used": 1567583,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3468644618988037,
      "objective/train/original_loss": 4.441183090209961,
      "objective/train/theoretical_loss": 3.340052435465383,
      "objective/train/tokens_used": 1142762976,
      "objective/train/unlikelihood_loss": 0.04538632929325104,
      "theoretical_loss": 3.340052435465383,
      "tokens_seen": 2783903744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015919191919191918,
      "loss": 1.3371,
      "theoretical_loss": 3.340052435465383,
      "tokens_seen": 2783903744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015917171717171717,
      "loss": 1.3188,
      "theoretical_loss": 3.3400462978633305,
      "tokens_seen": 2783969280
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015915151515151515,
      "loss": 1.3314,
      "theoretical_loss": 3.340040160446212,
      "tokens_seen": 2784034816
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015913131313131313,
      "loss": 1.152,
      "theoretical_loss": 3.3400340232140184,
      "tokens_seen": 2784100352
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015911111111111112,
      "loss": 1.4033,
      "theoretical_loss": 3.340027886166739,
      "tokens_seen": 2784165888
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001590909090909091,
      "loss": 1.2003,
      "theoretical_loss": 3.3400217493043645,
      "tokens_seen": 2784231424
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015907070707070706,
      "loss": 1.3563,
      "theoretical_loss": 3.3400156126268845,
      "tokens_seen": 2784296960
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015905050505050507,
      "loss": 1.358,
      "theoretical_loss": 3.340009476134289,
      "tokens_seen": 2784362496
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015903030303030303,
      "loss": 1.3072,
      "theoretical_loss": 3.340003339826569,
      "tokens_seen": 2784428032
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000159010101010101,
      "loss": 1.379,
      "theoretical_loss": 3.3399972037037133,
      "tokens_seen": 2784493568
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000158989898989899,
      "loss": 1.4402,
      "theoretical_loss": 3.3399910677657125,
      "tokens_seen": 2784559104
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015896969696969698,
      "loss": 1.3775,
      "theoretical_loss": 3.339984932012557,
      "tokens_seen": 2784624640
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015894949494949496,
      "loss": 1.177,
      "theoretical_loss": 3.339978796444237,
      "tokens_seen": 2784690176
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015892929292929292,
      "loss": 1.2751,
      "theoretical_loss": 3.339972661060742,
      "tokens_seen": 2784755712
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015890909090909093,
      "loss": 1.3293,
      "theoretical_loss": 3.3399665258620623,
      "tokens_seen": 2784821248
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001588888888888889,
      "loss": 1.2372,
      "theoretical_loss": 3.339960390848188,
      "tokens_seen": 2784886784
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001588686868686869,
      "loss": 1.2836,
      "theoretical_loss": 3.339954256019109,
      "tokens_seen": 2784952320
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015884848484848485,
      "loss": 1.2221,
      "theoretical_loss": 3.3399481213748157,
      "tokens_seen": 2785017856
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001588282828282828,
      "loss": 1.1923,
      "theoretical_loss": 3.3399419869152984,
      "tokens_seen": 2785083392
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015880808080808082,
      "loss": 1.1657,
      "theoretical_loss": 3.3399358526405463,
      "tokens_seen": 2785148928
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015878787878787878,
      "loss": 1.2665,
      "theoretical_loss": 3.3399297185505503,
      "tokens_seen": 2785214464
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001587676767676768,
      "loss": 1.494,
      "theoretical_loss": 3.3399235846453004,
      "tokens_seen": 2785280000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015874747474747475,
      "loss": 1.671,
      "theoretical_loss": 3.339917450924786,
      "tokens_seen": 2785345536
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015872727272727273,
      "loss": 1.244,
      "theoretical_loss": 3.3399113173889985,
      "tokens_seen": 2785411072
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015870707070707072,
      "loss": 1.1392,
      "theoretical_loss": 3.3399051840379266,
      "tokens_seen": 2785476608
    },
    {
      "epoch": 0.69,
      "objective/train/aligned_token_count": 0.3627810478210449,
      "objective/train/avg_token_score": 0.007816936820745468,
      "objective/train/docs_used": 1568939,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8979986310005188,
      "objective/train/original_loss": 4.103878021240234,
      "objective/train/theoretical_loss": 3.339899050871561,
      "objective/train/tokens_used": 1144401376,
      "objective/train/unlikelihood_loss": 0.06459350883960724,
      "theoretical_loss": 3.339899050871561,
      "tokens_seen": 2785542144
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015868686868686867,
      "loss": 1.2137,
      "theoretical_loss": 3.339899050871561,
      "tokens_seen": 2785542144
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015866666666666668,
      "loss": 1.2687,
      "theoretical_loss": 3.339892917889892,
      "tokens_seen": 2785607680
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015864646464646464,
      "loss": 1.3296,
      "theoretical_loss": 3.3398867850929093,
      "tokens_seen": 2785673216
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015862626262626262,
      "loss": 1.2131,
      "theoretical_loss": 3.339880652480603,
      "tokens_seen": 2785738752
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001586060606060606,
      "loss": 1.3039,
      "theoretical_loss": 3.339874520052964,
      "tokens_seen": 2785804288
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001585858585858586,
      "loss": 1.498,
      "theoretical_loss": 3.339868387809981,
      "tokens_seen": 2785869824
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015856565656565658,
      "loss": 1.2553,
      "theoretical_loss": 3.3398622557516453,
      "tokens_seen": 2785935360
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015854545454545456,
      "loss": 1.4648,
      "theoretical_loss": 3.3398561238779463,
      "tokens_seen": 2786000896
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015852525252525252,
      "loss": 1.2851,
      "theoretical_loss": 3.3398499921888742,
      "tokens_seen": 2786066432
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001585050505050505,
      "loss": 1.5818,
      "theoretical_loss": 3.33984386068442,
      "tokens_seen": 2786131968
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015848484848484848,
      "loss": 1.1784,
      "theoretical_loss": 3.339837729364572,
      "tokens_seen": 2786197504
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015846464646464647,
      "loss": 1.3366,
      "theoretical_loss": 3.3398315982293223,
      "tokens_seen": 2786263040
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015844444444444445,
      "loss": 1.3367,
      "theoretical_loss": 3.3398254672786596,
      "tokens_seen": 2786328576
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001584242424242424,
      "loss": 1.3543,
      "theoretical_loss": 3.339819336512574,
      "tokens_seen": 2786394112
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015840404040404042,
      "loss": 1.1903,
      "theoretical_loss": 3.3398132059310566,
      "tokens_seen": 2786459648
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015838383838383838,
      "loss": 1.4105,
      "theoretical_loss": 3.3398070755340967,
      "tokens_seen": 2786525184
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001583636363636364,
      "loss": 1.3985,
      "theoretical_loss": 3.3398009453216853,
      "tokens_seen": 2786590720
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015834343434343435,
      "loss": 1.4802,
      "theoretical_loss": 3.339794815293811,
      "tokens_seen": 2786656256
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001583232323232323,
      "loss": 1.2262,
      "theoretical_loss": 3.3397886854504653,
      "tokens_seen": 2786721792
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001583030303030303,
      "loss": 1.54,
      "theoretical_loss": 3.339782555791637,
      "tokens_seen": 2786787328
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015828282828282827,
      "loss": 1.2118,
      "theoretical_loss": 3.339776426317318,
      "tokens_seen": 2786852864
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015826262626262628,
      "loss": 1.265,
      "theoretical_loss": 3.339770297027497,
      "tokens_seen": 2786918400
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015824242424242424,
      "loss": 1.2209,
      "theoretical_loss": 3.3397641679221644,
      "tokens_seen": 2786983936
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015822222222222222,
      "loss": 1.3513,
      "theoretical_loss": 3.3397580390013104,
      "tokens_seen": 2787049472
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001582020202020202,
      "loss": 1.3894,
      "theoretical_loss": 3.339751910264925,
      "tokens_seen": 2787115008
    },
    {
      "epoch": 0.69,
      "objective/train/aligned_token_count": 0.33241692185401917,
      "objective/train/avg_token_score": 0.008783010765910149,
      "objective/train/docs_used": 1569381,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.022804617881775,
      "objective/train/original_loss": 5.36049747467041,
      "objective/train/theoretical_loss": 3.3397457817129985,
      "objective/train/tokens_used": 1146039776,
      "objective/train/unlikelihood_loss": 0.04629266634583473,
      "theoretical_loss": 3.3397457817129985,
      "tokens_seen": 2787180544
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001581818181818182,
      "loss": 1.3021,
      "theoretical_loss": 3.3397457817129985,
      "tokens_seen": 2787180544
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015816161616161617,
      "loss": 1.0854,
      "theoretical_loss": 3.339739653345521,
      "tokens_seen": 2787246080
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015814141414141413,
      "loss": 1.3358,
      "theoretical_loss": 3.3397335251624822,
      "tokens_seen": 2787311616
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015812121212121214,
      "loss": 1.3013,
      "theoretical_loss": 3.3397273971638732,
      "tokens_seen": 2787377152
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001581010101010101,
      "loss": 1.5044,
      "theoretical_loss": 3.339721269349683,
      "tokens_seen": 2787442688
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015808080808080808,
      "loss": 1.3587,
      "theoretical_loss": 3.339715141719902,
      "tokens_seen": 2787508224
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015806060606060607,
      "loss": 1.2283,
      "theoretical_loss": 3.339709014274521,
      "tokens_seen": 2787573760
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015804040404040405,
      "loss": 1.3059,
      "theoretical_loss": 3.3397028870135292,
      "tokens_seen": 2787639296
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015802020202020203,
      "loss": 1.3024,
      "theoretical_loss": 3.3396967599369174,
      "tokens_seen": 2787704832
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000158,
      "loss": 1.1818,
      "theoretical_loss": 3.3396906330446754,
      "tokens_seen": 2787770368
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015797979797979798,
      "loss": 1.3616,
      "theoretical_loss": 3.339684506336793,
      "tokens_seen": 2787835904
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015795959595959596,
      "loss": 1.365,
      "theoretical_loss": 3.339678379813261,
      "tokens_seen": 2787901440
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015793939393939394,
      "loss": 1.4373,
      "theoretical_loss": 3.3396722534740695,
      "tokens_seen": 2787966976
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015791919191919193,
      "loss": 1.5049,
      "theoretical_loss": 3.339666127319208,
      "tokens_seen": 2788032512
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001578989898989899,
      "loss": 1.3201,
      "theoretical_loss": 3.339660001348667,
      "tokens_seen": 2788098048
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015787878787878787,
      "loss": 1.266,
      "theoretical_loss": 3.3396538755624365,
      "tokens_seen": 2788163584
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015785858585858588,
      "loss": 1.4197,
      "theoretical_loss": 3.339647749960507,
      "tokens_seen": 2788229120
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015783838383838384,
      "loss": 1.3557,
      "theoretical_loss": 3.3396416245428675,
      "tokens_seen": 2788294656
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015781818181818182,
      "loss": 1.1967,
      "theoretical_loss": 3.3396354993095096,
      "tokens_seen": 2788360192
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001577979797979798,
      "loss": 1.182,
      "theoretical_loss": 3.3396293742604226,
      "tokens_seen": 2788425728
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015777777777777776,
      "loss": 1.4305,
      "theoretical_loss": 3.339623249395597,
      "tokens_seen": 2788491264
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015775757575757577,
      "loss": 1.4084,
      "theoretical_loss": 3.3396171247150224,
      "tokens_seen": 2788556800
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015773737373737373,
      "loss": 1.4017,
      "theoretical_loss": 3.339611000218689,
      "tokens_seen": 2788622336
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015771717171717174,
      "loss": 1.2216,
      "theoretical_loss": 3.3396048759065877,
      "tokens_seen": 2788687872
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001576969696969697,
      "loss": 1.3573,
      "theoretical_loss": 3.3395987517787082,
      "tokens_seen": 2788753408
    },
    {
      "epoch": 0.69,
      "objective/train/aligned_token_count": 0.49670088291168213,
      "objective/train/avg_token_score": 0.009816481731832027,
      "objective/train/docs_used": 1570879,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.535805583000183,
      "objective/train/original_loss": 4.2046799659729,
      "objective/train/theoretical_loss": 3.33959262783504,
      "objective/train/tokens_used": 1147678176,
      "objective/train/unlikelihood_loss": 0.03747493773698807,
      "theoretical_loss": 3.33959262783504,
      "tokens_seen": 2788818944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015767676767676768,
      "loss": 1.487,
      "theoretical_loss": 3.33959262783504,
      "tokens_seen": 2788818944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015765656565656566,
      "loss": 1.2795,
      "theoretical_loss": 3.339586504075574,
      "tokens_seen": 2788884480
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015763636363636362,
      "loss": 1.5,
      "theoretical_loss": 3.3395803805003004,
      "tokens_seen": 2788950016
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015761616161616163,
      "loss": 1.263,
      "theoretical_loss": 3.3395742571092084,
      "tokens_seen": 2789015552
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001575959595959596,
      "loss": 1.4752,
      "theoretical_loss": 3.339568133902289,
      "tokens_seen": 2789081088
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015757575757575757,
      "loss": 1.4413,
      "theoretical_loss": 3.3395620108795323,
      "tokens_seen": 2789146624
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015755555555555556,
      "loss": 0.9845,
      "theoretical_loss": 3.339555888040928,
      "tokens_seen": 2789212160
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015753535353535354,
      "loss": 1.4123,
      "theoretical_loss": 3.3395497653864665,
      "tokens_seen": 2789277696
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015751515151515153,
      "loss": 1.3997,
      "theoretical_loss": 3.339543642916138,
      "tokens_seen": 2789343232
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001574949494949495,
      "loss": 1.55,
      "theoretical_loss": 3.339537520629932,
      "tokens_seen": 2789408768
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015747474747474747,
      "loss": 1.3123,
      "theoretical_loss": 3.33953139852784,
      "tokens_seen": 2789474304
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015745454545454545,
      "loss": 1.562,
      "theoretical_loss": 3.3395252766098507,
      "tokens_seen": 2789539840
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015743434343434343,
      "loss": 1.4923,
      "theoretical_loss": 3.339519154875955,
      "tokens_seen": 2789605376
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015741414141414142,
      "loss": 1.1877,
      "theoretical_loss": 3.339513033326143,
      "tokens_seen": 2789670912
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001573939393939394,
      "loss": 1.3332,
      "theoretical_loss": 3.3395069119604046,
      "tokens_seen": 2789736448
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015737373737373736,
      "loss": 1.4296,
      "theoretical_loss": 3.3395007907787297,
      "tokens_seen": 2789801984
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015735353535353537,
      "loss": 1.2741,
      "theoretical_loss": 3.3394946697811094,
      "tokens_seen": 2789867520
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015733333333333333,
      "loss": 1.1734,
      "theoretical_loss": 3.339488548967533,
      "tokens_seen": 2789933056
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015731313131313134,
      "loss": 1.344,
      "theoretical_loss": 3.339482428337991,
      "tokens_seen": 2789998592
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001572929292929293,
      "loss": 1.156,
      "theoretical_loss": 3.3394763078924736,
      "tokens_seen": 2790064128
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015727272727272728,
      "loss": 1.303,
      "theoretical_loss": 3.3394701876309703,
      "tokens_seen": 2790129664
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015725252525252526,
      "loss": 1.4318,
      "theoretical_loss": 3.3394640675534717,
      "tokens_seen": 2790195200
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015723232323232322,
      "loss": 1.3489,
      "theoretical_loss": 3.3394579476599686,
      "tokens_seen": 2790260736
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015721212121212123,
      "loss": 1.2521,
      "theoretical_loss": 3.33945182795045,
      "tokens_seen": 2790326272
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001571919191919192,
      "loss": 1.4087,
      "theoretical_loss": 3.3394457084249067,
      "tokens_seen": 2790391808
    },
    {
      "epoch": 0.69,
      "objective/train/aligned_token_count": 0.45869991183280945,
      "objective/train/avg_token_score": 0.013077493757009506,
      "objective/train/docs_used": 1571508,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4797133207321167,
      "objective/train/original_loss": 3.772479295730591,
      "objective/train/theoretical_loss": 3.3394395890833284,
      "objective/train/tokens_used": 1149316576,
      "objective/train/unlikelihood_loss": 0.06250520795583725,
      "theoretical_loss": 3.3394395890833284,
      "tokens_seen": 2790457344
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001571717171717172,
      "loss": 1.3744,
      "theoretical_loss": 3.3394395890833284,
      "tokens_seen": 2790457344
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015715151515151516,
      "loss": 1.4704,
      "theoretical_loss": 3.339433469925706,
      "tokens_seen": 2790522880
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001571313131313131,
      "loss": 1.2891,
      "theoretical_loss": 3.339427350952029,
      "tokens_seen": 2790588416
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015711111111111112,
      "loss": 1.3928,
      "theoretical_loss": 3.339421232162288,
      "tokens_seen": 2790653952
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015709090909090908,
      "loss": 1.4008,
      "theoretical_loss": 3.3394151135564725,
      "tokens_seen": 2790719488
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001570707070707071,
      "loss": 1.2865,
      "theoretical_loss": 3.3394089951345736,
      "tokens_seen": 2790785024
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015705050505050505,
      "loss": 1.2833,
      "theoretical_loss": 3.3394028768965804,
      "tokens_seen": 2790850560
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015703030303030303,
      "loss": 1.3042,
      "theoretical_loss": 3.3393967588424838,
      "tokens_seen": 2790916096
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015701010101010102,
      "loss": 1.3958,
      "theoretical_loss": 3.3393906409722733,
      "tokens_seen": 2790981632
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000156989898989899,
      "loss": 1.4917,
      "theoretical_loss": 3.3393845232859403,
      "tokens_seen": 2791047168
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015696969696969698,
      "loss": 1.3302,
      "theoretical_loss": 3.3393784057834734,
      "tokens_seen": 2791112704
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015694949494949494,
      "loss": 1.319,
      "theoretical_loss": 3.339372288464864,
      "tokens_seen": 2791178240
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015692929292929292,
      "loss": 1.3901,
      "theoretical_loss": 3.3393661713301013,
      "tokens_seen": 2791243776
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001569090909090909,
      "loss": 1.2874,
      "theoretical_loss": 3.3393600543791764,
      "tokens_seen": 2791309312
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001568888888888889,
      "loss": 1.1418,
      "theoretical_loss": 3.3393539376120787,
      "tokens_seen": 2791374848
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015686868686868688,
      "loss": 1.5594,
      "theoretical_loss": 3.339347821028799,
      "tokens_seen": 2791440384
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015684848484848486,
      "loss": 1.3387,
      "theoretical_loss": 3.3393417046293266,
      "tokens_seen": 2791505920
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015682828282828282,
      "loss": 1.206,
      "theoretical_loss": 3.339335588413652,
      "tokens_seen": 2791571456
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015680808080808083,
      "loss": 1.4901,
      "theoretical_loss": 3.339329472381766,
      "tokens_seen": 2791636992
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015678787878787879,
      "loss": 1.5685,
      "theoretical_loss": 3.339323356533658,
      "tokens_seen": 2791702528
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015676767676767677,
      "loss": 1.3918,
      "theoretical_loss": 3.3393172408693186,
      "tokens_seen": 2791768064
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015674747474747475,
      "loss": 1.4478,
      "theoretical_loss": 3.3393111253887375,
      "tokens_seen": 2791833600
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001567272727272727,
      "loss": 1.3081,
      "theoretical_loss": 3.3393050100919055,
      "tokens_seen": 2791899136
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015670707070707072,
      "loss": 1.2395,
      "theoretical_loss": 3.3392988949788123,
      "tokens_seen": 2791964672
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015668686868686868,
      "loss": 1.298,
      "theoretical_loss": 3.3392927800494485,
      "tokens_seen": 2792030208
    },
    {
      "epoch": 0.69,
      "objective/train/aligned_token_count": 0.33290567994117737,
      "objective/train/avg_token_score": 0.011404612101614475,
      "objective/train/docs_used": 1572267,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0350443124771118,
      "objective/train/original_loss": 4.687503337860107,
      "objective/train/theoretical_loss": 3.3392866653038036,
      "objective/train/tokens_used": 1150954976,
      "objective/train/unlikelihood_loss": 0.06253455579280853,
      "theoretical_loss": 3.3392866653038036,
      "tokens_seen": 2792095744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001566666666666667,
      "loss": 1.1754,
      "theoretical_loss": 3.3392866653038036,
      "tokens_seen": 2792095744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015664646464646465,
      "loss": 1.3516,
      "theoretical_loss": 3.339280550741868,
      "tokens_seen": 2792161280
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015662626262626263,
      "loss": 1.4894,
      "theoretical_loss": 3.3392744363636324,
      "tokens_seen": 2792226816
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015660606060606061,
      "loss": 1.2806,
      "theoretical_loss": 3.3392683221690866,
      "tokens_seen": 2792292352
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015658585858585857,
      "loss": 1.534,
      "theoretical_loss": 3.33926220815822,
      "tokens_seen": 2792357888
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015656565656565658,
      "loss": 1.4322,
      "theoretical_loss": 3.3392560943310245,
      "tokens_seen": 2792423424
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015654545454545454,
      "loss": 1.3258,
      "theoretical_loss": 3.339249980687489,
      "tokens_seen": 2792488960
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015652525252525252,
      "loss": 1.3319,
      "theoretical_loss": 3.3392438672276037,
      "tokens_seen": 2792554496
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001565050505050505,
      "loss": 1.2482,
      "theoretical_loss": 3.339237753951359,
      "tokens_seen": 2792620032
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001564848484848485,
      "loss": 1.5186,
      "theoretical_loss": 3.3392316408587455,
      "tokens_seen": 2792685568
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015646464646464647,
      "loss": 1.1443,
      "theoretical_loss": 3.3392255279497527,
      "tokens_seen": 2792751104
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015644444444444443,
      "loss": 1.254,
      "theoretical_loss": 3.3392194152243713,
      "tokens_seen": 2792816640
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015642424242424244,
      "loss": 1.4624,
      "theoretical_loss": 3.339213302682591,
      "tokens_seen": 2792882176
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001564040404040404,
      "loss": 1.467,
      "theoretical_loss": 3.339207190324402,
      "tokens_seen": 2792947712
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015638383838383838,
      "loss": 1.3595,
      "theoretical_loss": 3.339201078149795,
      "tokens_seen": 2793013248
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015636363636363637,
      "loss": 1.3791,
      "theoretical_loss": 3.33919496615876,
      "tokens_seen": 2793078784
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015634343434343435,
      "loss": 1.31,
      "theoretical_loss": 3.3391888543512867,
      "tokens_seen": 2793144320
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015632323232323234,
      "loss": 1.3559,
      "theoretical_loss": 3.339182742727366,
      "tokens_seen": 2793209856
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015630303030303032,
      "loss": 1.5267,
      "theoretical_loss": 3.3391766312869873,
      "tokens_seen": 2793275392
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015628282828282828,
      "loss": 1.3837,
      "theoretical_loss": 3.339170520030142,
      "tokens_seen": 2793340928
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015626262626262626,
      "loss": 1.2358,
      "theoretical_loss": 3.339164408956819,
      "tokens_seen": 2793406464
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015624242424242424,
      "loss": 1.4166,
      "theoretical_loss": 3.3391582980670087,
      "tokens_seen": 2793472000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015622222222222223,
      "loss": 1.357,
      "theoretical_loss": 3.3391521873607015,
      "tokens_seen": 2793537536
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001562020202020202,
      "loss": 1.0003,
      "theoretical_loss": 3.3391460768378876,
      "tokens_seen": 2793603072
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015618181818181817,
      "loss": 1.3635,
      "theoretical_loss": 3.3391399664985575,
      "tokens_seen": 2793668608
    },
    {
      "epoch": 0.69,
      "objective/train/aligned_token_count": 0.31414955854415894,
      "objective/train/avg_token_score": 0.03534943610429764,
      "objective/train/docs_used": 1573597,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.969791829586029,
      "objective/train/original_loss": 4.974410057067871,
      "objective/train/theoretical_loss": 3.339133856342701,
      "objective/train/tokens_used": 1152593376,
      "objective/train/unlikelihood_loss": 0.0492885522544384,
      "theoretical_loss": 3.339133856342701,
      "tokens_seen": 2793734144
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015616161616161618,
      "loss": 1.2119,
      "theoretical_loss": 3.339133856342701,
      "tokens_seen": 2793734144
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015614141414141414,
      "loss": 1.229,
      "theoretical_loss": 3.3391277463703086,
      "tokens_seen": 2793799680
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015612121212121215,
      "loss": 1.5253,
      "theoretical_loss": 3.3391216365813703,
      "tokens_seen": 2793865216
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001561010101010101,
      "loss": 1.2297,
      "theoretical_loss": 3.339115526975876,
      "tokens_seen": 2793930752
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015608080808080806,
      "loss": 1.445,
      "theoretical_loss": 3.339109417553816,
      "tokens_seen": 2793996288
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015606060606060607,
      "loss": 1.4926,
      "theoretical_loss": 3.339103308315181,
      "tokens_seen": 2794061824
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015604040404040403,
      "loss": 1.2673,
      "theoretical_loss": 3.3390971992599607,
      "tokens_seen": 2794127360
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015602020202020204,
      "loss": 1.2798,
      "theoretical_loss": 3.3390910903881457,
      "tokens_seen": 2794192896
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000156,
      "loss": 1.2549,
      "theoretical_loss": 3.3390849816997257,
      "tokens_seen": 2794258432
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015597979797979798,
      "loss": 1.4011,
      "theoretical_loss": 3.339078873194691,
      "tokens_seen": 2794323968
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015595959595959597,
      "loss": 1.3134,
      "theoretical_loss": 3.339072764873032,
      "tokens_seen": 2794389504
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015593939393939395,
      "loss": 1.3295,
      "theoretical_loss": 3.3390666567347385,
      "tokens_seen": 2794455040
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015591919191919193,
      "loss": 1.4495,
      "theoretical_loss": 3.3390605487798015,
      "tokens_seen": 2794520576
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001558989898989899,
      "loss": 1.2826,
      "theoretical_loss": 3.3390544410082104,
      "tokens_seen": 2794586112
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015587878787878787,
      "loss": 1.1448,
      "theoretical_loss": 3.3390483334199557,
      "tokens_seen": 2794651648
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015585858585858586,
      "loss": 1.0617,
      "theoretical_loss": 3.3390422260150276,
      "tokens_seen": 2794717184
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015583838383838384,
      "loss": 1.3618,
      "theoretical_loss": 3.3390361187934166,
      "tokens_seen": 2794782720
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015581818181818183,
      "loss": 1.4365,
      "theoretical_loss": 3.339030011755112,
      "tokens_seen": 2794848256
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001557979797979798,
      "loss": 1.1648,
      "theoretical_loss": 3.339023904900105,
      "tokens_seen": 2794913792
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015577777777777777,
      "loss": 1.4769,
      "theoretical_loss": 3.3390177982283853,
      "tokens_seen": 2794979328
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015575757575757578,
      "loss": 1.3216,
      "theoretical_loss": 3.339011691739943,
      "tokens_seen": 2795044864
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015573737373737373,
      "loss": 1.3554,
      "theoretical_loss": 3.339005585434769,
      "tokens_seen": 2795110400
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015571717171717172,
      "loss": 1.5415,
      "theoretical_loss": 3.3389994793128523,
      "tokens_seen": 2795175936
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001556969696969697,
      "loss": 1.4016,
      "theoretical_loss": 3.3389933733741843,
      "tokens_seen": 2795241472
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015567676767676766,
      "loss": 1.2179,
      "theoretical_loss": 3.3389872676187546,
      "tokens_seen": 2795307008
    },
    {
      "epoch": 0.69,
      "objective/train/aligned_token_count": 0.42418134212493896,
      "objective/train/avg_token_score": 0.011046471074223518,
      "objective/train/docs_used": 1574190,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5125102996826172,
      "objective/train/original_loss": 4.507443904876709,
      "objective/train/theoretical_loss": 3.3389811620465535,
      "objective/train/tokens_used": 1154231776,
      "objective/train/unlikelihood_loss": 0.054854780435562134,
      "theoretical_loss": 3.3389811620465535,
      "tokens_seen": 2795372544
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015565656565656567,
      "loss": 1.5561,
      "theoretical_loss": 3.3389811620465535,
      "tokens_seen": 2795372544
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015563636363636363,
      "loss": 1.2515,
      "theoretical_loss": 3.338975056657571,
      "tokens_seen": 2795438080
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015561616161616164,
      "loss": 1.4478,
      "theoretical_loss": 3.3389689514517977,
      "tokens_seen": 2795503616
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001555959595959596,
      "loss": 1.3573,
      "theoretical_loss": 3.3389628464292236,
      "tokens_seen": 2795569152
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001555757575757576,
      "loss": 1.4338,
      "theoretical_loss": 3.338956741589839,
      "tokens_seen": 2795634688
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015555555555555556,
      "loss": 1.2784,
      "theoretical_loss": 3.338950636933634,
      "tokens_seen": 2795700224
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015553535353535352,
      "loss": 1.2033,
      "theoretical_loss": 3.338944532460599,
      "tokens_seen": 2795765760
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015551515151515153,
      "loss": 1.3537,
      "theoretical_loss": 3.3389384281707235,
      "tokens_seen": 2795831296
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001554949494949495,
      "loss": 1.461,
      "theoretical_loss": 3.338932324063999,
      "tokens_seen": 2795896832
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001554747474747475,
      "loss": 1.29,
      "theoretical_loss": 3.338926220140414,
      "tokens_seen": 2795962368
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015545454545454546,
      "loss": 1.2805,
      "theoretical_loss": 3.3389201163999607,
      "tokens_seen": 2796027904
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015543434343434344,
      "loss": 1.5694,
      "theoretical_loss": 3.3389140128426282,
      "tokens_seen": 2796093440
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015541414141414142,
      "loss": 1.3719,
      "theoretical_loss": 3.338907909468406,
      "tokens_seen": 2796158976
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015539393939393938,
      "loss": 1.1756,
      "theoretical_loss": 3.338901806277286,
      "tokens_seen": 2796224512
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001553737373737374,
      "loss": 1.3076,
      "theoretical_loss": 3.3388957032692574,
      "tokens_seen": 2796290048
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015535353535353535,
      "loss": 1.5614,
      "theoretical_loss": 3.3388896004443107,
      "tokens_seen": 2796355584
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015533333333333333,
      "loss": 1.2817,
      "theoretical_loss": 3.3388834978024358,
      "tokens_seen": 2796421120
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015531313131313132,
      "loss": 1.2148,
      "theoretical_loss": 3.338877395343623,
      "tokens_seen": 2796486656
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001552929292929293,
      "loss": 1.4247,
      "theoretical_loss": 3.3388712930678626,
      "tokens_seen": 2796552192
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015527272727272728,
      "loss": 1.3275,
      "theoretical_loss": 3.338865190975145,
      "tokens_seen": 2796617728
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015525252525252527,
      "loss": 1.2941,
      "theoretical_loss": 3.33885908906546,
      "tokens_seen": 2796683264
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015523232323232323,
      "loss": 1.2621,
      "theoretical_loss": 3.3388529873387984,
      "tokens_seen": 2796748800
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001552121212121212,
      "loss": 1.4011,
      "theoretical_loss": 3.33884688579515,
      "tokens_seen": 2796814336
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001551919191919192,
      "loss": 1.3632,
      "theoretical_loss": 3.338840784434505,
      "tokens_seen": 2796879872
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015517171717171718,
      "loss": 1.4015,
      "theoretical_loss": 3.338834683256854,
      "tokens_seen": 2796945408
    },
    {
      "epoch": 0.69,
      "objective/train/aligned_token_count": 0.2520161271095276,
      "objective/train/avg_token_score": 0.01029267255216837,
      "objective/train/docs_used": 1575609,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.7430685758590698,
      "objective/train/original_loss": 4.703075885772705,
      "objective/train/theoretical_loss": 3.338828582262187,
      "objective/train/tokens_used": 1155870176,
      "objective/train/unlikelihood_loss": 0.06561023741960526,
      "theoretical_loss": 3.338828582262187,
      "tokens_seen": 2797010944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015515151515151516,
      "loss": 1.1319,
      "theoretical_loss": 3.338828582262187,
      "tokens_seen": 2797010944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015513131313131312,
      "loss": 1.3522,
      "theoretical_loss": 3.338822481450494,
      "tokens_seen": 2797076480
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015511111111111113,
      "loss": 1.4057,
      "theoretical_loss": 3.3388163808217657,
      "tokens_seen": 2797142016
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015509090909090909,
      "loss": 1.5864,
      "theoretical_loss": 3.3388102803759914,
      "tokens_seen": 2797207552
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001550707070707071,
      "loss": 1.5442,
      "theoretical_loss": 3.3388041801131627,
      "tokens_seen": 2797273088
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015505050505050505,
      "loss": 1.3131,
      "theoretical_loss": 3.3387980800332686,
      "tokens_seen": 2797338624
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000155030303030303,
      "loss": 1.2279,
      "theoretical_loss": 3.3387919801363,
      "tokens_seen": 2797404160
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015501010101010102,
      "loss": 1.5173,
      "theoretical_loss": 3.338785880422247,
      "tokens_seen": 2797469696
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015498989898989898,
      "loss": 1.4331,
      "theoretical_loss": 3.3387797808911,
      "tokens_seen": 2797535232
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000154969696969697,
      "loss": 1.3453,
      "theoretical_loss": 3.338773681542849,
      "tokens_seen": 2797600768
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015494949494949495,
      "loss": 1.264,
      "theoretical_loss": 3.3387675823774843,
      "tokens_seen": 2797666304
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015492929292929293,
      "loss": 1.3757,
      "theoretical_loss": 3.338761483394996,
      "tokens_seen": 2797731840
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015490909090909091,
      "loss": 1.4889,
      "theoretical_loss": 3.338755384595374,
      "tokens_seen": 2797797376
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001548888888888889,
      "loss": 1.4085,
      "theoretical_loss": 3.3387492859786096,
      "tokens_seen": 2797862912
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015486868686868688,
      "loss": 1.4847,
      "theoretical_loss": 3.338743187544692,
      "tokens_seen": 2797928448
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015484848484848484,
      "loss": 1.3255,
      "theoretical_loss": 3.338737089293612,
      "tokens_seen": 2797993984
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015482828282828282,
      "loss": 1.3535,
      "theoretical_loss": 3.3387309912253595,
      "tokens_seen": 2798059520
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001548080808080808,
      "loss": 1.4089,
      "theoretical_loss": 3.338724893339925,
      "tokens_seen": 2798125056
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001547878787878788,
      "loss": 1.317,
      "theoretical_loss": 3.338718795637299,
      "tokens_seen": 2798190592
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015476767676767677,
      "loss": 1.3994,
      "theoretical_loss": 3.3387126981174706,
      "tokens_seen": 2798256128
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015474747474747476,
      "loss": 1.3552,
      "theoretical_loss": 3.338706600780432,
      "tokens_seen": 2798321664
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015472727272727274,
      "loss": 1.5407,
      "theoretical_loss": 3.338700503626171,
      "tokens_seen": 2798387200
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001547070707070707,
      "loss": 1.3881,
      "theoretical_loss": 3.3386944066546795,
      "tokens_seen": 2798452736
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015468686868686868,
      "loss": 1.3349,
      "theoretical_loss": 3.3386883098659474,
      "tokens_seen": 2798518272
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015466666666666667,
      "loss": 1.3671,
      "theoretical_loss": 3.338682213259965,
      "tokens_seen": 2798583808
    },
    {
      "epoch": 0.69,
      "objective/train/aligned_token_count": 0.4429985284805298,
      "objective/train/avg_token_score": 0.008923156186938286,
      "objective/train/docs_used": 1576275,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4102362394332886,
      "objective/train/original_loss": 3.9732553958892822,
      "objective/train/theoretical_loss": 3.3386761168367225,
      "objective/train/tokens_used": 1157508576,
      "objective/train/unlikelihood_loss": 0.07731222361326218,
      "theoretical_loss": 3.3386761168367225,
      "tokens_seen": 2798649344
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015464646464646465,
      "loss": 1.4476,
      "theoretical_loss": 3.3386761168367225,
      "tokens_seen": 2798649344
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015462626262626264,
      "loss": 1.2729,
      "theoretical_loss": 3.33867002059621,
      "tokens_seen": 2798714880
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015460606060606062,
      "loss": 1.3333,
      "theoretical_loss": 3.3386639245384178,
      "tokens_seen": 2798780416
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015458585858585858,
      "loss": 1.3993,
      "theoretical_loss": 3.3386578286633357,
      "tokens_seen": 2798845952
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001545656565656566,
      "loss": 1.5146,
      "theoretical_loss": 3.338651732970955,
      "tokens_seen": 2798911488
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015454545454545454,
      "loss": 1.1068,
      "theoretical_loss": 3.338645637461265,
      "tokens_seen": 2798977024
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015452525252525253,
      "loss": 1.1432,
      "theoretical_loss": 3.3386395421342567,
      "tokens_seen": 2799042560
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001545050505050505,
      "loss": 1.5691,
      "theoretical_loss": 3.3386334469899195,
      "tokens_seen": 2799108096
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015448484848484847,
      "loss": 1.4515,
      "theoretical_loss": 3.3386273520282446,
      "tokens_seen": 2799173632
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015446464646464648,
      "loss": 1.3918,
      "theoretical_loss": 3.3386212572492213,
      "tokens_seen": 2799239168
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015444444444444444,
      "loss": 1.2879,
      "theoretical_loss": 3.3386151626528404,
      "tokens_seen": 2799304704
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015442424242424245,
      "loss": 1.4091,
      "theoretical_loss": 3.338609068239092,
      "tokens_seen": 2799370240
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001544040404040404,
      "loss": 1.2981,
      "theoretical_loss": 3.3386029740079666,
      "tokens_seen": 2799435776
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001543838383838384,
      "loss": 1.2806,
      "theoretical_loss": 3.338596879959454,
      "tokens_seen": 2799501312
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015436363636363637,
      "loss": 1.3784,
      "theoretical_loss": 3.338590786093545,
      "tokens_seen": 2799566848
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015434343434343433,
      "loss": 1.1878,
      "theoretical_loss": 3.338584692410229,
      "tokens_seen": 2799632384
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015432323232323234,
      "loss": 1.3658,
      "theoretical_loss": 3.338578598909497,
      "tokens_seen": 2799697920
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0001543030303030303,
      "loss": 1.4088,
      "theoretical_loss": 3.3385725055913396,
      "tokens_seen": 2799763456
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015428282828282828,
      "loss": 1.3823,
      "theoretical_loss": 3.338566412455746,
      "tokens_seen": 2799828992
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015426262626262627,
      "loss": 1.2944,
      "theoretical_loss": 3.338560319502707,
      "tokens_seen": 2799894528
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015424242424242425,
      "loss": 1.3935,
      "theoretical_loss": 3.3385542267322132,
      "tokens_seen": 2799960064
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015422222222222223,
      "loss": 1.4693,
      "theoretical_loss": 3.338548134144254,
      "tokens_seen": 2800025600
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015420202020202022,
      "loss": 1.3504,
      "theoretical_loss": 3.3385420417388207,
      "tokens_seen": 2800091136
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015418181818181817,
      "loss": 1.5927,
      "theoretical_loss": 3.3385359495159026,
      "tokens_seen": 2800156672
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015416161616161616,
      "loss": 1.3216,
      "theoretical_loss": 3.3385298574754905,
      "tokens_seen": 2800222208
    },
    {
      "epoch": 0.69,
      "objective/train/aligned_token_count": 0.47562316060066223,
      "objective/train/avg_token_score": 0.007328839972615242,
      "objective/train/docs_used": 1577506,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4388517141342163,
      "objective/train/original_loss": 4.313477516174316,
      "objective/train/theoretical_loss": 3.3385237656175746,
      "objective/train/tokens_used": 1159146976,
      "objective/train/unlikelihood_loss": 0.06532740592956543,
      "theoretical_loss": 3.3385237656175746,
      "tokens_seen": 2800287744
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00015414141414141414,
      "loss": 1.2499,
      "theoretical_loss": 3.3385237656175746,
      "tokens_seen": 2800287744
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015412121212121213,
      "loss": 1.221,
      "theoretical_loss": 3.338517673942145,
      "tokens_seen": 2800353280
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001541010101010101,
      "loss": 1.2226,
      "theoretical_loss": 3.3385115824491924,
      "tokens_seen": 2800418816
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015408080808080807,
      "loss": 1.2586,
      "theoretical_loss": 3.338505491138706,
      "tokens_seen": 2800484352
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015406060606060608,
      "loss": 1.2528,
      "theoretical_loss": 3.3384994000106776,
      "tokens_seen": 2800549888
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015404040404040403,
      "loss": 1.2001,
      "theoretical_loss": 3.3384933090650963,
      "tokens_seen": 2800615424
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015402020202020205,
      "loss": 1.338,
      "theoretical_loss": 3.3384872183019527,
      "tokens_seen": 2800680960
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000154,
      "loss": 1.3775,
      "theoretical_loss": 3.338481127721237,
      "tokens_seen": 2800746496
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015397979797979796,
      "loss": 1.5114,
      "theoretical_loss": 3.3384750373229397,
      "tokens_seen": 2800812032
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015395959595959597,
      "loss": 1.3023,
      "theoretical_loss": 3.3384689471070508,
      "tokens_seen": 2800877568
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015393939393939393,
      "loss": 1.2318,
      "theoretical_loss": 3.338462857073561,
      "tokens_seen": 2800943104
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015391919191919194,
      "loss": 1.2859,
      "theoretical_loss": 3.3384567672224597,
      "tokens_seen": 2801008640
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001538989898989899,
      "loss": 1.411,
      "theoretical_loss": 3.3384506775537384,
      "tokens_seen": 2801074176
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001538787878787879,
      "loss": 1.3502,
      "theoretical_loss": 3.3384445880673863,
      "tokens_seen": 2801139712
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015385858585858586,
      "loss": 1.4493,
      "theoretical_loss": 3.338438498763394,
      "tokens_seen": 2801205248
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015383838383838382,
      "loss": 1.4021,
      "theoretical_loss": 3.3384324096417517,
      "tokens_seen": 2801270784
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015381818181818183,
      "loss": 1.3764,
      "theoretical_loss": 3.3384263207024505,
      "tokens_seen": 2801336320
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001537979797979798,
      "loss": 1.4999,
      "theoretical_loss": 3.3384202319454794,
      "tokens_seen": 2801401856
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001537777777777778,
      "loss": 1.554,
      "theoretical_loss": 3.3384141433708296,
      "tokens_seen": 2801467392
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015375757575757576,
      "loss": 1.1394,
      "theoretical_loss": 3.338408054978491,
      "tokens_seen": 2801532928
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015373737373737374,
      "loss": 1.3112,
      "theoretical_loss": 3.338401966768454,
      "tokens_seen": 2801598464
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015371717171717172,
      "loss": 1.3471,
      "theoretical_loss": 3.3383958787407084,
      "tokens_seen": 2801664000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001536969696969697,
      "loss": 1.5583,
      "theoretical_loss": 3.338389790895245,
      "tokens_seen": 2801729536
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001536767676767677,
      "loss": 1.3878,
      "theoretical_loss": 3.3383837032320542,
      "tokens_seen": 2801795072
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015365656565656565,
      "loss": 1.2737,
      "theoretical_loss": 3.338377615751126,
      "tokens_seen": 2801860608
    },
    {
      "epoch": 0.7,
      "objective/train/aligned_token_count": 0.4645039141178131,
      "objective/train/avg_token_score": 0.020974859595298767,
      "objective/train/docs_used": 1578159,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4087051153182983,
      "objective/train/original_loss": 3.74082088470459,
      "objective/train/theoretical_loss": 3.3383715284524507,
      "objective/train/tokens_used": 1160785376,
      "objective/train/unlikelihood_loss": 0.0750071257352829,
      "theoretical_loss": 3.3383715284524507,
      "tokens_seen": 2801926144
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015363636363636363,
      "loss": 1.4973,
      "theoretical_loss": 3.3383715284524507,
      "tokens_seen": 2801926144
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015361616161616162,
      "loss": 1.5222,
      "theoretical_loss": 3.3383654413360184,
      "tokens_seen": 2801991680
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001535959595959596,
      "loss": 1.5553,
      "theoretical_loss": 3.3383593544018195,
      "tokens_seen": 2802057216
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015357575757575758,
      "loss": 1.5415,
      "theoretical_loss": 3.3383532676498446,
      "tokens_seen": 2802122752
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015355555555555557,
      "loss": 1.6572,
      "theoretical_loss": 3.338347181080084,
      "tokens_seen": 2802188288
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015353535353535353,
      "loss": 1.3746,
      "theoretical_loss": 3.3383410946925274,
      "tokens_seen": 2802253824
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015351515151515154,
      "loss": 1.5411,
      "theoretical_loss": 3.3383350084871655,
      "tokens_seen": 2802319360
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001534949494949495,
      "loss": 1.4296,
      "theoretical_loss": 3.3383289224639885,
      "tokens_seen": 2802384896
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015347474747474748,
      "loss": 1.2255,
      "theoretical_loss": 3.3383228366229867,
      "tokens_seen": 2802450432
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015345454545454546,
      "loss": 1.2069,
      "theoretical_loss": 3.3383167509641503,
      "tokens_seen": 2802515968
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015343434343434342,
      "loss": 1.3096,
      "theoretical_loss": 3.3383106654874695,
      "tokens_seen": 2802581504
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015341414141414143,
      "loss": 1.5,
      "theoretical_loss": 3.338304580192935,
      "tokens_seen": 2802647040
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015339393939393939,
      "loss": 1.4053,
      "theoretical_loss": 3.3382984950805366,
      "tokens_seen": 2802712576
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001533737373737374,
      "loss": 1.2934,
      "theoretical_loss": 3.338292410150265,
      "tokens_seen": 2802778112
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015335353535353535,
      "loss": 1.4817,
      "theoretical_loss": 3.3382863254021107,
      "tokens_seen": 2802843648
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015333333333333334,
      "loss": 1.4444,
      "theoretical_loss": 3.338280240836063,
      "tokens_seen": 2802909184
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015331313131313132,
      "loss": 1.2157,
      "theoretical_loss": 3.338274156452113,
      "tokens_seen": 2802974720
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015329292929292928,
      "loss": 1.3833,
      "theoretical_loss": 3.3382680722502505,
      "tokens_seen": 2803040256
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001532727272727273,
      "loss": 1.2857,
      "theoretical_loss": 3.3382619882304665,
      "tokens_seen": 2803105792
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015325252525252525,
      "loss": 1.4648,
      "theoretical_loss": 3.3382559043927507,
      "tokens_seen": 2803171328
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015323232323232323,
      "loss": 1.2413,
      "theoretical_loss": 3.3382498207370936,
      "tokens_seen": 2803236864
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015321212121212121,
      "loss": 1.3599,
      "theoretical_loss": 3.3382437372634857,
      "tokens_seen": 2803302400
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001531919191919192,
      "loss": 1.2674,
      "theoretical_loss": 3.338237653971917,
      "tokens_seen": 2803367936
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015317171717171718,
      "loss": 1.3296,
      "theoretical_loss": 3.3382315708623778,
      "tokens_seen": 2803433472
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015315151515151514,
      "loss": 1.286,
      "theoretical_loss": 3.3382254879348583,
      "tokens_seen": 2803499008
    },
    {
      "epoch": 0.7,
      "objective/train/aligned_token_count": 0.4652370512485504,
      "objective/train/avg_token_score": 0.007319672033190727,
      "objective/train/docs_used": 1579013,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4044634103775024,
      "objective/train/original_loss": 4.015929222106934,
      "objective/train/theoretical_loss": 3.3382194051893492,
      "objective/train/tokens_used": 1162423776,
      "objective/train/unlikelihood_loss": 0.07955323904752731,
      "theoretical_loss": 3.3382194051893492,
      "tokens_seen": 2803564544
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015313131313131312,
      "loss": 1.3776,
      "theoretical_loss": 3.3382194051893492,
      "tokens_seen": 2803564544
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001531111111111111,
      "loss": 1.3868,
      "theoretical_loss": 3.3382133226258404,
      "tokens_seen": 2803630080
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001530909090909091,
      "loss": 1.2563,
      "theoretical_loss": 3.3382072402443224,
      "tokens_seen": 2803695616
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015307070707070708,
      "loss": 1.2552,
      "theoretical_loss": 3.3382011580447855,
      "tokens_seen": 2803761152
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015305050505050506,
      "loss": 1.3729,
      "theoretical_loss": 3.33819507602722,
      "tokens_seen": 2803826688
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015303030303030302,
      "loss": 1.4198,
      "theoretical_loss": 3.338188994191616,
      "tokens_seen": 2803892224
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015301010101010103,
      "loss": 1.1375,
      "theoretical_loss": 3.338182912537964,
      "tokens_seen": 2803957760
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015298989898989898,
      "loss": 1.2954,
      "theoretical_loss": 3.3381768310662543,
      "tokens_seen": 2804023296
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015296969696969697,
      "loss": 1.2648,
      "theoretical_loss": 3.338170749776477,
      "tokens_seen": 2804088832
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015294949494949495,
      "loss": 1.3844,
      "theoretical_loss": 3.338164668668623,
      "tokens_seen": 2804154368
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015292929292929294,
      "loss": 1.1601,
      "theoretical_loss": 3.338158587742682,
      "tokens_seen": 2804219904
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015290909090909092,
      "loss": 1.3522,
      "theoretical_loss": 3.338152506998645,
      "tokens_seen": 2804285440
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015288888888888888,
      "loss": 1.3245,
      "theoretical_loss": 3.338146426436501,
      "tokens_seen": 2804350976
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001528686868686869,
      "loss": 1.1541,
      "theoretical_loss": 3.3381403460562415,
      "tokens_seen": 2804416512
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015284848484848484,
      "loss": 1.2677,
      "theoretical_loss": 3.3381342658578563,
      "tokens_seen": 2804482048
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015282828282828286,
      "loss": 1.5069,
      "theoretical_loss": 3.338128185841336,
      "tokens_seen": 2804547584
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001528080808080808,
      "loss": 1.3448,
      "theoretical_loss": 3.3381221060066704,
      "tokens_seen": 2804613120
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015278787878787877,
      "loss": 1.4393,
      "theoretical_loss": 3.3381160263538505,
      "tokens_seen": 2804678656
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015276767676767678,
      "loss": 1.5491,
      "theoretical_loss": 3.338109946882866,
      "tokens_seen": 2804744192
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015274747474747474,
      "loss": 1.1451,
      "theoretical_loss": 3.338103867593708,
      "tokens_seen": 2804809728
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015272727272727275,
      "loss": 1.1888,
      "theoretical_loss": 3.338097788486366,
      "tokens_seen": 2804875264
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001527070707070707,
      "loss": 1.4121,
      "theoretical_loss": 3.33809170956083,
      "tokens_seen": 2804940800
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001526868686868687,
      "loss": 1.4133,
      "theoretical_loss": 3.3380856308170914,
      "tokens_seen": 2805006336
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015266666666666667,
      "loss": 1.3764,
      "theoretical_loss": 3.3380795522551403,
      "tokens_seen": 2805071872
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015264646464646466,
      "loss": 1.1343,
      "theoretical_loss": 3.3380734738749664,
      "tokens_seen": 2805137408
    },
    {
      "epoch": 0.7,
      "objective/train/aligned_token_count": 0.42247068881988525,
      "objective/train/avg_token_score": 0.025837045162916183,
      "objective/train/docs_used": 1579013,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.424133062362671,
      "objective/train/original_loss": 4.378701210021973,
      "objective/train/theoretical_loss": 3.3380673956765605,
      "objective/train/tokens_used": 1164062176,
      "objective/train/unlikelihood_loss": 0.050871342420578,
      "theoretical_loss": 3.3380673956765605,
      "tokens_seen": 2805202944
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015262626262626264,
      "loss": 1.3393,
      "theoretical_loss": 3.3380673956765605,
      "tokens_seen": 2805202944
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001526060606060606,
      "loss": 1.1949,
      "theoretical_loss": 3.338061317659913,
      "tokens_seen": 2805268480
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015258585858585858,
      "loss": 1.0152,
      "theoretical_loss": 3.338055239825014,
      "tokens_seen": 2805334016
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015256565656565657,
      "loss": 1.3767,
      "theoretical_loss": 3.3380491621718535,
      "tokens_seen": 2805399552
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015254545454545455,
      "loss": 1.3118,
      "theoretical_loss": 3.3380430847004225,
      "tokens_seen": 2805465088
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015252525252525253,
      "loss": 1.3854,
      "theoretical_loss": 3.3380370074107106,
      "tokens_seen": 2805530624
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015250505050505052,
      "loss": 1.1654,
      "theoretical_loss": 3.3380309303027085,
      "tokens_seen": 2805596160
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015248484848484847,
      "loss": 1.0913,
      "theoretical_loss": 3.3380248533764068,
      "tokens_seen": 2805661696
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015246464646464649,
      "loss": 1.0666,
      "theoretical_loss": 3.3380187766317957,
      "tokens_seen": 2805727232
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015244444444444444,
      "loss": 1.3145,
      "theoretical_loss": 3.338012700068865,
      "tokens_seen": 2805792768
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015242424242424243,
      "loss": 1.5342,
      "theoretical_loss": 3.3380066236876056,
      "tokens_seen": 2805858304
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001524040404040404,
      "loss": 1.4776,
      "theoretical_loss": 3.3380005474880075,
      "tokens_seen": 2805923840
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015238383838383837,
      "loss": 1.6249,
      "theoretical_loss": 3.337994471470061,
      "tokens_seen": 2805989376
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015236363636363638,
      "loss": 1.2415,
      "theoretical_loss": 3.3379883956337566,
      "tokens_seen": 2806054912
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015234343434343434,
      "loss": 1.5285,
      "theoretical_loss": 3.337982319979085,
      "tokens_seen": 2806120448
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015232323232323235,
      "loss": 1.2752,
      "theoretical_loss": 3.3379762445060353,
      "tokens_seen": 2806185984
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001523030303030303,
      "loss": 1.2645,
      "theoretical_loss": 3.3379701692145995,
      "tokens_seen": 2806251520
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015228282828282826,
      "loss": 1.3152,
      "theoretical_loss": 3.3379640941047666,
      "tokens_seen": 2806317056
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015226262626262627,
      "loss": 1.2928,
      "theoretical_loss": 3.3379580191765275,
      "tokens_seen": 2806382592
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015224242424242423,
      "loss": 1.4794,
      "theoretical_loss": 3.3379519444298724,
      "tokens_seen": 2806448128
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015222222222222224,
      "loss": 1.3131,
      "theoretical_loss": 3.337945869864792,
      "tokens_seen": 2806513664
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001522020202020202,
      "loss": 1.1301,
      "theoretical_loss": 3.337939795481276,
      "tokens_seen": 2806579200
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015218181818181818,
      "loss": 1.3662,
      "theoretical_loss": 3.337933721279315,
      "tokens_seen": 2806644736
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015216161616161616,
      "loss": 1.2308,
      "theoretical_loss": 3.3379276472589,
      "tokens_seen": 2806710272
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015214141414141415,
      "loss": 1.2512,
      "theoretical_loss": 3.33792157342002,
      "tokens_seen": 2806775808
    },
    {
      "epoch": 0.7,
      "objective/train/aligned_token_count": 0.36895161867141724,
      "objective/train/avg_token_score": 0.021854940801858902,
      "objective/train/docs_used": 1579588,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2226519584655762,
      "objective/train/original_loss": 4.2983927726745605,
      "objective/train/theoretical_loss": 3.337915499762666,
      "objective/train/tokens_used": 1165700576,
      "objective/train/unlikelihood_loss": 0.04864027351140976,
      "theoretical_loss": 3.337915499762666,
      "tokens_seen": 2806841344
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015212121212121213,
      "loss": 1.2497,
      "theoretical_loss": 3.337915499762666,
      "tokens_seen": 2806841344
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001521010101010101,
      "loss": 1.314,
      "theoretical_loss": 3.3379094262868287,
      "tokens_seen": 2806906880
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001520808080808081,
      "loss": 1.211,
      "theoretical_loss": 3.337903352992498,
      "tokens_seen": 2806972416
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015206060606060606,
      "loss": 1.3664,
      "theoretical_loss": 3.3378972798796647,
      "tokens_seen": 2807037952
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015204040404040404,
      "loss": 1.3346,
      "theoretical_loss": 3.3378912069483184,
      "tokens_seen": 2807103488
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015202020202020202,
      "loss": 1.281,
      "theoretical_loss": 3.33788513419845,
      "tokens_seen": 2807169024
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000152,
      "loss": 1.4048,
      "theoretical_loss": 3.3378790616300495,
      "tokens_seen": 2807234560
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000151979797979798,
      "loss": 1.4233,
      "theoretical_loss": 3.337872989243108,
      "tokens_seen": 2807300096
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015195959595959598,
      "loss": 1.3275,
      "theoretical_loss": 3.3378669170376147,
      "tokens_seen": 2807365632
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015193939393939393,
      "loss": 1.6049,
      "theoretical_loss": 3.3378608450135605,
      "tokens_seen": 2807431168
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015191919191919192,
      "loss": 1.3375,
      "theoretical_loss": 3.3378547731709363,
      "tokens_seen": 2807496704
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001518989898989899,
      "loss": 1.0499,
      "theoretical_loss": 3.3378487015097313,
      "tokens_seen": 2807562240
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015187878787878789,
      "loss": 1.1327,
      "theoretical_loss": 3.3378426300299364,
      "tokens_seen": 2807627776
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015185858585858587,
      "loss": 1.1514,
      "theoretical_loss": 3.337836558731542,
      "tokens_seen": 2807693312
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015183838383838383,
      "loss": 1.3891,
      "theoretical_loss": 3.337830487614539,
      "tokens_seen": 2807758848
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015181818181818184,
      "loss": 1.1442,
      "theoretical_loss": 3.3378244166789166,
      "tokens_seen": 2807824384
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001517979797979798,
      "loss": 1.3138,
      "theoretical_loss": 3.337818345924666,
      "tokens_seen": 2807889920
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001517777777777778,
      "loss": 1.093,
      "theoretical_loss": 3.3378122753517774,
      "tokens_seen": 2807955456
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015175757575757576,
      "loss": 1.1588,
      "theoretical_loss": 3.337806204960241,
      "tokens_seen": 2808020992
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015173737373737372,
      "loss": 1.1076,
      "theoretical_loss": 3.3378001347500468,
      "tokens_seen": 2808086528
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015171717171717173,
      "loss": 1.0627,
      "theoretical_loss": 3.3377940647211854,
      "tokens_seen": 2808152064
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001516969696969697,
      "loss": 1.3375,
      "theoretical_loss": 3.337787994873648,
      "tokens_seen": 2808217600
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001516767676767677,
      "loss": 1.2909,
      "theoretical_loss": 3.3377819252074237,
      "tokens_seen": 2808283136
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015165656565656565,
      "loss": 1.1421,
      "theoretical_loss": 3.3377758557225032,
      "tokens_seen": 2808348672
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015163636363636364,
      "loss": 1.2631,
      "theoretical_loss": 3.3377697864188773,
      "tokens_seen": 2808414208
    },
    {
      "debugging/Self-BLEU-5": 0.20303494455588592,
      "debugging/distinct-1-grams": 0.7960725957049486,
      "debugging/distinct-2-grams": 0.9561016342598115,
      "debugging/entropy-1-grams": 5.007351944943723,
      "debugging/entropy-2-grams": 5.529170180274562,
      "debugging/length": 479.0,
      "debugging/num_segments": 4,
      "epoch": 0.7,
      "objective/train/aligned_token_count": 0.5130742788314819,
      "objective/train/avg_token_score": 0.009924755431711674,
      "objective/train/docs_used": 1580206,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5729913711547852,
      "objective/train/original_loss": 4.162881374359131,
      "objective/train/theoretical_loss": 3.337763717296536,
      "objective/train/tokens_used": 1167338976,
      "objective/train/unlikelihood_loss": 0.058410946279764175,
      "theoretical_loss": 3.337763717296536,
      "tokens_seen": 2808479744
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015161616161616162,
      "loss": 1.3099,
      "theoretical_loss": 3.337763717296536,
      "tokens_seen": 2808479744
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001515959595959596,
      "loss": 1.3338,
      "theoretical_loss": 3.33775764835547,
      "tokens_seen": 2808545280
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001515757575757576,
      "loss": 1.2946,
      "theoretical_loss": 3.337751579595669,
      "tokens_seen": 2808610816
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015155555555555555,
      "loss": 1.1682,
      "theoretical_loss": 3.3377455110171246,
      "tokens_seen": 2808676352
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015153535353535353,
      "loss": 1.353,
      "theoretical_loss": 3.3377394426198252,
      "tokens_seen": 2808741888
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015151515151515152,
      "loss": 1.2111,
      "theoretical_loss": 3.337733374403763,
      "tokens_seen": 2808807424
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001514949494949495,
      "loss": 1.1083,
      "theoretical_loss": 3.3377273063689277,
      "tokens_seen": 2808872960
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015147474747474748,
      "loss": 1.2181,
      "theoretical_loss": 3.337721238515309,
      "tokens_seen": 2808938496
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015145454545454547,
      "loss": 1.398,
      "theoretical_loss": 3.3377151708428983,
      "tokens_seen": 2809004032
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015143434343434342,
      "loss": 1.3012,
      "theoretical_loss": 3.337709103351685,
      "tokens_seen": 2809069568
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001514141414141414,
      "loss": 1.3897,
      "theoretical_loss": 3.33770303604166,
      "tokens_seen": 2809135104
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001513939393939394,
      "loss": 1.0826,
      "theoretical_loss": 3.337696968912814,
      "tokens_seen": 2809200640
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015137373737373738,
      "loss": 1.2361,
      "theoretical_loss": 3.337690901965137,
      "tokens_seen": 2809266176
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015135353535353536,
      "loss": 1.2741,
      "theoretical_loss": 3.3376848351986195,
      "tokens_seen": 2809331712
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015133333333333332,
      "loss": 1.4364,
      "theoretical_loss": 3.3376787686132516,
      "tokens_seen": 2809397248
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015131313131313133,
      "loss": 1.2624,
      "theoretical_loss": 3.337672702209024,
      "tokens_seen": 2809462784
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015129292929292928,
      "loss": 1.3916,
      "theoretical_loss": 3.337666635985926,
      "tokens_seen": 2809528320
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001512727272727273,
      "loss": 1.243,
      "theoretical_loss": 3.3376605699439494,
      "tokens_seen": 2809593856
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015125252525252525,
      "loss": 1.1352,
      "theoretical_loss": 3.3376545040830843,
      "tokens_seen": 2809659392
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015123232323232324,
      "loss": 1.2055,
      "theoretical_loss": 3.3376484384033205,
      "tokens_seen": 2809724928
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015121212121212122,
      "loss": 1.3266,
      "theoretical_loss": 3.3376423729046483,
      "tokens_seen": 2809790464
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015119191919191918,
      "loss": 1.2254,
      "theoretical_loss": 3.3376363075870588,
      "tokens_seen": 2809856000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001511717171717172,
      "loss": 1.4516,
      "theoretical_loss": 3.3376302424505417,
      "tokens_seen": 2809921536
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015115151515151515,
      "loss": 1.2488,
      "theoretical_loss": 3.337624177495088,
      "tokens_seen": 2809987072
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015113131313131316,
      "loss": 1.2695,
      "theoretical_loss": 3.337618112720687,
      "tokens_seen": 2810052608
    },
    {
      "epoch": 0.7,
      "objective/train/aligned_token_count": 0.40835776925086975,
      "objective/train/avg_token_score": 0.012725447304546833,
      "objective/train/docs_used": 1581474,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2710609436035156,
      "objective/train/original_loss": 4.22544527053833,
      "objective/train/theoretical_loss": 3.3376120481273306,
      "objective/train/tokens_used": 1168977376,
      "objective/train/unlikelihood_loss": 0.05969413369894028,
      "theoretical_loss": 3.3376120481273306,
      "tokens_seen": 2810118144
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001511111111111111,
      "loss": 1.1924,
      "theoretical_loss": 3.3376120481273306,
      "tokens_seen": 2810118144
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001510909090909091,
      "loss": 1.1586,
      "theoretical_loss": 3.337605983715008,
      "tokens_seen": 2810183680
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015107070707070708,
      "loss": 1.3458,
      "theoretical_loss": 3.3375999194837096,
      "tokens_seen": 2810249216
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015105050505050504,
      "loss": 1.3431,
      "theoretical_loss": 3.337593855433427,
      "tokens_seen": 2810314752
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015103030303030305,
      "loss": 1.5137,
      "theoretical_loss": 3.337587791564149,
      "tokens_seen": 2810380288
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000151010101010101,
      "loss": 1.2253,
      "theoretical_loss": 3.3375817278758664,
      "tokens_seen": 2810445824
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000150989898989899,
      "loss": 1.3754,
      "theoretical_loss": 3.3375756643685706,
      "tokens_seen": 2810511360
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015096969696969697,
      "loss": 1.2829,
      "theoretical_loss": 3.337569601042251,
      "tokens_seen": 2810576896
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015094949494949496,
      "loss": 1.3458,
      "theoretical_loss": 3.3375635378968975,
      "tokens_seen": 2810642432
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015092929292929294,
      "loss": 1.4721,
      "theoretical_loss": 3.3375574749325017,
      "tokens_seen": 2810707968
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015090909090909093,
      "loss": 1.5239,
      "theoretical_loss": 3.3375514121490535,
      "tokens_seen": 2810773504
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015088888888888888,
      "loss": 1.5242,
      "theoretical_loss": 3.3375453495465432,
      "tokens_seen": 2810839040
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015086868686868687,
      "loss": 0.9798,
      "theoretical_loss": 3.3375392871249616,
      "tokens_seen": 2810904576
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015084848484848485,
      "loss": 1.3335,
      "theoretical_loss": 3.337533224884298,
      "tokens_seen": 2810970112
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015082828282828283,
      "loss": 1.3496,
      "theoretical_loss": 3.3375271628245438,
      "tokens_seen": 2811035648
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015080808080808082,
      "loss": 1.5589,
      "theoretical_loss": 3.337521100945689,
      "tokens_seen": 2811101184
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015078787878787877,
      "loss": 1.2884,
      "theoretical_loss": 3.337515039247724,
      "tokens_seen": 2811166720
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015076767676767679,
      "loss": 1.2521,
      "theoretical_loss": 3.3375089777306397,
      "tokens_seen": 2811232256
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015074747474747474,
      "loss": 1.5707,
      "theoretical_loss": 3.3375029163944254,
      "tokens_seen": 2811297792
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015072727272727275,
      "loss": 1.2869,
      "theoretical_loss": 3.3374968552390722,
      "tokens_seen": 2811363328
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001507070707070707,
      "loss": 1.3295,
      "theoretical_loss": 3.337490794264571,
      "tokens_seen": 2811428864
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015068686868686867,
      "loss": 1.3528,
      "theoretical_loss": 3.337484733470911,
      "tokens_seen": 2811494400
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015066666666666668,
      "loss": 1.3138,
      "theoretical_loss": 3.337478672858083,
      "tokens_seen": 2811559936
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015064646464646464,
      "loss": 1.4834,
      "theoretical_loss": 3.337472612426078,
      "tokens_seen": 2811625472
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015062626262626265,
      "loss": 1.2684,
      "theoretical_loss": 3.337466552174886,
      "tokens_seen": 2811691008
    },
    {
      "epoch": 0.7,
      "objective/train/aligned_token_count": 0.40163734555244446,
      "objective/train/avg_token_score": 0.016715794801712036,
      "objective/train/docs_used": 1582095,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1810110807418823,
      "objective/train/original_loss": 4.419708728790283,
      "objective/train/theoretical_loss": 3.337460492104497,
      "objective/train/tokens_used": 1170615776,
      "objective/train/unlikelihood_loss": 0.05445218086242676,
      "theoretical_loss": 3.337460492104497,
      "tokens_seen": 2811756544
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001506060606060606,
      "loss": 1.2961,
      "theoretical_loss": 3.337460492104497,
      "tokens_seen": 2811756544
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001505858585858586,
      "loss": 1.2588,
      "theoretical_loss": 3.337454432214902,
      "tokens_seen": 2811822080
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015056565656565657,
      "loss": 1.1438,
      "theoretical_loss": 3.337448372506091,
      "tokens_seen": 2811887616
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015054545454545453,
      "loss": 1.2045,
      "theoretical_loss": 3.3374423129780544,
      "tokens_seen": 2811953152
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015052525252525254,
      "loss": 1.212,
      "theoretical_loss": 3.337436253630783,
      "tokens_seen": 2812018688
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001505050505050505,
      "loss": 1.4709,
      "theoretical_loss": 3.337430194464267,
      "tokens_seen": 2812084224
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015048484848484848,
      "loss": 1.1654,
      "theoretical_loss": 3.3374241354784964,
      "tokens_seen": 2812149760
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015046464646464646,
      "loss": 1.3118,
      "theoretical_loss": 3.337418076673462,
      "tokens_seen": 2812215296
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015044444444444445,
      "loss": 1.1859,
      "theoretical_loss": 3.3374120180491538,
      "tokens_seen": 2812280832
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015042424242424243,
      "loss": 1.2973,
      "theoretical_loss": 3.3374059596055625,
      "tokens_seen": 2812346368
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015040404040404042,
      "loss": 1.5082,
      "theoretical_loss": 3.337399901342679,
      "tokens_seen": 2812411904
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001503838383838384,
      "loss": 1.3918,
      "theoretical_loss": 3.337393843260493,
      "tokens_seen": 2812477440
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015036363636363636,
      "loss": 1.3717,
      "theoretical_loss": 3.3373877853589953,
      "tokens_seen": 2812542976
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015034343434343434,
      "loss": 1.1706,
      "theoretical_loss": 3.3373817276381756,
      "tokens_seen": 2812608512
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015032323232323232,
      "loss": 1.1978,
      "theoretical_loss": 3.337375670098025,
      "tokens_seen": 2812674048
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001503030303030303,
      "loss": 1.3281,
      "theoretical_loss": 3.337369612738534,
      "tokens_seen": 2812739584
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001502828282828283,
      "loss": 1.3465,
      "theoretical_loss": 3.3373635555596923,
      "tokens_seen": 2812805120
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015026262626262628,
      "loss": 1.2663,
      "theoretical_loss": 3.3373574985614907,
      "tokens_seen": 2812870656
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015024242424242423,
      "loss": 1.4005,
      "theoretical_loss": 3.3373514417439196,
      "tokens_seen": 2812936192
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015022222222222224,
      "loss": 1.5569,
      "theoretical_loss": 3.33734538510697,
      "tokens_seen": 2813001728
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001502020202020202,
      "loss": 1.1421,
      "theoretical_loss": 3.337339328650631,
      "tokens_seen": 2813067264
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015018181818181819,
      "loss": 1.4376,
      "theoretical_loss": 3.3373332723748943,
      "tokens_seen": 2813132800
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015016161616161617,
      "loss": 1.229,
      "theoretical_loss": 3.3373272162797494,
      "tokens_seen": 2813198336
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015014141414141413,
      "loss": 1.1666,
      "theoretical_loss": 3.337321160365187,
      "tokens_seen": 2813263872
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015012121212121214,
      "loss": 1.2655,
      "theoretical_loss": 3.3373151046311977,
      "tokens_seen": 2813329408
    },
    {
      "epoch": 0.7,
      "objective/train/aligned_token_count": 0.4013318717479706,
      "objective/train/avg_token_score": 0.019104115664958954,
      "objective/train/docs_used": 1583341,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.16046142578125,
      "objective/train/original_loss": 4.79135799407959,
      "objective/train/theoretical_loss": 3.3373090490777715,
      "objective/train/tokens_used": 1172254176,
      "objective/train/unlikelihood_loss": 0.03227890282869339,
      "theoretical_loss": 3.3373090490777715,
      "tokens_seen": 2813394944
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001501010101010101,
      "loss": 1.4355,
      "theoretical_loss": 3.3373090490777715,
      "tokens_seen": 2813394944
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001500808080808081,
      "loss": 1.2089,
      "theoretical_loss": 3.3373029937048995,
      "tokens_seen": 2813460480
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015006060606060606,
      "loss": 1.1539,
      "theoretical_loss": 3.3372969385125715,
      "tokens_seen": 2813526016
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015004040404040405,
      "loss": 1.2506,
      "theoretical_loss": 3.3372908835007777,
      "tokens_seen": 2813591552
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015002020202020203,
      "loss": 1.3608,
      "theoretical_loss": 3.3372848286695094,
      "tokens_seen": 2813657088
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00015,
      "loss": 1.1247,
      "theoretical_loss": 3.337278774018756,
      "tokens_seen": 2813722624
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.000149979797979798,
      "loss": 1.3458,
      "theoretical_loss": 3.3372727195485092,
      "tokens_seen": 2813788160
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014995959595959595,
      "loss": 1.5431,
      "theoretical_loss": 3.3372666652587584,
      "tokens_seen": 2813853696
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014993939393939394,
      "loss": 1.3937,
      "theoretical_loss": 3.3372606111494942,
      "tokens_seen": 2813919232
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014991919191919192,
      "loss": 0.9856,
      "theoretical_loss": 3.337254557220707,
      "tokens_seen": 2813984768
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001498989898989899,
      "loss": 1.5882,
      "theoretical_loss": 3.337248503472387,
      "tokens_seen": 2814050304
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001498787878787879,
      "loss": 1.4216,
      "theoretical_loss": 3.3372424499045255,
      "tokens_seen": 2814115840
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014985858585858585,
      "loss": 1.0373,
      "theoretical_loss": 3.337236396517112,
      "tokens_seen": 2814181376
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014983838383838383,
      "loss": 1.5091,
      "theoretical_loss": 3.337230343310137,
      "tokens_seen": 2814246912
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014981818181818182,
      "loss": 1.3123,
      "theoretical_loss": 3.337224290283592,
      "tokens_seen": 2814312448
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001497979797979798,
      "loss": 1.3197,
      "theoretical_loss": 3.3372182374374657,
      "tokens_seen": 2814377984
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014977777777777778,
      "loss": 1.3569,
      "theoretical_loss": 3.33721218477175,
      "tokens_seen": 2814443520
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014975757575757577,
      "loss": 1.3055,
      "theoretical_loss": 3.3372061322864344,
      "tokens_seen": 2814509056
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014973737373737372,
      "loss": 1.2218,
      "theoretical_loss": 3.3372000799815096,
      "tokens_seen": 2814574592
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014971717171717174,
      "loss": 1.4153,
      "theoretical_loss": 3.3371940278569663,
      "tokens_seen": 2814640128
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001496969696969697,
      "loss": 1.4213,
      "theoretical_loss": 3.3371879759127947,
      "tokens_seen": 2814705664
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014967676767676768,
      "loss": 1.363,
      "theoretical_loss": 3.337181924148985,
      "tokens_seen": 2814771200
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014965656565656566,
      "loss": 1.2395,
      "theoretical_loss": 3.3371758725655285,
      "tokens_seen": 2814836736
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014963636363636362,
      "loss": 1.3557,
      "theoretical_loss": 3.3371698211624143,
      "tokens_seen": 2814902272
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014961616161616163,
      "loss": 1.3444,
      "theoretical_loss": 3.3371637699396337,
      "tokens_seen": 2814967808
    },
    {
      "epoch": 0.7,
      "objective/train/aligned_token_count": 0.38996821641921997,
      "objective/train/avg_token_score": 0.0225471593439579,
      "objective/train/docs_used": 1583792,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2803069353103638,
      "objective/train/original_loss": 4.408719062805176,
      "objective/train/theoretical_loss": 3.337157718897177,
      "objective/train/tokens_used": 1173892576,
      "objective/train/unlikelihood_loss": 0.06987224519252777,
      "theoretical_loss": 3.337157718897177,
      "tokens_seen": 2815033344
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014959595959595958,
      "loss": 1.105,
      "theoretical_loss": 3.337157718897177,
      "tokens_seen": 2815033344
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001495757575757576,
      "loss": 1.25,
      "theoretical_loss": 3.3371516680350344,
      "tokens_seen": 2815098880
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014955555555555555,
      "loss": 1.2268,
      "theoretical_loss": 3.3371456173531966,
      "tokens_seen": 2815164416
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014953535353535356,
      "loss": 1.5277,
      "theoretical_loss": 3.337139566851654,
      "tokens_seen": 2815229952
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014951515151515152,
      "loss": 1.277,
      "theoretical_loss": 3.337133516530397,
      "tokens_seen": 2815295488
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014949494949494948,
      "loss": 1.1678,
      "theoretical_loss": 3.3371274663894157,
      "tokens_seen": 2815361024
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001494747474747475,
      "loss": 1.3176,
      "theoretical_loss": 3.337121416428701,
      "tokens_seen": 2815426560
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014945454545454545,
      "loss": 1.4365,
      "theoretical_loss": 3.337115366648243,
      "tokens_seen": 2815492096
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014943434343434346,
      "loss": 1.4581,
      "theoretical_loss": 3.3371093170480326,
      "tokens_seen": 2815557632
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001494141414141414,
      "loss": 1.4485,
      "theoretical_loss": 3.33710326762806,
      "tokens_seen": 2815623168
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001493939393939394,
      "loss": 1.3542,
      "theoretical_loss": 3.337097218388315,
      "tokens_seen": 2815688704
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014937373737373738,
      "loss": 1.4033,
      "theoretical_loss": 3.3370911693287892,
      "tokens_seen": 2815754240
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014935353535353537,
      "loss": 1.3123,
      "theoretical_loss": 3.337085120449472,
      "tokens_seen": 2815819776
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014933333333333335,
      "loss": 1.2209,
      "theoretical_loss": 3.3370790717503542,
      "tokens_seen": 2815885312
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001493131313131313,
      "loss": 1.418,
      "theoretical_loss": 3.3370730232314267,
      "tokens_seen": 2815950848
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001492929292929293,
      "loss": 1.3754,
      "theoretical_loss": 3.3370669748926796,
      "tokens_seen": 2816016384
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014927272727272727,
      "loss": 1.3505,
      "theoretical_loss": 3.3370609267341025,
      "tokens_seen": 2816081920
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014925252525252526,
      "loss": 1.3739,
      "theoretical_loss": 3.3370548787556875,
      "tokens_seen": 2816147456
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014923232323232324,
      "loss": 1.3489,
      "theoretical_loss": 3.3370488309574236,
      "tokens_seen": 2816212992
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014921212121212123,
      "loss": 1.4489,
      "theoretical_loss": 3.337042783339302,
      "tokens_seen": 2816278528
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014919191919191918,
      "loss": 1.3362,
      "theoretical_loss": 3.337036735901313,
      "tokens_seen": 2816344064
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001491717171717172,
      "loss": 1.1407,
      "theoretical_loss": 3.337030688643447,
      "tokens_seen": 2816409600
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014915151515151515,
      "loss": 1.4749,
      "theoretical_loss": 3.3370246415656943,
      "tokens_seen": 2816475136
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014913131313131313,
      "loss": 1.3517,
      "theoretical_loss": 3.3370185946680455,
      "tokens_seen": 2816540672
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014911111111111112,
      "loss": 1.2781,
      "theoretical_loss": 3.3370125479504913,
      "tokens_seen": 2816606208
    },
    {
      "epoch": 0.7,
      "objective/train/aligned_token_count": 0.4633431136608124,
      "objective/train/avg_token_score": 0.017624158412218094,
      "objective/train/docs_used": 1585248,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4681633710861206,
      "objective/train/original_loss": 4.188065052032471,
      "objective/train/theoretical_loss": 3.3370065014130215,
      "objective/train/tokens_used": 1175530976,
      "objective/train/unlikelihood_loss": 0.06412944197654724,
      "theoretical_loss": 3.3370065014130215,
      "tokens_seen": 2816671744
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00014909090909090908,
      "loss": 1.3567,
      "theoretical_loss": 3.3370065014130215,
      "tokens_seen": 2816671744
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001490707070707071,
      "loss": 1.4902,
      "theoretical_loss": 3.337000455055627,
      "tokens_seen": 2816737280
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014905050505050504,
      "loss": 1.3484,
      "theoretical_loss": 3.336994408878298,
      "tokens_seen": 2816802816
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014903030303030305,
      "loss": 1.0793,
      "theoretical_loss": 3.3369883628810255,
      "tokens_seen": 2816868352
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000149010101010101,
      "loss": 1.4975,
      "theoretical_loss": 3.3369823170637996,
      "tokens_seen": 2816933888
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014898989898989897,
      "loss": 1.4961,
      "theoretical_loss": 3.3369762714266105,
      "tokens_seen": 2816999424
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014896969696969698,
      "loss": 1.1294,
      "theoretical_loss": 3.336970225969449,
      "tokens_seen": 2817064960
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014894949494949494,
      "loss": 1.0139,
      "theoretical_loss": 3.336964180692305,
      "tokens_seen": 2817130496
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014892929292929295,
      "loss": 1.6351,
      "theoretical_loss": 3.33695813559517,
      "tokens_seen": 2817196032
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001489090909090909,
      "loss": 1.3377,
      "theoretical_loss": 3.3369520906780332,
      "tokens_seen": 2817261568
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001488888888888889,
      "loss": 1.3356,
      "theoretical_loss": 3.336946045940886,
      "tokens_seen": 2817327104
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014886868686868687,
      "loss": 1.3046,
      "theoretical_loss": 3.3369400013837183,
      "tokens_seen": 2817392640
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014884848484848486,
      "loss": 1.4818,
      "theoretical_loss": 3.3369339570065213,
      "tokens_seen": 2817458176
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014882828282828284,
      "loss": 1.4147,
      "theoretical_loss": 3.3369279128092844,
      "tokens_seen": 2817523712
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001488080808080808,
      "loss": 1.3461,
      "theoretical_loss": 3.3369218687919986,
      "tokens_seen": 2817589248
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014878787878787878,
      "loss": 1.2945,
      "theoretical_loss": 3.3369158249546547,
      "tokens_seen": 2817654784
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014876767676767676,
      "loss": 1.3615,
      "theoretical_loss": 3.3369097812972424,
      "tokens_seen": 2817720320
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014874747474747475,
      "loss": 1.3776,
      "theoretical_loss": 3.336903737819753,
      "tokens_seen": 2817785856
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014872727272727273,
      "loss": 1.4626,
      "theoretical_loss": 3.336897694522176,
      "tokens_seen": 2817851392
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014870707070707072,
      "loss": 1.249,
      "theoretical_loss": 3.336891651404503,
      "tokens_seen": 2817916928
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001486868686868687,
      "loss": 1.5321,
      "theoretical_loss": 3.3368856084667233,
      "tokens_seen": 2817982464
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014866666666666668,
      "loss": 1.1701,
      "theoretical_loss": 3.3368795657088284,
      "tokens_seen": 2818048000
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014864646464646464,
      "loss": 1.2661,
      "theoretical_loss": 3.3368735231308078,
      "tokens_seen": 2818113536
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014862626262626263,
      "loss": 1.3201,
      "theoretical_loss": 3.3368674807326526,
      "tokens_seen": 2818179072
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001486060606060606,
      "loss": 1.2812,
      "theoretical_loss": 3.336861438514353,
      "tokens_seen": 2818244608
    },
    {
      "epoch": 0.71,
      "objective/train/aligned_token_count": 0.47385141253471375,
      "objective/train/avg_token_score": 0.03910459578037262,
      "objective/train/docs_used": 1585920,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4761865139007568,
      "objective/train/original_loss": 4.216108322143555,
      "objective/train/theoretical_loss": 3.3368553964759,
      "objective/train/tokens_used": 1177169376,
      "objective/train/unlikelihood_loss": 0.08022919297218323,
      "theoretical_loss": 3.3368553964759,
      "tokens_seen": 2818310144
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001485858585858586,
      "loss": 1.408,
      "theoretical_loss": 3.3368553964759,
      "tokens_seen": 2818310144
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014856565656565658,
      "loss": 1.3001,
      "theoretical_loss": 3.336849354617283,
      "tokens_seen": 2818375680
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014854545454545453,
      "loss": 1.2528,
      "theoretical_loss": 3.336843312938494,
      "tokens_seen": 2818441216
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014852525252525254,
      "loss": 1.3112,
      "theoretical_loss": 3.3368372714395216,
      "tokens_seen": 2818506752
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001485050505050505,
      "loss": 1.1277,
      "theoretical_loss": 3.3368312301203575,
      "tokens_seen": 2818572288
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001484848484848485,
      "loss": 1.4151,
      "theoretical_loss": 3.336825188980992,
      "tokens_seen": 2818637824
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014846464646464647,
      "loss": 1.2508,
      "theoretical_loss": 3.3368191480214153,
      "tokens_seen": 2818703360
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014844444444444443,
      "loss": 1.2202,
      "theoretical_loss": 3.3368131072416185,
      "tokens_seen": 2818768896
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014842424242424244,
      "loss": 1.356,
      "theoretical_loss": 3.336807066641591,
      "tokens_seen": 2818834432
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001484040404040404,
      "loss": 1.463,
      "theoretical_loss": 3.336801026221324,
      "tokens_seen": 2818899968
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001483838383838384,
      "loss": 1.1669,
      "theoretical_loss": 3.3367949859808084,
      "tokens_seen": 2818965504
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014836363636363636,
      "loss": 1.3556,
      "theoretical_loss": 3.3367889459200333,
      "tokens_seen": 2819031040
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014834343434343435,
      "loss": 1.478,
      "theoretical_loss": 3.3367829060389904,
      "tokens_seen": 2819096576
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014832323232323233,
      "loss": 1.3953,
      "theoretical_loss": 3.3367768663376696,
      "tokens_seen": 2819162112
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001483030303030303,
      "loss": 1.1896,
      "theoretical_loss": 3.336770826816062,
      "tokens_seen": 2819227648
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001482828282828283,
      "loss": 1.3177,
      "theoretical_loss": 3.336764787474157,
      "tokens_seen": 2819293184
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014826262626262626,
      "loss": 1.3633,
      "theoretical_loss": 3.3367587483119463,
      "tokens_seen": 2819358720
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014824242424242424,
      "loss": 1.4552,
      "theoretical_loss": 3.336752709329419,
      "tokens_seen": 2819424256
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014822222222222222,
      "loss": 1.428,
      "theoretical_loss": 3.336746670526567,
      "tokens_seen": 2819489792
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001482020202020202,
      "loss": 1.1741,
      "theoretical_loss": 3.3367406319033797,
      "tokens_seen": 2819555328
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001481818181818182,
      "loss": 1.3877,
      "theoretical_loss": 3.3367345934598482,
      "tokens_seen": 2819620864
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014816161616161617,
      "loss": 1.4327,
      "theoretical_loss": 3.336728555195963,
      "tokens_seen": 2819686400
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014814141414141413,
      "loss": 1.3169,
      "theoretical_loss": 3.336722517111714,
      "tokens_seen": 2819751936
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014812121212121212,
      "loss": 1.309,
      "theoretical_loss": 3.336716479207092,
      "tokens_seen": 2819817472
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001481010101010101,
      "loss": 1.3214,
      "theoretical_loss": 3.3367104414820874,
      "tokens_seen": 2819883008
    },
    {
      "epoch": 0.71,
      "objective/train/aligned_token_count": 0.43420088291168213,
      "objective/train/avg_token_score": 0.009425443597137928,
      "objective/train/docs_used": 1586975,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2513507604599,
      "objective/train/original_loss": 3.9455976486206055,
      "objective/train/theoretical_loss": 3.3367044039366913,
      "objective/train/tokens_used": 1178807776,
      "objective/train/unlikelihood_loss": 0.05780460312962532,
      "theoretical_loss": 3.3367044039366913,
      "tokens_seen": 2819948544
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014808080808080808,
      "loss": 1.4131,
      "theoretical_loss": 3.3367044039366913,
      "tokens_seen": 2819948544
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014806060606060607,
      "loss": 1.0645,
      "theoretical_loss": 3.3366983665708934,
      "tokens_seen": 2820014080
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014804040404040402,
      "loss": 1.227,
      "theoretical_loss": 3.3366923293846846,
      "tokens_seen": 2820079616
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014802020202020204,
      "loss": 1.3602,
      "theoretical_loss": 3.336686292378055,
      "tokens_seen": 2820145152
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000148,
      "loss": 1.2434,
      "theoretical_loss": 3.336680255550996,
      "tokens_seen": 2820210688
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000147979797979798,
      "loss": 1.2878,
      "theoretical_loss": 3.3366742189034966,
      "tokens_seen": 2820276224
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014795959595959596,
      "loss": 1.2562,
      "theoretical_loss": 3.336668182435548,
      "tokens_seen": 2820341760
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014793939393939392,
      "loss": 1.3305,
      "theoretical_loss": 3.3366621461471415,
      "tokens_seen": 2820407296
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014791919191919193,
      "loss": 1.415,
      "theoretical_loss": 3.3366561100382666,
      "tokens_seen": 2820472832
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014789898989898989,
      "loss": 1.514,
      "theoretical_loss": 3.3366500741089142,
      "tokens_seen": 2820538368
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001478787878787879,
      "loss": 1.532,
      "theoretical_loss": 3.3366440383590743,
      "tokens_seen": 2820603904
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014785858585858585,
      "loss": 1.4091,
      "theoretical_loss": 3.336638002788738,
      "tokens_seen": 2820669440
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014783838383838386,
      "loss": 1.4024,
      "theoretical_loss": 3.3366319673978957,
      "tokens_seen": 2820734976
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014781818181818182,
      "loss": 1.3878,
      "theoretical_loss": 3.3366259321865375,
      "tokens_seen": 2820800512
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001477979797979798,
      "loss": 1.398,
      "theoretical_loss": 3.3366198971546543,
      "tokens_seen": 2820866048
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001477777777777778,
      "loss": 1.2219,
      "theoretical_loss": 3.336613862302236,
      "tokens_seen": 2820931584
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014775757575757575,
      "loss": 1.4787,
      "theoretical_loss": 3.3366078276292743,
      "tokens_seen": 2820997120
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014773737373737376,
      "loss": 1.2437,
      "theoretical_loss": 3.3366017931357583,
      "tokens_seen": 2821062656
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014771717171717171,
      "loss": 1.2766,
      "theoretical_loss": 3.3365957588216792,
      "tokens_seen": 2821128192
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001476969696969697,
      "loss": 1.535,
      "theoretical_loss": 3.3365897246870277,
      "tokens_seen": 2821193728
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014767676767676768,
      "loss": 1.1821,
      "theoretical_loss": 3.3365836907317936,
      "tokens_seen": 2821259264
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014765656565656567,
      "loss": 1.2858,
      "theoretical_loss": 3.336577656955968,
      "tokens_seen": 2821324800
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014763636363636365,
      "loss": 1.2456,
      "theoretical_loss": 3.336571623359541,
      "tokens_seen": 2821390336
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014761616161616163,
      "loss": 1.0345,
      "theoretical_loss": 3.336565589942504,
      "tokens_seen": 2821455872
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001475959595959596,
      "loss": 1.3414,
      "theoretical_loss": 3.336559556704846,
      "tokens_seen": 2821521408
    },
    {
      "epoch": 0.71,
      "objective/train/aligned_token_count": 0.4074413478374481,
      "objective/train/avg_token_score": 0.01963413693010807,
      "objective/train/docs_used": 1587530,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2478983402252197,
      "objective/train/original_loss": 4.243338584899902,
      "objective/train/theoretical_loss": 3.3365535236465584,
      "objective/train/tokens_used": 1180446176,
      "objective/train/unlikelihood_loss": 0.05403592437505722,
      "theoretical_loss": 3.3365535236465584,
      "tokens_seen": 2821586944
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014757575757575757,
      "loss": 1.3206,
      "theoretical_loss": 3.3365535236465584,
      "tokens_seen": 2821586944
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014755555555555556,
      "loss": 1.1654,
      "theoretical_loss": 3.336547490767632,
      "tokens_seen": 2821652480
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014753535353535354,
      "loss": 1.2496,
      "theoretical_loss": 3.336541458068057,
      "tokens_seen": 2821718016
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014751515151515153,
      "loss": 1.5334,
      "theoretical_loss": 3.3365354255478232,
      "tokens_seen": 2821783552
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014749494949494948,
      "loss": 1.3862,
      "theoretical_loss": 3.3365293932069218,
      "tokens_seen": 2821849088
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001474747474747475,
      "loss": 1.2583,
      "theoretical_loss": 3.3365233610453435,
      "tokens_seen": 2821914624
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014745454545454545,
      "loss": 1.3117,
      "theoretical_loss": 3.3365173290630783,
      "tokens_seen": 2821980160
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014743434343434346,
      "loss": 1.3473,
      "theoretical_loss": 3.336511297260117,
      "tokens_seen": 2822045696
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014741414141414142,
      "loss": 1.3672,
      "theoretical_loss": 3.3365052656364504,
      "tokens_seen": 2822111232
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014739393939393938,
      "loss": 1.3273,
      "theoretical_loss": 3.3364992341920683,
      "tokens_seen": 2822176768
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001473737373737374,
      "loss": 1.3017,
      "theoretical_loss": 3.3364932029269614,
      "tokens_seen": 2822242304
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014735353535353534,
      "loss": 1.4127,
      "theoretical_loss": 3.3364871718411204,
      "tokens_seen": 2822307840
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014733333333333335,
      "loss": 1.5228,
      "theoretical_loss": 3.336481140934536,
      "tokens_seen": 2822373376
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001473131313131313,
      "loss": 1.3188,
      "theoretical_loss": 3.336475110207198,
      "tokens_seen": 2822438912
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001472929292929293,
      "loss": 1.4057,
      "theoretical_loss": 3.336469079659098,
      "tokens_seen": 2822504448
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014727272727272728,
      "loss": 1.3353,
      "theoretical_loss": 3.3364630492902254,
      "tokens_seen": 2822569984
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014725252525252524,
      "loss": 1.2642,
      "theoretical_loss": 3.3364570191005716,
      "tokens_seen": 2822635520
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014723232323232325,
      "loss": 1.3469,
      "theoretical_loss": 3.3364509890901264,
      "tokens_seen": 2822701056
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001472121212121212,
      "loss": 1.2631,
      "theoretical_loss": 3.3364449592588805,
      "tokens_seen": 2822766592
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001471919191919192,
      "loss": 1.4553,
      "theoretical_loss": 3.336438929606825,
      "tokens_seen": 2822832128
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014717171717171717,
      "loss": 1.3785,
      "theoretical_loss": 3.33643290013395,
      "tokens_seen": 2822897664
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014715151515151516,
      "loss": 1.359,
      "theoretical_loss": 3.3364268708402456,
      "tokens_seen": 2822963200
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014713131313131314,
      "loss": 1.4029,
      "theoretical_loss": 3.336420841725703,
      "tokens_seen": 2823028736
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014711111111111112,
      "loss": 1.0943,
      "theoretical_loss": 3.336414812790312,
      "tokens_seen": 2823094272
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014709090909090908,
      "loss": 1.2817,
      "theoretical_loss": 3.336408784034064,
      "tokens_seen": 2823159808
    },
    {
      "epoch": 0.71,
      "objective/train/aligned_token_count": 0.5137463212013245,
      "objective/train/avg_token_score": 0.01845177635550499,
      "objective/train/docs_used": 1588829,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7521332502365112,
      "objective/train/original_loss": 4.638481616973877,
      "objective/train/theoretical_loss": 3.336402755456948,
      "objective/train/tokens_used": 1182084576,
      "objective/train/unlikelihood_loss": 0.036979421973228455,
      "theoretical_loss": 3.336402755456948,
      "tokens_seen": 2823225344
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014707070707070706,
      "loss": 1.5173,
      "theoretical_loss": 3.336402755456948,
      "tokens_seen": 2823225344
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014705050505050505,
      "loss": 1.3695,
      "theoretical_loss": 3.3363967270589567,
      "tokens_seen": 2823290880
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014703030303030303,
      "loss": 1.3257,
      "theoretical_loss": 3.3363906988400793,
      "tokens_seen": 2823356416
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014701010101010102,
      "loss": 1.4203,
      "theoretical_loss": 3.336384670800306,
      "tokens_seen": 2823421952
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014698989898989897,
      "loss": 1.3491,
      "theoretical_loss": 3.3363786429396285,
      "tokens_seen": 2823487488
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014696969696969698,
      "loss": 1.2024,
      "theoretical_loss": 3.3363726152580364,
      "tokens_seen": 2823553024
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014694949494949494,
      "loss": 1.2189,
      "theoretical_loss": 3.3363665877555206,
      "tokens_seen": 2823618560
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014692929292929295,
      "loss": 1.2771,
      "theoretical_loss": 3.336360560432071,
      "tokens_seen": 2823684096
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001469090909090909,
      "loss": 1.3559,
      "theoretical_loss": 3.336354533287679,
      "tokens_seen": 2823749632
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001468888888888889,
      "loss": 1.4139,
      "theoretical_loss": 3.336348506322335,
      "tokens_seen": 2823815168
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014686868686868688,
      "loss": 1.4248,
      "theoretical_loss": 3.336342479536029,
      "tokens_seen": 2823880704
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014684848484848483,
      "loss": 1.2396,
      "theoretical_loss": 3.3363364529287516,
      "tokens_seen": 2823946240
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014682828282828285,
      "loss": 1.6102,
      "theoretical_loss": 3.3363304265004934,
      "tokens_seen": 2824011776
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001468080808080808,
      "loss": 1.184,
      "theoretical_loss": 3.3363244002512453,
      "tokens_seen": 2824077312
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001467878787878788,
      "loss": 1.4221,
      "theoretical_loss": 3.3363183741809976,
      "tokens_seen": 2824142848
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014676767676767677,
      "loss": 1.1299,
      "theoretical_loss": 3.3363123482897405,
      "tokens_seen": 2824208384
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014674747474747475,
      "loss": 1.3221,
      "theoretical_loss": 3.3363063225774656,
      "tokens_seen": 2824273920
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014672727272727274,
      "loss": 1.3569,
      "theoretical_loss": 3.336300297044162,
      "tokens_seen": 2824339456
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001467070707070707,
      "loss": 1.3502,
      "theoretical_loss": 3.3362942716898214,
      "tokens_seen": 2824404992
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001466868686868687,
      "loss": 1.2642,
      "theoretical_loss": 3.3362882465144335,
      "tokens_seen": 2824470528
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014666666666666666,
      "loss": 1.2273,
      "theoretical_loss": 3.336282221517989,
      "tokens_seen": 2824536064
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014664646464646465,
      "loss": 1.2864,
      "theoretical_loss": 3.336276196700479,
      "tokens_seen": 2824601600
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014662626262626263,
      "loss": 1.3,
      "theoretical_loss": 3.3362701720618935,
      "tokens_seen": 2824667136
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014660606060606061,
      "loss": 1.2164,
      "theoretical_loss": 3.3362641476022232,
      "tokens_seen": 2824732672
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001465858585858586,
      "loss": 1.4013,
      "theoretical_loss": 3.3362581233214583,
      "tokens_seen": 2824798208
    },
    {
      "epoch": 0.71,
      "objective/train/aligned_token_count": 0.45173507928848267,
      "objective/train/avg_token_score": 0.006888273172080517,
      "objective/train/docs_used": 1589488,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3781616687774658,
      "objective/train/original_loss": 4.038028717041016,
      "objective/train/theoretical_loss": 3.33625209921959,
      "objective/train/tokens_used": 1183722976,
      "objective/train/unlikelihood_loss": 0.05652686208486557,
      "theoretical_loss": 3.33625209921959,
      "tokens_seen": 2824863744
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014656565656565656,
      "loss": 1.4598,
      "theoretical_loss": 3.33625209921959,
      "tokens_seen": 2824863744
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014654545454545454,
      "loss": 1.3935,
      "theoretical_loss": 3.3362460752966085,
      "tokens_seen": 2824929280
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014652525252525252,
      "loss": 1.4697,
      "theoretical_loss": 3.336240051552504,
      "tokens_seen": 2824994816
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001465050505050505,
      "loss": 1.216,
      "theoretical_loss": 3.3362340279872678,
      "tokens_seen": 2825060352
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001464848484848485,
      "loss": 1.2208,
      "theoretical_loss": 3.3362280046008896,
      "tokens_seen": 2825125888
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014646464646464648,
      "loss": 1.1849,
      "theoretical_loss": 3.336221981393361,
      "tokens_seen": 2825191424
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014644444444444443,
      "loss": 1.5038,
      "theoretical_loss": 3.3362159583646713,
      "tokens_seen": 2825256960
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014642424242424244,
      "loss": 1.4536,
      "theoretical_loss": 3.3362099355148116,
      "tokens_seen": 2825322496
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001464040404040404,
      "loss": 1.4881,
      "theoretical_loss": 3.3362039128437724,
      "tokens_seen": 2825388032
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014638383838383838,
      "loss": 1.3768,
      "theoretical_loss": 3.336197890351545,
      "tokens_seen": 2825453568
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014636363636363637,
      "loss": 1.1894,
      "theoretical_loss": 3.336191868038118,
      "tokens_seen": 2825519104
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014634343434343432,
      "loss": 1.3234,
      "theoretical_loss": 3.3361858459034845,
      "tokens_seen": 2825584640
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014632323232323234,
      "loss": 0.8926,
      "theoretical_loss": 3.336179823947633,
      "tokens_seen": 2825650176
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001463030303030303,
      "loss": 1.4039,
      "theoretical_loss": 3.336173802170555,
      "tokens_seen": 2825715712
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001462828282828283,
      "loss": 1.208,
      "theoretical_loss": 3.3361677805722407,
      "tokens_seen": 2825781248
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014626262626262626,
      "loss": 1.3221,
      "theoretical_loss": 3.336161759152681,
      "tokens_seen": 2825846784
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014624242424242424,
      "loss": 1.291,
      "theoretical_loss": 3.336155737911866,
      "tokens_seen": 2825912320
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014622222222222223,
      "loss": 1.3059,
      "theoretical_loss": 3.3361497168497865,
      "tokens_seen": 2825977856
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014620202020202019,
      "loss": 1.3184,
      "theoretical_loss": 3.3361436959664332,
      "tokens_seen": 2826043392
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001461818181818182,
      "loss": 1.3612,
      "theoretical_loss": 3.3361376752617966,
      "tokens_seen": 2826108928
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014616161616161615,
      "loss": 1.6876,
      "theoretical_loss": 3.336131654735867,
      "tokens_seen": 2826174464
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014614141414141414,
      "loss": 1.4396,
      "theoretical_loss": 3.336125634388635,
      "tokens_seen": 2826240000
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014612121212121212,
      "loss": 1.2071,
      "theoretical_loss": 3.336119614220091,
      "tokens_seen": 2826305536
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001461010101010101,
      "loss": 1.3831,
      "theoretical_loss": 3.336113594230226,
      "tokens_seen": 2826371072
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001460808080808081,
      "loss": 1.4211,
      "theoretical_loss": 3.3361075744190307,
      "tokens_seen": 2826436608
    },
    {
      "epoch": 0.71,
      "objective/train/aligned_token_count": 0.37090665102005005,
      "objective/train/avg_token_score": 0.0135739641264081,
      "objective/train/docs_used": 1590914,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.102207064628601,
      "objective/train/original_loss": 4.1722893714904785,
      "objective/train/theoretical_loss": 3.3361015547864947,
      "objective/train/tokens_used": 1185361376,
      "objective/train/unlikelihood_loss": 0.08047188818454742,
      "theoretical_loss": 3.3361015547864947,
      "tokens_seen": 2826502144
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014606060606060607,
      "loss": 1.2911,
      "theoretical_loss": 3.3361015547864947,
      "tokens_seen": 2826502144
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014604040404040406,
      "loss": 1.2617,
      "theoretical_loss": 3.3360955353326096,
      "tokens_seen": 2826567680
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014602020202020201,
      "loss": 1.2433,
      "theoretical_loss": 3.336089516057365,
      "tokens_seen": 2826633216
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000146,
      "loss": 1.2128,
      "theoretical_loss": 3.3360834969607525,
      "tokens_seen": 2826698752
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014597979797979798,
      "loss": 1.5018,
      "theoretical_loss": 3.336077478042762,
      "tokens_seen": 2826764288
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014595959595959597,
      "loss": 1.3968,
      "theoretical_loss": 3.3360714593033842,
      "tokens_seen": 2826829824
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014593939393939395,
      "loss": 1.3035,
      "theoretical_loss": 3.336065440742609,
      "tokens_seen": 2826895360
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014591919191919193,
      "loss": 1.4439,
      "theoretical_loss": 3.3360594223604285,
      "tokens_seen": 2826960896
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001458989898989899,
      "loss": 1.3125,
      "theoretical_loss": 3.336053404156832,
      "tokens_seen": 2827026432
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001458787878787879,
      "loss": 1.5,
      "theoretical_loss": 3.3360473861318103,
      "tokens_seen": 2827091968
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014585858585858586,
      "loss": 1.2569,
      "theoretical_loss": 3.336041368285354,
      "tokens_seen": 2827157504
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014583838383838384,
      "loss": 1.0435,
      "theoretical_loss": 3.336035350617454,
      "tokens_seen": 2827223040
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014581818181818183,
      "loss": 1.3229,
      "theoretical_loss": 3.336029333128101,
      "tokens_seen": 2827288576
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014579797979797978,
      "loss": 1.1617,
      "theoretical_loss": 3.3360233158172843,
      "tokens_seen": 2827354112
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001457777777777778,
      "loss": 1.5801,
      "theoretical_loss": 3.336017298684996,
      "tokens_seen": 2827419648
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014575757575757575,
      "loss": 1.295,
      "theoretical_loss": 3.3360112817312255,
      "tokens_seen": 2827485184
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014573737373737376,
      "loss": 1.4329,
      "theoretical_loss": 3.3360052649559644,
      "tokens_seen": 2827550720
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014571717171717172,
      "loss": 1.5107,
      "theoretical_loss": 3.3359992483592023,
      "tokens_seen": 2827616256
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014569696969696968,
      "loss": 1.1784,
      "theoretical_loss": 3.3359932319409302,
      "tokens_seen": 2827681792
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001456767676767677,
      "loss": 1.1796,
      "theoretical_loss": 3.335987215701139,
      "tokens_seen": 2827747328
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014565656565656564,
      "loss": 1.2978,
      "theoretical_loss": 3.3359811996398188,
      "tokens_seen": 2827812864
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014563636363636366,
      "loss": 1.4196,
      "theoretical_loss": 3.3359751837569602,
      "tokens_seen": 2827878400
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001456161616161616,
      "loss": 1.2636,
      "theoretical_loss": 3.335969168052554,
      "tokens_seen": 2827943936
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001455959595959596,
      "loss": 1.1852,
      "theoretical_loss": 3.3359631525265905,
      "tokens_seen": 2828009472
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014557575757575758,
      "loss": 1.478,
      "theoretical_loss": 3.3359571371790606,
      "tokens_seen": 2828075008
    },
    {
      "epoch": 0.71,
      "objective/train/aligned_token_count": 0.581805944442749,
      "objective/train/avg_token_score": 0.011462582275271416,
      "objective/train/docs_used": 1591402,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7380067110061646,
      "objective/train/original_loss": 3.5706305503845215,
      "objective/train/theoretical_loss": 3.3359511220099547,
      "objective/train/tokens_used": 1186999776,
      "objective/train/unlikelihood_loss": 0.06461834907531738,
      "theoretical_loss": 3.3359511220099547,
      "tokens_seen": 2828140544
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014555555555555556,
      "loss": 1.4104,
      "theoretical_loss": 3.3359511220099547,
      "tokens_seen": 2828140544
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014553535353535355,
      "loss": 1.3359,
      "theoretical_loss": 3.3359451070192634,
      "tokens_seen": 2828206080
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001455151515151515,
      "loss": 1.3263,
      "theoretical_loss": 3.335939092206977,
      "tokens_seen": 2828271616
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001454949494949495,
      "loss": 1.1907,
      "theoretical_loss": 3.3359330775730864,
      "tokens_seen": 2828337152
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014547474747474747,
      "loss": 1.2964,
      "theoretical_loss": 3.335927063117582,
      "tokens_seen": 2828402688
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014545454545454546,
      "loss": 1.2761,
      "theoretical_loss": 3.3359210488404547,
      "tokens_seen": 2828468224
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014543434343434344,
      "loss": 1.2095,
      "theoretical_loss": 3.335915034741695,
      "tokens_seen": 2828533760
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014541414141414142,
      "loss": 1.4161,
      "theoretical_loss": 3.335909020821293,
      "tokens_seen": 2828599296
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014539393939393938,
      "loss": 1.3818,
      "theoretical_loss": 3.3359030070792395,
      "tokens_seen": 2828664832
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001453737373737374,
      "loss": 1.4501,
      "theoretical_loss": 3.3358969935155254,
      "tokens_seen": 2828730368
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014535353535353535,
      "loss": 1.3775,
      "theoretical_loss": 3.335890980130141,
      "tokens_seen": 2828795904
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014533333333333333,
      "loss": 1.4935,
      "theoretical_loss": 3.335884966923077,
      "tokens_seen": 2828861440
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014531313131313132,
      "loss": 1.4282,
      "theoretical_loss": 3.3358789538943237,
      "tokens_seen": 2828926976
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014529292929292927,
      "loss": 1.4305,
      "theoretical_loss": 3.3358729410438723,
      "tokens_seen": 2828992512
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014527272727272729,
      "loss": 1.3274,
      "theoretical_loss": 3.3358669283717126,
      "tokens_seen": 2829058048
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014525252525252524,
      "loss": 1.2721,
      "theoretical_loss": 3.335860915877836,
      "tokens_seen": 2829123584
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014523232323232325,
      "loss": 1.517,
      "theoretical_loss": 3.3358549035622325,
      "tokens_seen": 2829189120
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001452121212121212,
      "loss": 1.3946,
      "theoretical_loss": 3.3358488914248925,
      "tokens_seen": 2829254656
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014519191919191922,
      "loss": 1.2813,
      "theoretical_loss": 3.3358428794658073,
      "tokens_seen": 2829320192
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014517171717171718,
      "loss": 1.3193,
      "theoretical_loss": 3.335836867684967,
      "tokens_seen": 2829385728
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014515151515151513,
      "loss": 1.318,
      "theoretical_loss": 3.335830856082362,
      "tokens_seen": 2829451264
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014513131313131315,
      "loss": 1.4455,
      "theoretical_loss": 3.3358248446579837,
      "tokens_seen": 2829516800
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001451111111111111,
      "loss": 1.3762,
      "theoretical_loss": 3.335818833411822,
      "tokens_seen": 2829582336
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014509090909090911,
      "loss": 1.209,
      "theoretical_loss": 3.3358128223438674,
      "tokens_seen": 2829647872
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014507070707070707,
      "loss": 1.2996,
      "theoretical_loss": 3.335806811454111,
      "tokens_seen": 2829713408
    },
    {
      "epoch": 0.71,
      "objective/train/aligned_token_count": 0.4451979398727417,
      "objective/train/avg_token_score": 0.0071952566504478455,
      "objective/train/docs_used": 1592635,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5413883924484253,
      "objective/train/original_loss": 4.114957332611084,
      "objective/train/theoretical_loss": 3.335800800742543,
      "objective/train/tokens_used": 1188638176,
      "objective/train/unlikelihood_loss": 0.06121214106678963,
      "theoretical_loss": 3.335800800742543,
      "tokens_seen": 2829778944
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014505050505050505,
      "loss": 1.5296,
      "theoretical_loss": 3.335800800742543,
      "tokens_seen": 2829778944
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014503030303030304,
      "loss": 1.2443,
      "theoretical_loss": 3.335794790209154,
      "tokens_seen": 2829844480
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000145010101010101,
      "loss": 1.5062,
      "theoretical_loss": 3.335788779853935,
      "tokens_seen": 2829910016
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.000144989898989899,
      "loss": 1.2355,
      "theoretical_loss": 3.335782769676876,
      "tokens_seen": 2829975552
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014496969696969696,
      "loss": 1.2779,
      "theoretical_loss": 3.3357767596779686,
      "tokens_seen": 2830041088
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014494949494949495,
      "loss": 1.3506,
      "theoretical_loss": 3.3357707498572022,
      "tokens_seen": 2830106624
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014492929292929293,
      "loss": 1.2025,
      "theoretical_loss": 3.335764740214568,
      "tokens_seen": 2830172160
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014490909090909092,
      "loss": 1.4038,
      "theoretical_loss": 3.3357587307500562,
      "tokens_seen": 2830237696
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001448888888888889,
      "loss": 1.2915,
      "theoretical_loss": 3.335752721463658,
      "tokens_seen": 2830303232
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014486868686868688,
      "loss": 1.5775,
      "theoretical_loss": 3.3357467123553635,
      "tokens_seen": 2830368768
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014484848484848484,
      "loss": 1.4179,
      "theoretical_loss": 3.335740703425164,
      "tokens_seen": 2830434304
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014482828282828282,
      "loss": 1.1472,
      "theoretical_loss": 3.335734694673049,
      "tokens_seen": 2830499840
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001448080808080808,
      "loss": 1.2018,
      "theoretical_loss": 3.33572868609901,
      "tokens_seen": 2830565376
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001447878787878788,
      "loss": 1.4673,
      "theoretical_loss": 3.335722677703037,
      "tokens_seen": 2830630912
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014476767676767678,
      "loss": 1.2595,
      "theoretical_loss": 3.3357166694851212,
      "tokens_seen": 2830696448
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014474747474747473,
      "loss": 1.5099,
      "theoretical_loss": 3.3357106614452525,
      "tokens_seen": 2830761984
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014472727272727274,
      "loss": 1.1951,
      "theoretical_loss": 3.3357046535834223,
      "tokens_seen": 2830827520
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001447070707070707,
      "loss": 1.3651,
      "theoretical_loss": 3.3356986458996207,
      "tokens_seen": 2830893056
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001446868686868687,
      "loss": 1.4334,
      "theoretical_loss": 3.335692638393838,
      "tokens_seen": 2830958592
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014466666666666667,
      "loss": 1.3926,
      "theoretical_loss": 3.3356866310660656,
      "tokens_seen": 2831024128
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014464646464646463,
      "loss": 1.415,
      "theoretical_loss": 3.3356806239162937,
      "tokens_seen": 2831089664
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014462626262626264,
      "loss": 1.2716,
      "theoretical_loss": 3.335674616944513,
      "tokens_seen": 2831155200
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001446060606060606,
      "loss": 1.1613,
      "theoretical_loss": 3.3356686101507136,
      "tokens_seen": 2831220736
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001445858585858586,
      "loss": 1.3909,
      "theoretical_loss": 3.3356626035348866,
      "tokens_seen": 2831286272
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014456565656565656,
      "loss": 1.2548,
      "theoretical_loss": 3.3356565970970227,
      "tokens_seen": 2831351808
    },
    {
      "epoch": 0.71,
      "objective/train/aligned_token_count": 0.4493523836135864,
      "objective/train/avg_token_score": 0.006711914204061031,
      "objective/train/docs_used": 1593167,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5931365489959717,
      "objective/train/original_loss": 4.444538593292236,
      "objective/train/theoretical_loss": 3.3356505908371124,
      "objective/train/tokens_used": 1190276576,
      "objective/train/unlikelihood_loss": 0.05337769165635109,
      "theoretical_loss": 3.3356505908371124,
      "tokens_seen": 2831417344
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014454545454545455,
      "loss": 1.3012,
      "theoretical_loss": 3.3356505908371124,
      "tokens_seen": 2831417344
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014452525252525253,
      "loss": 1.7005,
      "theoretical_loss": 3.335644584755146,
      "tokens_seen": 2831482880
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001445050505050505,
      "loss": 1.5887,
      "theoretical_loss": 3.3356385788511145,
      "tokens_seen": 2831548416
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001444848484848485,
      "loss": 1.3102,
      "theoretical_loss": 3.3356325731250083,
      "tokens_seen": 2831613952
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014446464646464645,
      "loss": 1.4505,
      "theoretical_loss": 3.335626567576818,
      "tokens_seen": 2831679488
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014444444444444444,
      "loss": 1.3041,
      "theoretical_loss": 3.3356205622065347,
      "tokens_seen": 2831745024
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014442424242424242,
      "loss": 1.5999,
      "theoretical_loss": 3.335614557014148,
      "tokens_seen": 2831810560
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001444040404040404,
      "loss": 1.2106,
      "theoretical_loss": 3.3356085519996497,
      "tokens_seen": 2831876096
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001443838383838384,
      "loss": 1.4232,
      "theoretical_loss": 3.3356025471630297,
      "tokens_seen": 2831941632
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014436363636363637,
      "loss": 1.4649,
      "theoretical_loss": 3.3355965425042786,
      "tokens_seen": 2832007168
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014434343434343436,
      "loss": 1.3906,
      "theoretical_loss": 3.335590538023387,
      "tokens_seen": 2832072704
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014432323232323234,
      "loss": 1.3847,
      "theoretical_loss": 3.335584533720346,
      "tokens_seen": 2832138240
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001443030303030303,
      "loss": 1.3719,
      "theoretical_loss": 3.3355785295951454,
      "tokens_seen": 2832203776
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014428282828282828,
      "loss": 1.3076,
      "theoretical_loss": 3.335572525647777,
      "tokens_seen": 2832269312
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014426262626262627,
      "loss": 1.3434,
      "theoretical_loss": 3.3355665218782304,
      "tokens_seen": 2832334848
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014424242424242425,
      "loss": 1.5113,
      "theoretical_loss": 3.3355605182864965,
      "tokens_seen": 2832400384
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014422222222222223,
      "loss": 1.353,
      "theoretical_loss": 3.335554514872566,
      "tokens_seen": 2832465920
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001442020202020202,
      "loss": 1.4243,
      "theoretical_loss": 3.3355485116364294,
      "tokens_seen": 2832531456
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001441818181818182,
      "loss": 1.3183,
      "theoretical_loss": 3.3355425085780777,
      "tokens_seen": 2832596992
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014416161616161616,
      "loss": 1.3618,
      "theoretical_loss": 3.3355365056975006,
      "tokens_seen": 2832662528
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014414141414141417,
      "loss": 1.3609,
      "theoretical_loss": 3.33553050299469,
      "tokens_seen": 2832728064
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014412121212121213,
      "loss": 1.3177,
      "theoretical_loss": 3.3355245004696354,
      "tokens_seen": 2832793600
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014410101010101008,
      "loss": 1.3624,
      "theoretical_loss": 3.3355184981223283,
      "tokens_seen": 2832859136
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0001440808080808081,
      "loss": 1.4603,
      "theoretical_loss": 3.3355124959527584,
      "tokens_seen": 2832924672
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014406060606060605,
      "loss": 1.3997,
      "theoretical_loss": 3.335506493960917,
      "tokens_seen": 2832990208
    },
    {
      "epoch": 0.71,
      "objective/train/aligned_token_count": 0.3365102708339691,
      "objective/train/avg_token_score": 0.014186584390699863,
      "objective/train/docs_used": 1594447,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0691235065460205,
      "objective/train/original_loss": 4.446936130523682,
      "objective/train/theoretical_loss": 3.335500492146795,
      "objective/train/tokens_used": 1191914976,
      "objective/train/unlikelihood_loss": 0.0650569349527359,
      "theoretical_loss": 3.335500492146795,
      "tokens_seen": 2833055744
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00014404040404040406,
      "loss": 1.3615,
      "theoretical_loss": 3.335500492146795,
      "tokens_seen": 2833055744
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014402020202020202,
      "loss": 1.5673,
      "theoretical_loss": 3.3354944905103823,
      "tokens_seen": 2833121280
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000144,
      "loss": 1.2284,
      "theoretical_loss": 3.33548848905167,
      "tokens_seen": 2833186816
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000143979797979798,
      "loss": 1.2764,
      "theoretical_loss": 3.3354824877706486,
      "tokens_seen": 2833252352
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014395959595959594,
      "loss": 1.5904,
      "theoretical_loss": 3.335476486667308,
      "tokens_seen": 2833317888
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014393939393939396,
      "loss": 1.3335,
      "theoretical_loss": 3.3354704857416406,
      "tokens_seen": 2833383424
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001439191919191919,
      "loss": 1.3826,
      "theoretical_loss": 3.335464484993635,
      "tokens_seen": 2833448960
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001438989898989899,
      "loss": 1.2402,
      "theoretical_loss": 3.3354584844232833,
      "tokens_seen": 2833514496
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014387878787878788,
      "loss": 1.3476,
      "theoretical_loss": 3.3354524840305757,
      "tokens_seen": 2833580032
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014385858585858586,
      "loss": 1.4772,
      "theoretical_loss": 3.3354464838155025,
      "tokens_seen": 2833645568
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014383838383838385,
      "loss": 1.3196,
      "theoretical_loss": 3.3354404837780542,
      "tokens_seen": 2833711104
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014381818181818183,
      "loss": 1.3654,
      "theoretical_loss": 3.3354344839182226,
      "tokens_seen": 2833776640
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001437979797979798,
      "loss": 1.4337,
      "theoretical_loss": 3.335428484235997,
      "tokens_seen": 2833842176
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014377777777777777,
      "loss": 1.5555,
      "theoretical_loss": 3.3354224847313687,
      "tokens_seen": 2833907712
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014375757575757576,
      "loss": 1.3922,
      "theoretical_loss": 3.3354164854043282,
      "tokens_seen": 2833973248
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014373737373737374,
      "loss": 1.5223,
      "theoretical_loss": 3.335410486254866,
      "tokens_seen": 2834038784
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014371717171717172,
      "loss": 1.3983,
      "theoretical_loss": 3.3354044872829736,
      "tokens_seen": 2834104320
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014369696969696968,
      "loss": 1.2327,
      "theoretical_loss": 3.3353984884886403,
      "tokens_seen": 2834169856
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001436767676767677,
      "loss": 1.4261,
      "theoretical_loss": 3.3353924898718574,
      "tokens_seen": 2834235392
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014365656565656565,
      "loss": 1.135,
      "theoretical_loss": 3.3353864914326152,
      "tokens_seen": 2834300928
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014363636363636366,
      "loss": 1.141,
      "theoretical_loss": 3.3353804931709052,
      "tokens_seen": 2834366464
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014361616161616162,
      "loss": 1.3707,
      "theoretical_loss": 3.3353744950867172,
      "tokens_seen": 2834432000
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014359595959595957,
      "loss": 1.1818,
      "theoretical_loss": 3.3353684971800424,
      "tokens_seen": 2834497536
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014357575757575759,
      "loss": 1.3432,
      "theoretical_loss": 3.335362499450871,
      "tokens_seen": 2834563072
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014355555555555554,
      "loss": 1.4366,
      "theoretical_loss": 3.3353565018991937,
      "tokens_seen": 2834628608
    },
    {
      "epoch": 0.72,
      "objective/train/aligned_token_count": 0.38031524419784546,
      "objective/train/avg_token_score": 0.023266928270459175,
      "objective/train/docs_used": 1595629,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.241020917892456,
      "objective/train/original_loss": 4.531277656555176,
      "objective/train/theoretical_loss": 3.3353505045250014,
      "objective/train/tokens_used": 1193553376,
      "objective/train/unlikelihood_loss": 0.054239384829998016,
      "theoretical_loss": 3.3353505045250014,
      "tokens_seen": 2834694144
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014353535353535355,
      "loss": 1.3827,
      "theoretical_loss": 3.3353505045250014,
      "tokens_seen": 2834694144
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001435151515151515,
      "loss": 1.3153,
      "theoretical_loss": 3.335344507328285,
      "tokens_seen": 2834759680
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014349494949494952,
      "loss": 1.3549,
      "theoretical_loss": 3.3353385103090343,
      "tokens_seen": 2834825216
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014347474747474748,
      "loss": 1.3511,
      "theoretical_loss": 3.33533251346724,
      "tokens_seen": 2834890752
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014345454545454546,
      "loss": 1.3768,
      "theoretical_loss": 3.3353265168028936,
      "tokens_seen": 2834956288
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014343434343434345,
      "loss": 1.4675,
      "theoretical_loss": 3.3353205203159852,
      "tokens_seen": 2835021824
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001434141414141414,
      "loss": 1.4857,
      "theoretical_loss": 3.3353145240065056,
      "tokens_seen": 2835087360
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014339393939393941,
      "loss": 1.122,
      "theoretical_loss": 3.3353085278744454,
      "tokens_seen": 2835152896
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014337373737373737,
      "loss": 1.5888,
      "theoretical_loss": 3.3353025319197953,
      "tokens_seen": 2835218432
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014335353535353535,
      "loss": 1.1182,
      "theoretical_loss": 3.335296536142546,
      "tokens_seen": 2835283968
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014333333333333334,
      "loss": 1.3879,
      "theoretical_loss": 3.3352905405426876,
      "tokens_seen": 2835349504
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014331313131313132,
      "loss": 1.201,
      "theoretical_loss": 3.3352845451202113,
      "tokens_seen": 2835415040
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001432929292929293,
      "loss": 1.2512,
      "theoretical_loss": 3.335278549875108,
      "tokens_seen": 2835480576
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014327272727272726,
      "loss": 1.3637,
      "theoretical_loss": 3.335272554807368,
      "tokens_seen": 2835546112
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014325252525252525,
      "loss": 1.4071,
      "theoretical_loss": 3.335266559916981,
      "tokens_seen": 2835611648
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014323232323232323,
      "loss": 1.3687,
      "theoretical_loss": 3.3352605652039395,
      "tokens_seen": 2835677184
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014321212121212122,
      "loss": 1.3356,
      "theoretical_loss": 3.335254570668233,
      "tokens_seen": 2835742720
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001431919191919192,
      "loss": 1.2414,
      "theoretical_loss": 3.3352485763098527,
      "tokens_seen": 2835808256
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014317171717171718,
      "loss": 1.4534,
      "theoretical_loss": 3.3352425821287888,
      "tokens_seen": 2835873792
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014315151515151514,
      "loss": 1.2483,
      "theoretical_loss": 3.335236588125032,
      "tokens_seen": 2835939328
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014313131313131315,
      "loss": 1.4925,
      "theoretical_loss": 3.335230594298573,
      "tokens_seen": 2836004864
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001431111111111111,
      "loss": 1.4587,
      "theoretical_loss": 3.3352246006494024,
      "tokens_seen": 2836070400
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001430909090909091,
      "loss": 1.3345,
      "theoretical_loss": 3.3352186071775116,
      "tokens_seen": 2836135936
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014307070707070708,
      "loss": 1.4395,
      "theoretical_loss": 3.3352126138828906,
      "tokens_seen": 2836201472
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014305050505050503,
      "loss": 1.612,
      "theoretical_loss": 3.3352066207655295,
      "tokens_seen": 2836267008
    },
    {
      "epoch": 0.72,
      "objective/train/aligned_token_count": 0.4667644202709198,
      "objective/train/avg_token_score": 0.010480287484824657,
      "objective/train/docs_used": 1596277,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3762258291244507,
      "objective/train/original_loss": 4.085051536560059,
      "objective/train/theoretical_loss": 3.33520062782542,
      "objective/train/tokens_used": 1195191776,
      "objective/train/unlikelihood_loss": 0.06265047937631607,
      "theoretical_loss": 3.33520062782542,
      "tokens_seen": 2836332544
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014303030303030304,
      "loss": 1.5447,
      "theoretical_loss": 3.33520062782542,
      "tokens_seen": 2836332544
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000143010101010101,
      "loss": 1.3884,
      "theoretical_loss": 3.3351946350625523,
      "tokens_seen": 2836398080
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000142989898989899,
      "loss": 1.4843,
      "theoretical_loss": 3.335188642476917,
      "tokens_seen": 2836463616
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014296969696969697,
      "loss": 1.3952,
      "theoretical_loss": 3.3351826500685053,
      "tokens_seen": 2836529152
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014294949494949495,
      "loss": 1.4962,
      "theoretical_loss": 3.335176657837307,
      "tokens_seen": 2836594688
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014292929292929294,
      "loss": 1.2782,
      "theoretical_loss": 3.335170665783313,
      "tokens_seen": 2836660224
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001429090909090909,
      "loss": 1.3104,
      "theoretical_loss": 3.335164673906515,
      "tokens_seen": 2836725760
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001428888888888889,
      "loss": 1.387,
      "theoretical_loss": 3.335158682206902,
      "tokens_seen": 2836791296
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014286868686868686,
      "loss": 1.1837,
      "theoretical_loss": 3.335152690684466,
      "tokens_seen": 2836856832
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014284848484848485,
      "loss": 1.308,
      "theoretical_loss": 3.335146699339197,
      "tokens_seen": 2836922368
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014282828282828283,
      "loss": 1.4485,
      "theoretical_loss": 3.3351407081710858,
      "tokens_seen": 2836987904
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001428080808080808,
      "loss": 1.1854,
      "theoretical_loss": 3.3351347171801233,
      "tokens_seen": 2837053440
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001427878787878788,
      "loss": 1.3547,
      "theoretical_loss": 3.3351287263662996,
      "tokens_seen": 2837118976
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014276767676767678,
      "loss": 1.5567,
      "theoretical_loss": 3.3351227357296063,
      "tokens_seen": 2837184512
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014274747474747474,
      "loss": 1.5306,
      "theoretical_loss": 3.335116745270033,
      "tokens_seen": 2837250048
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014272727272727272,
      "loss": 1.4611,
      "theoretical_loss": 3.335110754987571,
      "tokens_seen": 2837315584
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001427070707070707,
      "loss": 1.3034,
      "theoretical_loss": 3.335104764882211,
      "tokens_seen": 2837381120
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001426868686868687,
      "loss": 1.2573,
      "theoretical_loss": 3.3350987749539436,
      "tokens_seen": 2837446656
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014266666666666667,
      "loss": 1.5539,
      "theoretical_loss": 3.335092785202759,
      "tokens_seen": 2837512192
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014264646464646466,
      "loss": 1.5854,
      "theoretical_loss": 3.3350867956286487,
      "tokens_seen": 2837577728
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014262626262626264,
      "loss": 1.2013,
      "theoretical_loss": 3.335080806231603,
      "tokens_seen": 2837643264
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001426060606060606,
      "loss": 1.4014,
      "theoretical_loss": 3.3350748170116127,
      "tokens_seen": 2837708800
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001425858585858586,
      "loss": 1.1707,
      "theoretical_loss": 3.3350688279686675,
      "tokens_seen": 2837774336
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014256565656565657,
      "loss": 1.216,
      "theoretical_loss": 3.3350628391027595,
      "tokens_seen": 2837839872
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014254545454545455,
      "loss": 1.4786,
      "theoretical_loss": 3.3350568504138787,
      "tokens_seen": 2837905408
    },
    {
      "epoch": 0.72,
      "objective/train/aligned_token_count": 0.3884408473968506,
      "objective/train/avg_token_score": 0.007494966499507427,
      "objective/train/docs_used": 1596937,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4393794536590576,
      "objective/train/original_loss": 5.007159233093262,
      "objective/train/theoretical_loss": 3.3350508619020163,
      "objective/train/tokens_used": 1196830176,
      "objective/train/unlikelihood_loss": 0.040895771235227585,
      "theoretical_loss": 3.3350508619020163,
      "tokens_seen": 2837970944
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014252525252525253,
      "loss": 1.4486,
      "theoretical_loss": 3.3350508619020163,
      "tokens_seen": 2837970944
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001425050505050505,
      "loss": 1.4116,
      "theoretical_loss": 3.3350448735671616,
      "tokens_seen": 2838036480
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001424848484848485,
      "loss": 1.2574,
      "theoretical_loss": 3.3350388854093067,
      "tokens_seen": 2838102016
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014246464646464646,
      "loss": 1.3689,
      "theoretical_loss": 3.3350328974284422,
      "tokens_seen": 2838167552
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014244444444444447,
      "loss": 1.2106,
      "theoretical_loss": 3.335026909624558,
      "tokens_seen": 2838233088
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014242424242424243,
      "loss": 1.3551,
      "theoretical_loss": 3.335020921997645,
      "tokens_seen": 2838298624
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014240404040404038,
      "loss": 1.3144,
      "theoretical_loss": 3.335014934547694,
      "tokens_seen": 2838364160
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001423838383838384,
      "loss": 1.4137,
      "theoretical_loss": 3.335008947274696,
      "tokens_seen": 2838429696
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014236363636363635,
      "loss": 1.4461,
      "theoretical_loss": 3.3350029601786417,
      "tokens_seen": 2838495232
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014234343434343436,
      "loss": 1.4057,
      "theoretical_loss": 3.334996973259521,
      "tokens_seen": 2838560768
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014232323232323232,
      "loss": 1.0623,
      "theoretical_loss": 3.334990986517325,
      "tokens_seen": 2838626304
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001423030303030303,
      "loss": 1.4617,
      "theoretical_loss": 3.3349849999520447,
      "tokens_seen": 2838691840
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001422828282828283,
      "loss": 1.3028,
      "theoretical_loss": 3.3349790135636708,
      "tokens_seen": 2838757376
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014226262626262627,
      "loss": 1.2935,
      "theoretical_loss": 3.334973027352193,
      "tokens_seen": 2838822912
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014224242424242426,
      "loss": 1.2633,
      "theoretical_loss": 3.3349670413176034,
      "tokens_seen": 2838888448
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001422222222222222,
      "loss": 1.5506,
      "theoretical_loss": 3.334961055459892,
      "tokens_seen": 2838953984
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001422020202020202,
      "loss": 1.1509,
      "theoretical_loss": 3.3349550697790495,
      "tokens_seen": 2839019520
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014218181818181818,
      "loss": 1.412,
      "theoretical_loss": 3.3349490842750664,
      "tokens_seen": 2839085056
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014216161616161616,
      "loss": 1.3422,
      "theoretical_loss": 3.334943098947934,
      "tokens_seen": 2839150592
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014214141414141415,
      "loss": 1.4522,
      "theoretical_loss": 3.334937113797642,
      "tokens_seen": 2839216128
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014212121212121213,
      "loss": 1.4725,
      "theoretical_loss": 3.334931128824182,
      "tokens_seen": 2839281664
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001421010101010101,
      "loss": 1.4475,
      "theoretical_loss": 3.3349251440275443,
      "tokens_seen": 2839347200
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001420808080808081,
      "loss": 1.4744,
      "theoretical_loss": 3.3349191594077197,
      "tokens_seen": 2839412736
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014206060606060606,
      "loss": 1.465,
      "theoretical_loss": 3.334913174964699,
      "tokens_seen": 2839478272
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014204040404040404,
      "loss": 1.3666,
      "theoretical_loss": 3.3349071906984724,
      "tokens_seen": 2839543808
    },
    {
      "epoch": 0.72,
      "objective/train/aligned_token_count": 0.5474095940589905,
      "objective/train/avg_token_score": 0.004180094227194786,
      "objective/train/docs_used": 1597942,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7658870220184326,
      "objective/train/original_loss": 3.7517788410186768,
      "objective/train/theoretical_loss": 3.3349012066090316,
      "objective/train/tokens_used": 1198468576,
      "objective/train/unlikelihood_loss": 0.05793384835124016,
      "theoretical_loss": 3.3349012066090316,
      "tokens_seen": 2839609344
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014202020202020203,
      "loss": 1.4424,
      "theoretical_loss": 3.3349012066090316,
      "tokens_seen": 2839609344
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014199999999999998,
      "loss": 1.3448,
      "theoretical_loss": 3.334895222696366,
      "tokens_seen": 2839674880
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000141979797979798,
      "loss": 1.4874,
      "theoretical_loss": 3.3348892389604674,
      "tokens_seen": 2839740416
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014195959595959595,
      "loss": 1.3417,
      "theoretical_loss": 3.334883255401326,
      "tokens_seen": 2839805952
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014193939393939396,
      "loss": 1.5585,
      "theoretical_loss": 3.3348772720189324,
      "tokens_seen": 2839871488
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014191919191919192,
      "loss": 1.3638,
      "theoretical_loss": 3.3348712888132774,
      "tokens_seen": 2839937024
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001418989898989899,
      "loss": 1.6219,
      "theoretical_loss": 3.3348653057843523,
      "tokens_seen": 2840002560
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014187878787878789,
      "loss": 1.575,
      "theoretical_loss": 3.3348593229321466,
      "tokens_seen": 2840068096
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014185858585858584,
      "loss": 1.508,
      "theoretical_loss": 3.334853340256652,
      "tokens_seen": 2840133632
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014183838383838385,
      "loss": 1.3082,
      "theoretical_loss": 3.334847357757859,
      "tokens_seen": 2840199168
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001418181818181818,
      "loss": 1.3049,
      "theoretical_loss": 3.334841375435758,
      "tokens_seen": 2840264704
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014179797979797982,
      "loss": 1.3032,
      "theoretical_loss": 3.33483539329034,
      "tokens_seen": 2840330240
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014177777777777778,
      "loss": 1.5375,
      "theoretical_loss": 3.334829411321595,
      "tokens_seen": 2840395776
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014175757575757576,
      "loss": 1.3666,
      "theoretical_loss": 3.3348234295295147,
      "tokens_seen": 2840461312
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014173737373737375,
      "loss": 1.4989,
      "theoretical_loss": 3.3348174479140895,
      "tokens_seen": 2840526848
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001417171717171717,
      "loss": 1.3705,
      "theoretical_loss": 3.33481146647531,
      "tokens_seen": 2840592384
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014169696969696971,
      "loss": 1.3963,
      "theoretical_loss": 3.3348054852131668,
      "tokens_seen": 2840657920
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014167676767676767,
      "loss": 1.4776,
      "theoretical_loss": 3.3347995041276506,
      "tokens_seen": 2840723456
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014165656565656566,
      "loss": 1.2202,
      "theoretical_loss": 3.334793523218752,
      "tokens_seen": 2840788992
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014163636363636364,
      "loss": 1.5552,
      "theoretical_loss": 3.334787542486463,
      "tokens_seen": 2840854528
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014161616161616162,
      "loss": 1.4102,
      "theoretical_loss": 3.3347815619307726,
      "tokens_seen": 2840920064
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001415959595959596,
      "loss": 1.5221,
      "theoretical_loss": 3.3347755815516718,
      "tokens_seen": 2840985600
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001415757575757576,
      "loss": 1.3518,
      "theoretical_loss": 3.334769601349152,
      "tokens_seen": 2841051136
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014155555555555555,
      "loss": 1.5227,
      "theoretical_loss": 3.3347636213232037,
      "tokens_seen": 2841116672
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014153535353535353,
      "loss": 1.5615,
      "theoretical_loss": 3.3347576414738174,
      "tokens_seen": 2841182208
    },
    {
      "epoch": 0.72,
      "objective/train/aligned_token_count": 0.564393937587738,
      "objective/train/avg_token_score": 0.008248298428952694,
      "objective/train/docs_used": 1598724,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7315202951431274,
      "objective/train/original_loss": 3.9295060634613037,
      "objective/train/theoretical_loss": 3.334751661800984,
      "objective/train/tokens_used": 1200106976,
      "objective/train/unlikelihood_loss": 0.040636446326971054,
      "theoretical_loss": 3.334751661800984,
      "tokens_seen": 2841247744
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014151515151515152,
      "loss": 1.4795,
      "theoretical_loss": 3.334751661800984,
      "tokens_seen": 2841247744
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001414949494949495,
      "loss": 1.4696,
      "theoretical_loss": 3.334745682304694,
      "tokens_seen": 2841313280
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014147474747474748,
      "loss": 1.2054,
      "theoretical_loss": 3.334739702984938,
      "tokens_seen": 2841378816
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014145454545454544,
      "loss": 1.2119,
      "theoretical_loss": 3.3347337238417074,
      "tokens_seen": 2841444352
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014143434343434345,
      "loss": 1.2222,
      "theoretical_loss": 3.3347277448749924,
      "tokens_seen": 2841509888
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001414141414141414,
      "loss": 1.3281,
      "theoretical_loss": 3.3347217660847837,
      "tokens_seen": 2841575424
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014139393939393942,
      "loss": 1.3561,
      "theoretical_loss": 3.334715787471072,
      "tokens_seen": 2841640960
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014137373737373738,
      "loss": 1.4208,
      "theoretical_loss": 3.3347098090338485,
      "tokens_seen": 2841706496
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014135353535353533,
      "loss": 1.6069,
      "theoretical_loss": 3.3347038307731034,
      "tokens_seen": 2841772032
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014133333333333334,
      "loss": 1.3665,
      "theoretical_loss": 3.3346978526888273,
      "tokens_seen": 2841837568
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001413131313131313,
      "loss": 1.3347,
      "theoretical_loss": 3.3346918747810115,
      "tokens_seen": 2841903104
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001412929292929293,
      "loss": 1.5192,
      "theoretical_loss": 3.3346858970496465,
      "tokens_seen": 2841968640
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014127272727272727,
      "loss": 1.4876,
      "theoretical_loss": 3.3346799194947225,
      "tokens_seen": 2842034176
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014125252525252525,
      "loss": 1.4171,
      "theoretical_loss": 3.334673942116231,
      "tokens_seen": 2842099712
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014123232323232324,
      "loss": 1.3254,
      "theoretical_loss": 3.3346679649141624,
      "tokens_seen": 2842165248
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014121212121212122,
      "loss": 1.4117,
      "theoretical_loss": 3.3346619878885075,
      "tokens_seen": 2842230784
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001411919191919192,
      "loss": 1.4987,
      "theoretical_loss": 3.3346560110392565,
      "tokens_seen": 2842296320
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014117171717171716,
      "loss": 1.2767,
      "theoretical_loss": 3.334650034366401,
      "tokens_seen": 2842361856
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014115151515151515,
      "loss": 1.2381,
      "theoretical_loss": 3.334644057869931,
      "tokens_seen": 2842427392
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014113131313131313,
      "loss": 1.5232,
      "theoretical_loss": 3.334638081549838,
      "tokens_seen": 2842492928
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014111111111111111,
      "loss": 1.2853,
      "theoretical_loss": 3.334632105406112,
      "tokens_seen": 2842558464
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001410909090909091,
      "loss": 1.402,
      "theoretical_loss": 3.334626129438744,
      "tokens_seen": 2842624000
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014107070707070708,
      "loss": 1.2657,
      "theoretical_loss": 3.3346201536477245,
      "tokens_seen": 2842689536
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014105050505050504,
      "loss": 1.4889,
      "theoretical_loss": 3.3346141780330445,
      "tokens_seen": 2842755072
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014103030303030305,
      "loss": 1.4427,
      "theoretical_loss": 3.334608202594695,
      "tokens_seen": 2842820608
    },
    {
      "epoch": 0.72,
      "objective/train/aligned_token_count": 0.38086509704589844,
      "objective/train/avg_token_score": 0.027381666004657745,
      "objective/train/docs_used": 1599978,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2302736043930054,
      "objective/train/original_loss": 5.1902923583984375,
      "objective/train/theoretical_loss": 3.334602227332666,
      "objective/train/tokens_used": 1201745376,
      "objective/train/unlikelihood_loss": 0.029726793989539146,
      "theoretical_loss": 3.334602227332666,
      "tokens_seen": 2842886144
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000141010101010101,
      "loss": 1.5439,
      "theoretical_loss": 3.334602227332666,
      "tokens_seen": 2842886144
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000140989898989899,
      "loss": 1.4553,
      "theoretical_loss": 3.334596252246949,
      "tokens_seen": 2842951680
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014096969696969697,
      "loss": 1.4114,
      "theoretical_loss": 3.3345902773375338,
      "tokens_seen": 2843017216
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014094949494949493,
      "loss": 1.3102,
      "theoretical_loss": 3.334584302604412,
      "tokens_seen": 2843082752
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014092929292929294,
      "loss": 1.2929,
      "theoretical_loss": 3.3345783280475745,
      "tokens_seen": 2843148288
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001409090909090909,
      "loss": 1.5343,
      "theoretical_loss": 3.334572353667011,
      "tokens_seen": 2843213824
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001408888888888889,
      "loss": 1.3327,
      "theoretical_loss": 3.334566379462713,
      "tokens_seen": 2843279360
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014086868686868687,
      "loss": 1.3107,
      "theoretical_loss": 3.334560405434671,
      "tokens_seen": 2843344896
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014084848484848488,
      "loss": 1.3141,
      "theoretical_loss": 3.334554431582876,
      "tokens_seen": 2843410432
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014082828282828284,
      "loss": 1.423,
      "theoretical_loss": 3.334548457907318,
      "tokens_seen": 2843475968
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001408080808080808,
      "loss": 1.3414,
      "theoretical_loss": 3.3345424844079887,
      "tokens_seen": 2843541504
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001407878787878788,
      "loss": 1.3768,
      "theoretical_loss": 3.3345365110848784,
      "tokens_seen": 2843607040
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014076767676767676,
      "loss": 1.5637,
      "theoretical_loss": 3.3345305379379777,
      "tokens_seen": 2843672576
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014074747474747477,
      "loss": 1.3948,
      "theoretical_loss": 3.334524564967278,
      "tokens_seen": 2843738112
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014072727272727273,
      "loss": 1.585,
      "theoretical_loss": 3.3345185921727687,
      "tokens_seen": 2843803648
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001407070707070707,
      "loss": 1.3812,
      "theoretical_loss": 3.3345126195544417,
      "tokens_seen": 2843869184
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001406868686868687,
      "loss": 1.3359,
      "theoretical_loss": 3.3345066471122875,
      "tokens_seen": 2843934720
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014066666666666665,
      "loss": 1.3431,
      "theoretical_loss": 3.334500674846297,
      "tokens_seen": 2844000256
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014064646464646466,
      "loss": 1.481,
      "theoretical_loss": 3.3344947027564604,
      "tokens_seen": 2844065792
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014062626262626262,
      "loss": 1.576,
      "theoretical_loss": 3.334488730842769,
      "tokens_seen": 2844131328
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001406060606060606,
      "loss": 1.4213,
      "theoretical_loss": 3.334482759105213,
      "tokens_seen": 2844196864
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001405858585858586,
      "loss": 1.7112,
      "theoretical_loss": 3.3344767875437835,
      "tokens_seen": 2844262400
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014056565656565657,
      "loss": 1.4548,
      "theoretical_loss": 3.334470816158471,
      "tokens_seen": 2844327936
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014054545454545456,
      "loss": 1.4886,
      "theoretical_loss": 3.3344648449492666,
      "tokens_seen": 2844393472
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014052525252525254,
      "loss": 1.5554,
      "theoretical_loss": 3.334458873916161,
      "tokens_seen": 2844459008
    },
    {
      "epoch": 0.72,
      "objective/train/aligned_token_count": 0.5327468514442444,
      "objective/train/avg_token_score": 0.007825806736946106,
      "objective/train/docs_used": 1601259,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6141077280044556,
      "objective/train/original_loss": 3.750728130340576,
      "objective/train/theoretical_loss": 3.3344529030591445,
      "objective/train/tokens_used": 1203383776,
      "objective/train/unlikelihood_loss": 0.09291576594114304,
      "theoretical_loss": 3.3344529030591445,
      "tokens_seen": 2844524544
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001405050505050505,
      "loss": 1.5557,
      "theoretical_loss": 3.3344529030591445,
      "tokens_seen": 2844524544
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001405050505050505,
      "loss": 1.2881,
      "theoretical_loss": 3.3344469323782087,
      "tokens_seen": 2844590080
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014048484848484848,
      "loss": 1.4262,
      "theoretical_loss": 3.3344409618733435,
      "tokens_seen": 2844655616
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014046464646464647,
      "loss": 1.3784,
      "theoretical_loss": 3.33443499154454,
      "tokens_seen": 2844721152
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014044444444444445,
      "loss": 1.4438,
      "theoretical_loss": 3.334429021391789,
      "tokens_seen": 2844786688
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014042424242424243,
      "loss": 1.3761,
      "theoretical_loss": 3.334423051415081,
      "tokens_seen": 2844852224
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001404040404040404,
      "loss": 1.5569,
      "theoretical_loss": 3.3344170816144074,
      "tokens_seen": 2844917760
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001403838383838384,
      "loss": 1.3698,
      "theoretical_loss": 3.334411111989758,
      "tokens_seen": 2844983296
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014036363636363636,
      "loss": 1.3072,
      "theoretical_loss": 3.3344051425411245,
      "tokens_seen": 2845048832
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014034343434343437,
      "loss": 1.4238,
      "theoretical_loss": 3.3343991732684968,
      "tokens_seen": 2845114368
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014032323232323233,
      "loss": 1.2105,
      "theoretical_loss": 3.3343932041718665,
      "tokens_seen": 2845179904
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014030303030303028,
      "loss": 1.3845,
      "theoretical_loss": 3.3343872352512234,
      "tokens_seen": 2845245440
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001402828282828283,
      "loss": 1.2971,
      "theoretical_loss": 3.3343812665065595,
      "tokens_seen": 2845310976
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014026262626262625,
      "loss": 1.4427,
      "theoretical_loss": 3.3343752979378642,
      "tokens_seen": 2845376512
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014024242424242426,
      "loss": 1.4497,
      "theoretical_loss": 3.334369329545129,
      "tokens_seen": 2845442048
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014022222222222222,
      "loss": 1.4583,
      "theoretical_loss": 3.3343633613283448,
      "tokens_seen": 2845507584
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001402020202020202,
      "loss": 1.3487,
      "theoretical_loss": 3.334357393287502,
      "tokens_seen": 2845573120
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014018181818181819,
      "loss": 1.3922,
      "theoretical_loss": 3.334351425422591,
      "tokens_seen": 2845638656
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014016161616161614,
      "loss": 1.5007,
      "theoretical_loss": 3.334345457733604,
      "tokens_seen": 2845704192
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014014141414141415,
      "loss": 1.3691,
      "theoretical_loss": 3.3343394902205303,
      "tokens_seen": 2845769728
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001401212121212121,
      "loss": 1.2984,
      "theoretical_loss": 3.334333522883361,
      "tokens_seen": 2845835264
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001401010101010101,
      "loss": 1.375,
      "theoretical_loss": 3.3343275557220875,
      "tokens_seen": 2845900800
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014008080808080808,
      "loss": 1.3575,
      "theoretical_loss": 3.3343215887366995,
      "tokens_seen": 2845966336
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014006060606060606,
      "loss": 1.4387,
      "theoretical_loss": 3.334315621927189,
      "tokens_seen": 2846031872
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014004040404040405,
      "loss": 1.562,
      "theoretical_loss": 3.334309655293546,
      "tokens_seen": 2846097408
    },
    {
      "epoch": 0.72,
      "objective/train/aligned_token_count": 0.4471529722213745,
      "objective/train/avg_token_score": 0.020442679524421692,
      "objective/train/docs_used": 1601933,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5117449760437012,
      "objective/train/original_loss": 4.455848693847656,
      "objective/train/theoretical_loss": 3.334303688835761,
      "objective/train/tokens_used": 1205022176,
      "objective/train/unlikelihood_loss": 0.06131961569190025,
      "theoretical_loss": 3.334303688835761,
      "tokens_seen": 2846162944
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014002020202020203,
      "loss": 1.3684,
      "theoretical_loss": 3.334303688835761,
      "tokens_seen": 2846162944
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00014000000000000001,
      "loss": 1.3533,
      "theoretical_loss": 3.334297722553825,
      "tokens_seen": 2846228480
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013997979797979797,
      "loss": 1.4447,
      "theoretical_loss": 3.3342917564477297,
      "tokens_seen": 2846294016
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013995959595959596,
      "loss": 1.4639,
      "theoretical_loss": 3.3342857905174648,
      "tokens_seen": 2846359552
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013993939393939394,
      "loss": 1.4556,
      "theoretical_loss": 3.334279824763021,
      "tokens_seen": 2846425088
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013991919191919192,
      "loss": 1.307,
      "theoretical_loss": 3.33427385918439,
      "tokens_seen": 2846490624
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001398989898989899,
      "loss": 1.3066,
      "theoretical_loss": 3.3342678937815617,
      "tokens_seen": 2846556160
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001398787878787879,
      "loss": 0.9397,
      "theoretical_loss": 3.334261928554527,
      "tokens_seen": 2846621696
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013985858585858585,
      "loss": 1.2422,
      "theoretical_loss": 3.3342559635032774,
      "tokens_seen": 2846687232
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013983838383838386,
      "loss": 1.4959,
      "theoretical_loss": 3.3342499986278025,
      "tokens_seen": 2846752768
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013981818181818182,
      "loss": 1.2207,
      "theoretical_loss": 3.334244033928094,
      "tokens_seen": 2846818304
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001397979797979798,
      "loss": 1.4675,
      "theoretical_loss": 3.334238069404143,
      "tokens_seen": 2846883840
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013977777777777778,
      "loss": 1.4771,
      "theoretical_loss": 3.3342321050559387,
      "tokens_seen": 2846949376
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013975757575757574,
      "loss": 1.443,
      "theoretical_loss": 3.334226140883473,
      "tokens_seen": 2847014912
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013973737373737375,
      "loss": 1.4907,
      "theoretical_loss": 3.334220176886737,
      "tokens_seen": 2847080448
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001397171717171717,
      "loss": 1.4411,
      "theoretical_loss": 3.334214213065721,
      "tokens_seen": 2847145984
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013969696969696972,
      "loss": 1.1973,
      "theoretical_loss": 3.3342082494204153,
      "tokens_seen": 2847211520
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013967676767676768,
      "loss": 1.4608,
      "theoretical_loss": 3.334202285950811,
      "tokens_seen": 2847277056
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013965656565656566,
      "loss": 1.3749,
      "theoretical_loss": 3.3341963226568994,
      "tokens_seen": 2847342592
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013963636363636364,
      "loss": 1.668,
      "theoretical_loss": 3.334190359538671,
      "tokens_seen": 2847408128
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001396161616161616,
      "loss": 1.4399,
      "theoretical_loss": 3.3341843965961164,
      "tokens_seen": 2847473664
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001395959595959596,
      "loss": 1.0869,
      "theoretical_loss": 3.3341784338292264,
      "tokens_seen": 2847539200
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013957575757575757,
      "loss": 1.1503,
      "theoretical_loss": 3.3341724712379914,
      "tokens_seen": 2847604736
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013955555555555555,
      "loss": 1.5041,
      "theoretical_loss": 3.3341665088224035,
      "tokens_seen": 2847670272
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013953535353535354,
      "loss": 1.2537,
      "theoretical_loss": 3.334160546582452,
      "tokens_seen": 2847735808
    },
    {
      "epoch": 0.72,
      "objective/train/aligned_token_count": 0.2728494703769684,
      "objective/train/avg_token_score": 0.013079564087092876,
      "objective/train/docs_used": 1603273,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.7297566533088684,
      "objective/train/original_loss": 4.730753421783447,
      "objective/train/theoretical_loss": 3.3341545845181284,
      "objective/train/tokens_used": 1206660576,
      "objective/train/unlikelihood_loss": 0.06316656619310379,
      "theoretical_loss": 3.3341545845181284,
      "tokens_seen": 2847801344
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013951515151515152,
      "loss": 1.0734,
      "theoretical_loss": 3.3341545845181284,
      "tokens_seen": 2847801344
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001394949494949495,
      "loss": 1.3352,
      "theoretical_loss": 3.3341486226294235,
      "tokens_seen": 2847866880
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001394747474747475,
      "loss": 1.3653,
      "theoretical_loss": 3.334142660916328,
      "tokens_seen": 2847932416
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013945454545454545,
      "loss": 1.2219,
      "theoretical_loss": 3.3341366993788326,
      "tokens_seen": 2847997952
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013943434343434343,
      "loss": 1.0735,
      "theoretical_loss": 3.334130738016928,
      "tokens_seen": 2848063488
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013941414141414141,
      "loss": 1.2599,
      "theoretical_loss": 3.3341247768306053,
      "tokens_seen": 2848129024
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001393939393939394,
      "loss": 1.6148,
      "theoretical_loss": 3.334118815819855,
      "tokens_seen": 2848194560
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013937373737373738,
      "loss": 1.3421,
      "theoretical_loss": 3.334112854984668,
      "tokens_seen": 2848260096
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013935353535353534,
      "loss": 1.4014,
      "theoretical_loss": 3.3341068943250356,
      "tokens_seen": 2848325632
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013933333333333335,
      "loss": 1.262,
      "theoretical_loss": 3.3341009338409475,
      "tokens_seen": 2848391168
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001393131313131313,
      "loss": 1.4435,
      "theoretical_loss": 3.334094973532395,
      "tokens_seen": 2848456704
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013929292929292932,
      "loss": 1.4013,
      "theoretical_loss": 3.3340890133993693,
      "tokens_seen": 2848522240
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013927272727272727,
      "loss": 1.5277,
      "theoretical_loss": 3.334083053441861,
      "tokens_seen": 2848587776
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013925252525252523,
      "loss": 1.3035,
      "theoretical_loss": 3.3340770936598605,
      "tokens_seen": 2848653312
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013923232323232324,
      "loss": 1.4822,
      "theoretical_loss": 3.334071134053359,
      "tokens_seen": 2848718848
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001392121212121212,
      "loss": 1.5706,
      "theoretical_loss": 3.3340651746223466,
      "tokens_seen": 2848784384
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001391919191919192,
      "loss": 1.3666,
      "theoretical_loss": 3.3340592153668154,
      "tokens_seen": 2848849920
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013917171717171717,
      "loss": 1.5048,
      "theoretical_loss": 3.3340532562867553,
      "tokens_seen": 2848915456
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013915151515151518,
      "loss": 1.2486,
      "theoretical_loss": 3.334047297382157,
      "tokens_seen": 2848980992
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013913131313131314,
      "loss": 1.5217,
      "theoretical_loss": 3.3340413386530114,
      "tokens_seen": 2849046528
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001391111111111111,
      "loss": 1.3773,
      "theoretical_loss": 3.33403538009931,
      "tokens_seen": 2849112064
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0001390909090909091,
      "loss": 1.4906,
      "theoretical_loss": 3.334029421721042,
      "tokens_seen": 2849177600
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013907070707070706,
      "loss": 1.4511,
      "theoretical_loss": 3.3340234635182004,
      "tokens_seen": 2849243136
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013905050505050507,
      "loss": 1.491,
      "theoretical_loss": 3.3340175054907744,
      "tokens_seen": 2849308672
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00013903030303030303,
      "loss": 1.4118,
      "theoretical_loss": 3.334011547638755,
      "tokens_seen": 2849374208
    },
    {
      "epoch": 0.72,
      "objective/train/aligned_token_count": 0.5618890523910522,
      "objective/train/avg_token_score": 0.006466910243034363,
      "objective/train/docs_used": 1603897,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7869397401809692,
      "objective/train/original_loss": 3.919118642807007,
      "objective/train/theoretical_loss": 3.3340055899621333,
      "objective/train/tokens_used": 1208298976,
      "objective/train/unlikelihood_loss": 0.05047599598765373,
      "theoretical_loss": 3.3340055899621333,
      "tokens_seen": 2849439744
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.000139010101010101,
      "loss": 1.5805,
      "theoretical_loss": 3.3340055899621333,
      "tokens_seen": 2849439744
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000138989898989899,
      "loss": 1.288,
      "theoretical_loss": 3.3339996324609,
      "tokens_seen": 2849505280
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013896969696969698,
      "loss": 1.3559,
      "theoretical_loss": 3.3339936751350465,
      "tokens_seen": 2849570816
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013894949494949496,
      "loss": 1.3093,
      "theoretical_loss": 3.3339877179845625,
      "tokens_seen": 2849636352
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013892929292929292,
      "loss": 1.635,
      "theoretical_loss": 3.3339817610094395,
      "tokens_seen": 2849701888
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001389090909090909,
      "loss": 1.2486,
      "theoretical_loss": 3.333975804209668,
      "tokens_seen": 2849767424
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001388888888888889,
      "loss": 1.4103,
      "theoretical_loss": 3.333969847585239,
      "tokens_seen": 2849832960
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013886868686868687,
      "loss": 1.3582,
      "theoretical_loss": 3.3339638911361433,
      "tokens_seen": 2849898496
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013884848484848486,
      "loss": 1.1812,
      "theoretical_loss": 3.333957934862372,
      "tokens_seen": 2849964032
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013882828282828284,
      "loss": 1.3836,
      "theoretical_loss": 3.333951978763915,
      "tokens_seen": 2850029568
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001388080808080808,
      "loss": 1.6113,
      "theoretical_loss": 3.333946022840764,
      "tokens_seen": 2850095104
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001387878787878788,
      "loss": 1.3925,
      "theoretical_loss": 3.3339400670929096,
      "tokens_seen": 2850160640
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013876767676767677,
      "loss": 1.4926,
      "theoretical_loss": 3.3339341115203425,
      "tokens_seen": 2850226176
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013874747474747475,
      "loss": 1.4349,
      "theoretical_loss": 3.3339281561230534,
      "tokens_seen": 2850291712
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013872727272727273,
      "loss": 1.4202,
      "theoretical_loss": 3.333922200901033,
      "tokens_seen": 2850357248
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001387070707070707,
      "loss": 1.3762,
      "theoretical_loss": 3.333916245854273,
      "tokens_seen": 2850422784
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001386868686868687,
      "loss": 1.3919,
      "theoretical_loss": 3.3339102909827627,
      "tokens_seen": 2850488320
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013866666666666666,
      "loss": 1.3801,
      "theoretical_loss": 3.333904336286494,
      "tokens_seen": 2850553856
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013864646464646467,
      "loss": 1.2305,
      "theoretical_loss": 3.333898381765458,
      "tokens_seen": 2850619392
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013862626262626263,
      "loss": 1.5273,
      "theoretical_loss": 3.3338924274196446,
      "tokens_seen": 2850684928
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001386060606060606,
      "loss": 1.3969,
      "theoretical_loss": 3.3338864732490454,
      "tokens_seen": 2850750464
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001385858585858586,
      "loss": 1.4875,
      "theoretical_loss": 3.3338805192536505,
      "tokens_seen": 2850816000
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013856565656565655,
      "loss": 1.388,
      "theoretical_loss": 3.333874565433451,
      "tokens_seen": 2850881536
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013854545454545456,
      "loss": 1.4704,
      "theoretical_loss": 3.333868611788438,
      "tokens_seen": 2850947072
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013852525252525252,
      "loss": 1.4214,
      "theoretical_loss": 3.3338626583186017,
      "tokens_seen": 2851012608
    },
    {
      "epoch": 0.73,
      "objective/train/aligned_token_count": 0.5618890523910522,
      "objective/train/avg_token_score": 0.023772623389959335,
      "objective/train/docs_used": 1605267,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8002686500549316,
      "objective/train/original_loss": 4.342200756072998,
      "objective/train/theoretical_loss": 3.333856705023934,
      "objective/train/tokens_used": 1209937376,
      "objective/train/unlikelihood_loss": 0.07499855756759644,
      "theoretical_loss": 3.333856705023934,
      "tokens_seen": 2851078144
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001385050505050505,
      "loss": 1.5487,
      "theoretical_loss": 3.333856705023934,
      "tokens_seen": 2851078144
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001384848484848485,
      "loss": 1.4944,
      "theoretical_loss": 3.333850751904424,
      "tokens_seen": 2851143680
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013846464646464647,
      "loss": 1.4193,
      "theoretical_loss": 3.3338447989600644,
      "tokens_seen": 2851209216
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013844444444444445,
      "loss": 1.2987,
      "theoretical_loss": 3.333838846190845,
      "tokens_seen": 2851274752
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001384242424242424,
      "loss": 1.4802,
      "theoretical_loss": 3.3338328935967567,
      "tokens_seen": 2851340288
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001384040404040404,
      "loss": 1.3267,
      "theoretical_loss": 3.33382694117779,
      "tokens_seen": 2851405824
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013838383838383838,
      "loss": 1.3333,
      "theoretical_loss": 3.3338209889339367,
      "tokens_seen": 2851471360
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013836363636363636,
      "loss": 1.317,
      "theoretical_loss": 3.333815036865187,
      "tokens_seen": 2851536896
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013834343434343435,
      "loss": 1.6755,
      "theoretical_loss": 3.3338090849715316,
      "tokens_seen": 2851602432
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013832323232323233,
      "loss": 1.4427,
      "theoretical_loss": 3.333803133252961,
      "tokens_seen": 2851667968
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013830303030303032,
      "loss": 1.4273,
      "theoretical_loss": 3.3337971817094676,
      "tokens_seen": 2851733504
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001382828282828283,
      "loss": 1.5208,
      "theoretical_loss": 3.3337912303410406,
      "tokens_seen": 2851799040
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013826262626262626,
      "loss": 1.3241,
      "theoretical_loss": 3.3337852791476714,
      "tokens_seen": 2851864576
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013824242424242424,
      "loss": 1.4649,
      "theoretical_loss": 3.3337793281293506,
      "tokens_seen": 2851930112
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013822222222222222,
      "loss": 1.1908,
      "theoretical_loss": 3.33377337728607,
      "tokens_seen": 2851995648
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001382020202020202,
      "loss": 1.3215,
      "theoretical_loss": 3.3337674266178188,
      "tokens_seen": 2852061184
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001381818181818182,
      "loss": 1.2244,
      "theoretical_loss": 3.333761476124589,
      "tokens_seen": 2852126720
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013816161616161615,
      "loss": 1.3845,
      "theoretical_loss": 3.3337555258063714,
      "tokens_seen": 2852192256
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013814141414141416,
      "loss": 1.3487,
      "theoretical_loss": 3.3337495756631563,
      "tokens_seen": 2852257792
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013812121212121212,
      "loss": 1.4117,
      "theoretical_loss": 3.3337436256949347,
      "tokens_seen": 2852323328
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013810101010101013,
      "loss": 1.3842,
      "theoretical_loss": 3.3337376759016975,
      "tokens_seen": 2852388864
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013808080808080808,
      "loss": 1.4457,
      "theoretical_loss": 3.333731726283436,
      "tokens_seen": 2852454400
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013806060606060604,
      "loss": 1.2958,
      "theoretical_loss": 3.33372577684014,
      "tokens_seen": 2852519936
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013804040404040405,
      "loss": 1.3959,
      "theoretical_loss": 3.3337198275718016,
      "tokens_seen": 2852585472
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000138020202020202,
      "loss": 1.4273,
      "theoretical_loss": 3.3337138784784104,
      "tokens_seen": 2852651008
    },
    {
      "epoch": 0.73,
      "objective/train/aligned_token_count": 0.4049975574016571,
      "objective/train/avg_token_score": 0.023587264120578766,
      "objective/train/docs_used": 1605428,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3373327255249023,
      "objective/train/original_loss": 4.581187725067139,
      "objective/train/theoretical_loss": 3.3337079295599583,
      "objective/train/tokens_used": 1211575776,
      "objective/train/unlikelihood_loss": 0.05914934352040291,
      "theoretical_loss": 3.3337079295599583,
      "tokens_seen": 2852716544
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013800000000000002,
      "loss": 1.3684,
      "theoretical_loss": 3.3337079295599583,
      "tokens_seen": 2852716544
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013797979797979798,
      "loss": 1.5799,
      "theoretical_loss": 3.333701980816435,
      "tokens_seen": 2852782080
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013795959595959596,
      "loss": 1.4151,
      "theoretical_loss": 3.3336960322478326,
      "tokens_seen": 2852847616
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013793939393939395,
      "loss": 1.3734,
      "theoretical_loss": 3.3336900838541412,
      "tokens_seen": 2852913152
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013791919191919193,
      "loss": 1.2902,
      "theoretical_loss": 3.3336841356353517,
      "tokens_seen": 2852978688
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001378989898989899,
      "loss": 1.3696,
      "theoretical_loss": 3.3336781875914547,
      "tokens_seen": 2853044224
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013787878787878787,
      "loss": 1.1776,
      "theoretical_loss": 3.333672239722442,
      "tokens_seen": 2853109760
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013785858585858585,
      "loss": 1.5409,
      "theoretical_loss": 3.333666292028303,
      "tokens_seen": 2853175296
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013783838383838384,
      "loss": 1.2602,
      "theoretical_loss": 3.33366034450903,
      "tokens_seen": 2853240832
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013781818181818182,
      "loss": 1.424,
      "theoretical_loss": 3.3336543971646124,
      "tokens_seen": 2853306368
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001377979797979798,
      "loss": 1.3797,
      "theoretical_loss": 3.3336484499950423,
      "tokens_seen": 2853371904
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001377777777777778,
      "loss": 1.4111,
      "theoretical_loss": 3.33364250300031,
      "tokens_seen": 2853437440
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013775757575757575,
      "loss": 1.357,
      "theoretical_loss": 3.3336365561804064,
      "tokens_seen": 2853502976
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013773737373737376,
      "loss": 1.249,
      "theoretical_loss": 3.333630609535322,
      "tokens_seen": 2853568512
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013771717171717171,
      "loss": 1.4024,
      "theoretical_loss": 3.3336246630650486,
      "tokens_seen": 2853634048
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001376969696969697,
      "loss": 1.2364,
      "theoretical_loss": 3.333618716769576,
      "tokens_seen": 2853699584
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013767676767676768,
      "loss": 1.306,
      "theoretical_loss": 3.333612770648896,
      "tokens_seen": 2853765120
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013765656565656564,
      "loss": 1.2849,
      "theoretical_loss": 3.3336068247029984,
      "tokens_seen": 2853830656
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013763636363636365,
      "loss": 1.4179,
      "theoretical_loss": 3.3336008789318745,
      "tokens_seen": 2853896192
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001376161616161616,
      "loss": 1.4963,
      "theoretical_loss": 3.3335949333355157,
      "tokens_seen": 2853961728
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013759595959595962,
      "loss": 1.4919,
      "theoretical_loss": 3.3335889879139122,
      "tokens_seen": 2854027264
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013757575757575758,
      "loss": 1.5105,
      "theoretical_loss": 3.3335830426670547,
      "tokens_seen": 2854092800
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013755555555555553,
      "loss": 1.7012,
      "theoretical_loss": 3.333577097594935,
      "tokens_seen": 2854158336
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013753535353535354,
      "loss": 1.3101,
      "theoretical_loss": 3.3335711526975427,
      "tokens_seen": 2854223872
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001375151515151515,
      "loss": 1.6195,
      "theoretical_loss": 3.3335652079748694,
      "tokens_seen": 2854289408
    },
    {
      "epoch": 0.73,
      "objective/train/aligned_token_count": 0.5020161271095276,
      "objective/train/avg_token_score": 0.028967617079615593,
      "objective/train/docs_used": 1606174,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5418790578842163,
      "objective/train/original_loss": 4.033195495605469,
      "objective/train/theoretical_loss": 3.3335592634269062,
      "objective/train/tokens_used": 1213214176,
      "objective/train/unlikelihood_loss": 0.07528182119131088,
      "theoretical_loss": 3.3335592634269062,
      "tokens_seen": 2854354944
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001374949494949495,
      "loss": 1.5791,
      "theoretical_loss": 3.3335592634269062,
      "tokens_seen": 2854354944
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013747474747474747,
      "loss": 1.2028,
      "theoretical_loss": 3.3335533190536433,
      "tokens_seen": 2854420480
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013745454545454548,
      "loss": 1.2133,
      "theoretical_loss": 3.333547374855072,
      "tokens_seen": 2854486016
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013743434343434344,
      "loss": 1.2906,
      "theoretical_loss": 3.333541430831183,
      "tokens_seen": 2854551552
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013741414141414142,
      "loss": 1.3195,
      "theoretical_loss": 3.333535486981967,
      "tokens_seen": 2854617088
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001373939393939394,
      "loss": 1.4484,
      "theoretical_loss": 3.333529543307415,
      "tokens_seen": 2854682624
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013737373737373736,
      "loss": 1.3672,
      "theoretical_loss": 3.333523599807518,
      "tokens_seen": 2854748160
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013735353535353537,
      "loss": 1.3261,
      "theoretical_loss": 3.333517656482267,
      "tokens_seen": 2854813696
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013733333333333333,
      "loss": 1.4179,
      "theoretical_loss": 3.333511713331652,
      "tokens_seen": 2854879232
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001373131313131313,
      "loss": 1.4899,
      "theoretical_loss": 3.3335057703556648,
      "tokens_seen": 2854944768
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001372929292929293,
      "loss": 1.4352,
      "theoretical_loss": 3.333499827554296,
      "tokens_seen": 2855010304
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013727272727272728,
      "loss": 1.3943,
      "theoretical_loss": 3.3334938849275364,
      "tokens_seen": 2855075840
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013725252525252526,
      "loss": 1.7579,
      "theoretical_loss": 3.3334879424753767,
      "tokens_seen": 2855141376
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013723232323232325,
      "loss": 1.421,
      "theoretical_loss": 3.3334820001978076,
      "tokens_seen": 2855206912
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001372121212121212,
      "loss": 1.2087,
      "theoretical_loss": 3.3334760580948206,
      "tokens_seen": 2855272448
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001371919191919192,
      "loss": 1.2964,
      "theoretical_loss": 3.333470116166406,
      "tokens_seen": 2855337984
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013717171717171717,
      "loss": 1.4189,
      "theoretical_loss": 3.3334641744125553,
      "tokens_seen": 2855403520
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013715151515151516,
      "loss": 1.4826,
      "theoretical_loss": 3.3334582328332587,
      "tokens_seen": 2855469056
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013713131313131314,
      "loss": 1.2041,
      "theoretical_loss": 3.3334522914285074,
      "tokens_seen": 2855534592
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001371111111111111,
      "loss": 1.374,
      "theoretical_loss": 3.333446350198292,
      "tokens_seen": 2855600128
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001370909090909091,
      "loss": 1.395,
      "theoretical_loss": 3.3334404091426038,
      "tokens_seen": 2855665664
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013707070707070707,
      "loss": 1.4105,
      "theoretical_loss": 3.333434468261433,
      "tokens_seen": 2855731200
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013705050505050508,
      "loss": 1.2937,
      "theoretical_loss": 3.333428527554771,
      "tokens_seen": 2855796736
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013703030303030303,
      "loss": 1.4094,
      "theoretical_loss": 3.333422587022609,
      "tokens_seen": 2855862272
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000137010101010101,
      "loss": 1.3973,
      "theoretical_loss": 3.3334166466649373,
      "tokens_seen": 2855927808
    },
    {
      "epoch": 0.73,
      "objective/train/aligned_token_count": 0.3335166275501251,
      "objective/train/avg_token_score": 0.01644147001206875,
      "objective/train/docs_used": 1607592,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9303444027900696,
      "objective/train/original_loss": 5.3227057456970215,
      "objective/train/theoretical_loss": 3.3334107064817466,
      "objective/train/tokens_used": 1214852576,
      "objective/train/unlikelihood_loss": 0.036250948905944824,
      "theoretical_loss": 3.3334107064817466,
      "tokens_seen": 2855993344
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000136989898989899,
      "loss": 1.2834,
      "theoretical_loss": 3.3334107064817466,
      "tokens_seen": 2855993344
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013696969696969696,
      "loss": 1.3908,
      "theoretical_loss": 3.3334047664730284,
      "tokens_seen": 2856058880
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013694949494949497,
      "loss": 1.2682,
      "theoretical_loss": 3.333398826638773,
      "tokens_seen": 2856124416
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013692929292929293,
      "loss": 1.4762,
      "theoretical_loss": 3.3333928869789715,
      "tokens_seen": 2856189952
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001369090909090909,
      "loss": 1.3519,
      "theoretical_loss": 3.333386947493615,
      "tokens_seen": 2856255488
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001368888888888889,
      "loss": 1.3994,
      "theoretical_loss": 3.333381008182694,
      "tokens_seen": 2856321024
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013686868686868685,
      "loss": 1.3777,
      "theoretical_loss": 3.3333750690461996,
      "tokens_seen": 2856386560
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013684848484848486,
      "loss": 1.3845,
      "theoretical_loss": 3.3333691300841224,
      "tokens_seen": 2856452096
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013682828282828282,
      "loss": 1.1977,
      "theoretical_loss": 3.3333631912964536,
      "tokens_seen": 2856517632
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001368080808080808,
      "loss": 1.5064,
      "theoretical_loss": 3.333357252683184,
      "tokens_seen": 2856583168
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001367878787878788,
      "loss": 1.5204,
      "theoretical_loss": 3.3333513142443048,
      "tokens_seen": 2856648704
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013676767676767677,
      "loss": 1.4104,
      "theoretical_loss": 3.3333453759798064,
      "tokens_seen": 2856714240
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013674747474747475,
      "loss": 1.5232,
      "theoretical_loss": 3.3333394378896797,
      "tokens_seen": 2856779776
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013672727272727274,
      "loss": 1.4372,
      "theoretical_loss": 3.3333334999739157,
      "tokens_seen": 2856845312
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001367070707070707,
      "loss": 1.5149,
      "theoretical_loss": 3.333327562232505,
      "tokens_seen": 2856910848
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013668686868686868,
      "loss": 1.2948,
      "theoretical_loss": 3.333321624665439,
      "tokens_seen": 2856976384
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013666666666666666,
      "loss": 1.4238,
      "theoretical_loss": 3.3333156872727083,
      "tokens_seen": 2857041920
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013664646464646465,
      "loss": 1.3208,
      "theoretical_loss": 3.333309750054304,
      "tokens_seen": 2857107456
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013662626262626263,
      "loss": 1.133,
      "theoretical_loss": 3.3333038130102164,
      "tokens_seen": 2857172992
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013660606060606062,
      "loss": 1.3102,
      "theoretical_loss": 3.333297876140437,
      "tokens_seen": 2857238528
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001365858585858586,
      "loss": 1.2855,
      "theoretical_loss": 3.3332919394449565,
      "tokens_seen": 2857304064
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013656565656565656,
      "loss": 1.4219,
      "theoretical_loss": 3.3332860029237654,
      "tokens_seen": 2857369600
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013654545454545457,
      "loss": 1.4383,
      "theoretical_loss": 3.3332800665768554,
      "tokens_seen": 2857435136
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013652525252525252,
      "loss": 1.4091,
      "theoretical_loss": 3.3332741304042166,
      "tokens_seen": 2857500672
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001365050505050505,
      "loss": 1.4938,
      "theoretical_loss": 3.33326819440584,
      "tokens_seen": 2857566208
    },
    {
      "epoch": 0.73,
      "objective/train/aligned_token_count": 0.3348607122898102,
      "objective/train/avg_token_score": 0.004023551009595394,
      "objective/train/docs_used": 1608324,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0356080532073975,
      "objective/train/original_loss": 4.3689446449279785,
      "objective/train/theoretical_loss": 3.3332622585817173,
      "objective/train/tokens_used": 1216490976,
      "objective/train/unlikelihood_loss": 0.06199241057038307,
      "theoretical_loss": 3.3332622585817173,
      "tokens_seen": 2857631744
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001364848484848485,
      "loss": 1.4605,
      "theoretical_loss": 3.3332622585817173,
      "tokens_seen": 2857631744
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013646464646464645,
      "loss": 1.4703,
      "theoretical_loss": 3.3332563229318386,
      "tokens_seen": 2857697280
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013644444444444446,
      "loss": 1.4351,
      "theoretical_loss": 3.3332503874561947,
      "tokens_seen": 2857762816
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013642424242424242,
      "loss": 1.4039,
      "theoretical_loss": 3.3332444521547773,
      "tokens_seen": 2857828352
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013640404040404043,
      "loss": 1.4615,
      "theoretical_loss": 3.333238517027576,
      "tokens_seen": 2857893888
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013638383838383838,
      "loss": 1.3011,
      "theoretical_loss": 3.333232582074583,
      "tokens_seen": 2857959424
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013636363636363637,
      "loss": 1.4028,
      "theoretical_loss": 3.3332266472957883,
      "tokens_seen": 2858024960
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013634343434343435,
      "loss": 1.2638,
      "theoretical_loss": 3.3332207126911833,
      "tokens_seen": 2858090496
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001363232323232323,
      "loss": 1.5191,
      "theoretical_loss": 3.333214778260759,
      "tokens_seen": 2858156032
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013630303030303032,
      "loss": 1.4628,
      "theoretical_loss": 3.333208844004506,
      "tokens_seen": 2858221568
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013628282828282828,
      "loss": 1.3494,
      "theoretical_loss": 3.3332029099224147,
      "tokens_seen": 2858287104
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013626262626262626,
      "loss": 1.3942,
      "theoretical_loss": 3.3331969760144764,
      "tokens_seen": 2858352640
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013624242424242425,
      "loss": 1.2211,
      "theoretical_loss": 3.3331910422806827,
      "tokens_seen": 2858418176
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013622222222222223,
      "loss": 1.2887,
      "theoretical_loss": 3.3331851087210236,
      "tokens_seen": 2858483712
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001362020202020202,
      "loss": 1.3706,
      "theoretical_loss": 3.3331791753354905,
      "tokens_seen": 2858549248
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001361818181818182,
      "loss": 1.273,
      "theoretical_loss": 3.333173242124074,
      "tokens_seen": 2858614784
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013616161616161615,
      "loss": 1.3381,
      "theoretical_loss": 3.333167309086765,
      "tokens_seen": 2858680320
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013614141414141414,
      "loss": 1.3797,
      "theoretical_loss": 3.3331613762235546,
      "tokens_seen": 2858745856
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013612121212121212,
      "loss": 1.3948,
      "theoretical_loss": 3.3331554435344333,
      "tokens_seen": 2858811392
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001361010101010101,
      "loss": 1.5687,
      "theoretical_loss": 3.3331495110193927,
      "tokens_seen": 2858876928
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001360808080808081,
      "loss": 1.3149,
      "theoretical_loss": 3.333143578678423,
      "tokens_seen": 2858942464
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013606060606060605,
      "loss": 1.3187,
      "theoretical_loss": 3.3331376465115157,
      "tokens_seen": 2859008000
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013604040404040406,
      "loss": 1.4451,
      "theoretical_loss": 3.333131714518661,
      "tokens_seen": 2859073536
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013602020202020201,
      "loss": 1.3703,
      "theoretical_loss": 3.3331257826998506,
      "tokens_seen": 2859139072
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013600000000000003,
      "loss": 1.4393,
      "theoretical_loss": 3.333119851055075,
      "tokens_seen": 2859204608
    },
    {
      "epoch": 0.73,
      "objective/train/aligned_token_count": 0.41428396105766296,
      "objective/train/avg_token_score": 0.020952189341187477,
      "objective/train/docs_used": 1609642,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4176836013793945,
      "objective/train/original_loss": 5.177004337310791,
      "objective/train/theoretical_loss": 3.333113919584325,
      "objective/train/tokens_used": 1218129376,
      "objective/train/unlikelihood_loss": 0.04139692708849907,
      "theoretical_loss": 3.333113919584325,
      "tokens_seen": 2859270144
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013597979797979798,
      "loss": 1.3652,
      "theoretical_loss": 3.333113919584325,
      "tokens_seen": 2859270144
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013595959595959594,
      "loss": 1.3844,
      "theoretical_loss": 3.3331079882875914,
      "tokens_seen": 2859335680
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013593939393939395,
      "loss": 1.2532,
      "theoretical_loss": 3.3331020571648655,
      "tokens_seen": 2859401216
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001359191919191919,
      "loss": 1.423,
      "theoretical_loss": 3.3330961262161383,
      "tokens_seen": 2859466752
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013589898989898992,
      "loss": 1.562,
      "theoretical_loss": 3.3330901954414,
      "tokens_seen": 2859532288
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013587878787878788,
      "loss": 1.3185,
      "theoretical_loss": 3.3330842648406422,
      "tokens_seen": 2859597824
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013585858585858586,
      "loss": 1.6572,
      "theoretical_loss": 3.3330783344138553,
      "tokens_seen": 2859663360
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013583838383838384,
      "loss": 1.3225,
      "theoretical_loss": 3.333072404161031,
      "tokens_seen": 2859728896
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001358181818181818,
      "loss": 1.4766,
      "theoretical_loss": 3.333066474082159,
      "tokens_seen": 2859794432
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001357979797979798,
      "loss": 1.4216,
      "theoretical_loss": 3.3330605441772314,
      "tokens_seen": 2859859968
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013577777777777777,
      "loss": 1.5624,
      "theoretical_loss": 3.3330546144462385,
      "tokens_seen": 2859925504
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013575757575757578,
      "loss": 1.2801,
      "theoretical_loss": 3.3330486848891714,
      "tokens_seen": 2859991040
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013573737373737374,
      "loss": 1.2708,
      "theoretical_loss": 3.3330427555060207,
      "tokens_seen": 2860056576
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013571717171717172,
      "loss": 1.1849,
      "theoretical_loss": 3.3330368262967776,
      "tokens_seen": 2860122112
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001356969696969697,
      "loss": 1.3902,
      "theoretical_loss": 3.333030897261433,
      "tokens_seen": 2860187648
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001356767676767677,
      "loss": 1.5186,
      "theoretical_loss": 3.333024968399978,
      "tokens_seen": 2860253184
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013565656565656567,
      "loss": 1.4506,
      "theoretical_loss": 3.333019039712403,
      "tokens_seen": 2860318720
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013563636363636363,
      "loss": 1.5698,
      "theoretical_loss": 3.3330131111986994,
      "tokens_seen": 2860384256
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001356161616161616,
      "loss": 1.5298,
      "theoretical_loss": 3.3330071828588577,
      "tokens_seen": 2860449792
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001355959595959596,
      "loss": 1.5859,
      "theoretical_loss": 3.333001254692869,
      "tokens_seen": 2860515328
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013557575757575758,
      "loss": 1.361,
      "theoretical_loss": 3.3329953267007246,
      "tokens_seen": 2860580864
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013555555555555556,
      "loss": 1.4118,
      "theoretical_loss": 3.3329893988824146,
      "tokens_seen": 2860646400
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013553535353535355,
      "loss": 1.5415,
      "theoretical_loss": 3.332983471237931,
      "tokens_seen": 2860711936
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001355151515151515,
      "loss": 1.5949,
      "theoretical_loss": 3.3329775437672637,
      "tokens_seen": 2860777472
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013549494949494952,
      "loss": 1.4612,
      "theoretical_loss": 3.332971616470404,
      "tokens_seen": 2860843008
    },
    {
      "epoch": 0.73,
      "objective/train/aligned_token_count": 0.5045821070671082,
      "objective/train/avg_token_score": 0.01372766587883234,
      "objective/train/docs_used": 1610230,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6445918083190918,
      "objective/train/original_loss": 4.009819984436035,
      "objective/train/theoretical_loss": 3.3329656893473434,
      "objective/train/tokens_used": 1219767776,
      "objective/train/unlikelihood_loss": 0.060272324830293655,
      "theoretical_loss": 3.3329656893473434,
      "tokens_seen": 2860908544
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013547474747474747,
      "loss": 1.5803,
      "theoretical_loss": 3.3329656893473434,
      "tokens_seen": 2860908544
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013545454545454546,
      "loss": 1.2734,
      "theoretical_loss": 3.3329597623980716,
      "tokens_seen": 2860974080
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013543434343434344,
      "loss": 1.3632,
      "theoretical_loss": 3.3329538356225807,
      "tokens_seen": 2861039616
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001354141414141414,
      "loss": 1.2544,
      "theoretical_loss": 3.332947909020861,
      "tokens_seen": 2861105152
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001353939393939394,
      "loss": 1.3535,
      "theoretical_loss": 3.3329419825929034,
      "tokens_seen": 2861170688
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013537373737373737,
      "loss": 1.4817,
      "theoretical_loss": 3.332936056338699,
      "tokens_seen": 2861236224
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013535353535353538,
      "loss": 1.3978,
      "theoretical_loss": 3.332930130258239,
      "tokens_seen": 2861301760
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013533333333333333,
      "loss": 1.531,
      "theoretical_loss": 3.3329242043515137,
      "tokens_seen": 2861367296
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013531313131313132,
      "loss": 1.3342,
      "theoretical_loss": 3.3329182786185143,
      "tokens_seen": 2861432832
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001352929292929293,
      "loss": 1.3505,
      "theoretical_loss": 3.332912353059232,
      "tokens_seen": 2861498368
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013527272727272726,
      "loss": 1.2855,
      "theoretical_loss": 3.3329064276736577,
      "tokens_seen": 2861563904
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013525252525252527,
      "loss": 1.6156,
      "theoretical_loss": 3.332900502461782,
      "tokens_seen": 2861629440
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013523232323232323,
      "loss": 1.6076,
      "theoretical_loss": 3.332894577423596,
      "tokens_seen": 2861694976
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001352121212121212,
      "loss": 1.396,
      "theoretical_loss": 3.332888652559091,
      "tokens_seen": 2861760512
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001351919191919192,
      "loss": 1.4046,
      "theoretical_loss": 3.332882727868257,
      "tokens_seen": 2861826048
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013517171717171718,
      "loss": 1.4818,
      "theoretical_loss": 3.3328768033510854,
      "tokens_seen": 2861891584
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013515151515151516,
      "loss": 1.5637,
      "theoretical_loss": 3.3328708790075674,
      "tokens_seen": 2861957120
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013513131313131312,
      "loss": 1.3683,
      "theoretical_loss": 3.3328649548376936,
      "tokens_seen": 2862022656
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001351111111111111,
      "loss": 1.1918,
      "theoretical_loss": 3.332859030841455,
      "tokens_seen": 2862088192
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001350909090909091,
      "loss": 1.5399,
      "theoretical_loss": 3.332853107018843,
      "tokens_seen": 2862153728
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013507070707070707,
      "loss": 1.4346,
      "theoretical_loss": 3.3328471833698483,
      "tokens_seen": 2862219264
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013505050505050506,
      "loss": 1.3506,
      "theoretical_loss": 3.3328412598944612,
      "tokens_seen": 2862284800
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013503030303030304,
      "loss": 1.3358,
      "theoretical_loss": 3.3328353365926735,
      "tokens_seen": 2862350336
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000135010101010101,
      "loss": 1.4801,
      "theoretical_loss": 3.3328294134644754,
      "tokens_seen": 2862415872
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000134989898989899,
      "loss": 1.4711,
      "theoretical_loss": 3.3328234905098584,
      "tokens_seen": 2862481408
    },
    {
      "epoch": 0.73,
      "objective/train/aligned_token_count": 0.4397605061531067,
      "objective/train/avg_token_score": 0.019638989120721817,
      "objective/train/docs_used": 1611371,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3919695615768433,
      "objective/train/original_loss": 4.483409881591797,
      "objective/train/theoretical_loss": 3.3328175677288137,
      "objective/train/tokens_used": 1221406176,
      "objective/train/unlikelihood_loss": 0.04912484437227249,
      "theoretical_loss": 3.3328175677288137,
      "tokens_seen": 2862546944
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013496969696969696,
      "loss": 1.5082,
      "theoretical_loss": 3.3328175677288137,
      "tokens_seen": 2862546944
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013494949494949495,
      "loss": 1.563,
      "theoretical_loss": 3.332811645121331,
      "tokens_seen": 2862612480
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013492929292929293,
      "loss": 1.4323,
      "theoretical_loss": 3.3328057226874024,
      "tokens_seen": 2862678016
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013490909090909092,
      "loss": 1.3784,
      "theoretical_loss": 3.3327998004270185,
      "tokens_seen": 2862743552
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001348888888888889,
      "loss": 1.1934,
      "theoretical_loss": 3.33279387834017,
      "tokens_seen": 2862809088
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013486868686868686,
      "loss": 1.435,
      "theoretical_loss": 3.332787956426848,
      "tokens_seen": 2862874624
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013484848484848487,
      "loss": 1.5101,
      "theoretical_loss": 3.332782034687044,
      "tokens_seen": 2862940160
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013482828282828282,
      "loss": 1.6229,
      "theoretical_loss": 3.332776113120748,
      "tokens_seen": 2863005696
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013480808080808084,
      "loss": 1.6573,
      "theoretical_loss": 3.3327701917279513,
      "tokens_seen": 2863071232
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001347878787878788,
      "loss": 1.3899,
      "theoretical_loss": 3.332764270508645,
      "tokens_seen": 2863136768
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013476767676767675,
      "loss": 1.3087,
      "theoretical_loss": 3.3327583494628197,
      "tokens_seen": 2863202304
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013474747474747476,
      "loss": 1.5153,
      "theoretical_loss": 3.3327524285904673,
      "tokens_seen": 2863267840
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013472727272727272,
      "loss": 1.5043,
      "theoretical_loss": 3.3327465078915774,
      "tokens_seen": 2863333376
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013470707070707073,
      "loss": 1.4416,
      "theoretical_loss": 3.332740587366142,
      "tokens_seen": 2863398912
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013468686868686869,
      "loss": 1.4237,
      "theoretical_loss": 3.3327346670141518,
      "tokens_seen": 2863464448
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013466666666666667,
      "loss": 1.611,
      "theoretical_loss": 3.3327287468355973,
      "tokens_seen": 2863529984
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013464646464646465,
      "loss": 1.4332,
      "theoretical_loss": 3.33272282683047,
      "tokens_seen": 2863595520
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013462626262626264,
      "loss": 1.2432,
      "theoretical_loss": 3.33271690699876,
      "tokens_seen": 2863661056
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013460606060606062,
      "loss": 1.5144,
      "theoretical_loss": 3.3327109873404597,
      "tokens_seen": 2863726592
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013458585858585858,
      "loss": 1.5399,
      "theoretical_loss": 3.3327050678555588,
      "tokens_seen": 2863792128
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013456565656565656,
      "loss": 1.4624,
      "theoretical_loss": 3.332699148544049,
      "tokens_seen": 2863857664
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013454545454545455,
      "loss": 1.3347,
      "theoretical_loss": 3.3326932294059204,
      "tokens_seen": 2863923200
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013452525252525253,
      "loss": 1.4745,
      "theoretical_loss": 3.3326873104411647,
      "tokens_seen": 2863988736
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013450505050505051,
      "loss": 1.4164,
      "theoretical_loss": 3.3326813916497726,
      "tokens_seen": 2864054272
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001344848484848485,
      "loss": 1.4982,
      "theoretical_loss": 3.332675473031735,
      "tokens_seen": 2864119808
    },
    {
      "epoch": 0.73,
      "objective/train/aligned_token_count": 0.6142473220825195,
      "objective/train/avg_token_score": 0.004512544255703688,
      "objective/train/docs_used": 1612003,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9063899517059326,
      "objective/train/original_loss": 3.893289804458618,
      "objective/train/theoretical_loss": 3.332669554587043,
      "objective/train/tokens_used": 1223044576,
      "objective/train/unlikelihood_loss": 0.049774009734392166,
      "theoretical_loss": 3.332669554587043,
      "tokens_seen": 2864185344
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013446464646464645,
      "loss": 1.7529,
      "theoretical_loss": 3.332669554587043,
      "tokens_seen": 2864185344
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013444444444444447,
      "loss": 1.5374,
      "theoretical_loss": 3.3326636363156874,
      "tokens_seen": 2864250880
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013442424242424242,
      "loss": 1.2859,
      "theoretical_loss": 3.3326577182176598,
      "tokens_seen": 2864316416
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001344040404040404,
      "loss": 1.4205,
      "theoretical_loss": 3.33265180029295,
      "tokens_seen": 2864381952
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001343838383838384,
      "loss": 1.6641,
      "theoretical_loss": 3.33264588254155,
      "tokens_seen": 2864447488
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013436363636363635,
      "loss": 1.4428,
      "theoretical_loss": 3.33263996496345,
      "tokens_seen": 2864513024
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013434343434343436,
      "loss": 1.4935,
      "theoretical_loss": 3.332634047558641,
      "tokens_seen": 2864578560
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013432323232323232,
      "loss": 1.4809,
      "theoretical_loss": 3.332628130327115,
      "tokens_seen": 2864644096
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013430303030303033,
      "loss": 1.5916,
      "theoretical_loss": 3.332622213268862,
      "tokens_seen": 2864709632
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013428282828282828,
      "loss": 1.4169,
      "theoretical_loss": 3.332616296383873,
      "tokens_seen": 2864775168
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013426262626262624,
      "loss": 1.4484,
      "theoretical_loss": 3.332610379672139,
      "tokens_seen": 2864840704
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013424242424242425,
      "loss": 1.3795,
      "theoretical_loss": 3.3326044631336513,
      "tokens_seen": 2864906240
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001342222222222222,
      "loss": 1.4894,
      "theoretical_loss": 3.3325985467684007,
      "tokens_seen": 2864971776
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013420202020202022,
      "loss": 1.4176,
      "theoretical_loss": 3.3325926305763782,
      "tokens_seen": 2865037312
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013418181818181818,
      "loss": 1.2716,
      "theoretical_loss": 3.3325867145575745,
      "tokens_seen": 2865102848
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013416161616161616,
      "loss": 1.3299,
      "theoretical_loss": 3.332580798711981,
      "tokens_seen": 2865168384
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013414141414141414,
      "loss": 1.3523,
      "theoretical_loss": 3.3325748830395883,
      "tokens_seen": 2865233920
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013412121212121213,
      "loss": 1.4522,
      "theoretical_loss": 3.3325689675403876,
      "tokens_seen": 2865299456
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0001341010101010101,
      "loss": 1.5478,
      "theoretical_loss": 3.3325630522143697,
      "tokens_seen": 2865364992
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013408080808080807,
      "loss": 1.4961,
      "theoretical_loss": 3.3325571370615252,
      "tokens_seen": 2865430528
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013406060606060605,
      "loss": 1.287,
      "theoretical_loss": 3.332551222081846,
      "tokens_seen": 2865496064
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013404040404040404,
      "loss": 1.2617,
      "theoretical_loss": 3.3325453072753226,
      "tokens_seen": 2865561600
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013402020202020202,
      "loss": 1.4672,
      "theoretical_loss": 3.3325393926419458,
      "tokens_seen": 2865627136
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000134,
      "loss": 1.3531,
      "theoretical_loss": 3.3325334781817064,
      "tokens_seen": 2865692672
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.000133979797979798,
      "loss": 1.3013,
      "theoretical_loss": 3.332527563894596,
      "tokens_seen": 2865758208
    },
    {
      "epoch": 0.73,
      "objective/train/aligned_token_count": 0.48503178358078003,
      "objective/train/avg_token_score": 0.006603041663765907,
      "objective/train/docs_used": 1613208,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4981690645217896,
      "objective/train/original_loss": 4.118851661682129,
      "objective/train/theoretical_loss": 3.3325216497806056,
      "objective/train/tokens_used": 1224682976,
      "objective/train/unlikelihood_loss": 0.05737055093050003,
      "theoretical_loss": 3.3325216497806056,
      "tokens_seen": 2865823744
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.00013395959595959597,
      "loss": 1.4911,
      "theoretical_loss": 3.3325216497806056,
      "tokens_seen": 2865823744
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013393939393939396,
      "loss": 1.3716,
      "theoretical_loss": 3.3325157358397255,
      "tokens_seen": 2865889280
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001339191919191919,
      "loss": 1.4722,
      "theoretical_loss": 3.3325098220719473,
      "tokens_seen": 2865954816
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001338989898989899,
      "loss": 1.4957,
      "theoretical_loss": 3.3325039084772614,
      "tokens_seen": 2866020352
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013387878787878788,
      "loss": 1.2737,
      "theoretical_loss": 3.332497995055659,
      "tokens_seen": 2866085888
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013385858585858587,
      "loss": 1.3298,
      "theoretical_loss": 3.332492081807131,
      "tokens_seen": 2866151424
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013383838383838385,
      "loss": 1.4002,
      "theoretical_loss": 3.332486168731669,
      "tokens_seen": 2866216960
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001338181818181818,
      "loss": 1.2594,
      "theoretical_loss": 3.332480255829263,
      "tokens_seen": 2866282496
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013379797979797982,
      "loss": 1.468,
      "theoretical_loss": 3.332474343099905,
      "tokens_seen": 2866348032
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013377777777777777,
      "loss": 1.5036,
      "theoretical_loss": 3.332468430543585,
      "tokens_seen": 2866413568
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013375757575757578,
      "loss": 1.2454,
      "theoretical_loss": 3.3324625181602947,
      "tokens_seen": 2866479104
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013373737373737374,
      "loss": 1.4206,
      "theoretical_loss": 3.332456605950025,
      "tokens_seen": 2866544640
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001337171717171717,
      "loss": 1.5417,
      "theoretical_loss": 3.332450693912766,
      "tokens_seen": 2866610176
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001336969696969697,
      "loss": 1.6483,
      "theoretical_loss": 3.33244478204851,
      "tokens_seen": 2866675712
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013367676767676767,
      "loss": 1.2698,
      "theoretical_loss": 3.3324388703572474,
      "tokens_seen": 2866741248
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013365656565656568,
      "loss": 1.378,
      "theoretical_loss": 3.3324329588389685,
      "tokens_seen": 2866806784
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013363636363636363,
      "loss": 1.462,
      "theoretical_loss": 3.3324270474936655,
      "tokens_seen": 2866872320
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013361616161616162,
      "loss": 1.5495,
      "theoretical_loss": 3.3324211363213285,
      "tokens_seen": 2866937856
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001335959595959596,
      "loss": 1.4409,
      "theoretical_loss": 3.332415225321949,
      "tokens_seen": 2867003392
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013357575757575756,
      "loss": 1.4302,
      "theoretical_loss": 3.332409314495518,
      "tokens_seen": 2867068928
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013355555555555557,
      "loss": 1.4543,
      "theoretical_loss": 3.332403403842026,
      "tokens_seen": 2867134464
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013353535353535353,
      "loss": 1.3807,
      "theoretical_loss": 3.332397493361464,
      "tokens_seen": 2867200000
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001335151515151515,
      "loss": 1.3095,
      "theoretical_loss": 3.3323915830538233,
      "tokens_seen": 2867265536
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001334949494949495,
      "loss": 1.3715,
      "theoretical_loss": 3.332385672919095,
      "tokens_seen": 2867331072
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013347474747474748,
      "loss": 1.3893,
      "theoretical_loss": 3.3323797629572702,
      "tokens_seen": 2867396608
    },
    {
      "epoch": 0.74,
      "objective/train/aligned_token_count": 0.363025426864624,
      "objective/train/avg_token_score": 0.015948623418807983,
      "objective/train/docs_used": 1613631,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2840282917022705,
      "objective/train/original_loss": 4.918742656707764,
      "objective/train/theoretical_loss": 3.332373853168339,
      "objective/train/tokens_used": 1226321376,
      "objective/train/unlikelihood_loss": 0.05649067461490631,
      "theoretical_loss": 3.332373853168339,
      "tokens_seen": 2867462144
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013345454545454546,
      "loss": 1.4224,
      "theoretical_loss": 3.332373853168339,
      "tokens_seen": 2867462144
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013343434343434345,
      "loss": 1.5302,
      "theoretical_loss": 3.3323679435522937,
      "tokens_seen": 2867527680
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001334141414141414,
      "loss": 1.4336,
      "theoretical_loss": 3.3323620341091242,
      "tokens_seen": 2867593216
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001333939393939394,
      "loss": 1.3525,
      "theoretical_loss": 3.332356124838822,
      "tokens_seen": 2867658752
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013337373737373737,
      "loss": 1.4815,
      "theoretical_loss": 3.3323502157413776,
      "tokens_seen": 2867724288
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013335353535353536,
      "loss": 1.3235,
      "theoretical_loss": 3.3323443068167826,
      "tokens_seen": 2867789824
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013333333333333334,
      "loss": 1.2176,
      "theoretical_loss": 3.332338398065028,
      "tokens_seen": 2867855360
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001333131313131313,
      "loss": 1.4493,
      "theoretical_loss": 3.3323324894861046,
      "tokens_seen": 2867920896
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001332929292929293,
      "loss": 1.4977,
      "theoretical_loss": 3.332326581080003,
      "tokens_seen": 2867986432
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013327272727272726,
      "loss": 1.6879,
      "theoretical_loss": 3.332320672846715,
      "tokens_seen": 2868051968
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013325252525252528,
      "loss": 1.6988,
      "theoretical_loss": 3.332314764786231,
      "tokens_seen": 2868117504
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013323232323232323,
      "loss": 1.4054,
      "theoretical_loss": 3.3323088568985417,
      "tokens_seen": 2868183040
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001332121212121212,
      "loss": 1.5071,
      "theoretical_loss": 3.332302949183639,
      "tokens_seen": 2868248576
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001331919191919192,
      "loss": 1.3817,
      "theoretical_loss": 3.3322970416415134,
      "tokens_seen": 2868314112
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013317171717171716,
      "loss": 1.2093,
      "theoretical_loss": 3.3322911342721557,
      "tokens_seen": 2868379648
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013315151515151517,
      "loss": 1.4045,
      "theoretical_loss": 3.332285227075557,
      "tokens_seen": 2868445184
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013313131313131313,
      "loss": 1.4832,
      "theoretical_loss": 3.3322793200517093,
      "tokens_seen": 2868510720
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013311111111111114,
      "loss": 1.6882,
      "theoretical_loss": 3.332273413200602,
      "tokens_seen": 2868576256
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001330909090909091,
      "loss": 1.1953,
      "theoretical_loss": 3.332267506522227,
      "tokens_seen": 2868641792
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013307070707070708,
      "loss": 1.2695,
      "theoretical_loss": 3.332261600016575,
      "tokens_seen": 2868707328
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013305050505050506,
      "loss": 1.3577,
      "theoretical_loss": 3.3322556936836376,
      "tokens_seen": 2868772864
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013303030303030302,
      "loss": 1.1894,
      "theoretical_loss": 3.332249787523405,
      "tokens_seen": 2868838400
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013301010101010103,
      "loss": 1.3544,
      "theoretical_loss": 3.332243881535869,
      "tokens_seen": 2868903936
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013298989898989899,
      "loss": 1.4897,
      "theoretical_loss": 3.3322379757210197,
      "tokens_seen": 2868969472
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013296969696969697,
      "loss": 1.2701,
      "theoretical_loss": 3.332232070078849,
      "tokens_seen": 2869035008
    },
    {
      "epoch": 0.74,
      "objective/train/aligned_token_count": 0.3455522954463959,
      "objective/train/avg_token_score": 0.011398487724363804,
      "objective/train/docs_used": 1614970,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1357951164245605,
      "objective/train/original_loss": 4.277784824371338,
      "objective/train/theoretical_loss": 3.332226164609347,
      "objective/train/tokens_used": 1227959776,
      "objective/train/unlikelihood_loss": 0.07077215611934662,
      "theoretical_loss": 3.332226164609347,
      "tokens_seen": 2869100544
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013294949494949495,
      "loss": 1.3083,
      "theoretical_loss": 3.332226164609347,
      "tokens_seen": 2869100544
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013292929292929294,
      "loss": 1.5281,
      "theoretical_loss": 3.332220259312505,
      "tokens_seen": 2869166080
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013290909090909092,
      "loss": 1.1941,
      "theoretical_loss": 3.3322143541883147,
      "tokens_seen": 2869231616
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001328888888888889,
      "loss": 1.5205,
      "theoretical_loss": 3.3322084492367665,
      "tokens_seen": 2869297152
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013286868686868686,
      "loss": 1.4289,
      "theoretical_loss": 3.332202544457852,
      "tokens_seen": 2869362688
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013284848484848485,
      "loss": 1.3902,
      "theoretical_loss": 3.332196639851561,
      "tokens_seen": 2869428224
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013282828282828283,
      "loss": 1.2486,
      "theoretical_loss": 3.3321907354178855,
      "tokens_seen": 2869493760
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013280808080808081,
      "loss": 1.5966,
      "theoretical_loss": 3.332184831156816,
      "tokens_seen": 2869559296
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001327878787878788,
      "loss": 1.428,
      "theoretical_loss": 3.332178927068344,
      "tokens_seen": 2869624832
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013276767676767676,
      "loss": 1.4618,
      "theoretical_loss": 3.3321730231524604,
      "tokens_seen": 2869690368
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013274747474747477,
      "loss": 1.5117,
      "theoretical_loss": 3.332167119409156,
      "tokens_seen": 2869755904
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013272727272727272,
      "loss": 1.4409,
      "theoretical_loss": 3.332161215838422,
      "tokens_seen": 2869821440
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013270707070707073,
      "loss": 1.5597,
      "theoretical_loss": 3.332155312440249,
      "tokens_seen": 2869886976
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001326868686868687,
      "loss": 1.3387,
      "theoretical_loss": 3.3321494092146287,
      "tokens_seen": 2869952512
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013266666666666665,
      "loss": 1.4096,
      "theoretical_loss": 3.332143506161551,
      "tokens_seen": 2870018048
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013264646464646466,
      "loss": 1.4253,
      "theoretical_loss": 3.3321376032810086,
      "tokens_seen": 2870083584
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013262626262626262,
      "loss": 1.4727,
      "theoretical_loss": 3.332131700572991,
      "tokens_seen": 2870149120
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013260606060606063,
      "loss": 1.5894,
      "theoretical_loss": 3.3321257980374903,
      "tokens_seen": 2870214656
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013258585858585858,
      "loss": 1.4134,
      "theoretical_loss": 3.332119895674497,
      "tokens_seen": 2870280192
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013256565656565657,
      "loss": 1.4726,
      "theoretical_loss": 3.3321139934840014,
      "tokens_seen": 2870345728
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013254545454545455,
      "loss": 1.2988,
      "theoretical_loss": 3.332108091465996,
      "tokens_seen": 2870411264
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001325252525252525,
      "loss": 1.3795,
      "theoretical_loss": 3.3321021896204708,
      "tokens_seen": 2870476800
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013250505050505052,
      "loss": 1.3318,
      "theoretical_loss": 3.332096287947417,
      "tokens_seen": 2870542336
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013248484848484848,
      "loss": 1.5716,
      "theoretical_loss": 3.3320903864468256,
      "tokens_seen": 2870607872
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013246464646464646,
      "loss": 1.4021,
      "theoretical_loss": 3.3320844851186884,
      "tokens_seen": 2870673408
    },
    {
      "epoch": 0.74,
      "objective/train/aligned_token_count": 0.5490591526031494,
      "objective/train/avg_token_score": 0.0059341429732739925,
      "objective/train/docs_used": 1615721,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6865599155426025,
      "objective/train/original_loss": 3.834667444229126,
      "objective/train/theoretical_loss": 3.3320785839629954,
      "objective/train/tokens_used": 1229598176,
      "objective/train/unlikelihood_loss": 0.0493159294128418,
      "theoretical_loss": 3.3320785839629954,
      "tokens_seen": 2870738944
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013244444444444444,
      "loss": 1.6325,
      "theoretical_loss": 3.3320785839629954,
      "tokens_seen": 2870738944
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013242424242424243,
      "loss": 1.4066,
      "theoretical_loss": 3.332072682979738,
      "tokens_seen": 2870804480
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001324040404040404,
      "loss": 1.3367,
      "theoretical_loss": 3.3320667821689067,
      "tokens_seen": 2870870016
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001323838383838384,
      "loss": 1.352,
      "theoretical_loss": 3.3320608815304937,
      "tokens_seen": 2870935552
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013236363636363635,
      "loss": 1.3645,
      "theoretical_loss": 3.3320549810644895,
      "tokens_seen": 2871001088
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013234343434343434,
      "loss": 1.3649,
      "theoretical_loss": 3.3320490807708847,
      "tokens_seen": 2871066624
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013232323232323232,
      "loss": 1.4374,
      "theoretical_loss": 3.3320431806496704,
      "tokens_seen": 2871132160
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001323030303030303,
      "loss": 1.6851,
      "theoretical_loss": 3.3320372807008383,
      "tokens_seen": 2871197696
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001322828282828283,
      "loss": 1.407,
      "theoretical_loss": 3.332031380924379,
      "tokens_seen": 2871263232
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013226262626262627,
      "loss": 1.4546,
      "theoretical_loss": 3.332025481320283,
      "tokens_seen": 2871328768
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013224242424242426,
      "loss": 1.5025,
      "theoretical_loss": 3.332019581888542,
      "tokens_seen": 2871394304
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013222222222222221,
      "loss": 1.5949,
      "theoretical_loss": 3.3320136826291473,
      "tokens_seen": 2871459840
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013220202020202022,
      "loss": 1.4908,
      "theoretical_loss": 3.332007783542089,
      "tokens_seen": 2871525376
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013218181818181818,
      "loss": 1.4009,
      "theoretical_loss": 3.332001884627359,
      "tokens_seen": 2871590912
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013216161616161617,
      "loss": 1.5133,
      "theoretical_loss": 3.331995985884948,
      "tokens_seen": 2871656448
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013214141414141415,
      "loss": 1.5598,
      "theoretical_loss": 3.331990087314847,
      "tokens_seen": 2871721984
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001321212121212121,
      "loss": 1.5601,
      "theoretical_loss": 3.3319841889170467,
      "tokens_seen": 2871787520
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013210101010101012,
      "loss": 1.3661,
      "theoretical_loss": 3.3319782906915387,
      "tokens_seen": 2871853056
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013208080808080807,
      "loss": 1.4139,
      "theoretical_loss": 3.331972392638314,
      "tokens_seen": 2871918592
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013206060606060609,
      "loss": 1.3716,
      "theoretical_loss": 3.331966494757363,
      "tokens_seen": 2871984128
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013204040404040404,
      "loss": 1.2764,
      "theoretical_loss": 3.3319605970486776,
      "tokens_seen": 2872049664
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013202020202020203,
      "loss": 1.572,
      "theoretical_loss": 3.331954699512248,
      "tokens_seen": 2872115200
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000132,
      "loss": 1.2374,
      "theoretical_loss": 3.3319488021480663,
      "tokens_seen": 2872180736
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013197979797979797,
      "loss": 1.5242,
      "theoretical_loss": 3.331942904956122,
      "tokens_seen": 2872246272
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013195959595959598,
      "loss": 1.4289,
      "theoretical_loss": 3.331937007936408,
      "tokens_seen": 2872311808
    },
    {
      "epoch": 0.74,
      "objective/train/aligned_token_count": 0.4449535608291626,
      "objective/train/avg_token_score": 0.008155014365911484,
      "objective/train/docs_used": 1616749,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4699586629867554,
      "objective/train/original_loss": 4.063313961029053,
      "objective/train/theoretical_loss": 3.331931111088914,
      "objective/train/tokens_used": 1231236576,
      "objective/train/unlikelihood_loss": 0.06177420914173126,
      "theoretical_loss": 3.331931111088914,
      "tokens_seen": 2872377344
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013193939393939393,
      "loss": 1.4474,
      "theoretical_loss": 3.331931111088914,
      "tokens_seen": 2872377344
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013191919191919192,
      "loss": 1.2008,
      "theoretical_loss": 3.3319252144136318,
      "tokens_seen": 2872442880
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001318989898989899,
      "loss": 1.2978,
      "theoretical_loss": 3.3319193179105513,
      "tokens_seen": 2872508416
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001318787878787879,
      "loss": 1.4862,
      "theoretical_loss": 3.3319134215796646,
      "tokens_seen": 2872573952
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013185858585858587,
      "loss": 1.5122,
      "theoretical_loss": 3.331907525420963,
      "tokens_seen": 2872639488
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013183838383838383,
      "loss": 1.4872,
      "theoretical_loss": 3.3319016294344364,
      "tokens_seen": 2872705024
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001318181818181818,
      "loss": 1.4375,
      "theoretical_loss": 3.3318957336200765,
      "tokens_seen": 2872770560
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001317979797979798,
      "loss": 1.2501,
      "theoretical_loss": 3.331889837977874,
      "tokens_seen": 2872836096
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013177777777777778,
      "loss": 1.5546,
      "theoretical_loss": 3.3318839425078206,
      "tokens_seen": 2872901632
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013175757575757576,
      "loss": 1.3291,
      "theoretical_loss": 3.331878047209907,
      "tokens_seen": 2872967168
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013173737373737375,
      "loss": 1.668,
      "theoretical_loss": 3.3318721520841246,
      "tokens_seen": 2873032704
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001317171717171717,
      "loss": 1.4137,
      "theoretical_loss": 3.3318662571304634,
      "tokens_seen": 2873098240
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013169696969696972,
      "loss": 1.2967,
      "theoretical_loss": 3.3318603623489156,
      "tokens_seen": 2873163776
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013167676767676767,
      "loss": 1.5448,
      "theoretical_loss": 3.3318544677394715,
      "tokens_seen": 2873229312
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013165656565656566,
      "loss": 1.52,
      "theoretical_loss": 3.3318485733021226,
      "tokens_seen": 2873294848
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013163636363636364,
      "loss": 1.2723,
      "theoretical_loss": 3.3318426790368596,
      "tokens_seen": 2873360384
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001316161616161616,
      "loss": 1.4005,
      "theoretical_loss": 3.3318367849436736,
      "tokens_seen": 2873425920
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001315959595959596,
      "loss": 1.4258,
      "theoretical_loss": 3.3318308910225563,
      "tokens_seen": 2873491456
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013157575757575756,
      "loss": 1.2023,
      "theoretical_loss": 3.3318249972734977,
      "tokens_seen": 2873556992
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013155555555555558,
      "loss": 1.2343,
      "theoretical_loss": 3.3318191036964895,
      "tokens_seen": 2873622528
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013153535353535353,
      "loss": 1.4833,
      "theoretical_loss": 3.3318132102915228,
      "tokens_seen": 2873688064
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013151515151515152,
      "loss": 1.3653,
      "theoretical_loss": 3.3318073170585882,
      "tokens_seen": 2873753600
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001314949494949495,
      "loss": 1.4504,
      "theoretical_loss": 3.3318014239976774,
      "tokens_seen": 2873819136
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013147474747474746,
      "loss": 1.4823,
      "theoretical_loss": 3.331795531108781,
      "tokens_seen": 2873884672
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013145454545454547,
      "loss": 1.3459,
      "theoretical_loss": 3.33178963839189,
      "tokens_seen": 2873950208
    },
    {
      "epoch": 0.74,
      "objective/train/aligned_token_count": 0.29747068881988525,
      "objective/train/avg_token_score": 0.0316876657307148,
      "objective/train/docs_used": 1617455,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9600871205329895,
      "objective/train/original_loss": 5.321650981903076,
      "objective/train/theoretical_loss": 3.331783745846996,
      "objective/train/tokens_used": 1232874976,
      "objective/train/unlikelihood_loss": 0.055472563952207565,
      "theoretical_loss": 3.331783745846996,
      "tokens_seen": 2874015744
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013143434343434343,
      "loss": 1.3161,
      "theoretical_loss": 3.331783745846996,
      "tokens_seen": 2874015744
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013141414141414144,
      "loss": 1.6569,
      "theoretical_loss": 3.3317778534740894,
      "tokens_seen": 2874081280
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001313939393939394,
      "loss": 1.4993,
      "theoretical_loss": 3.3317719612731613,
      "tokens_seen": 2874146816
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013137373737373738,
      "loss": 1.5303,
      "theoretical_loss": 3.3317660692442033,
      "tokens_seen": 2874212352
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013135353535353536,
      "loss": 1.5732,
      "theoretical_loss": 3.331760177387206,
      "tokens_seen": 2874277888
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013133333333333335,
      "loss": 1.3392,
      "theoretical_loss": 3.3317542857021607,
      "tokens_seen": 2874343424
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013131313131313133,
      "loss": 1.4656,
      "theoretical_loss": 3.3317483941890584,
      "tokens_seen": 2874408960
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013129292929292929,
      "loss": 1.4143,
      "theoretical_loss": 3.3317425028478898,
      "tokens_seen": 2874474496
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013127272727272727,
      "loss": 1.4465,
      "theoretical_loss": 3.3317366116786467,
      "tokens_seen": 2874540032
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013125252525252525,
      "loss": 1.5121,
      "theoretical_loss": 3.3317307206813194,
      "tokens_seen": 2874605568
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013123232323232324,
      "loss": 1.4426,
      "theoretical_loss": 3.3317248298559,
      "tokens_seen": 2874671104
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013121212121212122,
      "loss": 1.4426,
      "theoretical_loss": 3.3317189392023785,
      "tokens_seen": 2874736640
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001311919191919192,
      "loss": 1.4397,
      "theoretical_loss": 3.3317130487207463,
      "tokens_seen": 2874802176
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013117171717171716,
      "loss": 1.3894,
      "theoretical_loss": 3.3317071584109943,
      "tokens_seen": 2874867712
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013115151515151517,
      "loss": 1.3698,
      "theoretical_loss": 3.331701268273114,
      "tokens_seen": 2874933248
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013113131313131313,
      "loss": 1.3685,
      "theoretical_loss": 3.331695378307096,
      "tokens_seen": 2874998784
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013111111111111111,
      "loss": 1.4555,
      "theoretical_loss": 3.3316894885129322,
      "tokens_seen": 2875064320
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001310909090909091,
      "loss": 1.3759,
      "theoretical_loss": 3.3316835988906126,
      "tokens_seen": 2875129856
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013107070707070706,
      "loss": 1.324,
      "theoretical_loss": 3.331677709440129,
      "tokens_seen": 2875195392
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013105050505050507,
      "loss": 1.5093,
      "theoretical_loss": 3.331671820161472,
      "tokens_seen": 2875260928
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013103030303030302,
      "loss": 1.4597,
      "theoretical_loss": 3.3316659310546335,
      "tokens_seen": 2875326464
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013101010101010103,
      "loss": 1.4589,
      "theoretical_loss": 3.331660042119603,
      "tokens_seen": 2875392000
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000130989898989899,
      "loss": 1.3671,
      "theoretical_loss": 3.3316541533563733,
      "tokens_seen": 2875457536
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013096969696969695,
      "loss": 1.6802,
      "theoretical_loss": 3.331648264764935,
      "tokens_seen": 2875523072
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013094949494949496,
      "loss": 1.5583,
      "theoretical_loss": 3.331642376345278,
      "tokens_seen": 2875588608
    },
    {
      "epoch": 0.74,
      "objective/train/aligned_token_count": 0.30987292528152466,
      "objective/train/avg_token_score": 0.02147936262190342,
      "objective/train/docs_used": 1618853,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9461278319358826,
      "objective/train/original_loss": 4.69231653213501,
      "objective/train/theoretical_loss": 3.3316364880973945,
      "objective/train/tokens_used": 1234513376,
      "objective/train/unlikelihood_loss": 0.07146111130714417,
      "theoretical_loss": 3.3316364880973945,
      "tokens_seen": 2875654144
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013092929292929292,
      "loss": 1.2532,
      "theoretical_loss": 3.3316364880973945,
      "tokens_seen": 2875654144
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013090909090909093,
      "loss": 1.4952,
      "theoretical_loss": 3.331630600021276,
      "tokens_seen": 2875719680
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013088888888888888,
      "loss": 1.4348,
      "theoretical_loss": 3.331624712116912,
      "tokens_seen": 2875785216
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013086868686868687,
      "loss": 1.3405,
      "theoretical_loss": 3.331618824384295,
      "tokens_seen": 2875850752
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013084848484848485,
      "loss": 1.2612,
      "theoretical_loss": 3.331612936823415,
      "tokens_seen": 2875916288
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013082828282828284,
      "loss": 1.3832,
      "theoretical_loss": 3.3316070494342647,
      "tokens_seen": 2875981824
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013080808080808082,
      "loss": 1.5562,
      "theoretical_loss": 3.3316011622168333,
      "tokens_seen": 2876047360
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013078787878787878,
      "loss": 1.4983,
      "theoretical_loss": 3.3315952751711126,
      "tokens_seen": 2876112896
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013076767676767676,
      "loss": 1.2163,
      "theoretical_loss": 3.331589388297094,
      "tokens_seen": 2876178432
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013074747474747474,
      "loss": 1.5798,
      "theoretical_loss": 3.3315835015947686,
      "tokens_seen": 2876243968
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013072727272727273,
      "loss": 1.2272,
      "theoretical_loss": 3.331577615064127,
      "tokens_seen": 2876309504
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001307070707070707,
      "loss": 1.4694,
      "theoretical_loss": 3.331571728705161,
      "tokens_seen": 2876375040
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001306868686868687,
      "loss": 1.6351,
      "theoretical_loss": 3.3315658425178607,
      "tokens_seen": 2876440576
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013066666666666665,
      "loss": 1.3909,
      "theoretical_loss": 3.3315599565022174,
      "tokens_seen": 2876506112
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013064646464646466,
      "loss": 1.4349,
      "theoretical_loss": 3.331554070658223,
      "tokens_seen": 2876571648
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013062626262626262,
      "loss": 1.4379,
      "theoretical_loss": 3.331548184985868,
      "tokens_seen": 2876637184
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001306060606060606,
      "loss": 1.4404,
      "theoretical_loss": 3.3315422994851436,
      "tokens_seen": 2876702720
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001305858585858586,
      "loss": 1.223,
      "theoretical_loss": 3.331536414156041,
      "tokens_seen": 2876768256
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013056565656565657,
      "loss": 1.4349,
      "theoretical_loss": 3.3315305289985506,
      "tokens_seen": 2876833792
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013054545454545456,
      "loss": 1.3304,
      "theoretical_loss": 3.331524644012664,
      "tokens_seen": 2876899328
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013052525252525251,
      "loss": 1.2972,
      "theoretical_loss": 3.3315187591983726,
      "tokens_seen": 2876964864
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013050505050505053,
      "loss": 1.4142,
      "theoretical_loss": 3.331512874555667,
      "tokens_seen": 2877030400
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013048484848484848,
      "loss": 1.429,
      "theoretical_loss": 3.3315069900845384,
      "tokens_seen": 2877095936
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001304646464646465,
      "loss": 1.5903,
      "theoretical_loss": 3.331501105784978,
      "tokens_seen": 2877161472
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013044444444444445,
      "loss": 1.4811,
      "theoretical_loss": 3.3314952216569766,
      "tokens_seen": 2877227008
    },
    {
      "epoch": 0.74,
      "objective/train/aligned_token_count": 0.39400050044059753,
      "objective/train/avg_token_score": 0.019464677199721336,
      "objective/train/docs_used": 1619565,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2468276023864746,
      "objective/train/original_loss": 4.636404514312744,
      "objective/train/theoretical_loss": 3.3314893377005257,
      "objective/train/tokens_used": 1236151776,
      "objective/train/unlikelihood_loss": 0.04895918443799019,
      "theoretical_loss": 3.3314893377005257,
      "tokens_seen": 2877292544
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001304242424242424,
      "loss": 1.5165,
      "theoretical_loss": 3.3314893377005257,
      "tokens_seen": 2877292544
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013040404040404042,
      "loss": 1.3858,
      "theoretical_loss": 3.3314834539156166,
      "tokens_seen": 2877358080
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013038383838383837,
      "loss": 1.5157,
      "theoretical_loss": 3.3314775703022397,
      "tokens_seen": 2877423616
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013036363636363639,
      "loss": 1.2791,
      "theoretical_loss": 3.3314716868603864,
      "tokens_seen": 2877489152
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013034343434343434,
      "loss": 1.3666,
      "theoretical_loss": 3.331465803590048,
      "tokens_seen": 2877554688
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013032323232323233,
      "loss": 1.4815,
      "theoretical_loss": 3.331459920491215,
      "tokens_seen": 2877620224
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001303030303030303,
      "loss": 1.3976,
      "theoretical_loss": 3.331454037563879,
      "tokens_seen": 2877685760
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013028282828282827,
      "loss": 1.3364,
      "theoretical_loss": 3.3314481548080312,
      "tokens_seen": 2877751296
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013026262626262628,
      "loss": 1.5075,
      "theoretical_loss": 3.3314422722236623,
      "tokens_seen": 2877816832
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013024242424242424,
      "loss": 1.3318,
      "theoretical_loss": 3.3314363898107637,
      "tokens_seen": 2877882368
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013022222222222222,
      "loss": 1.5598,
      "theoretical_loss": 3.3314305075693262,
      "tokens_seen": 2877947904
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001302020202020202,
      "loss": 1.3421,
      "theoretical_loss": 3.3314246254993414,
      "tokens_seen": 2878013440
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001301818181818182,
      "loss": 1.3688,
      "theoretical_loss": 3.3314187436008,
      "tokens_seen": 2878078976
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013016161616161617,
      "loss": 1.2215,
      "theoretical_loss": 3.3314128618736927,
      "tokens_seen": 2878144512
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013014141414141416,
      "loss": 1.4814,
      "theoretical_loss": 3.3314069803180115,
      "tokens_seen": 2878210048
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001301212121212121,
      "loss": 1.4333,
      "theoretical_loss": 3.331401098933747,
      "tokens_seen": 2878275584
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001301010101010101,
      "loss": 1.379,
      "theoretical_loss": 3.33139521772089,
      "tokens_seen": 2878341120
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013008080808080808,
      "loss": 1.5401,
      "theoretical_loss": 3.3313893366794325,
      "tokens_seen": 2878406656
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013006060606060606,
      "loss": 1.3451,
      "theoretical_loss": 3.3313834558093647,
      "tokens_seen": 2878472192
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013004040404040405,
      "loss": 1.3607,
      "theoretical_loss": 3.3313775751106784,
      "tokens_seen": 2878537728
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000130020202020202,
      "loss": 1.2837,
      "theoretical_loss": 3.331371694583364,
      "tokens_seen": 2878603264
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00013000000000000002,
      "loss": 1.3241,
      "theoretical_loss": 3.3313658142274134,
      "tokens_seen": 2878668800
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012997979797979797,
      "loss": 1.1471,
      "theoretical_loss": 3.331359934042817,
      "tokens_seen": 2878734336
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012995959595959598,
      "loss": 1.3845,
      "theoretical_loss": 3.3313540540295663,
      "tokens_seen": 2878799872
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012993939393939394,
      "loss": 1.5415,
      "theoretical_loss": 3.331348174187652,
      "tokens_seen": 2878865408
    },
    {
      "epoch": 0.74,
      "objective/train/aligned_token_count": 0.5237047672271729,
      "objective/train/avg_token_score": 0.0086279371753335,
      "objective/train/docs_used": 1621029,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6239334344863892,
      "objective/train/original_loss": 4.021581172943115,
      "objective/train/theoretical_loss": 3.331342294517066,
      "objective/train/tokens_used": 1237790176,
      "objective/train/unlikelihood_loss": 0.05988957732915878,
      "theoretical_loss": 3.331342294517066,
      "tokens_seen": 2878930944
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001299191919191919,
      "loss": 1.4726,
      "theoretical_loss": 3.331342294517066,
      "tokens_seen": 2878930944
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001298989898989899,
      "loss": 1.4785,
      "theoretical_loss": 3.3313364150177986,
      "tokens_seen": 2878996480
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012987878787878787,
      "loss": 1.4085,
      "theoretical_loss": 3.3313305356898413,
      "tokens_seen": 2879062016
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012985858585858588,
      "loss": 1.413,
      "theoretical_loss": 3.3313246565331855,
      "tokens_seen": 2879127552
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012983838383838383,
      "loss": 1.6119,
      "theoretical_loss": 3.3313187775478212,
      "tokens_seen": 2879193088
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012981818181818182,
      "loss": 1.1979,
      "theoretical_loss": 3.331312898733741,
      "tokens_seen": 2879258624
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001297979797979798,
      "loss": 1.4167,
      "theoretical_loss": 3.331307020090935,
      "tokens_seen": 2879324160
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012977777777777779,
      "loss": 1.178,
      "theoretical_loss": 3.3313011416193943,
      "tokens_seen": 2879389696
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012975757575757577,
      "loss": 1.1624,
      "theoretical_loss": 3.3312952633191104,
      "tokens_seen": 2879455232
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012973737373737373,
      "loss": 1.3611,
      "theoretical_loss": 3.3312893851900744,
      "tokens_seen": 2879520768
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012971717171717174,
      "loss": 1.2997,
      "theoretical_loss": 3.3312835072322775,
      "tokens_seen": 2879586304
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001296969696969697,
      "loss": 1.4236,
      "theoretical_loss": 3.3312776294457103,
      "tokens_seen": 2879651840
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012967676767676768,
      "loss": 1.535,
      "theoretical_loss": 3.3312717518303643,
      "tokens_seen": 2879717376
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012965656565656566,
      "loss": 1.6617,
      "theoretical_loss": 3.3312658743862307,
      "tokens_seen": 2879782912
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012963636363636365,
      "loss": 1.2536,
      "theoretical_loss": 3.3312599971133006,
      "tokens_seen": 2879848448
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012961616161616163,
      "loss": 1.2688,
      "theoretical_loss": 3.3312541200115646,
      "tokens_seen": 2879913984
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012959595959595961,
      "loss": 1.4956,
      "theoretical_loss": 3.3312482430810144,
      "tokens_seen": 2879979520
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012957575757575757,
      "loss": 1.2648,
      "theoretical_loss": 3.331242366321641,
      "tokens_seen": 2880045056
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012955555555555555,
      "loss": 1.3443,
      "theoretical_loss": 3.3312364897334357,
      "tokens_seen": 2880110592
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012953535353535354,
      "loss": 1.333,
      "theoretical_loss": 3.3312306133163894,
      "tokens_seen": 2880176128
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012951515151515152,
      "loss": 1.292,
      "theoretical_loss": 3.3312247370704924,
      "tokens_seen": 2880241664
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001294949494949495,
      "loss": 1.5031,
      "theoretical_loss": 3.3312188609957376,
      "tokens_seen": 2880307200
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012947474747474746,
      "loss": 1.4841,
      "theoretical_loss": 3.3312129850921144,
      "tokens_seen": 2880372736
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012945454545454547,
      "loss": 1.3636,
      "theoretical_loss": 3.3312071093596147,
      "tokens_seen": 2880438272
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012943434343434343,
      "loss": 1.2896,
      "theoretical_loss": 3.3312012337982297,
      "tokens_seen": 2880503808
    },
    {
      "epoch": 0.74,
      "objective/train/aligned_token_count": 0.3004032373428345,
      "objective/train/avg_token_score": 0.02676505036652088,
      "objective/train/docs_used": 1621719,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9571623802185059,
      "objective/train/original_loss": 4.683320045471191,
      "objective/train/theoretical_loss": 3.3311953584079506,
      "objective/train/tokens_used": 1239428576,
      "objective/train/unlikelihood_loss": 0.07268504053354263,
      "theoretical_loss": 3.3311953584079506,
      "tokens_seen": 2880569344
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012941414141414142,
      "loss": 1.3808,
      "theoretical_loss": 3.3311953584079506,
      "tokens_seen": 2880569344
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001293939393939394,
      "loss": 1.564,
      "theoretical_loss": 3.3311894831887683,
      "tokens_seen": 2880634880
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012937373737373736,
      "loss": 1.3333,
      "theoretical_loss": 3.331183608140674,
      "tokens_seen": 2880700416
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012935353535353537,
      "loss": 1.1194,
      "theoretical_loss": 3.331177733263659,
      "tokens_seen": 2880765952
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012933333333333332,
      "loss": 1.3599,
      "theoretical_loss": 3.3311718585577137,
      "tokens_seen": 2880831488
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012931313131313133,
      "loss": 1.334,
      "theoretical_loss": 3.33116598402283,
      "tokens_seen": 2880897024
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001292929292929293,
      "loss": 1.402,
      "theoretical_loss": 3.3311601096589984,
      "tokens_seen": 2880962560
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012927272727272728,
      "loss": 1.4858,
      "theoretical_loss": 3.3311542354662107,
      "tokens_seen": 2881028096
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012925252525252526,
      "loss": 1.296,
      "theoretical_loss": 3.3311483614444577,
      "tokens_seen": 2881093632
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012923232323232322,
      "loss": 1.3135,
      "theoretical_loss": 3.3311424875937306,
      "tokens_seen": 2881159168
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012921212121212123,
      "loss": 1.5047,
      "theoretical_loss": 3.3311366139140204,
      "tokens_seen": 2881224704
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012919191919191918,
      "loss": 1.3139,
      "theoretical_loss": 3.3311307404053183,
      "tokens_seen": 2881290240
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012917171717171717,
      "loss": 1.4557,
      "theoretical_loss": 3.3311248670676155,
      "tokens_seen": 2881355776
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012915151515151515,
      "loss": 1.4596,
      "theoretical_loss": 3.331118993900903,
      "tokens_seen": 2881421312
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012913131313131314,
      "loss": 1.2756,
      "theoretical_loss": 3.331113120905172,
      "tokens_seen": 2881486848
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012911111111111112,
      "loss": 1.4123,
      "theoretical_loss": 3.331107248080414,
      "tokens_seen": 2881552384
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0001290909090909091,
      "loss": 1.6472,
      "theoretical_loss": 3.331101375426619,
      "tokens_seen": 2881617920
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012907070707070706,
      "loss": 1.1885,
      "theoretical_loss": 3.3310955029437794,
      "tokens_seen": 2881683456
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012905050505050505,
      "loss": 1.4107,
      "theoretical_loss": 3.331089630631886,
      "tokens_seen": 2881748992
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012903030303030303,
      "loss": 1.3699,
      "theoretical_loss": 3.3310837584909296,
      "tokens_seen": 2881814528
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000129010101010101,
      "loss": 1.5747,
      "theoretical_loss": 3.3310778865209016,
      "tokens_seen": 2881880064
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.000128989898989899,
      "loss": 1.4052,
      "theoretical_loss": 3.331072014721793,
      "tokens_seen": 2881945600
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012896969696969695,
      "loss": 1.4848,
      "theoretical_loss": 3.3310661430935946,
      "tokens_seen": 2882011136
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012894949494949496,
      "loss": 1.3173,
      "theoretical_loss": 3.3310602716362983,
      "tokens_seen": 2882076672
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012892929292929292,
      "loss": 1.3132,
      "theoretical_loss": 3.331054400349895,
      "tokens_seen": 2882142208
    },
    {
      "epoch": 0.74,
      "objective/train/aligned_token_count": 0.47611191868782043,
      "objective/train/avg_token_score": 0.00613713264465332,
      "objective/train/docs_used": 1622728,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.331666350364685,
      "objective/train/original_loss": 4.0757951736450195,
      "objective/train/theoretical_loss": 3.331048529234376,
      "objective/train/tokens_used": 1241066976,
      "objective/train/unlikelihood_loss": 0.06733284145593643,
      "theoretical_loss": 3.331048529234376,
      "tokens_seen": 2882207744
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00012890909090909093,
      "loss": 1.5065,
      "theoretical_loss": 3.331048529234376,
      "tokens_seen": 2882207744
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001288888888888889,
      "loss": 1.3256,
      "theoretical_loss": 3.3310426582897312,
      "tokens_seen": 2882273280
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012886868686868687,
      "loss": 1.4414,
      "theoretical_loss": 3.3310367875159534,
      "tokens_seen": 2882338816
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012884848484848486,
      "loss": 1.466,
      "theoretical_loss": 3.3310309169130328,
      "tokens_seen": 2882404352
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012882828282828281,
      "loss": 1.4402,
      "theoretical_loss": 3.331025046480961,
      "tokens_seen": 2882469888
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012880808080808083,
      "loss": 1.4143,
      "theoretical_loss": 3.331019176219729,
      "tokens_seen": 2882535424
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012878787878787878,
      "loss": 1.5579,
      "theoretical_loss": 3.3310133061293277,
      "tokens_seen": 2882600960
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001287676767676768,
      "loss": 1.5582,
      "theoretical_loss": 3.3310074362097484,
      "tokens_seen": 2882666496
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012874747474747475,
      "loss": 1.338,
      "theoretical_loss": 3.331001566460982,
      "tokens_seen": 2882732032
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001287272727272727,
      "loss": 1.3778,
      "theoretical_loss": 3.3309956968830203,
      "tokens_seen": 2882797568
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012870707070707072,
      "loss": 1.7407,
      "theoretical_loss": 3.330989827475854,
      "tokens_seen": 2882863104
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012868686868686868,
      "loss": 1.4556,
      "theoretical_loss": 3.3309839582394742,
      "tokens_seen": 2882928640
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012866666666666669,
      "loss": 1.2246,
      "theoretical_loss": 3.330978089173872,
      "tokens_seen": 2882994176
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012864646464646464,
      "loss": 1.3295,
      "theoretical_loss": 3.330972220279039,
      "tokens_seen": 2883059712
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012862626262626263,
      "loss": 1.3096,
      "theoretical_loss": 3.330966351554966,
      "tokens_seen": 2883125248
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001286060606060606,
      "loss": 1.2887,
      "theoretical_loss": 3.330960483001644,
      "tokens_seen": 2883190784
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001285858585858586,
      "loss": 1.5232,
      "theoretical_loss": 3.3309546146190647,
      "tokens_seen": 2883256320
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012856565656565658,
      "loss": 1.3381,
      "theoretical_loss": 3.3309487464072185,
      "tokens_seen": 2883321856
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012854545454545454,
      "loss": 1.457,
      "theoretical_loss": 3.330942878366097,
      "tokens_seen": 2883387392
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012852525252525252,
      "loss": 1.2819,
      "theoretical_loss": 3.330937010495692,
      "tokens_seen": 2883452928
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001285050505050505,
      "loss": 1.1017,
      "theoretical_loss": 3.330931142795993,
      "tokens_seen": 2883518464
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001284848484848485,
      "loss": 1.2266,
      "theoretical_loss": 3.3309252752669924,
      "tokens_seen": 2883584000
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012846464646464647,
      "loss": 1.4786,
      "theoretical_loss": 3.3309194079086812,
      "tokens_seen": 2883649536
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012844444444444446,
      "loss": 1.5005,
      "theoretical_loss": 3.3309135407210504,
      "tokens_seen": 2883715072
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001284242424242424,
      "loss": 1.5105,
      "theoretical_loss": 3.3309076737040915,
      "tokens_seen": 2883780608
    },
    {
      "epoch": 0.75,
      "objective/train/aligned_token_count": 0.40487536787986755,
      "objective/train/avg_token_score": 0.027302350848913193,
      "objective/train/docs_used": 1623211,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.315783977508545,
      "objective/train/original_loss": 4.442750453948975,
      "objective/train/theoretical_loss": 3.3309018068577947,
      "objective/train/tokens_used": 1242705376,
      "objective/train/unlikelihood_loss": 0.08278974145650864,
      "theoretical_loss": 3.3309018068577947,
      "tokens_seen": 2883846144
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012840404040404042,
      "loss": 1.3694,
      "theoretical_loss": 3.3309018068577947,
      "tokens_seen": 2883846144
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012838383838383838,
      "loss": 1.4768,
      "theoretical_loss": 3.330895940182152,
      "tokens_seen": 2883911680
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012836363636363636,
      "loss": 1.1925,
      "theoretical_loss": 3.3308900736771547,
      "tokens_seen": 2883977216
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012834343434343435,
      "loss": 1.5207,
      "theoretical_loss": 3.330884207342794,
      "tokens_seen": 2884042752
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001283232323232323,
      "loss": 1.4402,
      "theoretical_loss": 3.3308783411790595,
      "tokens_seen": 2884108288
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012830303030303032,
      "loss": 1.0385,
      "theoretical_loss": 3.330872475185944,
      "tokens_seen": 2884173824
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012828282828282827,
      "loss": 1.4208,
      "theoretical_loss": 3.3308666093634383,
      "tokens_seen": 2884239360
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012826262626262628,
      "loss": 1.3536,
      "theoretical_loss": 3.3308607437115336,
      "tokens_seen": 2884304896
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012824242424242424,
      "loss": 1.2665,
      "theoretical_loss": 3.3308548782302205,
      "tokens_seen": 2884370432
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012822222222222222,
      "loss": 1.1123,
      "theoretical_loss": 3.330849012919491,
      "tokens_seen": 2884435968
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001282020202020202,
      "loss": 1.3148,
      "theoretical_loss": 3.3308431477793357,
      "tokens_seen": 2884501504
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012818181818181817,
      "loss": 1.2901,
      "theoretical_loss": 3.330837282809746,
      "tokens_seen": 2884567040
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012816161616161618,
      "loss": 1.4227,
      "theoretical_loss": 3.330831418010713,
      "tokens_seen": 2884632576
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012814141414141413,
      "loss": 1.3632,
      "theoretical_loss": 3.3308255533822275,
      "tokens_seen": 2884698112
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012812121212121212,
      "loss": 1.1903,
      "theoretical_loss": 3.330819688924281,
      "tokens_seen": 2884763648
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001281010101010101,
      "loss": 1.4102,
      "theoretical_loss": 3.330813824636865,
      "tokens_seen": 2884829184
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012808080808080809,
      "loss": 1.3184,
      "theoretical_loss": 3.3308079605199703,
      "tokens_seen": 2884894720
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012806060606060607,
      "loss": 1.2563,
      "theoretical_loss": 3.330802096573588,
      "tokens_seen": 2884960256
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012804040404040405,
      "loss": 1.2489,
      "theoretical_loss": 3.3307962327977094,
      "tokens_seen": 2885025792
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000128020202020202,
      "loss": 1.3282,
      "theoretical_loss": 3.330790369192326,
      "tokens_seen": 2885091328
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000128,
      "loss": 1.4901,
      "theoretical_loss": 3.330784505757428,
      "tokens_seen": 2885156864
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012797979797979798,
      "loss": 1.3737,
      "theoretical_loss": 3.3307786424930073,
      "tokens_seen": 2885222400
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012795959595959596,
      "loss": 1.4207,
      "theoretical_loss": 3.330772779399055,
      "tokens_seen": 2885287936
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012793939393939395,
      "loss": 1.2956,
      "theoretical_loss": 3.3307669164755627,
      "tokens_seen": 2885353472
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012791919191919193,
      "loss": 1.4932,
      "theoretical_loss": 3.3307610537225205,
      "tokens_seen": 2885419008
    },
    {
      "epoch": 0.75,
      "objective/train/aligned_token_count": 0.44055473804473877,
      "objective/train/avg_token_score": 0.01748398132622242,
      "objective/train/docs_used": 1624561,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.414048433303833,
      "objective/train/original_loss": 4.665322303771973,
      "objective/train/theoretical_loss": 3.3307551911399207,
      "objective/train/tokens_used": 1244343776,
      "objective/train/unlikelihood_loss": 0.04736669734120369,
      "theoretical_loss": 3.3307551911399207,
      "tokens_seen": 2885484544
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012789898989898991,
      "loss": 1.085,
      "theoretical_loss": 3.3307551911399207,
      "tokens_seen": 2885484544
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012787878787878787,
      "loss": 1.4005,
      "theoretical_loss": 3.3307493287277534,
      "tokens_seen": 2885550080
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012785858585858588,
      "loss": 1.5179,
      "theoretical_loss": 3.3307434664860107,
      "tokens_seen": 2885615616
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012783838383838384,
      "loss": 1.4044,
      "theoretical_loss": 3.3307376044146833,
      "tokens_seen": 2885681152
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012781818181818182,
      "loss": 1.2689,
      "theoretical_loss": 3.330731742513762,
      "tokens_seen": 2885746688
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001277979797979798,
      "loss": 1.5742,
      "theoretical_loss": 3.3307258807832394,
      "tokens_seen": 2885812224
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012777777777777776,
      "loss": 1.2521,
      "theoretical_loss": 3.3307200192231052,
      "tokens_seen": 2885877760
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012775757575757577,
      "loss": 1.124,
      "theoretical_loss": 3.330714157833351,
      "tokens_seen": 2885943296
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012773737373737373,
      "loss": 1.3105,
      "theoretical_loss": 3.330708296613968,
      "tokens_seen": 2886008832
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012771717171717174,
      "loss": 1.1682,
      "theoretical_loss": 3.3307024355649477,
      "tokens_seen": 2886074368
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001276969696969697,
      "loss": 1.4036,
      "theoretical_loss": 3.330696574686281,
      "tokens_seen": 2886139904
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012767676767676766,
      "loss": 1.3037,
      "theoretical_loss": 3.330690713977959,
      "tokens_seen": 2886205440
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012765656565656567,
      "loss": 1.1759,
      "theoretical_loss": 3.3306848534399736,
      "tokens_seen": 2886270976
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012763636363636362,
      "loss": 1.4761,
      "theoretical_loss": 3.330678993072315,
      "tokens_seen": 2886336512
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012761616161616164,
      "loss": 1.2067,
      "theoretical_loss": 3.330673132874974,
      "tokens_seen": 2886402048
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001275959595959596,
      "loss": 1.3045,
      "theoretical_loss": 3.3306672728479434,
      "tokens_seen": 2886467584
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012757575757575758,
      "loss": 1.3918,
      "theoretical_loss": 3.3306614129912138,
      "tokens_seen": 2886533120
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012755555555555556,
      "loss": 1.2974,
      "theoretical_loss": 3.3306555533047755,
      "tokens_seen": 2886598656
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012753535353535354,
      "loss": 1.4167,
      "theoretical_loss": 3.3306496937886205,
      "tokens_seen": 2886664192
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012751515151515153,
      "loss": 1.2803,
      "theoretical_loss": 3.3306438344427396,
      "tokens_seen": 2886729728
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012749494949494948,
      "loss": 1.2774,
      "theoretical_loss": 3.3306379752671247,
      "tokens_seen": 2886795264
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012747474747474747,
      "loss": 1.4185,
      "theoretical_loss": 3.330632116261766,
      "tokens_seen": 2886860800
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012745454545454545,
      "loss": 1.4947,
      "theoretical_loss": 3.330626257426655,
      "tokens_seen": 2886926336
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012743434343434344,
      "loss": 1.3328,
      "theoretical_loss": 3.3306203987617833,
      "tokens_seen": 2886991872
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012741414141414142,
      "loss": 1.3411,
      "theoretical_loss": 3.330614540267142,
      "tokens_seen": 2887057408
    },
    {
      "epoch": 0.75,
      "objective/train/aligned_token_count": 0.4157502353191376,
      "objective/train/avg_token_score": 0.011384841054677963,
      "objective/train/docs_used": 1624832,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2586262226104736,
      "objective/train/original_loss": 4.5518012046813965,
      "objective/train/theoretical_loss": 3.330608681942722,
      "objective/train/tokens_used": 1245982176,
      "objective/train/unlikelihood_loss": 0.05512404441833496,
      "theoretical_loss": 3.330608681942722,
      "tokens_seen": 2887122944
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001273939393939394,
      "loss": 1.3522,
      "theoretical_loss": 3.330608681942722,
      "tokens_seen": 2887122944
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012737373737373736,
      "loss": 1.4584,
      "theoretical_loss": 3.3306028237885146,
      "tokens_seen": 2887188480
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012735353535353537,
      "loss": 1.2843,
      "theoretical_loss": 3.330596965804511,
      "tokens_seen": 2887254016
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012733333333333333,
      "loss": 1.3421,
      "theoretical_loss": 3.3305911079907027,
      "tokens_seen": 2887319552
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001273131313131313,
      "loss": 1.0335,
      "theoretical_loss": 3.33058525034708,
      "tokens_seen": 2887385088
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001272929292929293,
      "loss": 1.2788,
      "theoretical_loss": 3.330579392873635,
      "tokens_seen": 2887450624
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012727272727272725,
      "loss": 1.4974,
      "theoretical_loss": 3.3305735355703585,
      "tokens_seen": 2887516160
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012725252525252527,
      "loss": 1.1599,
      "theoretical_loss": 3.3305676784372418,
      "tokens_seen": 2887581696
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012723232323232322,
      "loss": 1.4186,
      "theoretical_loss": 3.330561821474276,
      "tokens_seen": 2887647232
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012721212121212123,
      "loss": 1.4257,
      "theoretical_loss": 3.3305559646814524,
      "tokens_seen": 2887712768
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001271919191919192,
      "loss": 1.4157,
      "theoretical_loss": 3.3305501080587625,
      "tokens_seen": 2887778304
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012717171717171717,
      "loss": 1.4393,
      "theoretical_loss": 3.3305442516061965,
      "tokens_seen": 2887843840
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012715151515151516,
      "loss": 1.6181,
      "theoretical_loss": 3.330538395323747,
      "tokens_seen": 2887909376
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012713131313131311,
      "loss": 1.2534,
      "theoretical_loss": 3.3305325392114034,
      "tokens_seen": 2887974912
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012711111111111113,
      "loss": 1.5086,
      "theoretical_loss": 3.3305266832691585,
      "tokens_seen": 2888040448
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012709090909090908,
      "loss": 1.323,
      "theoretical_loss": 3.3305208274970033,
      "tokens_seen": 2888105984
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001270707070707071,
      "loss": 1.2899,
      "theoretical_loss": 3.330514971894928,
      "tokens_seen": 2888171520
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012705050505050505,
      "loss": 1.2737,
      "theoretical_loss": 3.3305091164629252,
      "tokens_seen": 2888237056
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012703030303030303,
      "loss": 1.3216,
      "theoretical_loss": 3.3305032612009846,
      "tokens_seen": 2888302592
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012701010101010102,
      "loss": 1.3429,
      "theoretical_loss": 3.3304974061090986,
      "tokens_seen": 2888368128
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012698989898989898,
      "loss": 1.3737,
      "theoretical_loss": 3.330491551187258,
      "tokens_seen": 2888433664
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000126969696969697,
      "loss": 1.2255,
      "theoretical_loss": 3.3304856964354537,
      "tokens_seen": 2888499200
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012694949494949494,
      "loss": 1.3616,
      "theoretical_loss": 3.330479841853677,
      "tokens_seen": 2888564736
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012692929292929293,
      "loss": 1.3394,
      "theoretical_loss": 3.33047398744192,
      "tokens_seen": 2888630272
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001269090909090909,
      "loss": 1.3088,
      "theoretical_loss": 3.3304681332001724,
      "tokens_seen": 2888695808
    },
    {
      "epoch": 0.75,
      "objective/train/aligned_token_count": 0.4722629487514496,
      "objective/train/avg_token_score": 0.012102199718356133,
      "objective/train/docs_used": 1626153,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6171066761016846,
      "objective/train/original_loss": 4.58645486831665,
      "objective/train/theoretical_loss": 3.3304622791284264,
      "objective/train/tokens_used": 1247620576,
      "objective/train/unlikelihood_loss": 0.04844868928194046,
      "theoretical_loss": 3.3304622791284264,
      "tokens_seen": 2888761344
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001268888888888889,
      "loss": 1.4418,
      "theoretical_loss": 3.3304622791284264,
      "tokens_seen": 2888761344
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012686868686868688,
      "loss": 1.3073,
      "theoretical_loss": 3.3304564252266733,
      "tokens_seen": 2888826880
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012684848484848486,
      "loss": 1.372,
      "theoretical_loss": 3.3304505714949033,
      "tokens_seen": 2888892416
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012682828282828282,
      "loss": 1.3156,
      "theoretical_loss": 3.330444717933109,
      "tokens_seen": 2888957952
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001268080808080808,
      "loss": 1.3419,
      "theoretical_loss": 3.3304388645412804,
      "tokens_seen": 2889023488
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001267878787878788,
      "loss": 1.1005,
      "theoretical_loss": 3.3304330113194096,
      "tokens_seen": 2889089024
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012676767676767677,
      "loss": 1.3145,
      "theoretical_loss": 3.3304271582674874,
      "tokens_seen": 2889154560
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012674747474747476,
      "loss": 1.2181,
      "theoretical_loss": 3.3304213053855047,
      "tokens_seen": 2889220096
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001267272727272727,
      "loss": 1.5276,
      "theoretical_loss": 3.3304154526734533,
      "tokens_seen": 2889285632
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012670707070707072,
      "loss": 1.1642,
      "theoretical_loss": 3.330409600131324,
      "tokens_seen": 2889351168
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012668686868686868,
      "loss": 1.4378,
      "theoretical_loss": 3.3304037477591084,
      "tokens_seen": 2889416704
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001266666666666667,
      "loss": 1.4459,
      "theoretical_loss": 3.330397895556797,
      "tokens_seen": 2889482240
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012664646464646465,
      "loss": 1.0932,
      "theoretical_loss": 3.330392043524382,
      "tokens_seen": 2889547776
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001266262626262626,
      "loss": 1.467,
      "theoretical_loss": 3.3303861916618542,
      "tokens_seen": 2889613312
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012660606060606062,
      "loss": 1.327,
      "theoretical_loss": 3.3303803399692047,
      "tokens_seen": 2889678848
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012658585858585857,
      "loss": 1.1952,
      "theoretical_loss": 3.3303744884464246,
      "tokens_seen": 2889744384
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012656565656565658,
      "loss": 1.2312,
      "theoretical_loss": 3.330368637093505,
      "tokens_seen": 2889809920
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012654545454545454,
      "loss": 1.2778,
      "theoretical_loss": 3.3303627859104377,
      "tokens_seen": 2889875456
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012652525252525253,
      "loss": 1.2995,
      "theoretical_loss": 3.3303569348972135,
      "tokens_seen": 2889940992
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001265050505050505,
      "loss": 1.3202,
      "theoretical_loss": 3.3303510840538237,
      "tokens_seen": 2890006528
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001264848484848485,
      "loss": 1.1705,
      "theoretical_loss": 3.3303452333802595,
      "tokens_seen": 2890072064
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012646464646464648,
      "loss": 1.2466,
      "theoretical_loss": 3.330339382876512,
      "tokens_seen": 2890137600
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012644444444444443,
      "loss": 1.5237,
      "theoretical_loss": 3.330333532542573,
      "tokens_seen": 2890203136
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012642424242424242,
      "loss": 1.5285,
      "theoretical_loss": 3.330327682378433,
      "tokens_seen": 2890268672
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001264040404040404,
      "loss": 1.1832,
      "theoretical_loss": 3.330321832384084,
      "tokens_seen": 2890334208
    },
    {
      "epoch": 0.75,
      "objective/train/aligned_token_count": 0.3943059742450714,
      "objective/train/avg_token_score": 0.026357710361480713,
      "objective/train/docs_used": 1626814,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1431066989898682,
      "objective/train/original_loss": 4.6745171546936035,
      "objective/train/theoretical_loss": 3.330315982559516,
      "objective/train/tokens_used": 1249258976,
      "objective/train/unlikelihood_loss": 0.06599657237529755,
      "theoretical_loss": 3.330315982559516,
      "tokens_seen": 2890399744
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012638383838383839,
      "loss": 1.3447,
      "theoretical_loss": 3.330315982559516,
      "tokens_seen": 2890399744
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012636363636363637,
      "loss": 1.3457,
      "theoretical_loss": 3.3303101329047213,
      "tokens_seen": 2890465280
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012634343434343435,
      "loss": 1.2577,
      "theoretical_loss": 3.3303042834196908,
      "tokens_seen": 2890530816
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001263232323232323,
      "loss": 1.4478,
      "theoretical_loss": 3.3302984341044155,
      "tokens_seen": 2890596352
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012630303030303032,
      "loss": 1.3188,
      "theoretical_loss": 3.330292584958887,
      "tokens_seen": 2890661888
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012628282828282828,
      "loss": 1.3523,
      "theoretical_loss": 3.3302867359830968,
      "tokens_seen": 2890727424
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012626262626262626,
      "loss": 1.4365,
      "theoretical_loss": 3.330280887177035,
      "tokens_seen": 2890792960
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012624242424242425,
      "loss": 1.3245,
      "theoretical_loss": 3.330275038540694,
      "tokens_seen": 2890858496
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012622222222222223,
      "loss": 1.4644,
      "theoretical_loss": 3.3302691900740644,
      "tokens_seen": 2890924032
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012620202020202021,
      "loss": 1.4971,
      "theoretical_loss": 3.3302633417771377,
      "tokens_seen": 2890989568
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012618181818181817,
      "loss": 1.5727,
      "theoretical_loss": 3.3302574936499045,
      "tokens_seen": 2891055104
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012616161616161618,
      "loss": 1.5342,
      "theoretical_loss": 3.330251645692357,
      "tokens_seen": 2891120640
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012614141414141414,
      "loss": 1.324,
      "theoretical_loss": 3.3302457979044857,
      "tokens_seen": 2891186176
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012612121212121212,
      "loss": 1.2426,
      "theoretical_loss": 3.3302399502862823,
      "tokens_seen": 2891251712
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001261010101010101,
      "loss": 1.1038,
      "theoretical_loss": 3.3302341028377374,
      "tokens_seen": 2891317248
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012608080808080806,
      "loss": 1.3057,
      "theoretical_loss": 3.3302282555588434,
      "tokens_seen": 2891382784
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012606060606060608,
      "loss": 1.1801,
      "theoretical_loss": 3.33022240844959,
      "tokens_seen": 2891448320
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012604040404040403,
      "loss": 1.4974,
      "theoretical_loss": 3.3302165615099697,
      "tokens_seen": 2891513856
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012602020202020204,
      "loss": 1.4644,
      "theoretical_loss": 3.330210714739973,
      "tokens_seen": 2891579392
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000126,
      "loss": 1.3342,
      "theoretical_loss": 3.3302048681395915,
      "tokens_seen": 2891644928
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012597979797979798,
      "loss": 1.6403,
      "theoretical_loss": 3.3301990217088164,
      "tokens_seen": 2891710464
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012595959595959597,
      "loss": 1.5271,
      "theoretical_loss": 3.330193175447639,
      "tokens_seen": 2891776000
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012593939393939392,
      "loss": 1.6169,
      "theoretical_loss": 3.3301873293560504,
      "tokens_seen": 2891841536
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012591919191919194,
      "loss": 1.2298,
      "theoretical_loss": 3.3301814834340417,
      "tokens_seen": 2891907072
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001258989898989899,
      "loss": 1.3925,
      "theoretical_loss": 3.330175637681604,
      "tokens_seen": 2891972608
    },
    {
      "epoch": 0.75,
      "objective/train/aligned_token_count": 0.5057429075241089,
      "objective/train/avg_token_score": 0.026666805148124695,
      "objective/train/docs_used": 1628086,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5554147958755493,
      "objective/train/original_loss": 4.633951187133789,
      "objective/train/theoretical_loss": 3.3301697920987294,
      "objective/train/tokens_used": 1250897376,
      "objective/train/unlikelihood_loss": 0.040260396897792816,
      "theoretical_loss": 3.3301697920987294,
      "tokens_seen": 2892038144
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012587878787878788,
      "loss": 1.3422,
      "theoretical_loss": 3.3301697920987294,
      "tokens_seen": 2892038144
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012585858585858586,
      "loss": 1.1774,
      "theoretical_loss": 3.330163946685408,
      "tokens_seen": 2892103680
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012583838383838384,
      "loss": 1.603,
      "theoretical_loss": 3.330158101441632,
      "tokens_seen": 2892169216
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012581818181818183,
      "loss": 1.2025,
      "theoretical_loss": 3.3301522563673918,
      "tokens_seen": 2892234752
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001257979797979798,
      "loss": 1.2651,
      "theoretical_loss": 3.330146411462679,
      "tokens_seen": 2892300288
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012577777777777777,
      "loss": 1.2539,
      "theoretical_loss": 3.3301405667274855,
      "tokens_seen": 2892365824
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012575757575757575,
      "loss": 1.4266,
      "theoretical_loss": 3.3301347221618016,
      "tokens_seen": 2892431360
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012573737373737374,
      "loss": 1.616,
      "theoretical_loss": 3.330128877765619,
      "tokens_seen": 2892496896
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012571717171717172,
      "loss": 1.3199,
      "theoretical_loss": 3.330123033538929,
      "tokens_seen": 2892562432
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001256969696969697,
      "loss": 1.2791,
      "theoretical_loss": 3.3301171894817223,
      "tokens_seen": 2892627968
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012567676767676766,
      "loss": 1.2602,
      "theoretical_loss": 3.330111345593991,
      "tokens_seen": 2892693504
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012565656565656567,
      "loss": 1.34,
      "theoretical_loss": 3.3301055018757255,
      "tokens_seen": 2892759040
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012563636363636363,
      "loss": 1.3325,
      "theoretical_loss": 3.3300996583269176,
      "tokens_seen": 2892824576
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012561616161616164,
      "loss": 1.443,
      "theoretical_loss": 3.330093814947558,
      "tokens_seen": 2892890112
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001255959595959596,
      "loss": 1.4115,
      "theoretical_loss": 3.330087971737639,
      "tokens_seen": 2892955648
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012557575757575755,
      "loss": 1.5025,
      "theoretical_loss": 3.3300821286971507,
      "tokens_seen": 2893021184
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012555555555555557,
      "loss": 1.2013,
      "theoretical_loss": 3.330076285826085,
      "tokens_seen": 2893086720
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012553535353535352,
      "loss": 1.4466,
      "theoretical_loss": 3.3300704431244332,
      "tokens_seen": 2893152256
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012551515151515153,
      "loss": 1.3402,
      "theoretical_loss": 3.330064600592186,
      "tokens_seen": 2893217792
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001254949494949495,
      "loss": 1.49,
      "theoretical_loss": 3.330058758229335,
      "tokens_seen": 2893283328
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012547474747474747,
      "loss": 1.4112,
      "theoretical_loss": 3.3300529160358714,
      "tokens_seen": 2893348864
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012545454545454546,
      "loss": 1.4615,
      "theoretical_loss": 3.3300470740117865,
      "tokens_seen": 2893414400
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012543434343434342,
      "loss": 1.3142,
      "theoretical_loss": 3.330041232157072,
      "tokens_seen": 2893479936
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012541414141414143,
      "loss": 1.5414,
      "theoretical_loss": 3.3300353904717177,
      "tokens_seen": 2893545472
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012539393939393938,
      "loss": 1.0134,
      "theoretical_loss": 3.3300295489557166,
      "tokens_seen": 2893611008
    },
    {
      "epoch": 0.75,
      "objective/train/aligned_token_count": 0.3784824013710022,
      "objective/train/avg_token_score": 0.017129698768258095,
      "objective/train/docs_used": 1628768,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0766042470932007,
      "objective/train/original_loss": 4.819118976593018,
      "objective/train/theoretical_loss": 3.330023707609059,
      "objective/train/tokens_used": 1252535776,
      "objective/train/unlikelihood_loss": 0.08477392047643661,
      "theoretical_loss": 3.330023707609059,
      "tokens_seen": 2893676544
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001253737373737374,
      "loss": 1.3385,
      "theoretical_loss": 3.330023707609059,
      "tokens_seen": 2893676544
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012535353535353535,
      "loss": 1.6607,
      "theoretical_loss": 3.3300178664317364,
      "tokens_seen": 2893742080
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012533333333333334,
      "loss": 1.4889,
      "theoretical_loss": 3.33001202542374,
      "tokens_seen": 2893807616
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012531313131313132,
      "loss": 1.3552,
      "theoretical_loss": 3.330006184585061,
      "tokens_seen": 2893873152
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001252929292929293,
      "loss": 1.1332,
      "theoretical_loss": 3.3300003439156907,
      "tokens_seen": 2893938688
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001252727272727273,
      "loss": 1.2995,
      "theoretical_loss": 3.3299945034156204,
      "tokens_seen": 2894004224
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012525252525252524,
      "loss": 1.4272,
      "theoretical_loss": 3.3299886630848414,
      "tokens_seen": 2894069760
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012523232323232323,
      "loss": 1.2576,
      "theoretical_loss": 3.329982822923345,
      "tokens_seen": 2894135296
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001252121212121212,
      "loss": 1.4052,
      "theoretical_loss": 3.3299769829311217,
      "tokens_seen": 2894200832
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001251919191919192,
      "loss": 1.4468,
      "theoretical_loss": 3.329971143108164,
      "tokens_seen": 2894266368
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012517171717171718,
      "loss": 1.3687,
      "theoretical_loss": 3.3299653034544625,
      "tokens_seen": 2894331904
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012515151515151516,
      "loss": 1.4688,
      "theoretical_loss": 3.3299594639700087,
      "tokens_seen": 2894397440
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012513131313131312,
      "loss": 1.5589,
      "theoretical_loss": 3.3299536246547934,
      "tokens_seen": 2894462976
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012511111111111113,
      "loss": 1.3401,
      "theoretical_loss": 3.3299477855088084,
      "tokens_seen": 2894528512
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001250909090909091,
      "loss": 1.2155,
      "theoretical_loss": 3.3299419465320446,
      "tokens_seen": 2894594048
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012507070707070707,
      "loss": 1.3562,
      "theoretical_loss": 3.3299361077244933,
      "tokens_seen": 2894659584
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012505050505050506,
      "loss": 1.4695,
      "theoretical_loss": 3.3299302690861463,
      "tokens_seen": 2894725120
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000125030303030303,
      "loss": 1.4105,
      "theoretical_loss": 3.3299244306169937,
      "tokens_seen": 2894790656
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012501010101010102,
      "loss": 1.5248,
      "theoretical_loss": 3.329918592317028,
      "tokens_seen": 2894856192
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012498989898989898,
      "loss": 1.3916,
      "theoretical_loss": 3.32991275418624,
      "tokens_seen": 2894921728
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012496969696969696,
      "loss": 1.6402,
      "theoretical_loss": 3.329906916224621,
      "tokens_seen": 2894987264
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012494949494949495,
      "loss": 1.3386,
      "theoretical_loss": 3.3299010784321617,
      "tokens_seen": 2895052800
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012492929292929293,
      "loss": 1.1376,
      "theoretical_loss": 3.3298952408088542,
      "tokens_seen": 2895118336
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012490909090909092,
      "loss": 1.3639,
      "theoretical_loss": 3.3298894033546897,
      "tokens_seen": 2895183872
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012488888888888887,
      "loss": 1.3993,
      "theoretical_loss": 3.329883566069659,
      "tokens_seen": 2895249408
    },
    {
      "epoch": 0.75,
      "objective/train/aligned_token_count": 0.38202589750289917,
      "objective/train/avg_token_score": 0.017982229590415955,
      "objective/train/docs_used": 1629434,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2198495864868164,
      "objective/train/original_loss": 4.9608378410339355,
      "objective/train/theoretical_loss": 3.329877728953753,
      "objective/train/tokens_used": 1254174176,
      "objective/train/unlikelihood_loss": 0.052344419062137604,
      "theoretical_loss": 3.329877728953753,
      "tokens_seen": 2895314944
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012486868686868686,
      "loss": 1.1907,
      "theoretical_loss": 3.329877728953753,
      "tokens_seen": 2895314944
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012484848484848484,
      "loss": 1.3454,
      "theoretical_loss": 3.329871892006964,
      "tokens_seen": 2895380480
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012482828282828283,
      "loss": 1.4568,
      "theoretical_loss": 3.3298660552292834,
      "tokens_seen": 2895446016
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001248080808080808,
      "loss": 1.4857,
      "theoretical_loss": 3.329860218620701,
      "tokens_seen": 2895511552
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001247878787878788,
      "loss": 1.4069,
      "theoretical_loss": 3.3298543821812094,
      "tokens_seen": 2895577088
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012476767676767678,
      "loss": 1.6305,
      "theoretical_loss": 3.3298485459107994,
      "tokens_seen": 2895642624
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012474747474747476,
      "loss": 1.3936,
      "theoretical_loss": 3.3298427098094625,
      "tokens_seen": 2895708160
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012472727272727272,
      "loss": 1.2355,
      "theoretical_loss": 3.3298368738771895,
      "tokens_seen": 2895773696
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001247070707070707,
      "loss": 1.3726,
      "theoretical_loss": 3.329831038113972,
      "tokens_seen": 2895839232
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012468686868686869,
      "loss": 1.409,
      "theoretical_loss": 3.3298252025198014,
      "tokens_seen": 2895904768
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012466666666666667,
      "loss": 1.4153,
      "theoretical_loss": 3.329819367094669,
      "tokens_seen": 2895970304
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012464646464646465,
      "loss": 1.5785,
      "theoretical_loss": 3.3298135318385653,
      "tokens_seen": 2896035840
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012462626262626264,
      "loss": 1.4229,
      "theoretical_loss": 3.329807696751483,
      "tokens_seen": 2896101376
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012460606060606062,
      "loss": 1.5344,
      "theoretical_loss": 3.3298018618334115,
      "tokens_seen": 2896166912
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001245858585858586,
      "loss": 1.1929,
      "theoretical_loss": 3.329796027084344,
      "tokens_seen": 2896232448
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012456565656565656,
      "loss": 1.2407,
      "theoretical_loss": 3.3297901925042708,
      "tokens_seen": 2896297984
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012454545454545455,
      "loss": 1.2301,
      "theoretical_loss": 3.329784358093183,
      "tokens_seen": 2896363520
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012452525252525253,
      "loss": 1.3249,
      "theoretical_loss": 3.3297785238510724,
      "tokens_seen": 2896429056
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012450505050505051,
      "loss": 1.2611,
      "theoretical_loss": 3.32977268977793,
      "tokens_seen": 2896494592
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001244848484848485,
      "loss": 1.1716,
      "theoretical_loss": 3.329766855873747,
      "tokens_seen": 2896560128
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012446464646464646,
      "loss": 1.5149,
      "theoretical_loss": 3.3297610221385154,
      "tokens_seen": 2896625664
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012444444444444444,
      "loss": 1.1334,
      "theoretical_loss": 3.329755188572226,
      "tokens_seen": 2896691200
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012442424242424242,
      "loss": 1.4831,
      "theoretical_loss": 3.329749355174869,
      "tokens_seen": 2896756736
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001244040404040404,
      "loss": 1.2823,
      "theoretical_loss": 3.3297435219464377,
      "tokens_seen": 2896822272
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001243838383838384,
      "loss": 1.2953,
      "theoretical_loss": 3.329737688886922,
      "tokens_seen": 2896887808
    },
    {
      "epoch": 0.75,
      "objective/train/aligned_token_count": 0.369562566280365,
      "objective/train/avg_token_score": 0.0063091726042330265,
      "objective/train/docs_used": 1630757,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1443532705307007,
      "objective/train/original_loss": 4.169307708740234,
      "objective/train/theoretical_loss": 3.329731855996313,
      "objective/train/tokens_used": 1255812576,
      "objective/train/unlikelihood_loss": 0.06046897917985916,
      "theoretical_loss": 3.329731855996313,
      "tokens_seen": 2896953344
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012436363636363635,
      "loss": 1.3957,
      "theoretical_loss": 3.329731855996313,
      "tokens_seen": 2896953344
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012434343434343433,
      "loss": 1.2954,
      "theoretical_loss": 3.3297260232746035,
      "tokens_seen": 2897018880
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012432323232323232,
      "loss": 1.3524,
      "theoretical_loss": 3.3297201907217833,
      "tokens_seen": 2897084416
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001243030303030303,
      "loss": 1.396,
      "theoretical_loss": 3.3297143583378443,
      "tokens_seen": 2897149952
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012428282828282828,
      "loss": 1.2749,
      "theoretical_loss": 3.329708526122778,
      "tokens_seen": 2897215488
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012426262626262627,
      "loss": 1.3184,
      "theoretical_loss": 3.329702694076575,
      "tokens_seen": 2897281024
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012424242424242425,
      "loss": 1.369,
      "theoretical_loss": 3.3296968621992273,
      "tokens_seen": 2897346560
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012422222222222224,
      "loss": 1.3569,
      "theoretical_loss": 3.329691030490726,
      "tokens_seen": 2897412096
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001242020202020202,
      "loss": 1.5062,
      "theoretical_loss": 3.329685198951062,
      "tokens_seen": 2897477632
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012418181818181818,
      "loss": 1.3398,
      "theoretical_loss": 3.3296793675802268,
      "tokens_seen": 2897543168
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012416161616161616,
      "loss": 1.2619,
      "theoretical_loss": 3.329673536378212,
      "tokens_seen": 2897608704
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012414141414141414,
      "loss": 1.2399,
      "theoretical_loss": 3.3296677053450088,
      "tokens_seen": 2897674240
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012412121212121213,
      "loss": 1.4169,
      "theoretical_loss": 3.329661874480608,
      "tokens_seen": 2897739776
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001241010101010101,
      "loss": 1.4573,
      "theoretical_loss": 3.329656043785002,
      "tokens_seen": 2897805312
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001240808080808081,
      "loss": 1.3359,
      "theoretical_loss": 3.3296502132581804,
      "tokens_seen": 2897870848
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012406060606060608,
      "loss": 1.1994,
      "theoretical_loss": 3.329644382900136,
      "tokens_seen": 2897936384
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012404040404040404,
      "loss": 1.4342,
      "theoretical_loss": 3.3296385527108594,
      "tokens_seen": 2898001920
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012402020202020202,
      "loss": 1.2596,
      "theoretical_loss": 3.3296327226903424,
      "tokens_seen": 2898067456
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000124,
      "loss": 1.3762,
      "theoretical_loss": 3.3296268928385757,
      "tokens_seen": 2898132992
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.000123979797979798,
      "loss": 1.3989,
      "theoretical_loss": 3.3296210631555505,
      "tokens_seen": 2898198528
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012395959595959597,
      "loss": 1.5365,
      "theoretical_loss": 3.3296152336412588,
      "tokens_seen": 2898264064
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012393939393939393,
      "loss": 1.5379,
      "theoretical_loss": 3.3296094042956916,
      "tokens_seen": 2898329600
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012391919191919191,
      "loss": 1.535,
      "theoretical_loss": 3.32960357511884,
      "tokens_seen": 2898395136
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0001238989898989899,
      "loss": 1.3318,
      "theoretical_loss": 3.3295977461106956,
      "tokens_seen": 2898460672
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00012387878787878788,
      "loss": 1.2931,
      "theoretical_loss": 3.32959191727125,
      "tokens_seen": 2898526208
    },
    {
      "epoch": 0.75,
      "objective/train/aligned_token_count": 0.28879520297050476,
      "objective/train/avg_token_score": 0.012048942968249321,
      "objective/train/docs_used": 1631491,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9138427376747131,
      "objective/train/original_loss": 4.543020725250244,
      "objective/train/theoretical_loss": 3.329586088600493,
      "objective/train/tokens_used": 1257450976,
      "objective/train/unlikelihood_loss": 0.056786537170410156,
      "theoretical_loss": 3.329586088600493,
      "tokens_seen": 2898591744
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012385858585858587,
      "loss": 1.3359,
      "theoretical_loss": 3.329586088600493,
      "tokens_seen": 2898591744
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012383838383838385,
      "loss": 1.2155,
      "theoretical_loss": 3.329580260098418,
      "tokens_seen": 2898657280
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001238181818181818,
      "loss": 1.4043,
      "theoretical_loss": 3.329574431765015,
      "tokens_seen": 2898722816
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001237979797979798,
      "loss": 1.5907,
      "theoretical_loss": 3.329568603600275,
      "tokens_seen": 2898788352
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012377777777777777,
      "loss": 1.3556,
      "theoretical_loss": 3.3295627756041903,
      "tokens_seen": 2898853888
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012375757575757576,
      "loss": 1.6546,
      "theoretical_loss": 3.329556947776752,
      "tokens_seen": 2898919424
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012373737373737374,
      "loss": 1.4891,
      "theoretical_loss": 3.3295511201179506,
      "tokens_seen": 2898984960
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012371717171717173,
      "loss": 1.714,
      "theoretical_loss": 3.329545292627779,
      "tokens_seen": 2899050496
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012369696969696968,
      "loss": 1.5211,
      "theoretical_loss": 3.329539465306227,
      "tokens_seen": 2899116032
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012367676767676767,
      "loss": 1.137,
      "theoretical_loss": 3.329533638153286,
      "tokens_seen": 2899181568
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012365656565656565,
      "loss": 1.2994,
      "theoretical_loss": 3.329527811168948,
      "tokens_seen": 2899247104
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012363636363636364,
      "loss": 1.5951,
      "theoretical_loss": 3.329521984353204,
      "tokens_seen": 2899312640
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012361616161616162,
      "loss": 1.239,
      "theoretical_loss": 3.3295161577060455,
      "tokens_seen": 2899378176
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001235959595959596,
      "loss": 1.4808,
      "theoretical_loss": 3.329510331227464,
      "tokens_seen": 2899443712
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001235757575757576,
      "loss": 1.5078,
      "theoretical_loss": 3.32950450491745,
      "tokens_seen": 2899509248
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012355555555555557,
      "loss": 1.3809,
      "theoretical_loss": 3.3294986787759955,
      "tokens_seen": 2899574784
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012353535353535356,
      "loss": 1.5177,
      "theoretical_loss": 3.3294928528030914,
      "tokens_seen": 2899640320
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001235151515151515,
      "loss": 1.2928,
      "theoretical_loss": 3.3294870269987293,
      "tokens_seen": 2899705856
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001234949494949495,
      "loss": 1.2821,
      "theoretical_loss": 3.329481201362901,
      "tokens_seen": 2899771392
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012347474747474748,
      "loss": 1.651,
      "theoretical_loss": 3.3294753758955964,
      "tokens_seen": 2899836928
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012345454545454546,
      "loss": 1.4008,
      "theoretical_loss": 3.329469550596808,
      "tokens_seen": 2899902464
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012343434343434345,
      "loss": 1.4151,
      "theoretical_loss": 3.329463725466527,
      "tokens_seen": 2899968000
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012341414141414143,
      "loss": 1.283,
      "theoretical_loss": 3.329457900504744,
      "tokens_seen": 2900033536
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001233939393939394,
      "loss": 1.6276,
      "theoretical_loss": 3.329452075711451,
      "tokens_seen": 2900099072
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012337373737373737,
      "loss": 1.5842,
      "theoretical_loss": 3.32944625108664,
      "tokens_seen": 2900164608
    },
    {
      "epoch": 0.76,
      "objective/train/aligned_token_count": 0.2969208061695099,
      "objective/train/avg_token_score": 0.008718152530491352,
      "objective/train/docs_used": 1632851,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.998846709728241,
      "objective/train/original_loss": 4.306954383850098,
      "objective/train/theoretical_loss": 3.3294404266303004,
      "objective/train/tokens_used": 1259089376,
      "objective/train/unlikelihood_loss": 0.07249090075492859,
      "theoretical_loss": 3.3294404266303004,
      "tokens_seen": 2900230144
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012335353535353536,
      "loss": 1.1673,
      "theoretical_loss": 3.3294404266303004,
      "tokens_seen": 2900230144
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012333333333333334,
      "loss": 1.3236,
      "theoretical_loss": 3.329434602342425,
      "tokens_seen": 2900295680
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012331313131313132,
      "loss": 1.4864,
      "theoretical_loss": 3.3294287782230048,
      "tokens_seen": 2900361216
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012329292929292928,
      "loss": 1.5063,
      "theoretical_loss": 3.329422954272031,
      "tokens_seen": 2900426752
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012327272727272727,
      "loss": 1.4567,
      "theoretical_loss": 3.3294171304894946,
      "tokens_seen": 2900492288
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012325252525252525,
      "loss": 1.1983,
      "theoretical_loss": 3.3294113068753877,
      "tokens_seen": 2900557824
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012323232323232323,
      "loss": 1.2926,
      "theoretical_loss": 3.3294054834297007,
      "tokens_seen": 2900623360
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012321212121212122,
      "loss": 1.6331,
      "theoretical_loss": 3.3293996601524256,
      "tokens_seen": 2900688896
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001231919191919192,
      "loss": 1.374,
      "theoretical_loss": 3.329393837043554,
      "tokens_seen": 2900754432
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012317171717171716,
      "loss": 1.3147,
      "theoretical_loss": 3.3293880141030763,
      "tokens_seen": 2900819968
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012315151515151514,
      "loss": 1.2146,
      "theoretical_loss": 3.3293821913309847,
      "tokens_seen": 2900885504
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012313131313131313,
      "loss": 1.2881,
      "theoretical_loss": 3.3293763687272695,
      "tokens_seen": 2900951040
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001231111111111111,
      "loss": 1.1138,
      "theoretical_loss": 3.329370546291923,
      "tokens_seen": 2901016576
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001230909090909091,
      "loss": 1.2277,
      "theoretical_loss": 3.3293647240249364,
      "tokens_seen": 2901082112
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012307070707070708,
      "loss": 1.4692,
      "theoretical_loss": 3.3293589019263004,
      "tokens_seen": 2901147648
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012305050505050506,
      "loss": 1.2525,
      "theoretical_loss": 3.329353079996007,
      "tokens_seen": 2901213184
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012303030303030305,
      "loss": 1.6361,
      "theoretical_loss": 3.3293472582340473,
      "tokens_seen": 2901278720
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012301010101010103,
      "loss": 1.282,
      "theoretical_loss": 3.329341436640412,
      "tokens_seen": 2901344256
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000122989898989899,
      "loss": 1.5143,
      "theoretical_loss": 3.329335615215094,
      "tokens_seen": 2901409792
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012296969696969697,
      "loss": 1.3458,
      "theoretical_loss": 3.329329793958083,
      "tokens_seen": 2901475328
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012294949494949495,
      "loss": 1.3288,
      "theoretical_loss": 3.3293239728693713,
      "tokens_seen": 2901540864
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012292929292929294,
      "loss": 1.4722,
      "theoretical_loss": 3.3293181519489496,
      "tokens_seen": 2901606400
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012290909090909092,
      "loss": 1.4028,
      "theoretical_loss": 3.32931233119681,
      "tokens_seen": 2901671936
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001228888888888889,
      "loss": 1.3867,
      "theoretical_loss": 3.329306510612943,
      "tokens_seen": 2901737472
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012286868686868686,
      "loss": 1.168,
      "theoretical_loss": 3.3293006901973405,
      "tokens_seen": 2901803008
    },
    {
      "epoch": 0.76,
      "objective/train/aligned_token_count": 0.517412006855011,
      "objective/train/avg_token_score": 0.008600871078670025,
      "objective/train/docs_used": 1633283,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5087919235229492,
      "objective/train/original_loss": 3.996612548828125,
      "objective/train/theoretical_loss": 3.329294869949994,
      "objective/train/tokens_used": 1260727776,
      "objective/train/unlikelihood_loss": 0.0488428995013237,
      "theoretical_loss": 3.329294869949994,
      "tokens_seen": 2901868544
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012284848484848485,
      "loss": 1.3038,
      "theoretical_loss": 3.329294869949994,
      "tokens_seen": 2901868544
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012282828282828283,
      "loss": 1.3097,
      "theoretical_loss": 3.329289049870894,
      "tokens_seen": 2901934080
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012280808080808082,
      "loss": 1.5107,
      "theoretical_loss": 3.329283229960032,
      "tokens_seen": 2901999616
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001227878787878788,
      "loss": 1.0082,
      "theoretical_loss": 3.3292774102174003,
      "tokens_seen": 2902065152
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012276767676767676,
      "loss": 1.5301,
      "theoretical_loss": 3.3292715906429895,
      "tokens_seen": 2902130688
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012274747474747474,
      "loss": 1.4375,
      "theoretical_loss": 3.3292657712367912,
      "tokens_seen": 2902196224
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012272727272727272,
      "loss": 1.2983,
      "theoretical_loss": 3.329259951998796,
      "tokens_seen": 2902261760
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001227070707070707,
      "loss": 1.3268,
      "theoretical_loss": 3.3292541329289964,
      "tokens_seen": 2902327296
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001226868686868687,
      "loss": 1.4093,
      "theoretical_loss": 3.329248314027383,
      "tokens_seen": 2902392832
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012266666666666668,
      "loss": 1.249,
      "theoretical_loss": 3.3292424952939474,
      "tokens_seen": 2902458368
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012264646464646463,
      "loss": 1.2677,
      "theoretical_loss": 3.329236676728681,
      "tokens_seen": 2902523904
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012262626262626262,
      "loss": 1.5327,
      "theoretical_loss": 3.3292308583315746,
      "tokens_seen": 2902589440
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001226060606060606,
      "loss": 1.3449,
      "theoretical_loss": 3.32922504010262,
      "tokens_seen": 2902654976
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012258585858585858,
      "loss": 1.3588,
      "theoretical_loss": 3.3292192220418086,
      "tokens_seen": 2902720512
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012256565656565657,
      "loss": 1.4339,
      "theoretical_loss": 3.3292134041491317,
      "tokens_seen": 2902786048
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012254545454545455,
      "loss": 1.4646,
      "theoretical_loss": 3.3292075864245803,
      "tokens_seen": 2902851584
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012252525252525254,
      "loss": 1.4383,
      "theoretical_loss": 3.3292017688681463,
      "tokens_seen": 2902917120
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012250505050505052,
      "loss": 1.4962,
      "theoretical_loss": 3.3291959514798206,
      "tokens_seen": 2902982656
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012248484848484848,
      "loss": 1.2226,
      "theoretical_loss": 3.329190134259595,
      "tokens_seen": 2903048192
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012246464646464646,
      "loss": 1.3647,
      "theoretical_loss": 3.3291843172074604,
      "tokens_seen": 2903113728
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012244444444444445,
      "loss": 1.2005,
      "theoretical_loss": 3.3291785003234082,
      "tokens_seen": 2903179264
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012242424242424243,
      "loss": 1.4525,
      "theoretical_loss": 3.32917268360743,
      "tokens_seen": 2903244800
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001224040404040404,
      "loss": 1.2868,
      "theoretical_loss": 3.329166867059517,
      "tokens_seen": 2903310336
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001223838383838384,
      "loss": 1.2421,
      "theoretical_loss": 3.3291610506796605,
      "tokens_seen": 2903375872
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012236363636363638,
      "loss": 1.4229,
      "theoretical_loss": 3.329155234467852,
      "tokens_seen": 2903441408
    },
    {
      "epoch": 0.76,
      "objective/train/aligned_token_count": 0.37487781047821045,
      "objective/train/avg_token_score": 0.016955770552158356,
      "objective/train/docs_used": 1634619,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1970232725143433,
      "objective/train/original_loss": 4.37481164932251,
      "objective/train/theoretical_loss": 3.3291494184240826,
      "objective/train/tokens_used": 1262366176,
      "objective/train/unlikelihood_loss": 0.045552220195531845,
      "theoretical_loss": 3.3291494184240826,
      "tokens_seen": 2903506944
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012234343434343434,
      "loss": 1.4809,
      "theoretical_loss": 3.3291494184240826,
      "tokens_seen": 2903506944
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012232323232323232,
      "loss": 1.3215,
      "theoretical_loss": 3.329143602548344,
      "tokens_seen": 2903572480
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001223030303030303,
      "loss": 1.3038,
      "theoretical_loss": 3.3291377868406276,
      "tokens_seen": 2903638016
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001222828282828283,
      "loss": 1.1282,
      "theoretical_loss": 3.329131971300924,
      "tokens_seen": 2903703552
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012226262626262627,
      "loss": 1.445,
      "theoretical_loss": 3.329126155929225,
      "tokens_seen": 2903769088
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012224242424242423,
      "loss": 1.4945,
      "theoretical_loss": 3.3291203407255225,
      "tokens_seen": 2903834624
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012222222222222221,
      "loss": 1.4194,
      "theoretical_loss": 3.3291145256898074,
      "tokens_seen": 2903900160
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001222020202020202,
      "loss": 1.1363,
      "theoretical_loss": 3.3291087108220707,
      "tokens_seen": 2903965696
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012218181818181818,
      "loss": 1.453,
      "theoretical_loss": 3.329102896122304,
      "tokens_seen": 2904031232
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012216161616161617,
      "loss": 1.4442,
      "theoretical_loss": 3.3290970815904997,
      "tokens_seen": 2904096768
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012214141414141415,
      "loss": 1.3674,
      "theoretical_loss": 3.329091267226647,
      "tokens_seen": 2904162304
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001221212121212121,
      "loss": 1.5876,
      "theoretical_loss": 3.329085453030739,
      "tokens_seen": 2904227840
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001221010101010101,
      "loss": 1.4886,
      "theoretical_loss": 3.329079639002767,
      "tokens_seen": 2904293376
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012208080808080808,
      "loss": 1.5138,
      "theoretical_loss": 3.3290738251427214,
      "tokens_seen": 2904358912
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012206060606060607,
      "loss": 1.3857,
      "theoretical_loss": 3.3290680114505937,
      "tokens_seen": 2904424448
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012204040404040404,
      "loss": 1.5323,
      "theoretical_loss": 3.329062197926376,
      "tokens_seen": 2904489984
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012202020202020201,
      "loss": 1.2868,
      "theoretical_loss": 3.32905638457006,
      "tokens_seen": 2904555520
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000122,
      "loss": 1.3451,
      "theoretical_loss": 3.3290505713816354,
      "tokens_seen": 2904621056
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012197979797979798,
      "loss": 1.4017,
      "theoretical_loss": 3.3290447583610945,
      "tokens_seen": 2904686592
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012195959595959597,
      "loss": 1.405,
      "theoretical_loss": 3.329038945508429,
      "tokens_seen": 2904752128
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012193939393939395,
      "loss": 1.6027,
      "theoretical_loss": 3.3290331328236302,
      "tokens_seen": 2904817664
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012191919191919192,
      "loss": 1.2877,
      "theoretical_loss": 3.3290273203066887,
      "tokens_seen": 2904883200
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001218989898989899,
      "loss": 1.3259,
      "theoretical_loss": 3.3290215079575964,
      "tokens_seen": 2904948736
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012187878787878787,
      "loss": 1.4502,
      "theoretical_loss": 3.329015695776345,
      "tokens_seen": 2905014272
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012185858585858586,
      "loss": 1.438,
      "theoretical_loss": 3.329009883762925,
      "tokens_seen": 2905079808
    },
    {
      "epoch": 0.76,
      "objective/train/aligned_token_count": 0.35477760434150696,
      "objective/train/avg_token_score": 0.01436271145939827,
      "objective/train/docs_used": 1635246,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0864481925964355,
      "objective/train/original_loss": 5.180546283721924,
      "objective/train/theoretical_loss": 3.329004071917329,
      "objective/train/tokens_used": 1264004576,
      "objective/train/unlikelihood_loss": 0.04380493611097336,
      "theoretical_loss": 3.329004071917329,
      "tokens_seen": 2905145344
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012183838383838384,
      "loss": 1.4356,
      "theoretical_loss": 3.329004071917329,
      "tokens_seen": 2905145344
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012181818181818181,
      "loss": 1.3461,
      "theoretical_loss": 3.3289982602395467,
      "tokens_seen": 2905210880
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001217979797979798,
      "loss": 1.1827,
      "theoretical_loss": 3.3289924487295712,
      "tokens_seen": 2905276416
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012177777777777778,
      "loss": 1.1449,
      "theoretical_loss": 3.3289866373873926,
      "tokens_seen": 2905341952
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012175757575757576,
      "loss": 1.1932,
      "theoretical_loss": 3.328980826213003,
      "tokens_seen": 2905407488
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012173737373737375,
      "loss": 1.408,
      "theoretical_loss": 3.3289750152063933,
      "tokens_seen": 2905473024
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012171717171717173,
      "loss": 1.5223,
      "theoretical_loss": 3.328969204367555,
      "tokens_seen": 2905538560
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012169696969696969,
      "loss": 1.3024,
      "theoretical_loss": 3.3289633936964798,
      "tokens_seen": 2905604096
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012167676767676767,
      "loss": 1.4117,
      "theoretical_loss": 3.328957583193159,
      "tokens_seen": 2905669632
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012165656565656566,
      "loss": 1.2635,
      "theoretical_loss": 3.3289517728575833,
      "tokens_seen": 2905735168
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012163636363636364,
      "loss": 1.372,
      "theoretical_loss": 3.328945962689745,
      "tokens_seen": 2905800704
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012161616161616162,
      "loss": 1.4532,
      "theoretical_loss": 3.328940152689635,
      "tokens_seen": 2905866240
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001215959595959596,
      "loss": 1.1856,
      "theoretical_loss": 3.3289343428572447,
      "tokens_seen": 2905931776
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012157575757575758,
      "loss": 1.2965,
      "theoretical_loss": 3.3289285331925655,
      "tokens_seen": 2905997312
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012155555555555556,
      "loss": 1.3139,
      "theoretical_loss": 3.3289227236955887,
      "tokens_seen": 2906062848
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012153535353535355,
      "loss": 0.9621,
      "theoretical_loss": 3.3289169143663058,
      "tokens_seen": 2906128384
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012151515151515152,
      "loss": 1.3944,
      "theoretical_loss": 3.3289111052047082,
      "tokens_seen": 2906193920
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012149494949494949,
      "loss": 1.3016,
      "theoretical_loss": 3.3289052962107872,
      "tokens_seen": 2906259456
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012147474747474747,
      "loss": 1.5274,
      "theoretical_loss": 3.328899487384534,
      "tokens_seen": 2906324992
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012145454545454546,
      "loss": 1.6171,
      "theoretical_loss": 3.32889367872594,
      "tokens_seen": 2906390528
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012143434343434344,
      "loss": 1.4677,
      "theoretical_loss": 3.3288878702349973,
      "tokens_seen": 2906456064
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012141414141414142,
      "loss": 1.3505,
      "theoretical_loss": 3.3288820619116963,
      "tokens_seen": 2906521600
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001213939393939394,
      "loss": 1.401,
      "theoretical_loss": 3.3288762537560292,
      "tokens_seen": 2906587136
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012137373737373738,
      "loss": 1.4775,
      "theoretical_loss": 3.328870445767987,
      "tokens_seen": 2906652672
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012135353535353535,
      "loss": 1.4728,
      "theoretical_loss": 3.328864637947561,
      "tokens_seen": 2906718208
    },
    {
      "epoch": 0.76,
      "objective/train/aligned_token_count": 0.3222140669822693,
      "objective/train/avg_token_score": 0.028952177613973618,
      "objective/train/docs_used": 1636404,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.060826063156128,
      "objective/train/original_loss": 5.384815692901611,
      "objective/train/theoretical_loss": 3.3288588302947426,
      "objective/train/tokens_used": 1265642976,
      "objective/train/unlikelihood_loss": 0.03957049176096916,
      "theoretical_loss": 3.3288588302947426,
      "tokens_seen": 2906783744
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012133333333333333,
      "loss": 1.4017,
      "theoretical_loss": 3.3288588302947426,
      "tokens_seen": 2906783744
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012131313131313132,
      "loss": 1.3234,
      "theoretical_loss": 3.3288530228095228,
      "tokens_seen": 2906849280
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001212929292929293,
      "loss": 1.3931,
      "theoretical_loss": 3.328847215491894,
      "tokens_seen": 2906914816
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012127272727272727,
      "loss": 1.4995,
      "theoretical_loss": 3.328841408341847,
      "tokens_seen": 2906980352
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012125252525252525,
      "loss": 1.3414,
      "theoretical_loss": 3.328835601359373,
      "tokens_seen": 2907045888
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012123232323232324,
      "loss": 1.1814,
      "theoretical_loss": 3.328829794544464,
      "tokens_seen": 2907111424
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012121212121212122,
      "loss": 1.4023,
      "theoretical_loss": 3.3288239878971106,
      "tokens_seen": 2907176960
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012119191919191921,
      "loss": 1.3493,
      "theoretical_loss": 3.3288181814173043,
      "tokens_seen": 2907242496
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012117171717171716,
      "loss": 1.3004,
      "theoretical_loss": 3.3288123751050374,
      "tokens_seen": 2907308032
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012115151515151515,
      "loss": 1.4083,
      "theoretical_loss": 3.3288065689603004,
      "tokens_seen": 2907373568
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012113131313131313,
      "loss": 1.4602,
      "theoretical_loss": 3.328800762983085,
      "tokens_seen": 2907439104
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012111111111111112,
      "loss": 1.2887,
      "theoretical_loss": 3.328794957173382,
      "tokens_seen": 2907504640
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001210909090909091,
      "loss": 1.319,
      "theoretical_loss": 3.328789151531184,
      "tokens_seen": 2907570176
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012107070707070707,
      "loss": 1.1903,
      "theoretical_loss": 3.3287833460564817,
      "tokens_seen": 2907635712
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012105050505050505,
      "loss": 1.5063,
      "theoretical_loss": 3.328777540749266,
      "tokens_seen": 2907701248
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012103030303030304,
      "loss": 1.3932,
      "theoretical_loss": 3.328771735609529,
      "tokens_seen": 2907766784
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012101010101010101,
      "loss": 1.4193,
      "theoretical_loss": 3.328765930637262,
      "tokens_seen": 2907832320
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012098989898989899,
      "loss": 1.2604,
      "theoretical_loss": 3.3287601258324564,
      "tokens_seen": 2907897856
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012096969696969696,
      "loss": 1.3749,
      "theoretical_loss": 3.3287543211951034,
      "tokens_seen": 2907963392
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012094949494949495,
      "loss": 1.4452,
      "theoretical_loss": 3.3287485167251942,
      "tokens_seen": 2908028928
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012092929292929293,
      "loss": 1.1316,
      "theoretical_loss": 3.328742712422721,
      "tokens_seen": 2908094464
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012090909090909091,
      "loss": 1.2911,
      "theoretical_loss": 3.328736908287674,
      "tokens_seen": 2908160000
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001208888888888889,
      "loss": 1.3571,
      "theoretical_loss": 3.328731104320046,
      "tokens_seen": 2908225536
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012086868686868688,
      "loss": 1.3155,
      "theoretical_loss": 3.328725300519827,
      "tokens_seen": 2908291072
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012084848484848485,
      "loss": 1.457,
      "theoretical_loss": 3.3287194968870093,
      "tokens_seen": 2908356608
    },
    {
      "epoch": 0.76,
      "objective/train/aligned_token_count": 0.6178519129753113,
      "objective/train/avg_token_score": 0.005269635934382677,
      "objective/train/docs_used": 1636998,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9221341609954834,
      "objective/train/original_loss": 3.631531238555908,
      "objective/train/theoretical_loss": 3.328713693421584,
      "objective/train/tokens_used": 1267281376,
      "objective/train/unlikelihood_loss": 0.042909570038318634,
      "theoretical_loss": 3.328713693421584,
      "tokens_seen": 2908422144
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012082828282828282,
      "loss": 1.7077,
      "theoretical_loss": 3.328713693421584,
      "tokens_seen": 2908422144
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012080808080808081,
      "loss": 1.3965,
      "theoretical_loss": 3.328707890123543,
      "tokens_seen": 2908487680
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012078787878787879,
      "loss": 1.4491,
      "theoretical_loss": 3.3287020869928767,
      "tokens_seen": 2908553216
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012076767676767678,
      "loss": 1.4142,
      "theoretical_loss": 3.3286962840295775,
      "tokens_seen": 2908618752
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012074747474747475,
      "loss": 1.3293,
      "theoretical_loss": 3.328690481233636,
      "tokens_seen": 2908684288
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012072727272727273,
      "loss": 1.4569,
      "theoretical_loss": 3.328684678605044,
      "tokens_seen": 2908749824
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012070707070707071,
      "loss": 1.4697,
      "theoretical_loss": 3.3286788761437935,
      "tokens_seen": 2908815360
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001206868686868687,
      "loss": 1.4084,
      "theoretical_loss": 3.328673073849875,
      "tokens_seen": 2908880896
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012066666666666667,
      "loss": 1.4723,
      "theoretical_loss": 3.3286672717232797,
      "tokens_seen": 2908946432
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012064646464646464,
      "loss": 1.2527,
      "theoretical_loss": 3.328661469764,
      "tokens_seen": 2909011968
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012062626262626262,
      "loss": 1.2759,
      "theoretical_loss": 3.3286556679720265,
      "tokens_seen": 2909077504
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001206060606060606,
      "loss": 1.3398,
      "theoretical_loss": 3.328649866347351,
      "tokens_seen": 2909143040
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012058585858585859,
      "loss": 1.1779,
      "theoretical_loss": 3.328644064889965,
      "tokens_seen": 2909208576
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012056565656565657,
      "loss": 1.5498,
      "theoretical_loss": 3.3286382635998595,
      "tokens_seen": 2909274112
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012054545454545454,
      "loss": 1.4247,
      "theoretical_loss": 3.3286324624770263,
      "tokens_seen": 2909339648
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012052525252525253,
      "loss": 1.4379,
      "theoretical_loss": 3.328626661521456,
      "tokens_seen": 2909405184
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012050505050505051,
      "loss": 1.1482,
      "theoretical_loss": 3.3286208607331416,
      "tokens_seen": 2909470720
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012048484848484848,
      "loss": 1.471,
      "theoretical_loss": 3.3286150601120728,
      "tokens_seen": 2909536256
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012046464646464647,
      "loss": 1.3276,
      "theoretical_loss": 3.328609259658242,
      "tokens_seen": 2909601792
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012044444444444445,
      "loss": 1.3642,
      "theoretical_loss": 3.3286034593716405,
      "tokens_seen": 2909667328
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012042424242424242,
      "loss": 1.0517,
      "theoretical_loss": 3.3285976592522597,
      "tokens_seen": 2909732864
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001204040404040404,
      "loss": 1.3823,
      "theoretical_loss": 3.328591859300091,
      "tokens_seen": 2909798400
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012038383838383839,
      "loss": 1.4378,
      "theoretical_loss": 3.3285860595151253,
      "tokens_seen": 2909863936
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012036363636363637,
      "loss": 1.2924,
      "theoretical_loss": 3.328580259897355,
      "tokens_seen": 2909929472
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012034343434343436,
      "loss": 1.1948,
      "theoretical_loss": 3.32857446044677,
      "tokens_seen": 2909995008
    },
    {
      "epoch": 0.76,
      "objective/train/aligned_token_count": 0.4632209241390228,
      "objective/train/avg_token_score": 0.007904615253210068,
      "objective/train/docs_used": 1638307,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6616572141647339,
      "objective/train/original_loss": 4.532105922698975,
      "objective/train/theoretical_loss": 3.3285686611633634,
      "objective/train/tokens_used": 1268919776,
      "objective/train/unlikelihood_loss": 0.044959358870983124,
      "theoretical_loss": 3.3285686611633634,
      "tokens_seen": 2910060544
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012032323232323231,
      "loss": 1.4186,
      "theoretical_loss": 3.3285686611633634,
      "tokens_seen": 2910060544
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001203030303030303,
      "loss": 1.329,
      "theoretical_loss": 3.328562862047126,
      "tokens_seen": 2910126080
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012028282828282828,
      "loss": 1.5167,
      "theoretical_loss": 3.3285570630980486,
      "tokens_seen": 2910191616
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012026262626262627,
      "loss": 1.1487,
      "theoretical_loss": 3.3285512643161232,
      "tokens_seen": 2910257152
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012024242424242425,
      "loss": 1.3674,
      "theoretical_loss": 3.328545465701341,
      "tokens_seen": 2910322688
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012022222222222222,
      "loss": 1.4603,
      "theoretical_loss": 3.328539667253694,
      "tokens_seen": 2910388224
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001202020202020202,
      "loss": 1.3188,
      "theoretical_loss": 3.3285338689731727,
      "tokens_seen": 2910453760
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012018181818181819,
      "loss": 1.5872,
      "theoretical_loss": 3.3285280708597695,
      "tokens_seen": 2910519296
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012016161616161617,
      "loss": 1.2552,
      "theoretical_loss": 3.3285222729134754,
      "tokens_seen": 2910584832
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012014141414141414,
      "loss": 1.2788,
      "theoretical_loss": 3.328516475134281,
      "tokens_seen": 2910650368
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012012121212121211,
      "loss": 1.5589,
      "theoretical_loss": 3.328510677522179,
      "tokens_seen": 2910715904
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001201010101010101,
      "loss": 1.2582,
      "theoretical_loss": 3.32850488007716,
      "tokens_seen": 2910781440
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012008080808080808,
      "loss": 1.2842,
      "theoretical_loss": 3.328499082799216,
      "tokens_seen": 2910846976
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012006060606060606,
      "loss": 1.4277,
      "theoretical_loss": 3.328493285688338,
      "tokens_seen": 2910912512
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012004040404040405,
      "loss": 1.4309,
      "theoretical_loss": 3.3284874887445173,
      "tokens_seen": 2910978048
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012002020202020203,
      "loss": 1.3535,
      "theoretical_loss": 3.3284816919677453,
      "tokens_seen": 2911043584
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00012,
      "loss": 1.3121,
      "theoretical_loss": 3.3284758953580145,
      "tokens_seen": 2911109120
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011997979797979799,
      "loss": 1.371,
      "theoretical_loss": 3.3284700989153153,
      "tokens_seen": 2911174656
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011995959595959596,
      "loss": 1.2125,
      "theoretical_loss": 3.328464302639639,
      "tokens_seen": 2911240192
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011993939393939394,
      "loss": 1.3237,
      "theoretical_loss": 3.3284585065309775,
      "tokens_seen": 2911305728
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011991919191919193,
      "loss": 1.2295,
      "theoretical_loss": 3.328452710589322,
      "tokens_seen": 2911371264
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001198989898989899,
      "loss": 1.2763,
      "theoretical_loss": 3.3284469148146645,
      "tokens_seen": 2911436800
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011987878787878788,
      "loss": 1.3906,
      "theoretical_loss": 3.3284411192069956,
      "tokens_seen": 2911502336
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011985858585858586,
      "loss": 1.2835,
      "theoretical_loss": 3.3284353237663074,
      "tokens_seen": 2911567872
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011983838383838385,
      "loss": 1.2969,
      "theoretical_loss": 3.328429528492591,
      "tokens_seen": 2911633408
    },
    {
      "epoch": 0.76,
      "objective/train/aligned_token_count": 0.333272248506546,
      "objective/train/avg_token_score": 0.011366506107151508,
      "objective/train/docs_used": 1639054,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9280815124511719,
      "objective/train/original_loss": 4.8467254638671875,
      "objective/train/theoretical_loss": 3.3284237333858373,
      "objective/train/tokens_used": 1270558176,
      "objective/train/unlikelihood_loss": 0.046744778752326965,
      "theoretical_loss": 3.3284237333858373,
      "tokens_seen": 2911698944
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011981818181818183,
      "loss": 1.3385,
      "theoretical_loss": 3.3284237333858373,
      "tokens_seen": 2911698944
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011979797979797979,
      "loss": 1.4718,
      "theoretical_loss": 3.3284179384460386,
      "tokens_seen": 2911764480
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011977777777777777,
      "loss": 1.3848,
      "theoretical_loss": 3.3284121436731864,
      "tokens_seen": 2911830016
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011975757575757576,
      "loss": 1.3186,
      "theoretical_loss": 3.3284063490672713,
      "tokens_seen": 2911895552
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011973737373737374,
      "loss": 1.1902,
      "theoretical_loss": 3.3284005546282853,
      "tokens_seen": 2911961088
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011971717171717172,
      "loss": 1.1674,
      "theoretical_loss": 3.3283947603562196,
      "tokens_seen": 2912026624
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001196969696969697,
      "loss": 1.1647,
      "theoretical_loss": 3.328388966251066,
      "tokens_seen": 2912092160
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011967676767676768,
      "loss": 1.2248,
      "theoretical_loss": 3.3283831723128157,
      "tokens_seen": 2912157696
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011965656565656566,
      "loss": 1.2146,
      "theoretical_loss": 3.3283773785414597,
      "tokens_seen": 2912223232
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011963636363636365,
      "loss": 1.3143,
      "theoretical_loss": 3.32837158493699,
      "tokens_seen": 2912288768
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011961616161616162,
      "loss": 1.4311,
      "theoretical_loss": 3.3283657914993983,
      "tokens_seen": 2912354304
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001195959595959596,
      "loss": 1.329,
      "theoretical_loss": 3.3283599982286756,
      "tokens_seen": 2912419840
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011957575757575757,
      "loss": 1.3246,
      "theoretical_loss": 3.328354205124813,
      "tokens_seen": 2912485376
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011955555555555556,
      "loss": 1.3001,
      "theoretical_loss": 3.3283484121878026,
      "tokens_seen": 2912550912
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011953535353535354,
      "loss": 1.1373,
      "theoretical_loss": 3.3283426194176355,
      "tokens_seen": 2912616448
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011951515151515152,
      "loss": 1.1765,
      "theoretical_loss": 3.328336826814303,
      "tokens_seen": 2912681984
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011949494949494951,
      "loss": 1.2722,
      "theoretical_loss": 3.328331034377797,
      "tokens_seen": 2912747520
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011947474747474748,
      "loss": 1.4046,
      "theoretical_loss": 3.3283252421081087,
      "tokens_seen": 2912813056
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011945454545454545,
      "loss": 1.3259,
      "theoretical_loss": 3.3283194500052296,
      "tokens_seen": 2912878592
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011943434343434343,
      "loss": 1.3819,
      "theoretical_loss": 3.328313658069151,
      "tokens_seen": 2912944128
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011941414141414142,
      "loss": 1.4146,
      "theoretical_loss": 3.328307866299864,
      "tokens_seen": 2913009664
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001193939393939394,
      "loss": 1.368,
      "theoretical_loss": 3.3283020746973606,
      "tokens_seen": 2913075200
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011937373737373737,
      "loss": 1.3396,
      "theoretical_loss": 3.3282962832616323,
      "tokens_seen": 2913140736
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011935353535353535,
      "loss": 1.1199,
      "theoretical_loss": 3.3282904919926706,
      "tokens_seen": 2913206272
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011933333333333334,
      "loss": 1.385,
      "theoretical_loss": 3.3282847008904666,
      "tokens_seen": 2913271808
    },
    {
      "epoch": 0.76,
      "objective/train/aligned_token_count": 0.4161168038845062,
      "objective/train/avg_token_score": 0.007869567722082138,
      "objective/train/docs_used": 1640560,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.19307279586792,
      "objective/train/original_loss": 4.442509174346924,
      "objective/train/theoretical_loss": 3.3282789099550114,
      "objective/train/tokens_used": 1272196576,
      "objective/train/unlikelihood_loss": 0.07080573588609695,
      "theoretical_loss": 3.3282789099550114,
      "tokens_seen": 2913337344
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011931313131313132,
      "loss": 1.3707,
      "theoretical_loss": 3.3282789099550114,
      "tokens_seen": 2913337344
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001192929292929293,
      "loss": 1.2679,
      "theoretical_loss": 3.328273119186297,
      "tokens_seen": 2913402880
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011927272727272726,
      "loss": 1.2793,
      "theoretical_loss": 3.328267328584315,
      "tokens_seen": 2913468416
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011925252525252525,
      "loss": 1.3199,
      "theoretical_loss": 3.3282615381490563,
      "tokens_seen": 2913533952
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011923232323232323,
      "loss": 1.3026,
      "theoretical_loss": 3.328255747880513,
      "tokens_seen": 2913599488
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011921212121212121,
      "loss": 1.4916,
      "theoretical_loss": 3.328249957778676,
      "tokens_seen": 2913665024
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001191919191919192,
      "loss": 1.4645,
      "theoretical_loss": 3.328244167843537,
      "tokens_seen": 2913730560
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011917171717171718,
      "loss": 1.3616,
      "theoretical_loss": 3.3282383780750875,
      "tokens_seen": 2913796096
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011915151515151515,
      "loss": 1.374,
      "theoretical_loss": 3.3282325884733184,
      "tokens_seen": 2913861632
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011913131313131314,
      "loss": 1.1925,
      "theoretical_loss": 3.328226799038222,
      "tokens_seen": 2913927168
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011911111111111112,
      "loss": 1.3626,
      "theoretical_loss": 3.328221009769789,
      "tokens_seen": 2913992704
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011909090909090909,
      "loss": 1.4488,
      "theoretical_loss": 3.3282152206680116,
      "tokens_seen": 2914058240
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011907070707070708,
      "loss": 1.2531,
      "theoretical_loss": 3.3282094317328808,
      "tokens_seen": 2914123776
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011905050505050505,
      "loss": 1.3788,
      "theoretical_loss": 3.3282036429643878,
      "tokens_seen": 2914189312
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011903030303030303,
      "loss": 1.3345,
      "theoretical_loss": 3.3281978543625246,
      "tokens_seen": 2914254848
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011901010101010101,
      "loss": 1.3213,
      "theoretical_loss": 3.3281920659272823,
      "tokens_seen": 2914320384
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.000118989898989899,
      "loss": 1.394,
      "theoretical_loss": 3.3281862776586526,
      "tokens_seen": 2914385920
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011896969696969698,
      "loss": 1.3166,
      "theoretical_loss": 3.328180489556627,
      "tokens_seen": 2914451456
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011894949494949495,
      "loss": 1.3019,
      "theoretical_loss": 3.3281747016211964,
      "tokens_seen": 2914516992
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011892929292929292,
      "loss": 1.3244,
      "theoretical_loss": 3.328168913852353,
      "tokens_seen": 2914582528
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0001189090909090909,
      "loss": 1.3222,
      "theoretical_loss": 3.3281631262500877,
      "tokens_seen": 2914648064
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011888888888888889,
      "loss": 1.313,
      "theoretical_loss": 3.3281573388143926,
      "tokens_seen": 2914713600
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011886868686868687,
      "loss": 1.2984,
      "theoretical_loss": 3.328151551545258,
      "tokens_seen": 2914779136
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011884848484848484,
      "loss": 1.4229,
      "theoretical_loss": 3.3281457644426764,
      "tokens_seen": 2914844672
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.00011882828282828283,
      "loss": 1.1716,
      "theoretical_loss": 3.3281399775066394,
      "tokens_seen": 2914910208
    },
    {
      "epoch": 0.76,
      "objective/train/aligned_token_count": 0.3142717480659485,
      "objective/train/avg_token_score": 0.0385635681450367,
      "objective/train/docs_used": 1641100,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.033934235572815,
      "objective/train/original_loss": 4.547698497772217,
      "objective/train/theoretical_loss": 3.3281341907371376,
      "objective/train/tokens_used": 1273834976,
      "objective/train/unlikelihood_loss": 0.06379666924476624,
      "theoretical_loss": 3.3281341907371376,
      "tokens_seen": 2914975744
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011880808080808081,
      "loss": 1.2648,
      "theoretical_loss": 3.3281341907371376,
      "tokens_seen": 2914975744
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001187878787878788,
      "loss": 1.0098,
      "theoretical_loss": 3.328128404134163,
      "tokens_seen": 2915041280
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011876767676767678,
      "loss": 1.3524,
      "theoretical_loss": 3.328122617697707,
      "tokens_seen": 2915106816
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011874747474747475,
      "loss": 1.4009,
      "theoretical_loss": 3.328116831427761,
      "tokens_seen": 2915172352
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011872727272727272,
      "loss": 1.4222,
      "theoretical_loss": 3.328111045324316,
      "tokens_seen": 2915237888
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001187070707070707,
      "loss": 1.4376,
      "theoretical_loss": 3.328105259387365,
      "tokens_seen": 2915303424
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011868686868686869,
      "loss": 1.3433,
      "theoretical_loss": 3.3280994736168976,
      "tokens_seen": 2915368960
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011866666666666667,
      "loss": 1.227,
      "theoretical_loss": 3.328093688012906,
      "tokens_seen": 2915434496
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011864646464646466,
      "loss": 1.42,
      "theoretical_loss": 3.3280879025753825,
      "tokens_seen": 2915500032
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011862626262626263,
      "loss": 1.4698,
      "theoretical_loss": 3.328082117304317,
      "tokens_seen": 2915565568
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011860606060606061,
      "loss": 1.3609,
      "theoretical_loss": 3.3280763321997022,
      "tokens_seen": 2915631104
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011858585858585858,
      "loss": 1.3905,
      "theoretical_loss": 3.328070547261529,
      "tokens_seen": 2915696640
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011856565656565657,
      "loss": 1.1667,
      "theoretical_loss": 3.328064762489789,
      "tokens_seen": 2915762176
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011854545454545455,
      "loss": 1.1494,
      "theoretical_loss": 3.3280589778844742,
      "tokens_seen": 2915827712
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011852525252525252,
      "loss": 1.2372,
      "theoretical_loss": 3.328053193445575,
      "tokens_seen": 2915893248
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001185050505050505,
      "loss": 1.3978,
      "theoretical_loss": 3.3280474091730836,
      "tokens_seen": 2915958784
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011848484848484849,
      "loss": 1.4037,
      "theoretical_loss": 3.3280416250669913,
      "tokens_seen": 2916024320
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011846464646464647,
      "loss": 1.2853,
      "theoretical_loss": 3.32803584112729,
      "tokens_seen": 2916089856
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011844444444444446,
      "loss": 1.4669,
      "theoretical_loss": 3.32803005735397,
      "tokens_seen": 2916155392
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011842424242424243,
      "loss": 1.3542,
      "theoretical_loss": 3.3280242737470243,
      "tokens_seen": 2916220928
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001184040404040404,
      "loss": 1.3027,
      "theoretical_loss": 3.328018490306443,
      "tokens_seen": 2916286464
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011838383838383838,
      "loss": 1.1813,
      "theoretical_loss": 3.3280127070322187,
      "tokens_seen": 2916352000
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011836363636363637,
      "loss": 1.3425,
      "theoretical_loss": 3.328006923924342,
      "tokens_seen": 2916417536
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011834343434343435,
      "loss": 1.466,
      "theoretical_loss": 3.328001140982805,
      "tokens_seen": 2916483072
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011832323232323232,
      "loss": 1.4026,
      "theoretical_loss": 3.3279953582075987,
      "tokens_seen": 2916548608
    },
    {
      "epoch": 0.77,
      "objective/train/aligned_token_count": 0.33840420842170715,
      "objective/train/avg_token_score": 0.0381217859685421,
      "objective/train/docs_used": 1641732,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1339952945709229,
      "objective/train/original_loss": 4.617783069610596,
      "objective/train/theoretical_loss": 3.327989575598715,
      "objective/train/tokens_used": 1275473376,
      "objective/train/unlikelihood_loss": 0.05570315569639206,
      "theoretical_loss": 3.327989575598715,
      "tokens_seen": 2916614144
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001183030303030303,
      "loss": 1.2647,
      "theoretical_loss": 3.327989575598715,
      "tokens_seen": 2916614144
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011828282828282829,
      "loss": 1.4936,
      "theoretical_loss": 3.327983793156145,
      "tokens_seen": 2916679680
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011826262626262627,
      "loss": 1.4126,
      "theoretical_loss": 3.3279780108798804,
      "tokens_seen": 2916745216
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011824242424242426,
      "loss": 1.5292,
      "theoretical_loss": 3.3279722287699127,
      "tokens_seen": 2916810752
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011822222222222223,
      "loss": 1.418,
      "theoretical_loss": 3.3279664468262333,
      "tokens_seen": 2916876288
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001182020202020202,
      "loss": 1.4738,
      "theoretical_loss": 3.327960665048834,
      "tokens_seen": 2916941824
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011818181818181818,
      "loss": 1.4806,
      "theoretical_loss": 3.3279548834377053,
      "tokens_seen": 2917007360
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011816161616161616,
      "loss": 1.1239,
      "theoretical_loss": 3.3279491019928398,
      "tokens_seen": 2917072896
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011814141414141415,
      "loss": 1.259,
      "theoretical_loss": 3.3279433207142284,
      "tokens_seen": 2917138432
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011812121212121213,
      "loss": 1.3471,
      "theoretical_loss": 3.327937539601863,
      "tokens_seen": 2917203968
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001181010101010101,
      "loss": 1.5332,
      "theoretical_loss": 3.3279317586557347,
      "tokens_seen": 2917269504
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011808080808080809,
      "loss": 1.2034,
      "theoretical_loss": 3.327925977875835,
      "tokens_seen": 2917335040
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011806060606060606,
      "loss": 1.2885,
      "theoretical_loss": 3.3279201972621557,
      "tokens_seen": 2917400576
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011804040404040404,
      "loss": 1.3867,
      "theoretical_loss": 3.3279144168146875,
      "tokens_seen": 2917466112
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011802020202020202,
      "loss": 1.2848,
      "theoretical_loss": 3.327908636533423,
      "tokens_seen": 2917531648
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000118,
      "loss": 1.5675,
      "theoretical_loss": 3.327902856418353,
      "tokens_seen": 2917597184
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011797979797979798,
      "loss": 1.3199,
      "theoretical_loss": 3.3278970764694695,
      "tokens_seen": 2917662720
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011795959595959596,
      "loss": 1.229,
      "theoretical_loss": 3.3278912966867633,
      "tokens_seen": 2917728256
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011793939393939395,
      "loss": 1.201,
      "theoretical_loss": 3.3278855170702264,
      "tokens_seen": 2917793792
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011791919191919193,
      "loss": 1.1916,
      "theoretical_loss": 3.32787973761985,
      "tokens_seen": 2917859328
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001178989898989899,
      "loss": 1.324,
      "theoretical_loss": 3.3278739583356254,
      "tokens_seen": 2917924864
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011787878787878787,
      "loss": 1.3174,
      "theoretical_loss": 3.327868179217545,
      "tokens_seen": 2917990400
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011785858585858586,
      "loss": 1.2735,
      "theoretical_loss": 3.327862400265599,
      "tokens_seen": 2918055936
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011783838383838384,
      "loss": 1.096,
      "theoretical_loss": 3.3278566214797802,
      "tokens_seen": 2918121472
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011781818181818182,
      "loss": 1.4031,
      "theoretical_loss": 3.3278508428600793,
      "tokens_seen": 2918187008
    },
    {
      "epoch": 0.77,
      "objective/train/aligned_token_count": 0.4462365508079529,
      "objective/train/avg_token_score": 0.014469507150352001,
      "objective/train/docs_used": 1643189,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4661800861358643,
      "objective/train/original_loss": 4.172431945800781,
      "objective/train/theoretical_loss": 3.3278450644064876,
      "objective/train/tokens_used": 1277111776,
      "objective/train/unlikelihood_loss": 0.05419435724616051,
      "theoretical_loss": 3.3278450644064876,
      "tokens_seen": 2918252544
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011779797979797981,
      "loss": 1.5586,
      "theoretical_loss": 3.3278450644064876,
      "tokens_seen": 2918252544
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011777777777777778,
      "loss": 1.3228,
      "theoretical_loss": 3.3278392861189974,
      "tokens_seen": 2918318080
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011775757575757576,
      "loss": 1.3007,
      "theoretical_loss": 3.3278335079975996,
      "tokens_seen": 2918383616
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011773737373737375,
      "loss": 1.1845,
      "theoretical_loss": 3.327827730042286,
      "tokens_seen": 2918449152
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011771717171717172,
      "loss": 1.2303,
      "theoretical_loss": 3.3278219522530477,
      "tokens_seen": 2918514688
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001176969696969697,
      "loss": 1.4687,
      "theoretical_loss": 3.3278161746298762,
      "tokens_seen": 2918580224
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011767676767676767,
      "loss": 1.3414,
      "theoretical_loss": 3.327810397172764,
      "tokens_seen": 2918645760
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011765656565656565,
      "loss": 1.2394,
      "theoretical_loss": 3.3278046198817015,
      "tokens_seen": 2918711296
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011763636363636364,
      "loss": 1.3544,
      "theoretical_loss": 3.3277988427566805,
      "tokens_seen": 2918776832
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011761616161616162,
      "loss": 1.1507,
      "theoretical_loss": 3.3277930657976924,
      "tokens_seen": 2918842368
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001175959595959596,
      "loss": 1.3418,
      "theoretical_loss": 3.3277872890047293,
      "tokens_seen": 2918907904
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011757575757575758,
      "loss": 1.3334,
      "theoretical_loss": 3.327781512377782,
      "tokens_seen": 2918973440
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011755555555555556,
      "loss": 1.3058,
      "theoretical_loss": 3.327775735916842,
      "tokens_seen": 2919038976
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011753535353535353,
      "loss": 1.2024,
      "theoretical_loss": 3.327769959621901,
      "tokens_seen": 2919104512
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011751515151515152,
      "loss": 1.2118,
      "theoretical_loss": 3.327764183492951,
      "tokens_seen": 2919170048
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001174949494949495,
      "loss": 1.1738,
      "theoretical_loss": 3.327758407529983,
      "tokens_seen": 2919235584
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011747474747474747,
      "loss": 1.2844,
      "theoretical_loss": 3.3277526317329884,
      "tokens_seen": 2919301120
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011745454545454545,
      "loss": 1.3885,
      "theoretical_loss": 3.327746856101959,
      "tokens_seen": 2919366656
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011743434343434344,
      "loss": 1.3358,
      "theoretical_loss": 3.327741080636886,
      "tokens_seen": 2919432192
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011741414141414142,
      "loss": 1.2946,
      "theoretical_loss": 3.327735305337761,
      "tokens_seen": 2919497728
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001173939393939394,
      "loss": 1.5314,
      "theoretical_loss": 3.327729530204576,
      "tokens_seen": 2919563264
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011737373737373738,
      "loss": 1.2649,
      "theoretical_loss": 3.327723755237322,
      "tokens_seen": 2919628800
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011735353535353535,
      "loss": 1.441,
      "theoretical_loss": 3.32771798043599,
      "tokens_seen": 2919694336
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011733333333333333,
      "loss": 1.3269,
      "theoretical_loss": 3.3277122058005726,
      "tokens_seen": 2919759872
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011731313131313131,
      "loss": 1.432,
      "theoretical_loss": 3.3277064313310607,
      "tokens_seen": 2919825408
    },
    {
      "epoch": 0.77,
      "objective/train/aligned_token_count": 0.4379887580871582,
      "objective/train/avg_token_score": 0.0051958863623440266,
      "objective/train/docs_used": 1643758,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.327188491821289,
      "objective/train/original_loss": 4.355626106262207,
      "objective/train/theoretical_loss": 3.327700657027446,
      "objective/train/tokens_used": 1278750176,
      "objective/train/unlikelihood_loss": 0.0585920512676239,
      "theoretical_loss": 3.327700657027446,
      "tokens_seen": 2919890944
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001172929292929293,
      "loss": 1.3838,
      "theoretical_loss": 3.327700657027446,
      "tokens_seen": 2919890944
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011727272727272728,
      "loss": 0.9864,
      "theoretical_loss": 3.32769488288972,
      "tokens_seen": 2919956480
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011725252525252525,
      "loss": 1.4167,
      "theoretical_loss": 3.327689108917874,
      "tokens_seen": 2920022016
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011723232323232324,
      "loss": 1.3385,
      "theoretical_loss": 3.3276833351119,
      "tokens_seen": 2920087552
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011721212121212122,
      "loss": 1.4297,
      "theoretical_loss": 3.3276775614717886,
      "tokens_seen": 2920153088
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011719191919191919,
      "loss": 1.188,
      "theoretical_loss": 3.3276717879975326,
      "tokens_seen": 2920218624
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011717171717171717,
      "loss": 1.2859,
      "theoretical_loss": 3.327666014689122,
      "tokens_seen": 2920284160
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011715151515151515,
      "loss": 1.4935,
      "theoretical_loss": 3.3276602415465497,
      "tokens_seen": 2920349696
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011713131313131313,
      "loss": 1.5815,
      "theoretical_loss": 3.3276544685698064,
      "tokens_seen": 2920415232
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011711111111111111,
      "loss": 1.1246,
      "theoretical_loss": 3.3276486957588842,
      "tokens_seen": 2920480768
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001170909090909091,
      "loss": 1.2203,
      "theoretical_loss": 3.327642923113774,
      "tokens_seen": 2920546304
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011707070707070708,
      "loss": 1.5048,
      "theoretical_loss": 3.3276371506344673,
      "tokens_seen": 2920611840
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011705050505050505,
      "loss": 1.2619,
      "theoretical_loss": 3.327631378320956,
      "tokens_seen": 2920677376
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011703030303030302,
      "loss": 1.3731,
      "theoretical_loss": 3.327625606173232,
      "tokens_seen": 2920742912
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000117010101010101,
      "loss": 1.4127,
      "theoretical_loss": 3.3276198341912857,
      "tokens_seen": 2920808448
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011698989898989899,
      "loss": 1.4524,
      "theoretical_loss": 3.3276140623751096,
      "tokens_seen": 2920873984
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011696969696969697,
      "loss": 1.3899,
      "theoretical_loss": 3.3276082907246947,
      "tokens_seen": 2920939520
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011694949494949496,
      "loss": 1.3922,
      "theoretical_loss": 3.327602519240033,
      "tokens_seen": 2921005056
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011692929292929293,
      "loss": 1.5587,
      "theoretical_loss": 3.3275967479211155,
      "tokens_seen": 2921070592
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011690909090909091,
      "loss": 1.4017,
      "theoretical_loss": 3.327590976767934,
      "tokens_seen": 2921136128
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001168888888888889,
      "loss": 1.3692,
      "theoretical_loss": 3.32758520578048,
      "tokens_seen": 2921201664
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011686868686868688,
      "loss": 1.4005,
      "theoretical_loss": 3.3275794349587446,
      "tokens_seen": 2921267200
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011684848484848485,
      "loss": 1.279,
      "theoretical_loss": 3.3275736643027205,
      "tokens_seen": 2921332736
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011682828282828282,
      "loss": 1.2125,
      "theoretical_loss": 3.3275678938123976,
      "tokens_seen": 2921398272
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001168080808080808,
      "loss": 1.2192,
      "theoretical_loss": 3.327562123487769,
      "tokens_seen": 2921463808
    },
    {
      "epoch": 0.77,
      "objective/train/aligned_token_count": 0.36149805784225464,
      "objective/train/avg_token_score": 0.022337336093187332,
      "objective/train/docs_used": 1645135,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0525131225585938,
      "objective/train/original_loss": 4.370710849761963,
      "objective/train/theoretical_loss": 3.327556353328825,
      "objective/train/tokens_used": 1280388576,
      "objective/train/unlikelihood_loss": 0.05943276733160019,
      "theoretical_loss": 3.327556353328825,
      "tokens_seen": 2921529344
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011678787878787879,
      "loss": 1.1656,
      "theoretical_loss": 3.327556353328825,
      "tokens_seen": 2921529344
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011676767676767677,
      "loss": 1.5121,
      "theoretical_loss": 3.327550583335557,
      "tokens_seen": 2921594880
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011674747474747476,
      "loss": 1.4742,
      "theoretical_loss": 3.3275448135079584,
      "tokens_seen": 2921660416
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011672727272727273,
      "loss": 1.3094,
      "theoretical_loss": 3.3275390438460186,
      "tokens_seen": 2921725952
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011670707070707071,
      "loss": 1.2032,
      "theoretical_loss": 3.3275332743497303,
      "tokens_seen": 2921791488
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001166868686868687,
      "loss": 1.3947,
      "theoretical_loss": 3.3275275050190847,
      "tokens_seen": 2921857024
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011666666666666667,
      "loss": 1.2327,
      "theoretical_loss": 3.3275217358540736,
      "tokens_seen": 2921922560
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011664646464646465,
      "loss": 1.4832,
      "theoretical_loss": 3.327515966854688,
      "tokens_seen": 2921988096
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011662626262626262,
      "loss": 1.2811,
      "theoretical_loss": 3.3275101980209194,
      "tokens_seen": 2922053632
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001166060606060606,
      "loss": 1.3605,
      "theoretical_loss": 3.32750442935276,
      "tokens_seen": 2922119168
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011658585858585859,
      "loss": 1.2985,
      "theoretical_loss": 3.3274986608502006,
      "tokens_seen": 2922184704
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011656565656565657,
      "loss": 1.4249,
      "theoretical_loss": 3.327492892513234,
      "tokens_seen": 2922250240
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011654545454545456,
      "loss": 1.338,
      "theoretical_loss": 3.3274871243418502,
      "tokens_seen": 2922315776
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011652525252525254,
      "loss": 1.386,
      "theoretical_loss": 3.3274813563360413,
      "tokens_seen": 2922381312
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001165050505050505,
      "loss": 1.4831,
      "theoretical_loss": 3.327475588495799,
      "tokens_seen": 2922446848
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011648484848484848,
      "loss": 1.415,
      "theoretical_loss": 3.3274698208211144,
      "tokens_seen": 2922512384
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011646464646464646,
      "loss": 1.3801,
      "theoretical_loss": 3.32746405331198,
      "tokens_seen": 2922577920
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011644444444444445,
      "loss": 1.4768,
      "theoretical_loss": 3.3274582859683863,
      "tokens_seen": 2922643456
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011642424242424243,
      "loss": 1.5656,
      "theoretical_loss": 3.3274525187903254,
      "tokens_seen": 2922708992
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001164040404040404,
      "loss": 1.0739,
      "theoretical_loss": 3.3274467517777886,
      "tokens_seen": 2922774528
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011638383838383839,
      "loss": 1.4058,
      "theoretical_loss": 3.327440984930768,
      "tokens_seen": 2922840064
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011636363636363637,
      "loss": 1.1661,
      "theoretical_loss": 3.3274352182492537,
      "tokens_seen": 2922905600
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011634343434343435,
      "loss": 1.2057,
      "theoretical_loss": 3.327429451733239,
      "tokens_seen": 2922971136
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011632323232323233,
      "loss": 1.2807,
      "theoretical_loss": 3.3274236853827146,
      "tokens_seen": 2923036672
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001163030303030303,
      "loss": 1.3487,
      "theoretical_loss": 3.3274179191976714,
      "tokens_seen": 2923102208
    },
    {
      "epoch": 0.77,
      "objective/train/aligned_token_count": 0.35202834010124207,
      "objective/train/avg_token_score": 0.03710121288895607,
      "objective/train/docs_used": 1645886,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0146808624267578,
      "objective/train/original_loss": 4.969164848327637,
      "objective/train/theoretical_loss": 3.327412153178102,
      "objective/train/tokens_used": 1282026976,
      "objective/train/unlikelihood_loss": 0.047619786113500595,
      "theoretical_loss": 3.327412153178102,
      "tokens_seen": 2923167744
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011628282828282828,
      "loss": 1.2155,
      "theoretical_loss": 3.327412153178102,
      "tokens_seen": 2923167744
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011626262626262626,
      "loss": 1.4012,
      "theoretical_loss": 3.327406387323998,
      "tokens_seen": 2923233280
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011624242424242425,
      "loss": 1.2015,
      "theoretical_loss": 3.32740062163535,
      "tokens_seen": 2923298816
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011622222222222223,
      "loss": 1.3287,
      "theoretical_loss": 3.32739485611215,
      "tokens_seen": 2923364352
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001162020202020202,
      "loss": 1.2173,
      "theoretical_loss": 3.32738909075439,
      "tokens_seen": 2923429888
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011618181818181819,
      "loss": 1.2555,
      "theoretical_loss": 3.327383325562061,
      "tokens_seen": 2923495424
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011616161616161616,
      "loss": 1.4709,
      "theoretical_loss": 3.3273775605351545,
      "tokens_seen": 2923560960
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011614141414141414,
      "loss": 1.4193,
      "theoretical_loss": 3.3273717956736624,
      "tokens_seen": 2923626496
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011612121212121212,
      "loss": 1.499,
      "theoretical_loss": 3.3273660309775757,
      "tokens_seen": 2923692032
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011610101010101011,
      "loss": 1.2752,
      "theoretical_loss": 3.327360266446887,
      "tokens_seen": 2923757568
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011608080808080808,
      "loss": 1.1587,
      "theoretical_loss": 3.3273545020815867,
      "tokens_seen": 2923823104
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011606060606060606,
      "loss": 1.4771,
      "theoretical_loss": 3.3273487378816666,
      "tokens_seen": 2923888640
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011604040404040405,
      "loss": 1.3549,
      "theoretical_loss": 3.327342973847119,
      "tokens_seen": 2923954176
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011602020202020203,
      "loss": 1.1157,
      "theoretical_loss": 3.3273372099779346,
      "tokens_seen": 2924019712
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011600000000000001,
      "loss": 1.3372,
      "theoretical_loss": 3.3273314462741053,
      "tokens_seen": 2924085248
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011597979797979797,
      "loss": 1.256,
      "theoretical_loss": 3.327325682735623,
      "tokens_seen": 2924150784
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011595959595959596,
      "loss": 1.5366,
      "theoretical_loss": 3.3273199193624783,
      "tokens_seen": 2924216320
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011593939393939394,
      "loss": 1.345,
      "theoretical_loss": 3.3273141561546633,
      "tokens_seen": 2924281856
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011591919191919192,
      "loss": 1.1627,
      "theoretical_loss": 3.32730839311217,
      "tokens_seen": 2924347392
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011589898989898991,
      "loss": 1.1473,
      "theoretical_loss": 3.327302630234989,
      "tokens_seen": 2924412928
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011587878787878788,
      "loss": 1.2985,
      "theoretical_loss": 3.327296867523113,
      "tokens_seen": 2924478464
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011585858585858586,
      "loss": 1.3326,
      "theoretical_loss": 3.327291104976532,
      "tokens_seen": 2924544000
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011583838383838385,
      "loss": 0.9696,
      "theoretical_loss": 3.3272853425952396,
      "tokens_seen": 2924609536
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011581818181818183,
      "loss": 1.2402,
      "theoretical_loss": 3.3272795803792254,
      "tokens_seen": 2924675072
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001157979797979798,
      "loss": 1.4072,
      "theoretical_loss": 3.3272738183284822,
      "tokens_seen": 2924740608
    },
    {
      "epoch": 0.77,
      "objective/train/aligned_token_count": 0.3955889642238617,
      "objective/train/avg_token_score": 0.013041549362242222,
      "objective/train/docs_used": 1647115,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3574528694152832,
      "objective/train/original_loss": 4.626835823059082,
      "objective/train/theoretical_loss": 3.327268056443001,
      "objective/train/tokens_used": 1283665376,
      "objective/train/unlikelihood_loss": 0.05217356234788895,
      "theoretical_loss": 3.327268056443001,
      "tokens_seen": 2924806144
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011577777777777777,
      "loss": 1.3735,
      "theoretical_loss": 3.327268056443001,
      "tokens_seen": 2924806144
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011575757575757575,
      "loss": 1.5447,
      "theoretical_loss": 3.3272622947227735,
      "tokens_seen": 2924871680
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011573737373737374,
      "loss": 1.2607,
      "theoretical_loss": 3.3272565331677915,
      "tokens_seen": 2924937216
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011571717171717172,
      "loss": 1.2694,
      "theoretical_loss": 3.327250771778046,
      "tokens_seen": 2925002752
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001156969696969697,
      "loss": 1.2255,
      "theoretical_loss": 3.327245010553529,
      "tokens_seen": 2925068288
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011567676767676769,
      "loss": 1.0279,
      "theoretical_loss": 3.327239249494232,
      "tokens_seen": 2925133824
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011565656565656566,
      "loss": 1.1494,
      "theoretical_loss": 3.3272334886001467,
      "tokens_seen": 2925199360
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011563636363636363,
      "loss": 1.3003,
      "theoretical_loss": 3.3272277278712643,
      "tokens_seen": 2925264896
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011561616161616161,
      "loss": 1.3917,
      "theoretical_loss": 3.327221967307576,
      "tokens_seen": 2925330432
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001155959595959596,
      "loss": 1.6169,
      "theoretical_loss": 3.3272162069090747,
      "tokens_seen": 2925395968
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011557575757575758,
      "loss": 1.4942,
      "theoretical_loss": 3.3272104466757506,
      "tokens_seen": 2925461504
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011555555555555555,
      "loss": 1.2343,
      "theoretical_loss": 3.327204686607596,
      "tokens_seen": 2925527040
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011553535353535354,
      "loss": 1.3453,
      "theoretical_loss": 3.3271989267046025,
      "tokens_seen": 2925592576
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011551515151515152,
      "loss": 1.1045,
      "theoretical_loss": 3.327193166966761,
      "tokens_seen": 2925658112
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001154949494949495,
      "loss": 1.3023,
      "theoretical_loss": 3.3271874073940637,
      "tokens_seen": 2925723648
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011547474747474749,
      "loss": 1.2773,
      "theoretical_loss": 3.327181647986502,
      "tokens_seen": 2925789184
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011545454545454545,
      "loss": 1.3667,
      "theoretical_loss": 3.3271758887440677,
      "tokens_seen": 2925854720
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011543434343434343,
      "loss": 0.9601,
      "theoretical_loss": 3.3271701296667517,
      "tokens_seen": 2925920256
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011541414141414141,
      "loss": 1.1993,
      "theoretical_loss": 3.327164370754546,
      "tokens_seen": 2925985792
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001153939393939394,
      "loss": 1.2968,
      "theoretical_loss": 3.3271586120074423,
      "tokens_seen": 2926051328
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011537373737373738,
      "loss": 1.4353,
      "theoretical_loss": 3.327152853425432,
      "tokens_seen": 2926116864
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011535353535353535,
      "loss": 1.0833,
      "theoretical_loss": 3.327147095008507,
      "tokens_seen": 2926182400
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011533333333333334,
      "loss": 1.3177,
      "theoretical_loss": 3.327141336756658,
      "tokens_seen": 2926247936
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011531313131313132,
      "loss": 1.2599,
      "theoretical_loss": 3.327135578669877,
      "tokens_seen": 2926313472
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011529292929292929,
      "loss": 1.135,
      "theoretical_loss": 3.3271298207481563,
      "tokens_seen": 2926379008
    },
    {
      "epoch": 0.77,
      "objective/train/aligned_token_count": 0.45149070024490356,
      "objective/train/avg_token_score": 0.006710331421345472,
      "objective/train/docs_used": 1647827,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.298624038696289,
      "objective/train/original_loss": 4.395281791687012,
      "objective/train/theoretical_loss": 3.3271240629914867,
      "objective/train/tokens_used": 1285303776,
      "objective/train/unlikelihood_loss": 0.07609055191278458,
      "theoretical_loss": 3.3271240629914867,
      "tokens_seen": 2926444544
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011527272727272727,
      "loss": 1.2894,
      "theoretical_loss": 3.3271240629914867,
      "tokens_seen": 2926444544
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011525252525252526,
      "loss": 1.347,
      "theoretical_loss": 3.3271183053998596,
      "tokens_seen": 2926510080
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011523232323232323,
      "loss": 1.4109,
      "theoretical_loss": 3.3271125479732673,
      "tokens_seen": 2926575616
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011521212121212121,
      "loss": 1.1682,
      "theoretical_loss": 3.3271067907117007,
      "tokens_seen": 2926641152
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001151919191919192,
      "loss": 1.1187,
      "theoretical_loss": 3.327101033615152,
      "tokens_seen": 2926706688
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011517171717171718,
      "loss": 1.2381,
      "theoretical_loss": 3.327095276683612,
      "tokens_seen": 2926772224
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011515151515151516,
      "loss": 1.2507,
      "theoretical_loss": 3.327089519917073,
      "tokens_seen": 2926837760
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011513131313131313,
      "loss": 1.2176,
      "theoretical_loss": 3.3270837633155264,
      "tokens_seen": 2926903296
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001151111111111111,
      "loss": 1.4239,
      "theoretical_loss": 3.3270780068789634,
      "tokens_seen": 2926968832
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011509090909090909,
      "loss": 1.3031,
      "theoretical_loss": 3.327072250607376,
      "tokens_seen": 2927034368
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011507070707070707,
      "loss": 1.1505,
      "theoretical_loss": 3.3270664945007553,
      "tokens_seen": 2927099904
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011505050505050506,
      "loss": 1.4785,
      "theoretical_loss": 3.3270607385590933,
      "tokens_seen": 2927165440
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011503030303030303,
      "loss": 1.1813,
      "theoretical_loss": 3.3270549827823817,
      "tokens_seen": 2927230976
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011501010101010101,
      "loss": 1.4041,
      "theoretical_loss": 3.327049227170612,
      "tokens_seen": 2927296512
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000114989898989899,
      "loss": 1.2892,
      "theoretical_loss": 3.3270434717237753,
      "tokens_seen": 2927362048
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011496969696969698,
      "loss": 1.2134,
      "theoretical_loss": 3.327037716441864,
      "tokens_seen": 2927427584
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011494949494949495,
      "loss": 1.1795,
      "theoretical_loss": 3.3270319613248684,
      "tokens_seen": 2927493120
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011492929292929292,
      "loss": 1.3698,
      "theoretical_loss": 3.3270262063727816,
      "tokens_seen": 2927558656
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001149090909090909,
      "loss": 1.1746,
      "theoretical_loss": 3.327020451585594,
      "tokens_seen": 2927624192
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011488888888888889,
      "loss": 1.2699,
      "theoretical_loss": 3.327014696963298,
      "tokens_seen": 2927689728
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011486868686868687,
      "loss": 1.2311,
      "theoretical_loss": 3.3270089425058846,
      "tokens_seen": 2927755264
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011484848484848486,
      "loss": 1.1994,
      "theoretical_loss": 3.3270031882133457,
      "tokens_seen": 2927820800
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011482828282828284,
      "loss": 1.2419,
      "theoretical_loss": 3.3269974340856727,
      "tokens_seen": 2927886336
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011480808080808081,
      "loss": 1.2452,
      "theoretical_loss": 3.3269916801228576,
      "tokens_seen": 2927951872
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001147878787878788,
      "loss": 1.2938,
      "theoretical_loss": 3.3269859263248915,
      "tokens_seen": 2928017408
    },
    {
      "epoch": 0.77,
      "objective/train/aligned_token_count": 0.37805473804473877,
      "objective/train/avg_token_score": 0.01525930967181921,
      "objective/train/docs_used": 1648971,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1057424545288086,
      "objective/train/original_loss": 4.303940773010254,
      "objective/train/theoretical_loss": 3.326980172691766,
      "objective/train/tokens_used": 1286942176,
      "objective/train/unlikelihood_loss": 0.0826089084148407,
      "theoretical_loss": 3.326980172691766,
      "tokens_seen": 2928082944
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011476767676767676,
      "loss": 1.3121,
      "theoretical_loss": 3.326980172691766,
      "tokens_seen": 2928082944
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011474747474747475,
      "loss": 1.064,
      "theoretical_loss": 3.326974419223473,
      "tokens_seen": 2928148480
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011472727272727273,
      "loss": 1.1454,
      "theoretical_loss": 3.3269686659200044,
      "tokens_seen": 2928214016
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001147070707070707,
      "loss": 1.285,
      "theoretical_loss": 3.326962912781351,
      "tokens_seen": 2928279552
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011468686868686869,
      "loss": 1.1555,
      "theoretical_loss": 3.3269571598075043,
      "tokens_seen": 2928345088
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011466666666666667,
      "loss": 1.0705,
      "theoretical_loss": 3.3269514069984565,
      "tokens_seen": 2928410624
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011464646464646466,
      "loss": 1.4592,
      "theoretical_loss": 3.3269456543541995,
      "tokens_seen": 2928476160
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011462626262626264,
      "loss": 1.4872,
      "theoretical_loss": 3.3269399018747237,
      "tokens_seen": 2928541696
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001146060606060606,
      "loss": 1.0539,
      "theoretical_loss": 3.3269341495600218,
      "tokens_seen": 2928607232
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011458585858585858,
      "loss": 1.1764,
      "theoretical_loss": 3.326928397410085,
      "tokens_seen": 2928672768
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011456565656565656,
      "loss": 1.4534,
      "theoretical_loss": 3.326922645424905,
      "tokens_seen": 2928738304
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011454545454545455,
      "loss": 1.4367,
      "theoretical_loss": 3.326916893604473,
      "tokens_seen": 2928803840
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011452525252525253,
      "loss": 1.2759,
      "theoretical_loss": 3.326911141948781,
      "tokens_seen": 2928869376
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001145050505050505,
      "loss": 1.3464,
      "theoretical_loss": 3.3269053904578203,
      "tokens_seen": 2928934912
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011448484848484849,
      "loss": 1.3166,
      "theoretical_loss": 3.326899639131583,
      "tokens_seen": 2929000448
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011446464646464647,
      "loss": 1.3931,
      "theoretical_loss": 3.3268938879700602,
      "tokens_seen": 2929065984
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011444444444444445,
      "loss": 1.1659,
      "theoretical_loss": 3.3268881369732437,
      "tokens_seen": 2929131520
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011442424242424242,
      "loss": 1.3945,
      "theoretical_loss": 3.3268823861411247,
      "tokens_seen": 2929197056
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011440404040404041,
      "loss": 1.4628,
      "theoretical_loss": 3.3268766354736954,
      "tokens_seen": 2929262592
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011438383838383838,
      "loss": 1.4469,
      "theoretical_loss": 3.326870884970947,
      "tokens_seen": 2929328128
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011436363636363636,
      "loss": 1.1935,
      "theoretical_loss": 3.3268651346328717,
      "tokens_seen": 2929393664
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011434343434343435,
      "loss": 1.4261,
      "theoretical_loss": 3.32685938445946,
      "tokens_seen": 2929459200
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011432323232323233,
      "loss": 1.1915,
      "theoretical_loss": 3.3268536344507047,
      "tokens_seen": 2929524736
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011430303030303031,
      "loss": 1.3084,
      "theoretical_loss": 3.3268478846065968,
      "tokens_seen": 2929590272
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011428282828282829,
      "loss": 1.3722,
      "theoretical_loss": 3.3268421349271278,
      "tokens_seen": 2929655808
    },
    {
      "epoch": 0.77,
      "objective/train/aligned_token_count": 0.536107063293457,
      "objective/train/avg_token_score": 0.017638999968767166,
      "objective/train/docs_used": 1649763,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7608067989349365,
      "objective/train/original_loss": 4.418849945068359,
      "objective/train/theoretical_loss": 3.3268363854122893,
      "objective/train/tokens_used": 1288580576,
      "objective/train/unlikelihood_loss": 0.0383111834526062,
      "theoretical_loss": 3.3268363854122893,
      "tokens_seen": 2929721344
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011426262626262627,
      "loss": 1.4549,
      "theoretical_loss": 3.3268363854122893,
      "tokens_seen": 2929721344
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011424242424242424,
      "loss": 1.3088,
      "theoretical_loss": 3.3268306360620734,
      "tokens_seen": 2929786880
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011422222222222222,
      "loss": 1.1988,
      "theoretical_loss": 3.326824886876471,
      "tokens_seen": 2929852416
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011420202020202021,
      "loss": 1.3284,
      "theoretical_loss": 3.326819137855474,
      "tokens_seen": 2929917952
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011418181818181818,
      "loss": 1.4609,
      "theoretical_loss": 3.3268133889990747,
      "tokens_seen": 2929983488
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011416161616161616,
      "loss": 1.1678,
      "theoretical_loss": 3.3268076403072637,
      "tokens_seen": 2930049024
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011414141414141415,
      "loss": 1.4086,
      "theoretical_loss": 3.326801891780033,
      "tokens_seen": 2930114560
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011412121212121213,
      "loss": 1.3572,
      "theoretical_loss": 3.3267961434173743,
      "tokens_seen": 2930180096
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011410101010101011,
      "loss": 1.5328,
      "theoretical_loss": 3.326790395219279,
      "tokens_seen": 2930245632
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011408080808080807,
      "loss": 1.419,
      "theoretical_loss": 3.3267846471857387,
      "tokens_seen": 2930311168
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011406060606060605,
      "loss": 1.0531,
      "theoretical_loss": 3.3267788993167455,
      "tokens_seen": 2930376704
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011404040404040404,
      "loss": 1.1767,
      "theoretical_loss": 3.32677315161229,
      "tokens_seen": 2930442240
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011402020202020202,
      "loss": 1.2018,
      "theoretical_loss": 3.326767404072365,
      "tokens_seen": 2930507776
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.000114,
      "loss": 1.2292,
      "theoretical_loss": 3.3267616566969616,
      "tokens_seen": 2930573312
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011397979797979799,
      "loss": 1.4086,
      "theoretical_loss": 3.3267559094860712,
      "tokens_seen": 2930638848
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011395959595959596,
      "loss": 1.1789,
      "theoretical_loss": 3.3267501624396854,
      "tokens_seen": 2930704384
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011393939393939394,
      "loss": 1.2395,
      "theoretical_loss": 3.3267444155577963,
      "tokens_seen": 2930769920
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011391919191919193,
      "loss": 1.3685,
      "theoretical_loss": 3.326738668840395,
      "tokens_seen": 2930835456
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001138989898989899,
      "loss": 1.3576,
      "theoretical_loss": 3.3267329222874733,
      "tokens_seen": 2930900992
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011387878787878788,
      "loss": 1.32,
      "theoretical_loss": 3.326727175899023,
      "tokens_seen": 2930966528
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011385858585858585,
      "loss": 1.1685,
      "theoretical_loss": 3.326721429675035,
      "tokens_seen": 2931032064
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011383838383838384,
      "loss": 1.1124,
      "theoretical_loss": 3.326715683615502,
      "tokens_seen": 2931097600
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011381818181818182,
      "loss": 1.5209,
      "theoretical_loss": 3.3267099377204152,
      "tokens_seen": 2931163136
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001137979797979798,
      "loss": 1.3882,
      "theoretical_loss": 3.3267041919897657,
      "tokens_seen": 2931228672
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.00011377777777777779,
      "loss": 1.4366,
      "theoretical_loss": 3.3266984464235456,
      "tokens_seen": 2931294208
    },
    {
      "epoch": 0.77,
      "objective/train/aligned_token_count": 0.4514296054840088,
      "objective/train/avg_token_score": 0.0031480479519814253,
      "objective/train/docs_used": 1650742,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2340714931488037,
      "objective/train/original_loss": 3.725032329559326,
      "objective/train/theoretical_loss": 3.326692701021747,
      "objective/train/tokens_used": 1290218976,
      "objective/train/unlikelihood_loss": 0.06366879492998123,
      "theoretical_loss": 3.326692701021747,
      "tokens_seen": 2931359744
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011375757575757576,
      "loss": 1.1452,
      "theoretical_loss": 3.326692701021747,
      "tokens_seen": 2931359744
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011373737373737373,
      "loss": 1.2594,
      "theoretical_loss": 3.32668695578436,
      "tokens_seen": 2931425280
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011371717171717171,
      "loss": 1.247,
      "theoretical_loss": 3.326681210711378,
      "tokens_seen": 2931490816
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001136969696969697,
      "loss": 1.3234,
      "theoretical_loss": 3.3266754658027913,
      "tokens_seen": 2931556352
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011367676767676768,
      "loss": 1.2214,
      "theoretical_loss": 3.3266697210585923,
      "tokens_seen": 2931621888
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011365656565656565,
      "loss": 1.3702,
      "theoretical_loss": 3.326663976478772,
      "tokens_seen": 2931687424
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011363636363636364,
      "loss": 1.3252,
      "theoretical_loss": 3.3266582320633225,
      "tokens_seen": 2931752960
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011361616161616162,
      "loss": 1.5328,
      "theoretical_loss": 3.326652487812235,
      "tokens_seen": 2931818496
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001135959595959596,
      "loss": 1.4894,
      "theoretical_loss": 3.3266467437255023,
      "tokens_seen": 2931884032
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011357575757575759,
      "loss": 1.2264,
      "theoretical_loss": 3.3266409998031143,
      "tokens_seen": 2931949568
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011355555555555556,
      "loss": 1.421,
      "theoretical_loss": 3.326635256045064,
      "tokens_seen": 2932015104
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011353535353535353,
      "loss": 1.2753,
      "theoretical_loss": 3.3266295124513423,
      "tokens_seen": 2932080640
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011351515151515151,
      "loss": 1.405,
      "theoretical_loss": 3.326623769021941,
      "tokens_seen": 2932146176
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001134949494949495,
      "loss": 1.3692,
      "theoretical_loss": 3.326618025756851,
      "tokens_seen": 2932211712
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011347474747474748,
      "loss": 1.1984,
      "theoretical_loss": 3.3266122826560656,
      "tokens_seen": 2932277248
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011345454545454546,
      "loss": 1.2337,
      "theoretical_loss": 3.326606539719575,
      "tokens_seen": 2932342784
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011343434343434344,
      "loss": 1.2255,
      "theoretical_loss": 3.3266007969473717,
      "tokens_seen": 2932408320
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011341414141414142,
      "loss": 1.2705,
      "theoretical_loss": 3.3265950543394465,
      "tokens_seen": 2932473856
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001133939393939394,
      "loss": 1.346,
      "theoretical_loss": 3.3265893118957917,
      "tokens_seen": 2932539392
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011337373737373737,
      "loss": 1.4676,
      "theoretical_loss": 3.326583569616399,
      "tokens_seen": 2932604928
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011335353535353536,
      "loss": 1.3567,
      "theoretical_loss": 3.326577827501259,
      "tokens_seen": 2932670464
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011333333333333333,
      "loss": 1.2122,
      "theoretical_loss": 3.3265720855503647,
      "tokens_seen": 2932736000
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011331313131313131,
      "loss": 1.5273,
      "theoretical_loss": 3.326566343763707,
      "tokens_seen": 2932801536
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001132929292929293,
      "loss": 1.4506,
      "theoretical_loss": 3.326560602141277,
      "tokens_seen": 2932867072
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011327272727272728,
      "loss": 1.1522,
      "theoretical_loss": 3.3265548606830677,
      "tokens_seen": 2932932608
    },
    {
      "epoch": 0.78,
      "objective/train/aligned_token_count": 0.3075513243675232,
      "objective/train/avg_token_score": 0.028925929218530655,
      "objective/train/docs_used": 1651324,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9664291739463806,
      "objective/train/original_loss": 4.7379560470581055,
      "objective/train/theoretical_loss": 3.3265491193890697,
      "objective/train/tokens_used": 1291857376,
      "objective/train/unlikelihood_loss": 0.07130970805883408,
      "theoretical_loss": 3.3265491193890697,
      "tokens_seen": 2932998144
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011325252525252526,
      "loss": 1.1371,
      "theoretical_loss": 3.3265491193890697,
      "tokens_seen": 2932998144
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011323232323232323,
      "loss": 1.2314,
      "theoretical_loss": 3.326543378259275,
      "tokens_seen": 2933063680
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001132121212121212,
      "loss": 1.2213,
      "theoretical_loss": 3.3265376372936752,
      "tokens_seen": 2933129216
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011319191919191919,
      "loss": 1.4611,
      "theoretical_loss": 3.326531896492262,
      "tokens_seen": 2933194752
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011317171717171717,
      "loss": 1.3179,
      "theoretical_loss": 3.3265261558550265,
      "tokens_seen": 2933260288
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011315151515151516,
      "loss": 1.3279,
      "theoretical_loss": 3.326520415381961,
      "tokens_seen": 2933325824
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011313131313131314,
      "loss": 1.3474,
      "theoretical_loss": 3.326514675073057,
      "tokens_seen": 2933391360
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011311111111111111,
      "loss": 1.3783,
      "theoretical_loss": 3.326508934928306,
      "tokens_seen": 2933456896
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001130909090909091,
      "loss": 1.1869,
      "theoretical_loss": 3.3265031949476995,
      "tokens_seen": 2933522432
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011307070707070708,
      "loss": 1.3224,
      "theoretical_loss": 3.3264974551312294,
      "tokens_seen": 2933587968
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011305050505050506,
      "loss": 1.1274,
      "theoretical_loss": 3.3264917154788876,
      "tokens_seen": 2933653504
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011303030303030303,
      "loss": 1.2907,
      "theoretical_loss": 3.326485975990665,
      "tokens_seen": 2933719040
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000113010101010101,
      "loss": 1.5231,
      "theoretical_loss": 3.326480236666554,
      "tokens_seen": 2933784576
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011298989898989899,
      "loss": 1.3316,
      "theoretical_loss": 3.3264744975065454,
      "tokens_seen": 2933850112
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011296969696969697,
      "loss": 1.1625,
      "theoretical_loss": 3.326468758510632,
      "tokens_seen": 2933915648
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011294949494949496,
      "loss": 1.4381,
      "theoretical_loss": 3.3264630196788043,
      "tokens_seen": 2933981184
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011292929292929294,
      "loss": 1.5519,
      "theoretical_loss": 3.3264572810110544,
      "tokens_seen": 2934046720
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011290909090909091,
      "loss": 1.1511,
      "theoretical_loss": 3.326451542507374,
      "tokens_seen": 2934112256
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001128888888888889,
      "loss": 1.0999,
      "theoretical_loss": 3.326445804167755,
      "tokens_seen": 2934177792
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011286868686868686,
      "loss": 1.1075,
      "theoretical_loss": 3.3264400659921884,
      "tokens_seen": 2934243328
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011284848484848485,
      "loss": 1.3952,
      "theoretical_loss": 3.3264343279806665,
      "tokens_seen": 2934308864
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011282828282828283,
      "loss": 1.2661,
      "theoretical_loss": 3.3264285901331805,
      "tokens_seen": 2934374400
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001128080808080808,
      "loss": 1.2166,
      "theoretical_loss": 3.3264228524497224,
      "tokens_seen": 2934439936
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011278787878787879,
      "loss": 1.1065,
      "theoretical_loss": 3.3264171149302832,
      "tokens_seen": 2934505472
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011276767676767677,
      "loss": 1.1994,
      "theoretical_loss": 3.326411377574855,
      "tokens_seen": 2934571008
    },
    {
      "epoch": 0.78,
      "objective/train/aligned_token_count": 0.27095553278923035,
      "objective/train/avg_token_score": 0.029841726645827293,
      "objective/train/docs_used": 1652566,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.7994075417518616,
      "objective/train/original_loss": 5.477022171020508,
      "objective/train/theoretical_loss": 3.3264056403834297,
      "objective/train/tokens_used": 1293495776,
      "objective/train/unlikelihood_loss": 0.060542307794094086,
      "theoretical_loss": 3.3264056403834297,
      "tokens_seen": 2934636544
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011274747474747475,
      "loss": 1.3017,
      "theoretical_loss": 3.3264056403834297,
      "tokens_seen": 2934636544
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011272727272727274,
      "loss": 1.424,
      "theoretical_loss": 3.3263999033559983,
      "tokens_seen": 2934702080
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011270707070707072,
      "loss": 1.3928,
      "theoretical_loss": 3.3263941664925536,
      "tokens_seen": 2934767616
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011268686868686868,
      "loss": 1.2939,
      "theoretical_loss": 3.3263884297930857,
      "tokens_seen": 2934833152
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011266666666666666,
      "loss": 1.3314,
      "theoretical_loss": 3.3263826932575875,
      "tokens_seen": 2934898688
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011264646464646465,
      "loss": 1.4789,
      "theoretical_loss": 3.32637695688605,
      "tokens_seen": 2934964224
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011262626262626263,
      "loss": 1.5895,
      "theoretical_loss": 3.3263712206784652,
      "tokens_seen": 2935029760
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011260606060606062,
      "loss": 1.7056,
      "theoretical_loss": 3.3263654846348243,
      "tokens_seen": 2935095296
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011258585858585859,
      "loss": 1.1306,
      "theoretical_loss": 3.3263597487551193,
      "tokens_seen": 2935160832
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011256565656565657,
      "loss": 1.1455,
      "theoretical_loss": 3.3263540130393423,
      "tokens_seen": 2935226368
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011254545454545455,
      "loss": 1.33,
      "theoretical_loss": 3.326348277487484,
      "tokens_seen": 2935291904
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011252525252525254,
      "loss": 1.2942,
      "theoretical_loss": 3.3263425420995367,
      "tokens_seen": 2935357440
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011250505050505051,
      "loss": 1.3864,
      "theoretical_loss": 3.3263368068754917,
      "tokens_seen": 2935422976
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011248484848484848,
      "loss": 1.5215,
      "theoretical_loss": 3.326331071815341,
      "tokens_seen": 2935488512
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011246464646464646,
      "loss": 1.4133,
      "theoretical_loss": 3.326325336919076,
      "tokens_seen": 2935554048
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011244444444444445,
      "loss": 1.3337,
      "theoretical_loss": 3.3263196021866883,
      "tokens_seen": 2935619584
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011242424242424243,
      "loss": 1.4078,
      "theoretical_loss": 3.32631386761817,
      "tokens_seen": 2935685120
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011240404040404041,
      "loss": 1.1468,
      "theoretical_loss": 3.3263081332135123,
      "tokens_seen": 2935750656
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011238383838383838,
      "loss": 1.3531,
      "theoretical_loss": 3.326302398972707,
      "tokens_seen": 2935816192
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011236363636363637,
      "loss": 1.2207,
      "theoretical_loss": 3.3262966648957457,
      "tokens_seen": 2935881728
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011234343434343434,
      "loss": 1.3147,
      "theoretical_loss": 3.32629093098262,
      "tokens_seen": 2935947264
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011232323232323232,
      "loss": 1.3787,
      "theoretical_loss": 3.326285197233322,
      "tokens_seen": 2936012800
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001123030303030303,
      "loss": 1.2587,
      "theoretical_loss": 3.326279463647843,
      "tokens_seen": 2936078336
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011228282828282829,
      "loss": 1.3754,
      "theoretical_loss": 3.326273730226175,
      "tokens_seen": 2936143872
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011226262626262626,
      "loss": 1.2981,
      "theoretical_loss": 3.326267996968309,
      "tokens_seen": 2936209408
    },
    {
      "epoch": 0.78,
      "objective/train/aligned_token_count": 0.5315859913825989,
      "objective/train/avg_token_score": 0.013074967078864574,
      "objective/train/docs_used": 1653144,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.717998743057251,
      "objective/train/original_loss": 4.254833698272705,
      "objective/train/theoretical_loss": 3.3262622638742374,
      "objective/train/tokens_used": 1295134176,
      "objective/train/unlikelihood_loss": 0.06617125123739243,
      "theoretical_loss": 3.3262622638742374,
      "tokens_seen": 2936274944
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011224242424242424,
      "loss": 1.4052,
      "theoretical_loss": 3.3262622638742374,
      "tokens_seen": 2936274944
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011222222222222223,
      "loss": 1.1425,
      "theoretical_loss": 3.3262565309439514,
      "tokens_seen": 2936340480
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011220202020202021,
      "loss": 1.239,
      "theoretical_loss": 3.3262507981774423,
      "tokens_seen": 2936406016
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001121818181818182,
      "loss": 1.3038,
      "theoretical_loss": 3.326245065574703,
      "tokens_seen": 2936471552
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011216161616161615,
      "loss": 1.2943,
      "theoretical_loss": 3.326239333135724,
      "tokens_seen": 2936537088
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011214141414141414,
      "loss": 1.2188,
      "theoretical_loss": 3.3262336008604976,
      "tokens_seen": 2936602624
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011212121212121212,
      "loss": 1.2161,
      "theoretical_loss": 3.326227868749015,
      "tokens_seen": 2936668160
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001121010101010101,
      "loss": 1.3468,
      "theoretical_loss": 3.3262221368012685,
      "tokens_seen": 2936733696
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011208080808080809,
      "loss": 1.1255,
      "theoretical_loss": 3.3262164050172487,
      "tokens_seen": 2936799232
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011206060606060606,
      "loss": 1.457,
      "theoretical_loss": 3.3262106733969485,
      "tokens_seen": 2936864768
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011204040404040404,
      "loss": 1.3876,
      "theoretical_loss": 3.326204941940359,
      "tokens_seen": 2936930304
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011202020202020203,
      "loss": 1.2074,
      "theoretical_loss": 3.326199210647472,
      "tokens_seen": 2936995840
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.000112,
      "loss": 1.4252,
      "theoretical_loss": 3.3261934795182793,
      "tokens_seen": 2937061376
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011197979797979798,
      "loss": 1.4169,
      "theoretical_loss": 3.3261877485527718,
      "tokens_seen": 2937126912
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011195959595959595,
      "loss": 1.5753,
      "theoretical_loss": 3.326182017750942,
      "tokens_seen": 2937192448
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011193939393939394,
      "loss": 1.4413,
      "theoretical_loss": 3.326176287112781,
      "tokens_seen": 2937257984
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011191919191919192,
      "loss": 1.1858,
      "theoretical_loss": 3.326170556638281,
      "tokens_seen": 2937323520
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001118989898989899,
      "loss": 1.2877,
      "theoretical_loss": 3.3261648263274335,
      "tokens_seen": 2937389056
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011187878787878789,
      "loss": 1.4049,
      "theoretical_loss": 3.32615909618023,
      "tokens_seen": 2937454592
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011185858585858586,
      "loss": 1.3824,
      "theoretical_loss": 3.3261533661966625,
      "tokens_seen": 2937520128
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011183838383838384,
      "loss": 1.4457,
      "theoretical_loss": 3.3261476363767226,
      "tokens_seen": 2937585664
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011181818181818181,
      "loss": 1.2035,
      "theoretical_loss": 3.3261419067204017,
      "tokens_seen": 2937651200
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001117979797979798,
      "loss": 1.2727,
      "theoretical_loss": 3.326136177227691,
      "tokens_seen": 2937716736
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011177777777777778,
      "loss": 1.2173,
      "theoretical_loss": 3.3261304478985836,
      "tokens_seen": 2937782272
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011175757575757577,
      "loss": 1.3212,
      "theoretical_loss": 3.32612471873307,
      "tokens_seen": 2937847808
    },
    {
      "epoch": 0.78,
      "objective/train/aligned_token_count": 0.4171554148197174,
      "objective/train/avg_token_score": 0.023244494572281837,
      "objective/train/docs_used": 1654272,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4244202375411987,
      "objective/train/original_loss": 4.734897613525391,
      "objective/train/theoretical_loss": 3.3261189897311425,
      "objective/train/tokens_used": 1296772576,
      "objective/train/unlikelihood_loss": 0.04013212025165558,
      "theoretical_loss": 3.3261189897311425,
      "tokens_seen": 2937913344
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011173737373737374,
      "loss": 1.461,
      "theoretical_loss": 3.3261189897311425,
      "tokens_seen": 2937913344
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011171717171717172,
      "loss": 1.1222,
      "theoretical_loss": 3.326113260892792,
      "tokens_seen": 2937978880
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001116969696969697,
      "loss": 1.4428,
      "theoretical_loss": 3.3261075322180114,
      "tokens_seen": 2938044416
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011167676767676769,
      "loss": 1.2375,
      "theoretical_loss": 3.3261018037067913,
      "tokens_seen": 2938109952
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011165656565656566,
      "loss": 1.2678,
      "theoretical_loss": 3.326096075359124,
      "tokens_seen": 2938175488
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011163636363636363,
      "loss": 1.2433,
      "theoretical_loss": 3.326090347175001,
      "tokens_seen": 2938241024
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011161616161616161,
      "loss": 1.2062,
      "theoretical_loss": 3.326084619154414,
      "tokens_seen": 2938306560
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001115959595959596,
      "loss": 1.4709,
      "theoretical_loss": 3.3260788912973545,
      "tokens_seen": 2938372096
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011157575757575758,
      "loss": 1.1825,
      "theoretical_loss": 3.326073163603814,
      "tokens_seen": 2938437632
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011155555555555556,
      "loss": 1.1894,
      "theoretical_loss": 3.326067436073785,
      "tokens_seen": 2938503168
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011153535353535353,
      "loss": 1.3184,
      "theoretical_loss": 3.326061708707259,
      "tokens_seen": 2938568704
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011151515151515152,
      "loss": 1.3042,
      "theoretical_loss": 3.3260559815042265,
      "tokens_seen": 2938634240
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001114949494949495,
      "loss": 1.242,
      "theoretical_loss": 3.326050254464681,
      "tokens_seen": 2938699776
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011147474747474747,
      "loss": 1.3865,
      "theoretical_loss": 3.3260445275886124,
      "tokens_seen": 2938765312
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011145454545454546,
      "loss": 1.2749,
      "theoretical_loss": 3.3260388008760136,
      "tokens_seen": 2938830848
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011143434343434343,
      "loss": 1.422,
      "theoretical_loss": 3.326033074326876,
      "tokens_seen": 2938896384
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011141414141414141,
      "loss": 1.466,
      "theoretical_loss": 3.326027347941191,
      "tokens_seen": 2938961920
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001113939393939394,
      "loss": 1.2697,
      "theoretical_loss": 3.326021621718951,
      "tokens_seen": 2939027456
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011137373737373738,
      "loss": 1.1858,
      "theoretical_loss": 3.3260158956601464,
      "tokens_seen": 2939092992
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011135353535353536,
      "loss": 1.3365,
      "theoretical_loss": 3.3260101697647704,
      "tokens_seen": 2939158528
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011133333333333335,
      "loss": 1.3167,
      "theoretical_loss": 3.3260044440328134,
      "tokens_seen": 2939224064
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001113131313131313,
      "loss": 1.2657,
      "theoretical_loss": 3.325998718464268,
      "tokens_seen": 2939289600
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011129292929292929,
      "loss": 1.5063,
      "theoretical_loss": 3.325992993059126,
      "tokens_seen": 2939355136
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011127272727272727,
      "loss": 1.3772,
      "theoretical_loss": 3.3259872678173776,
      "tokens_seen": 2939420672
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011125252525252526,
      "loss": 1.3735,
      "theoretical_loss": 3.3259815427390165,
      "tokens_seen": 2939486208
    },
    {
      "epoch": 0.78,
      "objective/train/aligned_token_count": 0.354227751493454,
      "objective/train/avg_token_score": 0.009427751414477825,
      "objective/train/docs_used": 1655060,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1929844617843628,
      "objective/train/original_loss": 4.970573425292969,
      "objective/train/theoretical_loss": 3.325975817824033,
      "objective/train/tokens_used": 1298410976,
      "objective/train/unlikelihood_loss": 0.054175928235054016,
      "theoretical_loss": 3.325975817824033,
      "tokens_seen": 2939551744
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011123232323232324,
      "loss": 1.2796,
      "theoretical_loss": 3.325975817824033,
      "tokens_seen": 2939551744
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011121212121212121,
      "loss": 1.4607,
      "theoretical_loss": 3.3259700930724194,
      "tokens_seen": 2939617280
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001111919191919192,
      "loss": 1.5013,
      "theoretical_loss": 3.325964368484167,
      "tokens_seen": 2939682816
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011117171717171718,
      "loss": 1.3804,
      "theoretical_loss": 3.325958644059268,
      "tokens_seen": 2939748352
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011115151515151516,
      "loss": 1.2811,
      "theoretical_loss": 3.325952919797714,
      "tokens_seen": 2939813888
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011113131313131313,
      "loss": 1.4146,
      "theoretical_loss": 3.3259471956994964,
      "tokens_seen": 2939879424
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001111111111111111,
      "loss": 1.3443,
      "theoretical_loss": 3.325941471764607,
      "tokens_seen": 2939944960
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011109090909090909,
      "loss": 1.2513,
      "theoretical_loss": 3.3259357479930376,
      "tokens_seen": 2940010496
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011107070707070707,
      "loss": 1.2912,
      "theoretical_loss": 3.3259300243847796,
      "tokens_seen": 2940076032
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011105050505050505,
      "loss": 1.3124,
      "theoretical_loss": 3.325924300939825,
      "tokens_seen": 2940141568
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011103030303030304,
      "loss": 1.2854,
      "theoretical_loss": 3.3259185776581655,
      "tokens_seen": 2940207104
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011101010101010101,
      "loss": 1.2642,
      "theoretical_loss": 3.325912854539793,
      "tokens_seen": 2940272640
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011098989898989899,
      "loss": 1.2761,
      "theoretical_loss": 3.3259071315846986,
      "tokens_seen": 2940338176
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011096969696969698,
      "loss": 1.5014,
      "theoretical_loss": 3.325901408792874,
      "tokens_seen": 2940403712
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011094949494949495,
      "loss": 1.4013,
      "theoretical_loss": 3.325895686164312,
      "tokens_seen": 2940469248
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011092929292929293,
      "loss": 1.393,
      "theoretical_loss": 3.3258899636990034,
      "tokens_seen": 2940534784
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011090909090909092,
      "loss": 1.5266,
      "theoretical_loss": 3.3258842413969396,
      "tokens_seen": 2940600320
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011088888888888889,
      "loss": 1.3975,
      "theoretical_loss": 3.325878519258113,
      "tokens_seen": 2940665856
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011086868686868687,
      "loss": 1.1827,
      "theoretical_loss": 3.325872797282515,
      "tokens_seen": 2940731392
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011084848484848485,
      "loss": 1.2843,
      "theoretical_loss": 3.3258670754701374,
      "tokens_seen": 2940796928
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011082828282828284,
      "loss": 1.2711,
      "theoretical_loss": 3.325861353820972,
      "tokens_seen": 2940862464
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011080808080808082,
      "loss": 1.0316,
      "theoretical_loss": 3.32585563233501,
      "tokens_seen": 2940928000
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011078787878787878,
      "loss": 1.4188,
      "theoretical_loss": 3.325849911012244,
      "tokens_seen": 2940993536
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011076767676767676,
      "loss": 1.4046,
      "theoretical_loss": 3.325844189852665,
      "tokens_seen": 2941059072
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011074747474747475,
      "loss": 1.2963,
      "theoretical_loss": 3.325838468856265,
      "tokens_seen": 2941124608
    },
    {
      "epoch": 0.78,
      "objective/train/aligned_token_count": 0.5701368451118469,
      "objective/train/avg_token_score": 0.005313211586326361,
      "objective/train/docs_used": 1656552,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.740568995475769,
      "objective/train/original_loss": 3.841562509536743,
      "objective/train/theoretical_loss": 3.325832748023035,
      "objective/train/tokens_used": 1300049376,
      "objective/train/unlikelihood_loss": 0.054371800273656845,
      "theoretical_loss": 3.325832748023035,
      "tokens_seen": 2941190144
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011072727272727273,
      "loss": 1.4256,
      "theoretical_loss": 3.325832748023035,
      "tokens_seen": 2941190144
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011070707070707071,
      "loss": 1.1278,
      "theoretical_loss": 3.325827027352968,
      "tokens_seen": 2941255680
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011068686868686868,
      "loss": 1.1786,
      "theoretical_loss": 3.3258213068460547,
      "tokens_seen": 2941321216
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011066666666666667,
      "loss": 1.4182,
      "theoretical_loss": 3.3258155865022876,
      "tokens_seen": 2941386752
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011064646464646465,
      "loss": 1.2564,
      "theoretical_loss": 3.3258098663216575,
      "tokens_seen": 2941452288
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011062626262626264,
      "loss": 1.4437,
      "theoretical_loss": 3.325804146304157,
      "tokens_seen": 2941517824
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011060606060606061,
      "loss": 1.4086,
      "theoretical_loss": 3.325798426449777,
      "tokens_seen": 2941583360
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011058585858585858,
      "loss": 1.3321,
      "theoretical_loss": 3.32579270675851,
      "tokens_seen": 2941648896
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011056565656565656,
      "loss": 1.44,
      "theoretical_loss": 3.3257869872303467,
      "tokens_seen": 2941714432
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011054545454545455,
      "loss": 1.4533,
      "theoretical_loss": 3.32578126786528,
      "tokens_seen": 2941779968
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011052525252525253,
      "loss": 1.4176,
      "theoretical_loss": 3.3257755486633007,
      "tokens_seen": 2941845504
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011050505050505051,
      "loss": 1.4248,
      "theoretical_loss": 3.325769829624401,
      "tokens_seen": 2941911040
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001104848484848485,
      "loss": 1.3555,
      "theoretical_loss": 3.3257641107485725,
      "tokens_seen": 2941976576
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011046464646464647,
      "loss": 1.5539,
      "theoretical_loss": 3.325758392035807,
      "tokens_seen": 2942042112
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011044444444444444,
      "loss": 1.4869,
      "theoretical_loss": 3.325752673486096,
      "tokens_seen": 2942107648
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011042424242424242,
      "loss": 1.3654,
      "theoretical_loss": 3.325746955099431,
      "tokens_seen": 2942173184
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001104040404040404,
      "loss": 1.2843,
      "theoretical_loss": 3.3257412368758046,
      "tokens_seen": 2942238720
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011038383838383839,
      "loss": 1.3165,
      "theoretical_loss": 3.325735518815208,
      "tokens_seen": 2942304256
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011036363636363636,
      "loss": 1.3861,
      "theoretical_loss": 3.3257298009176326,
      "tokens_seen": 2942369792
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011034343434343434,
      "loss": 1.0643,
      "theoretical_loss": 3.3257240831830703,
      "tokens_seen": 2942435328
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011032323232323233,
      "loss": 1.4772,
      "theoretical_loss": 3.3257183656115132,
      "tokens_seen": 2942500864
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011030303030303031,
      "loss": 1.089,
      "theoretical_loss": 3.325712648202953,
      "tokens_seen": 2942566400
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001102828282828283,
      "loss": 1.4858,
      "theoretical_loss": 3.325706930957381,
      "tokens_seen": 2942631936
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011026262626262625,
      "loss": 1.2373,
      "theoretical_loss": 3.3257012138747886,
      "tokens_seen": 2942697472
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011024242424242424,
      "loss": 1.2442,
      "theoretical_loss": 3.3256954969551686,
      "tokens_seen": 2942763008
    },
    {
      "epoch": 0.78,
      "objective/train/aligned_token_count": 0.3373045027256012,
      "objective/train/avg_token_score": 0.019415628165006638,
      "objective/train/docs_used": 1657166,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.085178017616272,
      "objective/train/original_loss": 4.345276832580566,
      "objective/train/theoretical_loss": 3.325689780198512,
      "objective/train/tokens_used": 1301687776,
      "objective/train/unlikelihood_loss": 0.06209184601902962,
      "theoretical_loss": 3.325689780198512,
      "tokens_seen": 2942828544
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011022222222222222,
      "loss": 1.274,
      "theoretical_loss": 3.325689780198512,
      "tokens_seen": 2942828544
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001102020202020202,
      "loss": 1.3425,
      "theoretical_loss": 3.3256840636048106,
      "tokens_seen": 2942894080
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011018181818181819,
      "loss": 1.1447,
      "theoretical_loss": 3.3256783471740565,
      "tokens_seen": 2942959616
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011016161616161616,
      "loss": 1.2917,
      "theoretical_loss": 3.325672630906241,
      "tokens_seen": 2943025152
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011014141414141414,
      "loss": 1.3173,
      "theoretical_loss": 3.325666914801356,
      "tokens_seen": 2943090688
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011012121212121213,
      "loss": 1.2523,
      "theoretical_loss": 3.325661198859393,
      "tokens_seen": 2943156224
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011010101010101011,
      "loss": 1.3767,
      "theoretical_loss": 3.325655483080344,
      "tokens_seen": 2943221760
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011008080808080808,
      "loss": 1.2294,
      "theoretical_loss": 3.325649767464201,
      "tokens_seen": 2943287296
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011006060606060607,
      "loss": 1.3401,
      "theoretical_loss": 3.3256440520109547,
      "tokens_seen": 2943352832
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011004040404040404,
      "loss": 1.2976,
      "theoretical_loss": 3.325638336720598,
      "tokens_seen": 2943418368
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011002020202020202,
      "loss": 1.2636,
      "theoretical_loss": 3.325632621593122,
      "tokens_seen": 2943483904
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00011,
      "loss": 1.2071,
      "theoretical_loss": 3.3256269066285187,
      "tokens_seen": 2943549440
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010997979797979799,
      "loss": 1.4183,
      "theoretical_loss": 3.3256211918267793,
      "tokens_seen": 2943614976
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010995959595959597,
      "loss": 1.0494,
      "theoretical_loss": 3.3256154771878963,
      "tokens_seen": 2943680512
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010993939393939394,
      "loss": 1.2574,
      "theoretical_loss": 3.3256097627118613,
      "tokens_seen": 2943746048
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010991919191919191,
      "loss": 1.21,
      "theoretical_loss": 3.3256040483986657,
      "tokens_seen": 2943811584
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001098989898989899,
      "loss": 1.3701,
      "theoretical_loss": 3.3255983342483013,
      "tokens_seen": 2943877120
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010987878787878788,
      "loss": 1.3274,
      "theoretical_loss": 3.3255926202607595,
      "tokens_seen": 2943942656
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010985858585858586,
      "loss": 1.5611,
      "theoretical_loss": 3.325586906436033,
      "tokens_seen": 2944008192
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010983838383838383,
      "loss": 1.3271,
      "theoretical_loss": 3.3255811927741123,
      "tokens_seen": 2944073728
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010981818181818182,
      "loss": 1.4015,
      "theoretical_loss": 3.32557547927499,
      "tokens_seen": 2944139264
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001097979797979798,
      "loss": 1.4944,
      "theoretical_loss": 3.325569765938658,
      "tokens_seen": 2944204800
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010977777777777779,
      "loss": 1.3183,
      "theoretical_loss": 3.3255640527651074,
      "tokens_seen": 2944270336
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010975757575757577,
      "loss": 1.139,
      "theoretical_loss": 3.3255583397543305,
      "tokens_seen": 2944335872
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010973737373737373,
      "loss": 1.2923,
      "theoretical_loss": 3.325552626906318,
      "tokens_seen": 2944401408
    },
    {
      "epoch": 0.78,
      "objective/train/aligned_token_count": 0.3830034136772156,
      "objective/train/avg_token_score": 0.043330609798431396,
      "objective/train/docs_used": 1658433,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2741467952728271,
      "objective/train/original_loss": 4.3888115882873535,
      "objective/train/theoretical_loss": 3.325546914221063,
      "objective/train/tokens_used": 1303326176,
      "objective/train/unlikelihood_loss": 0.06988922506570816,
      "theoretical_loss": 3.325546914221063,
      "tokens_seen": 2944466944
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010971717171717171,
      "loss": 1.3522,
      "theoretical_loss": 3.325546914221063,
      "tokens_seen": 2944466944
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001096969696969697,
      "loss": 1.3851,
      "theoretical_loss": 3.325541201698557,
      "tokens_seen": 2944532480
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010967676767676768,
      "loss": 1.3239,
      "theoretical_loss": 3.3255354893387903,
      "tokens_seen": 2944598016
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010965656565656566,
      "loss": 1.3534,
      "theoretical_loss": 3.3255297771417562,
      "tokens_seen": 2944663552
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010963636363636365,
      "loss": 1.4597,
      "theoretical_loss": 3.3255240651074462,
      "tokens_seen": 2944729088
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010961616161616162,
      "loss": 1.2022,
      "theoretical_loss": 3.325518353235852,
      "tokens_seen": 2944794624
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001095959595959596,
      "loss": 1.3296,
      "theoretical_loss": 3.3255126415269647,
      "tokens_seen": 2944860160
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010957575757575757,
      "loss": 1.2412,
      "theoretical_loss": 3.3255069299807762,
      "tokens_seen": 2944925696
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010955555555555556,
      "loss": 1.2641,
      "theoretical_loss": 3.325501218597279,
      "tokens_seen": 2944991232
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010953535353535354,
      "loss": 1.2162,
      "theoretical_loss": 3.325495507376464,
      "tokens_seen": 2945056768
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010951515151515151,
      "loss": 1.2972,
      "theoretical_loss": 3.3254897963183234,
      "tokens_seen": 2945122304
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001094949494949495,
      "loss": 1.07,
      "theoretical_loss": 3.3254840854228496,
      "tokens_seen": 2945187840
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010947474747474748,
      "loss": 1.3273,
      "theoretical_loss": 3.3254783746900327,
      "tokens_seen": 2945253376
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010945454545454546,
      "loss": 1.2131,
      "theoretical_loss": 3.3254726641198658,
      "tokens_seen": 2945318912
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010943434343434345,
      "loss": 1.2708,
      "theoretical_loss": 3.32546695371234,
      "tokens_seen": 2945384448
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010941414141414142,
      "loss": 1.3771,
      "theoretical_loss": 3.3254612434674473,
      "tokens_seen": 2945449984
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010939393939393939,
      "loss": 1.3109,
      "theoretical_loss": 3.3254555333851794,
      "tokens_seen": 2945515520
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010937373737373737,
      "loss": 1.0942,
      "theoretical_loss": 3.325449823465528,
      "tokens_seen": 2945581056
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010935353535353536,
      "loss": 1.4153,
      "theoretical_loss": 3.3254441137084854,
      "tokens_seen": 2945646592
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010933333333333334,
      "loss": 1.3557,
      "theoretical_loss": 3.3254384041140423,
      "tokens_seen": 2945712128
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010931313131313131,
      "loss": 1.3888,
      "theoretical_loss": 3.325432694682191,
      "tokens_seen": 2945777664
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001092929292929293,
      "loss": 1.3083,
      "theoretical_loss": 3.325426985412924,
      "tokens_seen": 2945843200
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010927272727272728,
      "loss": 1.0459,
      "theoretical_loss": 3.3254212763062316,
      "tokens_seen": 2945908736
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010925252525252526,
      "loss": 1.2988,
      "theoretical_loss": 3.325415567362106,
      "tokens_seen": 2945974272
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010923232323232323,
      "loss": 1.3818,
      "theoretical_loss": 3.32540985858054,
      "tokens_seen": 2946039808
    },
    {
      "epoch": 0.78,
      "objective/train/aligned_token_count": 0.3323558270931244,
      "objective/train/avg_token_score": 0.0156298466026783,
      "objective/train/docs_used": 1659139,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.068486213684082,
      "objective/train/original_loss": 4.949840068817139,
      "objective/train/theoretical_loss": 3.3254041499615243,
      "objective/train/tokens_used": 1304964576,
      "objective/train/unlikelihood_loss": 0.055758144706487656,
      "theoretical_loss": 3.3254041499615243,
      "tokens_seen": 2946105344
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010921212121212122,
      "loss": 1.2288,
      "theoretical_loss": 3.3254041499615243,
      "tokens_seen": 2946105344
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010919191919191919,
      "loss": 1.4164,
      "theoretical_loss": 3.325398441505051,
      "tokens_seen": 2946170880
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010917171717171717,
      "loss": 1.4386,
      "theoretical_loss": 3.3253927332111117,
      "tokens_seen": 2946236416
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010915151515151515,
      "loss": 1.4894,
      "theoretical_loss": 3.3253870250796984,
      "tokens_seen": 2946301952
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010913131313131314,
      "loss": 1.4774,
      "theoretical_loss": 3.3253813171108026,
      "tokens_seen": 2946367488
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010911111111111112,
      "loss": 1.1815,
      "theoretical_loss": 3.325375609304416,
      "tokens_seen": 2946433024
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010909090909090909,
      "loss": 1.1274,
      "theoretical_loss": 3.3253699016605305,
      "tokens_seen": 2946498560
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010907070707070708,
      "loss": 1.2645,
      "theoretical_loss": 3.325364194179138,
      "tokens_seen": 2946564096
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010905050505050505,
      "loss": 1.4316,
      "theoretical_loss": 3.3253584868602304,
      "tokens_seen": 2946629632
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010903030303030303,
      "loss": 1.0972,
      "theoretical_loss": 3.325352779703799,
      "tokens_seen": 2946695168
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010901010101010101,
      "loss": 1.2517,
      "theoretical_loss": 3.3253470727098358,
      "tokens_seen": 2946760704
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010898989898989899,
      "loss": 1.2088,
      "theoretical_loss": 3.3253413658783324,
      "tokens_seen": 2946826240
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010896969696969697,
      "loss": 1.3712,
      "theoretical_loss": 3.325335659209281,
      "tokens_seen": 2946891776
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010894949494949495,
      "loss": 1.3785,
      "theoretical_loss": 3.3253299527026727,
      "tokens_seen": 2946957312
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010892929292929294,
      "loss": 1.1871,
      "theoretical_loss": 3.3253242463585,
      "tokens_seen": 2947022848
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010890909090909092,
      "loss": 1.2196,
      "theoretical_loss": 3.3253185401767538,
      "tokens_seen": 2947088384
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010888888888888888,
      "loss": 1.4067,
      "theoretical_loss": 3.325312834157427,
      "tokens_seen": 2947153920
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010886868686868686,
      "loss": 1.3231,
      "theoretical_loss": 3.3253071283005102,
      "tokens_seen": 2947219456
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010884848484848485,
      "loss": 1.395,
      "theoretical_loss": 3.3253014226059956,
      "tokens_seen": 2947284992
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010882828282828283,
      "loss": 1.2348,
      "theoretical_loss": 3.3252957170738755,
      "tokens_seen": 2947350528
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010880808080808081,
      "loss": 1.3754,
      "theoretical_loss": 3.3252900117041406,
      "tokens_seen": 2947416064
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0001087878787878788,
      "loss": 1.4798,
      "theoretical_loss": 3.325284306496784,
      "tokens_seen": 2947481600
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010876767676767677,
      "loss": 1.3922,
      "theoretical_loss": 3.325278601451796,
      "tokens_seen": 2947547136
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010874747474747475,
      "loss": 1.5074,
      "theoretical_loss": 3.3252728965691696,
      "tokens_seen": 2947612672
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00010872727272727274,
      "loss": 1.3688,
      "theoretical_loss": 3.325267191848896,
      "tokens_seen": 2947678208
    },
    {
      "epoch": 0.78,
      "objective/train/aligned_token_count": 0.36620235443115234,
      "objective/train/avg_token_score": 0.006512366235256195,
      "objective/train/docs_used": 1660399,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0767982006072998,
      "objective/train/original_loss": 4.570614814758301,
      "objective/train/theoretical_loss": 3.325261487290967,
      "objective/train/tokens_used": 1306602976,
      "objective/train/unlikelihood_loss": 0.04781334102153778,
      "theoretical_loss": 3.325261487290967,
      "tokens_seen": 2947743744
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001087070707070707,
      "loss": 1.456,
      "theoretical_loss": 3.325261487290967,
      "tokens_seen": 2947743744
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010868686868686869,
      "loss": 1.3174,
      "theoretical_loss": 3.3252557828953746,
      "tokens_seen": 2947809280
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010866666666666666,
      "loss": 1.3865,
      "theoretical_loss": 3.32525007866211,
      "tokens_seen": 2947874816
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010864646464646464,
      "loss": 1.2515,
      "theoretical_loss": 3.3252443745911657,
      "tokens_seen": 2947940352
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010862626262626263,
      "loss": 1.5159,
      "theoretical_loss": 3.325238670682533,
      "tokens_seen": 2948005888
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010860606060606061,
      "loss": 1.4313,
      "theoretical_loss": 3.325232966936204,
      "tokens_seen": 2948071424
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001085858585858586,
      "loss": 1.2266,
      "theoretical_loss": 3.3252272633521707,
      "tokens_seen": 2948136960
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010856565656565657,
      "loss": 1.347,
      "theoretical_loss": 3.325221559930424,
      "tokens_seen": 2948202496
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010854545454545455,
      "loss": 1.3195,
      "theoretical_loss": 3.325215856670956,
      "tokens_seen": 2948268032
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010852525252525252,
      "loss": 1.3133,
      "theoretical_loss": 3.3252101535737584,
      "tokens_seen": 2948333568
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001085050505050505,
      "loss": 1.3475,
      "theoretical_loss": 3.325204450638824,
      "tokens_seen": 2948399104
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010848484848484849,
      "loss": 1.2819,
      "theoretical_loss": 3.325198747866143,
      "tokens_seen": 2948464640
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010846464646464646,
      "loss": 1.4966,
      "theoretical_loss": 3.325193045255708,
      "tokens_seen": 2948530176
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010844444444444444,
      "loss": 1.4933,
      "theoretical_loss": 3.3251873428075114,
      "tokens_seen": 2948595712
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010842424242424243,
      "loss": 1.1803,
      "theoretical_loss": 3.325181640521544,
      "tokens_seen": 2948661248
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010840404040404041,
      "loss": 1.2872,
      "theoretical_loss": 3.3251759383977975,
      "tokens_seen": 2948726784
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001083838383838384,
      "loss": 1.4935,
      "theoretical_loss": 3.3251702364362643,
      "tokens_seen": 2948792320
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010836363636363637,
      "loss": 1.2505,
      "theoretical_loss": 3.325164534636936,
      "tokens_seen": 2948857856
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010834343434343434,
      "loss": 1.2632,
      "theoretical_loss": 3.3251588329998043,
      "tokens_seen": 2948923392
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010832323232323232,
      "loss": 1.3968,
      "theoretical_loss": 3.325153131524861,
      "tokens_seen": 2948988928
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001083030303030303,
      "loss": 1.2356,
      "theoretical_loss": 3.325147430212098,
      "tokens_seen": 2949054464
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010828282828282829,
      "loss": 1.1149,
      "theoretical_loss": 3.3251417290615066,
      "tokens_seen": 2949120000
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010826262626262627,
      "loss": 1.4985,
      "theoretical_loss": 3.3251360280730795,
      "tokens_seen": 2949185536
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010824242424242424,
      "loss": 1.3493,
      "theoretical_loss": 3.3251303272468076,
      "tokens_seen": 2949251072
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010822222222222223,
      "loss": 1.3086,
      "theoretical_loss": 3.3251246265826833,
      "tokens_seen": 2949316608
    },
    {
      "epoch": 0.79,
      "objective/train/aligned_token_count": 0.3878909945487976,
      "objective/train/avg_token_score": 0.023629309609532356,
      "objective/train/docs_used": 1661100,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2736353874206543,
      "objective/train/original_loss": 5.044515132904053,
      "objective/train/theoretical_loss": 3.3251189260806977,
      "objective/train/tokens_used": 1308241376,
      "objective/train/unlikelihood_loss": 0.04076236113905907,
      "theoretical_loss": 3.3251189260806977,
      "tokens_seen": 2949382144
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010820202020202021,
      "loss": 1.268,
      "theoretical_loss": 3.3251189260806977,
      "tokens_seen": 2949382144
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010818181818181818,
      "loss": 1.21,
      "theoretical_loss": 3.325113225740843,
      "tokens_seen": 2949447680
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010816161616161616,
      "loss": 1.1447,
      "theoretical_loss": 3.3251075255631113,
      "tokens_seen": 2949513216
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010814141414141414,
      "loss": 1.4666,
      "theoretical_loss": 3.325101825547494,
      "tokens_seen": 2949578752
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010812121212121212,
      "loss": 1.1961,
      "theoretical_loss": 3.3250961256939826,
      "tokens_seen": 2949644288
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001081010101010101,
      "loss": 1.3613,
      "theoretical_loss": 3.3250904260025695,
      "tokens_seen": 2949709824
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010808080808080809,
      "loss": 1.2615,
      "theoretical_loss": 3.3250847264732464,
      "tokens_seen": 2949775360
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010806060606060607,
      "loss": 1.4327,
      "theoretical_loss": 3.325079027106005,
      "tokens_seen": 2949840896
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010804040404040404,
      "loss": 1.2862,
      "theoretical_loss": 3.325073327900837,
      "tokens_seen": 2949906432
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010802020202020201,
      "loss": 1.284,
      "theoretical_loss": 3.3250676288577337,
      "tokens_seen": 2949971968
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000108,
      "loss": 1.3118,
      "theoretical_loss": 3.325061929976688,
      "tokens_seen": 2950037504
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010797979797979798,
      "loss": 1.1654,
      "theoretical_loss": 3.3250562312576903,
      "tokens_seen": 2950103040
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010795959595959596,
      "loss": 1.4669,
      "theoretical_loss": 3.325050532700734,
      "tokens_seen": 2950168576
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010793939393939395,
      "loss": 1.1205,
      "theoretical_loss": 3.3250448343058094,
      "tokens_seen": 2950234112
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010791919191919192,
      "loss": 1.3107,
      "theoretical_loss": 3.3250391360729097,
      "tokens_seen": 2950299648
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001078989898989899,
      "loss": 1.3838,
      "theoretical_loss": 3.325033438002025,
      "tokens_seen": 2950365184
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010787878787878789,
      "loss": 1.3952,
      "theoretical_loss": 3.325027740093149,
      "tokens_seen": 2950430720
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010785858585858587,
      "loss": 1.5181,
      "theoretical_loss": 3.325022042346272,
      "tokens_seen": 2950496256
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010783838383838384,
      "loss": 1.3844,
      "theoretical_loss": 3.3250163447613867,
      "tokens_seen": 2950561792
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010781818181818181,
      "loss": 1.4498,
      "theoretical_loss": 3.3250106473384844,
      "tokens_seen": 2950627328
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001077979797979798,
      "loss": 1.6577,
      "theoretical_loss": 3.325004950077557,
      "tokens_seen": 2950692864
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010777777777777778,
      "loss": 1.192,
      "theoretical_loss": 3.3249992529785963,
      "tokens_seen": 2950758400
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010775757575757576,
      "loss": 1.4892,
      "theoretical_loss": 3.324993556041594,
      "tokens_seen": 2950823936
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010773737373737375,
      "loss": 1.4564,
      "theoretical_loss": 3.3249878592665425,
      "tokens_seen": 2950889472
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010771717171717172,
      "loss": 1.3584,
      "theoretical_loss": 3.3249821626534324,
      "tokens_seen": 2950955008
    },
    {
      "epoch": 0.79,
      "objective/train/aligned_token_count": 0.423081636428833,
      "objective/train/avg_token_score": 0.009568899869918823,
      "objective/train/docs_used": 1661802,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3021057844161987,
      "objective/train/original_loss": 4.949382305145264,
      "objective/train/theoretical_loss": 3.3249764662022567,
      "objective/train/tokens_used": 1309879776,
      "objective/train/unlikelihood_loss": 0.04517403990030289,
      "theoretical_loss": 3.3249764662022567,
      "tokens_seen": 2951020544
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001076969696969697,
      "loss": 1.2796,
      "theoretical_loss": 3.3249764662022567,
      "tokens_seen": 2951020544
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010767676767676769,
      "loss": 1.2018,
      "theoretical_loss": 3.3249707699130067,
      "tokens_seen": 2951086080
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010765656565656566,
      "loss": 1.3593,
      "theoretical_loss": 3.324965073785674,
      "tokens_seen": 2951151616
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010763636363636364,
      "loss": 1.2855,
      "theoretical_loss": 3.3249593778202513,
      "tokens_seen": 2951217152
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010761616161616161,
      "loss": 1.4704,
      "theoretical_loss": 3.324953682016729,
      "tokens_seen": 2951282688
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001075959595959596,
      "loss": 1.3065,
      "theoretical_loss": 3.3249479863751,
      "tokens_seen": 2951348224
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010757575757575758,
      "loss": 1.3727,
      "theoretical_loss": 3.324942290895356,
      "tokens_seen": 2951413760
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010755555555555556,
      "loss": 1.2297,
      "theoretical_loss": 3.324936595577488,
      "tokens_seen": 2951479296
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010753535353535355,
      "loss": 1.5737,
      "theoretical_loss": 3.3249309004214886,
      "tokens_seen": 2951544832
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010751515151515153,
      "loss": 1.3141,
      "theoretical_loss": 3.3249252054273493,
      "tokens_seen": 2951610368
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010749494949494949,
      "loss": 1.4092,
      "theoretical_loss": 3.324919510595062,
      "tokens_seen": 2951675904
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010747474747474747,
      "loss": 1.2935,
      "theoretical_loss": 3.3249138159246185,
      "tokens_seen": 2951741440
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010745454545454545,
      "loss": 1.2144,
      "theoretical_loss": 3.3249081214160108,
      "tokens_seen": 2951806976
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010743434343434344,
      "loss": 1.217,
      "theoretical_loss": 3.3249024270692304,
      "tokens_seen": 2951872512
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010741414141414142,
      "loss": 1.2602,
      "theoretical_loss": 3.3248967328842687,
      "tokens_seen": 2951938048
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010739393939393939,
      "loss": 1.2752,
      "theoretical_loss": 3.3248910388611184,
      "tokens_seen": 2952003584
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010737373737373738,
      "loss": 1.136,
      "theoretical_loss": 3.324885344999771,
      "tokens_seen": 2952069120
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010735353535353536,
      "loss": 1.1196,
      "theoretical_loss": 3.3248796513002183,
      "tokens_seen": 2952134656
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010733333333333334,
      "loss": 1.2459,
      "theoretical_loss": 3.3248739577624518,
      "tokens_seen": 2952200192
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010731313131313132,
      "loss": 1.3636,
      "theoretical_loss": 3.3248682643864633,
      "tokens_seen": 2952265728
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010729292929292929,
      "loss": 1.2777,
      "theoretical_loss": 3.3248625711722455,
      "tokens_seen": 2952331264
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010727272727272727,
      "loss": 1.3953,
      "theoretical_loss": 3.324856878119789,
      "tokens_seen": 2952396800
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010725252525252525,
      "loss": 1.3895,
      "theoretical_loss": 3.3248511852290865,
      "tokens_seen": 2952462336
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010723232323232324,
      "loss": 1.341,
      "theoretical_loss": 3.3248454925001294,
      "tokens_seen": 2952527872
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010721212121212122,
      "loss": 1.4925,
      "theoretical_loss": 3.3248397999329096,
      "tokens_seen": 2952593408
    },
    {
      "epoch": 0.79,
      "objective/train/aligned_token_count": 0.5428885817527771,
      "objective/train/avg_token_score": 0.06465417146682739,
      "objective/train/docs_used": 1662642,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6591153144836426,
      "objective/train/original_loss": 4.4160380363464355,
      "objective/train/theoretical_loss": 3.324834107527419,
      "objective/train/tokens_used": 1311518176,
      "objective/train/unlikelihood_loss": 0.04000373184680939,
      "theoretical_loss": 3.324834107527419,
      "tokens_seen": 2952658944
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010719191919191919,
      "loss": 1.5172,
      "theoretical_loss": 3.324834107527419,
      "tokens_seen": 2952658944
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010717171717171718,
      "loss": 1.4558,
      "theoretical_loss": 3.3248284152836494,
      "tokens_seen": 2952724480
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010715151515151515,
      "loss": 1.4341,
      "theoretical_loss": 3.3248227232015926,
      "tokens_seen": 2952790016
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010713131313131313,
      "loss": 1.3417,
      "theoretical_loss": 3.32481703128124,
      "tokens_seen": 2952855552
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010711111111111111,
      "loss": 1.3078,
      "theoretical_loss": 3.3248113395225842,
      "tokens_seen": 2952921088
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001070909090909091,
      "loss": 1.1271,
      "theoretical_loss": 3.3248056479256167,
      "tokens_seen": 2952986624
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010707070707070707,
      "loss": 1.2658,
      "theoretical_loss": 3.324799956490329,
      "tokens_seen": 2953052160
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010705050505050505,
      "loss": 1.2097,
      "theoretical_loss": 3.324794265216713,
      "tokens_seen": 2953117696
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010703030303030304,
      "loss": 1.3666,
      "theoretical_loss": 3.324788574104761,
      "tokens_seen": 2953183232
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010701010101010102,
      "loss": 1.5094,
      "theoretical_loss": 3.324782883154464,
      "tokens_seen": 2953248768
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000106989898989899,
      "loss": 1.3815,
      "theoretical_loss": 3.324777192365815,
      "tokens_seen": 2953314304
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010696969696969696,
      "loss": 1.3432,
      "theoretical_loss": 3.3247715017388044,
      "tokens_seen": 2953379840
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010694949494949495,
      "loss": 1.3004,
      "theoretical_loss": 3.3247658112734255,
      "tokens_seen": 2953445376
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010692929292929293,
      "loss": 1.3824,
      "theoretical_loss": 3.324760120969669,
      "tokens_seen": 2953510912
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010690909090909091,
      "loss": 1.1506,
      "theoretical_loss": 3.324754430827527,
      "tokens_seen": 2953576448
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001068888888888889,
      "loss": 1.2658,
      "theoretical_loss": 3.3247487408469913,
      "tokens_seen": 2953641984
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010686868686868687,
      "loss": 1.253,
      "theoretical_loss": 3.3247430510280545,
      "tokens_seen": 2953707520
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010684848484848485,
      "loss": 1.1357,
      "theoretical_loss": 3.324737361370707,
      "tokens_seen": 2953773056
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010682828282828284,
      "loss": 1.1573,
      "theoretical_loss": 3.324731671874942,
      "tokens_seen": 2953838592
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010680808080808082,
      "loss": 1.3234,
      "theoretical_loss": 3.3247259825407505,
      "tokens_seen": 2953904128
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010678787878787879,
      "loss": 1.4632,
      "theoretical_loss": 3.3247202933681246,
      "tokens_seen": 2953969664
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010676767676767676,
      "loss": 1.408,
      "theoretical_loss": 3.324714604357056,
      "tokens_seen": 2954035200
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010674747474747474,
      "loss": 1.3363,
      "theoretical_loss": 3.3247089155075367,
      "tokens_seen": 2954100736
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010672727272727273,
      "loss": 1.3863,
      "theoretical_loss": 3.324703226819558,
      "tokens_seen": 2954166272
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010670707070707071,
      "loss": 1.267,
      "theoretical_loss": 3.3246975382931128,
      "tokens_seen": 2954231808
    },
    {
      "epoch": 0.79,
      "objective/train/aligned_token_count": 0.43890517950057983,
      "objective/train/avg_token_score": 0.013032070361077785,
      "objective/train/docs_used": 1663271,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1984424591064453,
      "objective/train/original_loss": 3.840472459793091,
      "objective/train/theoretical_loss": 3.3246918499281923,
      "objective/train/tokens_used": 1313156576,
      "objective/train/unlikelihood_loss": 0.0668654590845108,
      "theoretical_loss": 3.3246918499281923,
      "tokens_seen": 2954297344
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001066868686868687,
      "loss": 1.3858,
      "theoretical_loss": 3.3246918499281923,
      "tokens_seen": 2954297344
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010666666666666668,
      "loss": 1.3232,
      "theoretical_loss": 3.324686161724788,
      "tokens_seen": 2954362880
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010664646464646465,
      "loss": 1.4144,
      "theoretical_loss": 3.3246804736828923,
      "tokens_seen": 2954428416
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010662626262626262,
      "loss": 1.282,
      "theoretical_loss": 3.3246747858024968,
      "tokens_seen": 2954493952
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001066060606060606,
      "loss": 1.2766,
      "theoretical_loss": 3.324669098083593,
      "tokens_seen": 2954559488
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010658585858585859,
      "loss": 1.269,
      "theoretical_loss": 3.324663410526173,
      "tokens_seen": 2954625024
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010656565656565657,
      "loss": 1.4704,
      "theoretical_loss": 3.3246577231302292,
      "tokens_seen": 2954690560
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010654545454545454,
      "loss": 1.2287,
      "theoretical_loss": 3.3246520358957525,
      "tokens_seen": 2954756096
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010652525252525253,
      "loss": 1.5479,
      "theoretical_loss": 3.3246463488227356,
      "tokens_seen": 2954821632
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010650505050505051,
      "loss": 1.3815,
      "theoretical_loss": 3.32464066191117,
      "tokens_seen": 2954887168
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001064848484848485,
      "loss": 1.7539,
      "theoretical_loss": 3.3246349751610467,
      "tokens_seen": 2954952704
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010646464646464648,
      "loss": 1.5069,
      "theoretical_loss": 3.3246292885723587,
      "tokens_seen": 2955018240
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010644444444444444,
      "loss": 1.5404,
      "theoretical_loss": 3.3246236021450972,
      "tokens_seen": 2955083776
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010642424242424242,
      "loss": 1.2803,
      "theoretical_loss": 3.3246179158792546,
      "tokens_seen": 2955149312
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001064040404040404,
      "loss": 1.4878,
      "theoretical_loss": 3.3246122297748224,
      "tokens_seen": 2955214848
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010638383838383839,
      "loss": 1.6155,
      "theoretical_loss": 3.3246065438317918,
      "tokens_seen": 2955280384
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010636363636363637,
      "loss": 1.3065,
      "theoretical_loss": 3.3246008580501556,
      "tokens_seen": 2955345920
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010634343434343434,
      "loss": 1.5306,
      "theoretical_loss": 3.3245951724299054,
      "tokens_seen": 2955411456
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010632323232323233,
      "loss": 1.2421,
      "theoretical_loss": 3.324589486971033,
      "tokens_seen": 2955476992
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010630303030303031,
      "loss": 1.3525,
      "theoretical_loss": 3.3245838016735303,
      "tokens_seen": 2955542528
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010628282828282828,
      "loss": 1.2858,
      "theoretical_loss": 3.3245781165373884,
      "tokens_seen": 2955608064
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010626262626262626,
      "loss": 1.3125,
      "theoretical_loss": 3.3245724315626006,
      "tokens_seen": 2955673600
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010624242424242425,
      "loss": 1.3903,
      "theoretical_loss": 3.3245667467491575,
      "tokens_seen": 2955739136
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010622222222222222,
      "loss": 1.433,
      "theoretical_loss": 3.324561062097051,
      "tokens_seen": 2955804672
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001062020202020202,
      "loss": 1.5595,
      "theoretical_loss": 3.3245553776062735,
      "tokens_seen": 2955870208
    },
    {
      "epoch": 0.79,
      "objective/train/aligned_token_count": 0.32502442598342896,
      "objective/train/avg_token_score": 0.026489894837141037,
      "objective/train/docs_used": 1664408,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0113320350646973,
      "objective/train/original_loss": 4.9697465896606445,
      "objective/train/theoretical_loss": 3.324549693276817,
      "objective/train/tokens_used": 1314794976,
      "objective/train/unlikelihood_loss": 0.05897453427314758,
      "theoretical_loss": 3.324549693276817,
      "tokens_seen": 2955935744
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010618181818181819,
      "loss": 1.3421,
      "theoretical_loss": 3.324549693276817,
      "tokens_seen": 2955935744
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010616161616161617,
      "loss": 1.3824,
      "theoretical_loss": 3.324544009108673,
      "tokens_seen": 2956001280
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010614141414141415,
      "loss": 1.2417,
      "theoretical_loss": 3.324538325101833,
      "tokens_seen": 2956066816
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010612121212121212,
      "loss": 1.1869,
      "theoretical_loss": 3.3245326412562894,
      "tokens_seen": 2956132352
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001061010101010101,
      "loss": 1.4383,
      "theoretical_loss": 3.3245269575720338,
      "tokens_seen": 2956197888
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010608080808080808,
      "loss": 1.4414,
      "theoretical_loss": 3.324521274049058,
      "tokens_seen": 2956263424
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010606060606060606,
      "loss": 1.5342,
      "theoretical_loss": 3.3245155906873536,
      "tokens_seen": 2956328960
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010604040404040405,
      "loss": 1.3268,
      "theoretical_loss": 3.324509907486913,
      "tokens_seen": 2956394496
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010602020202020202,
      "loss": 1.4667,
      "theoretical_loss": 3.324504224447728,
      "tokens_seen": 2956460032
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000106,
      "loss": 1.3654,
      "theoretical_loss": 3.32449854156979,
      "tokens_seen": 2956525568
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010597979797979799,
      "loss": 1.2638,
      "theoretical_loss": 3.3244928588530915,
      "tokens_seen": 2956591104
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010595959595959597,
      "loss": 1.3503,
      "theoretical_loss": 3.324487176297624,
      "tokens_seen": 2956656640
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010593939393939394,
      "loss": 1.5502,
      "theoretical_loss": 3.324481493903379,
      "tokens_seen": 2956722176
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010591919191919191,
      "loss": 1.3622,
      "theoretical_loss": 3.324475811670349,
      "tokens_seen": 2956787712
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001058989898989899,
      "loss": 1.5,
      "theoretical_loss": 3.324470129598525,
      "tokens_seen": 2956853248
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010587878787878788,
      "loss": 1.6171,
      "theoretical_loss": 3.3244644476878995,
      "tokens_seen": 2956918784
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010585858585858586,
      "loss": 1.3055,
      "theoretical_loss": 3.3244587659384646,
      "tokens_seen": 2956984320
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010583838383838385,
      "loss": 1.4038,
      "theoretical_loss": 3.324453084350212,
      "tokens_seen": 2957049856
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010581818181818183,
      "loss": 1.0934,
      "theoretical_loss": 3.324447402923133,
      "tokens_seen": 2957115392
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001057979797979798,
      "loss": 1.3241,
      "theoretical_loss": 3.3244417216572195,
      "tokens_seen": 2957180928
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010577777777777778,
      "loss": 1.5832,
      "theoretical_loss": 3.324436040552464,
      "tokens_seen": 2957246464
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010575757575757575,
      "loss": 1.5407,
      "theoretical_loss": 3.324430359608858,
      "tokens_seen": 2957312000
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010573737373737374,
      "loss": 1.2874,
      "theoretical_loss": 3.3244246788263934,
      "tokens_seen": 2957377536
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010571717171717172,
      "loss": 1.1738,
      "theoretical_loss": 3.324418998205062,
      "tokens_seen": 2957443072
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010569696969696969,
      "loss": 1.5023,
      "theoretical_loss": 3.3244133177448556,
      "tokens_seen": 2957508608
    },
    {
      "epoch": 0.79,
      "objective/train/aligned_token_count": 0.43786656856536865,
      "objective/train/avg_token_score": 0.007747194729745388,
      "objective/train/docs_used": 1665006,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3715813159942627,
      "objective/train/original_loss": 4.28681755065918,
      "objective/train/theoretical_loss": 3.3244076374457663,
      "objective/train/tokens_used": 1316433376,
      "objective/train/unlikelihood_loss": 0.05123388394713402,
      "theoretical_loss": 3.3244076374457663,
      "tokens_seen": 2957574144
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010567676767676768,
      "loss": 1.2819,
      "theoretical_loss": 3.3244076374457663,
      "tokens_seen": 2957574144
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010565656565656566,
      "loss": 1.1941,
      "theoretical_loss": 3.324401957307786,
      "tokens_seen": 2957639680
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010563636363636365,
      "loss": 1.3012,
      "theoretical_loss": 3.324396277330906,
      "tokens_seen": 2957705216
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010561616161616163,
      "loss": 1.3563,
      "theoretical_loss": 3.3243905975151185,
      "tokens_seen": 2957770752
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010559595959595959,
      "loss": 1.4871,
      "theoretical_loss": 3.3243849178604155,
      "tokens_seen": 2957836288
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010557575757575757,
      "loss": 1.3676,
      "theoretical_loss": 3.324379238366789,
      "tokens_seen": 2957901824
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010555555555555555,
      "loss": 1.4963,
      "theoretical_loss": 3.32437355903423,
      "tokens_seen": 2957967360
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010553535353535354,
      "loss": 1.4024,
      "theoretical_loss": 3.3243678798627316,
      "tokens_seen": 2958032896
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010551515151515152,
      "loss": 1.3615,
      "theoretical_loss": 3.324362200852285,
      "tokens_seen": 2958098432
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010549494949494949,
      "loss": 1.4837,
      "theoretical_loss": 3.324356522002882,
      "tokens_seen": 2958163968
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010547474747474748,
      "loss": 1.6538,
      "theoretical_loss": 3.3243508433145146,
      "tokens_seen": 2958229504
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010545454545454546,
      "loss": 1.4512,
      "theoretical_loss": 3.3243451647871747,
      "tokens_seen": 2958295040
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010543434343434344,
      "loss": 1.303,
      "theoretical_loss": 3.324339486420854,
      "tokens_seen": 2958360576
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010541414141414141,
      "loss": 1.2841,
      "theoretical_loss": 3.3243338082155445,
      "tokens_seen": 2958426112
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010539393939393938,
      "loss": 1.4541,
      "theoretical_loss": 3.324328130171238,
      "tokens_seen": 2958491648
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010537373737373737,
      "loss": 1.308,
      "theoretical_loss": 3.3243224522879267,
      "tokens_seen": 2958557184
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010535353535353535,
      "loss": 1.5002,
      "theoretical_loss": 3.324316774565602,
      "tokens_seen": 2958622720
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010533333333333334,
      "loss": 1.4219,
      "theoretical_loss": 3.324311097004256,
      "tokens_seen": 2958688256
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010531313131313132,
      "loss": 1.2459,
      "theoretical_loss": 3.3243054196038804,
      "tokens_seen": 2958753792
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001052929292929293,
      "loss": 1.3743,
      "theoretical_loss": 3.3242997423644676,
      "tokens_seen": 2958819328
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010527272727272728,
      "loss": 1.5237,
      "theoretical_loss": 3.3242940652860087,
      "tokens_seen": 2958884864
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010525252525252526,
      "loss": 1.3888,
      "theoretical_loss": 3.3242883883684957,
      "tokens_seen": 2958950400
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010523232323232323,
      "loss": 1.4883,
      "theoretical_loss": 3.3242827116119216,
      "tokens_seen": 2959015936
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010521212121212121,
      "loss": 1.3597,
      "theoretical_loss": 3.3242770350162765,
      "tokens_seen": 2959081472
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001051919191919192,
      "loss": 1.4519,
      "theoretical_loss": 3.3242713585815533,
      "tokens_seen": 2959147008
    },
    {
      "epoch": 0.79,
      "objective/train/aligned_token_count": 0.4176441729068756,
      "objective/train/avg_token_score": 0.02708185650408268,
      "objective/train/docs_used": 1665825,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2349313497543335,
      "objective/train/original_loss": 4.842739582061768,
      "objective/train/theoretical_loss": 3.324265682307744,
      "objective/train/tokens_used": 1318071776,
      "objective/train/unlikelihood_loss": 0.04712604731321335,
      "theoretical_loss": 3.324265682307744,
      "tokens_seen": 2959212544
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010517171717171717,
      "loss": 1.3223,
      "theoretical_loss": 3.324265682307744,
      "tokens_seen": 2959212544
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010515151515151515,
      "loss": 1.1893,
      "theoretical_loss": 3.3242600061948404,
      "tokens_seen": 2959278080
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010513131313131314,
      "loss": 1.2526,
      "theoretical_loss": 3.324254330242834,
      "tokens_seen": 2959343616
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010511111111111112,
      "loss": 1.1573,
      "theoretical_loss": 3.3242486544517167,
      "tokens_seen": 2959409152
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001050909090909091,
      "loss": 1.5111,
      "theoretical_loss": 3.3242429788214807,
      "tokens_seen": 2959474688
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010507070707070706,
      "loss": 1.2601,
      "theoretical_loss": 3.3242373033521173,
      "tokens_seen": 2959540224
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010505050505050504,
      "loss": 1.2991,
      "theoretical_loss": 3.324231628043619,
      "tokens_seen": 2959605760
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010503030303030303,
      "loss": 1.3896,
      "theoretical_loss": 3.324225952895978,
      "tokens_seen": 2959671296
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010501010101010101,
      "loss": 1.3376,
      "theoretical_loss": 3.324220277909185,
      "tokens_seen": 2959736832
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000104989898989899,
      "loss": 1.4392,
      "theoretical_loss": 3.324214603083233,
      "tokens_seen": 2959802368
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010496969696969697,
      "loss": 1.4528,
      "theoretical_loss": 3.324208928418113,
      "tokens_seen": 2959867904
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010494949494949495,
      "loss": 1.2713,
      "theoretical_loss": 3.3242032539138178,
      "tokens_seen": 2959933440
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010492929292929293,
      "loss": 1.4724,
      "theoretical_loss": 3.3241975795703382,
      "tokens_seen": 2959998976
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010490909090909092,
      "loss": 1.2496,
      "theoretical_loss": 3.324191905387667,
      "tokens_seen": 2960064512
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010488888888888889,
      "loss": 1.3042,
      "theoretical_loss": 3.324186231365796,
      "tokens_seen": 2960130048
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010486868686868687,
      "loss": 1.6522,
      "theoretical_loss": 3.3241805575047163,
      "tokens_seen": 2960195584
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010484848484848484,
      "loss": 1.4307,
      "theoretical_loss": 3.3241748838044205,
      "tokens_seen": 2960261120
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010482828282828283,
      "loss": 1.5456,
      "theoretical_loss": 3.3241692102649,
      "tokens_seen": 2960326656
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010480808080808081,
      "loss": 1.4454,
      "theoretical_loss": 3.3241635368861475,
      "tokens_seen": 2960392192
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001047878787878788,
      "loss": 1.2906,
      "theoretical_loss": 3.3241578636681544,
      "tokens_seen": 2960457728
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010476767676767678,
      "loss": 1.4572,
      "theoretical_loss": 3.324152190610912,
      "tokens_seen": 2960523264
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010474747474747475,
      "loss": 1.2789,
      "theoretical_loss": 3.3241465177144134,
      "tokens_seen": 2960588800
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010472727272727272,
      "loss": 1.4232,
      "theoretical_loss": 3.324140844978649,
      "tokens_seen": 2960654336
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001047070707070707,
      "loss": 1.6694,
      "theoretical_loss": 3.3241351724036123,
      "tokens_seen": 2960719872
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010468686868686869,
      "loss": 1.4521,
      "theoretical_loss": 3.3241294999892936,
      "tokens_seen": 2960785408
    },
    {
      "epoch": 0.79,
      "objective/train/aligned_token_count": 0.3751832842826843,
      "objective/train/avg_token_score": 0.009580383077263832,
      "objective/train/docs_used": 1666526,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1593621969223022,
      "objective/train/original_loss": 4.589208602905273,
      "objective/train/theoretical_loss": 3.3241238277356864,
      "objective/train/tokens_used": 1319710176,
      "objective/train/unlikelihood_loss": 0.04558636620640755,
      "theoretical_loss": 3.3241238277356864,
      "tokens_seen": 2960850944
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010466666666666667,
      "loss": 1.3061,
      "theoretical_loss": 3.3241238277356864,
      "tokens_seen": 2960850944
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010464646464646464,
      "loss": 1.5874,
      "theoretical_loss": 3.3241181556427812,
      "tokens_seen": 2960916480
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010462626262626263,
      "loss": 1.3785,
      "theoretical_loss": 3.3241124837105707,
      "tokens_seen": 2960982016
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010460606060606061,
      "loss": 1.4866,
      "theoretical_loss": 3.3241068119390467,
      "tokens_seen": 2961047552
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001045858585858586,
      "loss": 1.397,
      "theoretical_loss": 3.324101140328201,
      "tokens_seen": 2961113088
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010456565656565658,
      "loss": 1.5389,
      "theoretical_loss": 3.324095468878025,
      "tokens_seen": 2961178624
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010454545454545454,
      "loss": 1.4071,
      "theoretical_loss": 3.3240897975885115,
      "tokens_seen": 2961244160
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010452525252525252,
      "loss": 1.4463,
      "theoretical_loss": 3.3240841264596517,
      "tokens_seen": 2961309696
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001045050505050505,
      "loss": 1.4973,
      "theoretical_loss": 3.3240784554914375,
      "tokens_seen": 2961375232
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010448484848484849,
      "loss": 1.2087,
      "theoretical_loss": 3.324072784683861,
      "tokens_seen": 2961440768
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010446464646464647,
      "loss": 1.1266,
      "theoretical_loss": 3.324067114036915,
      "tokens_seen": 2961506304
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010444444444444445,
      "loss": 1.2277,
      "theoretical_loss": 3.3240614435505895,
      "tokens_seen": 2961571840
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010442424242424243,
      "loss": 1.4208,
      "theoretical_loss": 3.324055773224878,
      "tokens_seen": 2961637376
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010440404040404041,
      "loss": 1.4487,
      "theoretical_loss": 3.3240501030597716,
      "tokens_seen": 2961702912
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010438383838383839,
      "loss": 1.2548,
      "theoretical_loss": 3.324044433055262,
      "tokens_seen": 2961768448
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010436363636363636,
      "loss": 1.188,
      "theoretical_loss": 3.324038763211342,
      "tokens_seen": 2961833984
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010434343434343435,
      "loss": 1.4241,
      "theoretical_loss": 3.324033093528003,
      "tokens_seen": 2961899520
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010432323232323232,
      "loss": 1.3626,
      "theoretical_loss": 3.3240274240052368,
      "tokens_seen": 2961965056
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001043030303030303,
      "loss": 1.2705,
      "theoretical_loss": 3.3240217546430353,
      "tokens_seen": 2962030592
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010428282828282829,
      "loss": 1.3059,
      "theoretical_loss": 3.3240160854413903,
      "tokens_seen": 2962096128
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010426262626262627,
      "loss": 1.3593,
      "theoretical_loss": 3.324010416400294,
      "tokens_seen": 2962161664
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010424242424242425,
      "loss": 1.7058,
      "theoretical_loss": 3.3240047475197385,
      "tokens_seen": 2962227200
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010422222222222222,
      "loss": 1.3203,
      "theoretical_loss": 3.3239990787997153,
      "tokens_seen": 2962292736
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001042020202020202,
      "loss": 1.4137,
      "theoretical_loss": 3.323993410240216,
      "tokens_seen": 2962358272
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010418181818181818,
      "loss": 1.3253,
      "theoretical_loss": 3.3239877418412336,
      "tokens_seen": 2962423808
    },
    {
      "epoch": 0.79,
      "objective/train/aligned_token_count": 0.45369014143943787,
      "objective/train/avg_token_score": 0.015634510666131973,
      "objective/train/docs_used": 1667691,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.258712649345398,
      "objective/train/original_loss": 4.262852668762207,
      "objective/train/theoretical_loss": 3.323982073602759,
      "objective/train/tokens_used": 1321348576,
      "objective/train/unlikelihood_loss": 0.06873971223831177,
      "theoretical_loss": 3.323982073602759,
      "tokens_seen": 2962489344
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010416161616161616,
      "loss": 1.2707,
      "theoretical_loss": 3.323982073602759,
      "tokens_seen": 2962489344
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010414141414141415,
      "loss": 1.3516,
      "theoretical_loss": 3.3239764055247845,
      "tokens_seen": 2962554880
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010412121212121212,
      "loss": 1.3804,
      "theoretical_loss": 3.3239707376073016,
      "tokens_seen": 2962620416
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001041010101010101,
      "loss": 1.6925,
      "theoretical_loss": 3.3239650698503027,
      "tokens_seen": 2962685952
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010408080808080808,
      "loss": 1.2574,
      "theoretical_loss": 3.3239594022537795,
      "tokens_seen": 2962751488
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010406060606060607,
      "loss": 1.4149,
      "theoretical_loss": 3.323953734817724,
      "tokens_seen": 2962817024
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010404040404040405,
      "loss": 1.4165,
      "theoretical_loss": 3.323948067542128,
      "tokens_seen": 2962882560
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010402020202020202,
      "loss": 1.3864,
      "theoretical_loss": 3.3239424004269833,
      "tokens_seen": 2962948096
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.000104,
      "loss": 1.1942,
      "theoretical_loss": 3.3239367334722822,
      "tokens_seen": 2963013632
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010397979797979798,
      "loss": 1.3938,
      "theoretical_loss": 3.323931066678016,
      "tokens_seen": 2963079168
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010395959595959596,
      "loss": 1.2031,
      "theoretical_loss": 3.3239254000441774,
      "tokens_seen": 2963144704
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010393939393939395,
      "loss": 1.5086,
      "theoretical_loss": 3.323919733570758,
      "tokens_seen": 2963210240
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010391919191919193,
      "loss": 1.3934,
      "theoretical_loss": 3.323914067257749,
      "tokens_seen": 2963275776
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001038989898989899,
      "loss": 1.5544,
      "theoretical_loss": 3.323908401105143,
      "tokens_seen": 2963341312
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010387878787878788,
      "loss": 1.286,
      "theoretical_loss": 3.3239027351129327,
      "tokens_seen": 2963406848
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010385858585858585,
      "loss": 1.4522,
      "theoretical_loss": 3.3238970692811085,
      "tokens_seen": 2963472384
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010383838383838384,
      "loss": 1.1648,
      "theoretical_loss": 3.3238914036096627,
      "tokens_seen": 2963537920
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010381818181818182,
      "loss": 1.4567,
      "theoretical_loss": 3.323885738098588,
      "tokens_seen": 2963603456
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010379797979797979,
      "loss": 1.4166,
      "theoretical_loss": 3.3238800727478752,
      "tokens_seen": 2963668992
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010377777777777778,
      "loss": 1.5321,
      "theoretical_loss": 3.323874407557517,
      "tokens_seen": 2963734528
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010375757575757576,
      "loss": 1.6323,
      "theoretical_loss": 3.3238687425275053,
      "tokens_seen": 2963800064
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010373737373737374,
      "loss": 1.1779,
      "theoretical_loss": 3.3238630776578315,
      "tokens_seen": 2963865600
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010371717171717173,
      "loss": 1.3972,
      "theoretical_loss": 3.323857412948488,
      "tokens_seen": 2963931136
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0001036969696969697,
      "loss": 1.5209,
      "theoretical_loss": 3.3238517483994667,
      "tokens_seen": 2963996672
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.00010367676767676767,
      "loss": 1.5207,
      "theoretical_loss": 3.3238460840107593,
      "tokens_seen": 2964062208
    },
    {
      "epoch": 0.79,
      "objective/train/aligned_token_count": 0.306512713432312,
      "objective/train/avg_token_score": 0.016950635239481926,
      "objective/train/docs_used": 1668204,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0386186838150024,
      "objective/train/original_loss": 4.652601718902588,
      "objective/train/theoretical_loss": 3.323840419782358,
      "objective/train/tokens_used": 1322986976,
      "objective/train/unlikelihood_loss": 0.07792497426271439,
      "theoretical_loss": 3.323840419782358,
      "tokens_seen": 2964127744
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010365656565656565,
      "loss": 1.4223,
      "theoretical_loss": 3.323840419782358,
      "tokens_seen": 2964127744
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010363636363636364,
      "loss": 1.3606,
      "theoretical_loss": 3.3238347557142545,
      "tokens_seen": 2964193280
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010361616161616162,
      "loss": 1.3364,
      "theoretical_loss": 3.32382909180644,
      "tokens_seen": 2964258816
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001035959595959596,
      "loss": 1.2795,
      "theoretical_loss": 3.323823428058908,
      "tokens_seen": 2964324352
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010357575757575758,
      "loss": 1.3264,
      "theoretical_loss": 3.323817764471649,
      "tokens_seen": 2964389888
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010355555555555556,
      "loss": 1.4101,
      "theoretical_loss": 3.323812101044656,
      "tokens_seen": 2964455424
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010353535353535354,
      "loss": 1.3571,
      "theoretical_loss": 3.32380643777792,
      "tokens_seen": 2964520960
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010351515151515151,
      "loss": 1.3377,
      "theoretical_loss": 3.3238007746714335,
      "tokens_seen": 2964586496
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001034949494949495,
      "loss": 1.3213,
      "theoretical_loss": 3.3237951117251883,
      "tokens_seen": 2964652032
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010347474747474747,
      "loss": 1.1371,
      "theoretical_loss": 3.323789448939176,
      "tokens_seen": 2964717568
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010345454545454545,
      "loss": 1.3136,
      "theoretical_loss": 3.323783786313389,
      "tokens_seen": 2964783104
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010343434343434344,
      "loss": 1.4192,
      "theoretical_loss": 3.3237781238478195,
      "tokens_seen": 2964848640
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010341414141414142,
      "loss": 1.5206,
      "theoretical_loss": 3.3237724615424584,
      "tokens_seen": 2964914176
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001033939393939394,
      "loss": 1.2142,
      "theoretical_loss": 3.323766799397298,
      "tokens_seen": 2964979712
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010337373737373737,
      "loss": 1.4622,
      "theoretical_loss": 3.323761137412331,
      "tokens_seen": 2965045248
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010335353535353536,
      "loss": 1.3893,
      "theoretical_loss": 3.3237554755875487,
      "tokens_seen": 2965110784
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010333333333333333,
      "loss": 1.1259,
      "theoretical_loss": 3.3237498139229427,
      "tokens_seen": 2965176320
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010331313131313131,
      "loss": 1.3201,
      "theoretical_loss": 3.3237441524185054,
      "tokens_seen": 2965241856
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001032929292929293,
      "loss": 1.5733,
      "theoretical_loss": 3.323738491074229,
      "tokens_seen": 2965307392
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010327272727272727,
      "loss": 1.3323,
      "theoretical_loss": 3.3237328298901048,
      "tokens_seen": 2965372928
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010325252525252525,
      "loss": 1.3332,
      "theoretical_loss": 3.3237271688661245,
      "tokens_seen": 2965438464
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010323232323232324,
      "loss": 1.3649,
      "theoretical_loss": 3.323721508002281,
      "tokens_seen": 2965504000
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010321212121212122,
      "loss": 1.527,
      "theoretical_loss": 3.3237158472985655,
      "tokens_seen": 2965569536
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001031919191919192,
      "loss": 1.3101,
      "theoretical_loss": 3.3237101867549708,
      "tokens_seen": 2965635072
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010317171717171719,
      "loss": 1.3427,
      "theoretical_loss": 3.3237045263714875,
      "tokens_seen": 2965700608
    },
    {
      "epoch": 0.8,
      "objective/train/aligned_token_count": 0.369012713432312,
      "objective/train/avg_token_score": 0.01784663274884224,
      "objective/train/docs_used": 1669270,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1621402502059937,
      "objective/train/original_loss": 4.396885871887207,
      "objective/train/theoretical_loss": 3.3236988661481086,
      "objective/train/tokens_used": 1324625376,
      "objective/train/unlikelihood_loss": 0.05419619753956795,
      "theoretical_loss": 3.3236988661481086,
      "tokens_seen": 2965766144
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010315151515151514,
      "loss": 1.3127,
      "theoretical_loss": 3.3236988661481086,
      "tokens_seen": 2965766144
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010313131313131313,
      "loss": 1.2725,
      "theoretical_loss": 3.3236932060848257,
      "tokens_seen": 2965831680
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010311111111111111,
      "loss": 1.3168,
      "theoretical_loss": 3.3236875461816306,
      "tokens_seen": 2965897216
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001030909090909091,
      "loss": 1.3249,
      "theoretical_loss": 3.3236818864385156,
      "tokens_seen": 2965962752
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010307070707070708,
      "loss": 1.3701,
      "theoretical_loss": 3.323676226855472,
      "tokens_seen": 2966028288
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010305050505050505,
      "loss": 1.5618,
      "theoretical_loss": 3.3236705674324925,
      "tokens_seen": 2966093824
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010303030303030303,
      "loss": 1.2203,
      "theoretical_loss": 3.323664908169568,
      "tokens_seen": 2966159360
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010301010101010102,
      "loss": 1.2221,
      "theoretical_loss": 3.323659249066692,
      "tokens_seen": 2966224896
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010298989898989899,
      "loss": 1.3259,
      "theoretical_loss": 3.323653590123855,
      "tokens_seen": 2966290432
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010296969696969697,
      "loss": 1.3137,
      "theoretical_loss": 3.3236479313410494,
      "tokens_seen": 2966355968
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010294949494949494,
      "loss": 1.34,
      "theoretical_loss": 3.3236422727182675,
      "tokens_seen": 2966421504
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010292929292929293,
      "loss": 1.2733,
      "theoretical_loss": 3.3236366142555007,
      "tokens_seen": 2966487040
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010290909090909091,
      "loss": 1.2963,
      "theoretical_loss": 3.323630955952741,
      "tokens_seen": 2966552576
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001028888888888889,
      "loss": 1.3637,
      "theoretical_loss": 3.323625297809981,
      "tokens_seen": 2966618112
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010286868686868688,
      "loss": 1.5224,
      "theoretical_loss": 3.323619639827212,
      "tokens_seen": 2966683648
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010284848484848485,
      "loss": 1.3755,
      "theoretical_loss": 3.323613982004426,
      "tokens_seen": 2966749184
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010282828282828283,
      "loss": 1.4515,
      "theoretical_loss": 3.323608324341615,
      "tokens_seen": 2966814720
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001028080808080808,
      "loss": 1.4621,
      "theoretical_loss": 3.3236026668387715,
      "tokens_seen": 2966880256
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010278787878787879,
      "loss": 1.3276,
      "theoretical_loss": 3.323597009495886,
      "tokens_seen": 2966945792
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010276767676767677,
      "loss": 1.3547,
      "theoretical_loss": 3.3235913523129526,
      "tokens_seen": 2967011328
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010274747474747476,
      "loss": 1.4275,
      "theoretical_loss": 3.323585695289961,
      "tokens_seen": 2967076864
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010272727272727273,
      "loss": 1.4313,
      "theoretical_loss": 3.323580038426905,
      "tokens_seen": 2967142400
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010270707070707071,
      "loss": 1.2929,
      "theoretical_loss": 3.323574381723775,
      "tokens_seen": 2967207936
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001026868686868687,
      "loss": 1.1566,
      "theoretical_loss": 3.3235687251805643,
      "tokens_seen": 2967273472
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010266666666666668,
      "loss": 1.3659,
      "theoretical_loss": 3.323563068797264,
      "tokens_seen": 2967339008
    },
    {
      "epoch": 0.8,
      "objective/train/aligned_token_count": 0.33736559748649597,
      "objective/train/avg_token_score": 0.020665505900979042,
      "objective/train/docs_used": 1669763,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0949420928955078,
      "objective/train/original_loss": 4.880209445953369,
      "objective/train/theoretical_loss": 3.323557412573866,
      "objective/train/tokens_used": 1326263776,
      "objective/train/unlikelihood_loss": 0.051140666007995605,
      "theoretical_loss": 3.323557412573866,
      "tokens_seen": 2967404544
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010264646464646465,
      "loss": 1.4631,
      "theoretical_loss": 3.323557412573866,
      "tokens_seen": 2967404544
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010262626262626262,
      "loss": 1.378,
      "theoretical_loss": 3.3235517565103625,
      "tokens_seen": 2967470080
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001026060606060606,
      "loss": 1.3608,
      "theoretical_loss": 3.3235461006067455,
      "tokens_seen": 2967535616
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010258585858585859,
      "loss": 1.3051,
      "theoretical_loss": 3.3235404448630073,
      "tokens_seen": 2967601152
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010256565656565657,
      "loss": 1.5907,
      "theoretical_loss": 3.323534789279139,
      "tokens_seen": 2967666688
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010254545454545455,
      "loss": 1.2707,
      "theoretical_loss": 3.3235291338551334,
      "tokens_seen": 2967732224
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010252525252525252,
      "loss": 1.4683,
      "theoretical_loss": 3.3235234785909817,
      "tokens_seen": 2967797760
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010250505050505051,
      "loss": 1.4119,
      "theoretical_loss": 3.3235178234866765,
      "tokens_seen": 2967863296
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010248484848484849,
      "loss": 1.5817,
      "theoretical_loss": 3.3235121685422095,
      "tokens_seen": 2967928832
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010246464646464646,
      "loss": 1.398,
      "theoretical_loss": 3.3235065137575726,
      "tokens_seen": 2967994368
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010244444444444445,
      "loss": 1.3762,
      "theoretical_loss": 3.323500859132757,
      "tokens_seen": 2968059904
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010242424242424242,
      "loss": 1.3458,
      "theoretical_loss": 3.3234952046677564,
      "tokens_seen": 2968125440
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001024040404040404,
      "loss": 1.1188,
      "theoretical_loss": 3.3234895503625617,
      "tokens_seen": 2968190976
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010238383838383839,
      "loss": 1.5154,
      "theoretical_loss": 3.3234838962171644,
      "tokens_seen": 2968256512
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010236363636363637,
      "loss": 1.4473,
      "theoretical_loss": 3.323478242231557,
      "tokens_seen": 2968322048
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010234343434343435,
      "loss": 1.3123,
      "theoretical_loss": 3.323472588405732,
      "tokens_seen": 2968387584
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010232323232323234,
      "loss": 1.4632,
      "theoretical_loss": 3.3234669347396806,
      "tokens_seen": 2968453120
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001023030303030303,
      "loss": 1.358,
      "theoretical_loss": 3.323461281233395,
      "tokens_seen": 2968518656
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010228282828282828,
      "loss": 1.2382,
      "theoretical_loss": 3.323455627886867,
      "tokens_seen": 2968584192
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010226262626262626,
      "loss": 1.4494,
      "theoretical_loss": 3.3234499747000887,
      "tokens_seen": 2968649728
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010224242424242425,
      "loss": 1.4915,
      "theoretical_loss": 3.323444321673052,
      "tokens_seen": 2968715264
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010222222222222223,
      "loss": 1.3489,
      "theoretical_loss": 3.323438668805749,
      "tokens_seen": 2968780800
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001022020202020202,
      "loss": 1.291,
      "theoretical_loss": 3.3234330160981713,
      "tokens_seen": 2968846336
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010218181818181818,
      "loss": 1.477,
      "theoretical_loss": 3.3234273635503118,
      "tokens_seen": 2968911872
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010216161616161617,
      "loss": 1.2157,
      "theoretical_loss": 3.323421711162161,
      "tokens_seen": 2968977408
    },
    {
      "epoch": 0.8,
      "objective/train/aligned_token_count": 0.3861803412437439,
      "objective/train/avg_token_score": 0.009217427112162113,
      "objective/train/docs_used": 1671048,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.314888596534729,
      "objective/train/original_loss": 4.597879409790039,
      "objective/train/theoretical_loss": 3.323416058933712,
      "objective/train/tokens_used": 1327902176,
      "objective/train/unlikelihood_loss": 0.053666114807128906,
      "theoretical_loss": 3.323416058933712,
      "tokens_seen": 2969042944
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010214141414141415,
      "loss": 1.2512,
      "theoretical_loss": 3.323416058933712,
      "tokens_seen": 2969042944
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010212121212121212,
      "loss": 1.436,
      "theoretical_loss": 3.3234104068649564,
      "tokens_seen": 2969108480
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010210101010101009,
      "loss": 1.611,
      "theoretical_loss": 3.3234047549558863,
      "tokens_seen": 2969174016
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010208080808080808,
      "loss": 1.4365,
      "theoretical_loss": 3.3233991032064933,
      "tokens_seen": 2969239552
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010206060606060606,
      "loss": 1.4733,
      "theoretical_loss": 3.3233934516167696,
      "tokens_seen": 2969305088
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010204040404040404,
      "loss": 1.2585,
      "theoretical_loss": 3.3233878001867074,
      "tokens_seen": 2969370624
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010202020202020203,
      "loss": 1.3087,
      "theoretical_loss": 3.3233821489162985,
      "tokens_seen": 2969436160
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000102,
      "loss": 1.1712,
      "theoretical_loss": 3.3233764978055342,
      "tokens_seen": 2969501696
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010197979797979798,
      "loss": 1.6276,
      "theoretical_loss": 3.3233708468544076,
      "tokens_seen": 2969567232
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010195959595959597,
      "loss": 1.2894,
      "theoretical_loss": 3.32336519606291,
      "tokens_seen": 2969632768
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010193939393939394,
      "loss": 1.344,
      "theoretical_loss": 3.3233595454310336,
      "tokens_seen": 2969698304
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010191919191919192,
      "loss": 1.396,
      "theoretical_loss": 3.32335389495877,
      "tokens_seen": 2969763840
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001018989898989899,
      "loss": 1.2764,
      "theoretical_loss": 3.3233482446461116,
      "tokens_seen": 2969829376
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010187878787878788,
      "loss": 1.428,
      "theoretical_loss": 3.32334259449305,
      "tokens_seen": 2969894912
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010185858585858586,
      "loss": 1.5178,
      "theoretical_loss": 3.323336944499578,
      "tokens_seen": 2969960448
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010183838383838384,
      "loss": 1.6021,
      "theoretical_loss": 3.323331294665686,
      "tokens_seen": 2970025984
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010181818181818183,
      "loss": 1.5229,
      "theoretical_loss": 3.3233256449913675,
      "tokens_seen": 2970091520
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010179797979797981,
      "loss": 1.5942,
      "theoretical_loss": 3.323319995476614,
      "tokens_seen": 2970157056
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010177777777777777,
      "loss": 1.3239,
      "theoretical_loss": 3.323314346121417,
      "tokens_seen": 2970222592
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010175757575757575,
      "loss": 1.5264,
      "theoretical_loss": 3.3233086969257695,
      "tokens_seen": 2970288128
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010173737373737374,
      "loss": 1.3362,
      "theoretical_loss": 3.323303047889662,
      "tokens_seen": 2970353664
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010171717171717172,
      "loss": 1.4657,
      "theoretical_loss": 3.323297399013088,
      "tokens_seen": 2970419200
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001016969696969697,
      "loss": 1.3749,
      "theoretical_loss": 3.3232917502960384,
      "tokens_seen": 2970484736
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010167676767676767,
      "loss": 1.4549,
      "theoretical_loss": 3.3232861017385056,
      "tokens_seen": 2970550272
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010165656565656566,
      "loss": 1.4505,
      "theoretical_loss": 3.323280453340481,
      "tokens_seen": 2970615808
    },
    {
      "epoch": 0.8,
      "objective/train/aligned_token_count": 0.4253421425819397,
      "objective/train/avg_token_score": 0.02136329375207424,
      "objective/train/docs_used": 1671649,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4400073289871216,
      "objective/train/original_loss": 4.193417549133301,
      "objective/train/theoretical_loss": 3.323274805101958,
      "objective/train/tokens_used": 1329540576,
      "objective/train/unlikelihood_loss": 0.05755547061562538,
      "theoretical_loss": 3.323274805101958,
      "tokens_seen": 2970681344
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010163636363636364,
      "loss": 1.3894,
      "theoretical_loss": 3.323274805101958,
      "tokens_seen": 2970681344
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010161616161616163,
      "loss": 1.3712,
      "theoretical_loss": 3.3232691570229274,
      "tokens_seen": 2970746880
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001015959595959596,
      "loss": 1.3569,
      "theoretical_loss": 3.323263509103381,
      "tokens_seen": 2970812416
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010157575757575757,
      "loss": 1.4848,
      "theoretical_loss": 3.3232578613433117,
      "tokens_seen": 2970877952
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010155555555555555,
      "loss": 1.1974,
      "theoretical_loss": 3.3232522137427107,
      "tokens_seen": 2970943488
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010153535353535354,
      "loss": 1.3539,
      "theoretical_loss": 3.3232465663015702,
      "tokens_seen": 2971009024
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010151515151515152,
      "loss": 1.3032,
      "theoretical_loss": 3.3232409190198826,
      "tokens_seen": 2971074560
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001014949494949495,
      "loss": 1.3983,
      "theoretical_loss": 3.3232352718976395,
      "tokens_seen": 2971140096
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010147474747474749,
      "loss": 1.4769,
      "theoretical_loss": 3.3232296249348328,
      "tokens_seen": 2971205632
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010145454545454546,
      "loss": 1.2501,
      "theoretical_loss": 3.3232239781314545,
      "tokens_seen": 2971271168
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010143434343434343,
      "loss": 1.4559,
      "theoretical_loss": 3.3232183314874972,
      "tokens_seen": 2971336704
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010141414141414141,
      "loss": 1.2563,
      "theoretical_loss": 3.323212685002952,
      "tokens_seen": 2971402240
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001013939393939394,
      "loss": 1.4884,
      "theoretical_loss": 3.323207038677811,
      "tokens_seen": 2971467776
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010137373737373738,
      "loss": 1.3421,
      "theoretical_loss": 3.323201392512067,
      "tokens_seen": 2971533312
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010135353535353535,
      "loss": 1.4471,
      "theoretical_loss": 3.323195746505711,
      "tokens_seen": 2971598848
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010133333333333333,
      "loss": 1.3489,
      "theoretical_loss": 3.323190100658736,
      "tokens_seen": 2971664384
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010131313131313132,
      "loss": 1.2899,
      "theoretical_loss": 3.3231844549711327,
      "tokens_seen": 2971729920
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001012929292929293,
      "loss": 1.5399,
      "theoretical_loss": 3.323178809442894,
      "tokens_seen": 2971795456
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010127272727272729,
      "loss": 1.3118,
      "theoretical_loss": 3.323173164074012,
      "tokens_seen": 2971860992
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010125252525252524,
      "loss": 1.5175,
      "theoretical_loss": 3.323167518864478,
      "tokens_seen": 2971926528
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010123232323232323,
      "loss": 1.3159,
      "theoretical_loss": 3.3231618738142843,
      "tokens_seen": 2971992064
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010121212121212121,
      "loss": 1.2985,
      "theoretical_loss": 3.3231562289234233,
      "tokens_seen": 2972057600
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001011919191919192,
      "loss": 1.4402,
      "theoretical_loss": 3.323150584191886,
      "tokens_seen": 2972123136
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010117171717171718,
      "loss": 1.3113,
      "theoretical_loss": 3.3231449396196657,
      "tokens_seen": 2972188672
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010115151515151515,
      "loss": 1.5483,
      "theoretical_loss": 3.3231392952067536,
      "tokens_seen": 2972254208
    },
    {
      "debugging/Self-BLEU-5": 0.29714909839637155,
      "debugging/distinct-1-grams": 0.7583905665726303,
      "debugging/distinct-2-grams": 0.9373430156197423,
      "debugging/entropy-1-grams": 5.38260279658836,
      "debugging/entropy-2-grams": 6.052160411565846,
      "debugging/length": 452.2857142857143,
      "debugging/num_segments": 7,
      "epoch": 0.8,
      "objective/train/aligned_token_count": 0.41458943486213684,
      "objective/train/avg_token_score": 0.012768231332302094,
      "objective/train/docs_used": 1672806,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2828956842422485,
      "objective/train/original_loss": 4.09173583984375,
      "objective/train/theoretical_loss": 3.3231336509531415,
      "objective/train/tokens_used": 1331178976,
      "objective/train/unlikelihood_loss": 0.06537236273288727,
      "theoretical_loss": 3.3231336509531415,
      "tokens_seen": 2972319744
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010113131313131313,
      "loss": 1.2673,
      "theoretical_loss": 3.3231336509531415,
      "tokens_seen": 2972319744
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010111111111111112,
      "loss": 1.1543,
      "theoretical_loss": 3.3231280068588216,
      "tokens_seen": 2972385280
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010109090909090909,
      "loss": 1.4984,
      "theoretical_loss": 3.3231223629237863,
      "tokens_seen": 2972450816
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010107070707070707,
      "loss": 1.142,
      "theoretical_loss": 3.323116719148027,
      "tokens_seen": 2972516352
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010105050505050506,
      "loss": 1.3047,
      "theoretical_loss": 3.3231110755315365,
      "tokens_seen": 2972581888
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010103030303030303,
      "loss": 1.469,
      "theoretical_loss": 3.323105432074306,
      "tokens_seen": 2972647424
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010101010101010101,
      "loss": 1.2073,
      "theoretical_loss": 3.323099788776328,
      "tokens_seen": 2972712960
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.000100989898989899,
      "loss": 1.1751,
      "theoretical_loss": 3.3230941456375938,
      "tokens_seen": 2972778496
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010096969696969698,
      "loss": 1.415,
      "theoretical_loss": 3.323088502658096,
      "tokens_seen": 2972844032
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010094949494949496,
      "loss": 1.3462,
      "theoretical_loss": 3.323082859837826,
      "tokens_seen": 2972909568
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010092929292929293,
      "loss": 1.6212,
      "theoretical_loss": 3.323077217176777,
      "tokens_seen": 2972975104
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001009090909090909,
      "loss": 1.3808,
      "theoretical_loss": 3.32307157467494,
      "tokens_seen": 2973040640
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010088888888888889,
      "loss": 1.5956,
      "theoretical_loss": 3.323065932332307,
      "tokens_seen": 2973106176
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010086868686868687,
      "loss": 1.2569,
      "theoretical_loss": 3.3230602901488706,
      "tokens_seen": 2973171712
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010084848484848485,
      "loss": 1.3898,
      "theoretical_loss": 3.3230546481246224,
      "tokens_seen": 2973237248
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010082828282828283,
      "loss": 1.4701,
      "theoretical_loss": 3.3230490062595543,
      "tokens_seen": 2973302784
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010080808080808081,
      "loss": 1.4075,
      "theoretical_loss": 3.3230433645536586,
      "tokens_seen": 2973368320
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010078787878787879,
      "loss": 1.2295,
      "theoretical_loss": 3.3230377230069266,
      "tokens_seen": 2973433856
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010076767676767678,
      "loss": 1.2309,
      "theoretical_loss": 3.3230320816193517,
      "tokens_seen": 2973499392
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010074747474747476,
      "loss": 1.3545,
      "theoretical_loss": 3.323026440390924,
      "tokens_seen": 2973564928
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010072727272727272,
      "loss": 1.1342,
      "theoretical_loss": 3.3230207993216374,
      "tokens_seen": 2973630464
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001007070707070707,
      "loss": 1.3134,
      "theoretical_loss": 3.323015158411483,
      "tokens_seen": 2973696000
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010068686868686869,
      "loss": 1.4248,
      "theoretical_loss": 3.3230095176604526,
      "tokens_seen": 2973761536
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010066666666666667,
      "loss": 1.3571,
      "theoretical_loss": 3.3230038770685386,
      "tokens_seen": 2973827072
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010064646464646465,
      "loss": 1.5409,
      "theoretical_loss": 3.3229982366357325,
      "tokens_seen": 2973892608
    },
    {
      "epoch": 0.8,
      "objective/train/aligned_token_count": 0.47831133008003235,
      "objective/train/avg_token_score": 0.009169208817183971,
      "objective/train/docs_used": 1673458,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.504703164100647,
      "objective/train/original_loss": 4.242588996887207,
      "objective/train/theoretical_loss": 3.322992596362027,
      "objective/train/tokens_used": 1332817376,
      "objective/train/unlikelihood_loss": 0.05326914042234421,
      "theoretical_loss": 3.322992596362027,
      "tokens_seen": 2973958144
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010062626262626264,
      "loss": 1.4154,
      "theoretical_loss": 3.322992596362027,
      "tokens_seen": 2973958144
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010060606060606061,
      "loss": 1.4973,
      "theoretical_loss": 3.322986956247414,
      "tokens_seen": 2974023680
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010058585858585859,
      "loss": 1.2663,
      "theoretical_loss": 3.3229813162918846,
      "tokens_seen": 2974089216
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010056565656565656,
      "loss": 1.3585,
      "theoretical_loss": 3.322975676495432,
      "tokens_seen": 2974154752
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010054545454545455,
      "loss": 1.3282,
      "theoretical_loss": 3.322970036858048,
      "tokens_seen": 2974220288
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010052525252525253,
      "loss": 1.1974,
      "theoretical_loss": 3.3229643973797236,
      "tokens_seen": 2974285824
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001005050505050505,
      "loss": 1.3804,
      "theoretical_loss": 3.322958758060452,
      "tokens_seen": 2974351360
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010048484848484848,
      "loss": 1.2636,
      "theoretical_loss": 3.3229531189002244,
      "tokens_seen": 2974416896
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010046464646464647,
      "loss": 1.3342,
      "theoretical_loss": 3.3229474798990335,
      "tokens_seen": 2974482432
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010044444444444445,
      "loss": 1.4867,
      "theoretical_loss": 3.3229418410568705,
      "tokens_seen": 2974547968
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010042424242424244,
      "loss": 1.3053,
      "theoretical_loss": 3.3229362023737283,
      "tokens_seen": 2974613504
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010040404040404041,
      "loss": 1.2139,
      "theoretical_loss": 3.3229305638495985,
      "tokens_seen": 2974679040
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010038383838383838,
      "loss": 1.2145,
      "theoretical_loss": 3.3229249254844726,
      "tokens_seen": 2974744576
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010036363636363636,
      "loss": 1.3826,
      "theoretical_loss": 3.3229192872783435,
      "tokens_seen": 2974810112
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010034343434343435,
      "loss": 1.3434,
      "theoretical_loss": 3.322913649231203,
      "tokens_seen": 2974875648
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010032323232323233,
      "loss": 1.2935,
      "theoretical_loss": 3.3229080113430425,
      "tokens_seen": 2974941184
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001003030303030303,
      "loss": 1.1327,
      "theoretical_loss": 3.3229023736138545,
      "tokens_seen": 2975006720
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010028282828282828,
      "loss": 1.2369,
      "theoretical_loss": 3.322896736043631,
      "tokens_seen": 2975072256
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010026262626262627,
      "loss": 1.2574,
      "theoretical_loss": 3.322891098632364,
      "tokens_seen": 2975137792
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010024242424242425,
      "loss": 1.5274,
      "theoretical_loss": 3.3228854613800456,
      "tokens_seen": 2975203328
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010022222222222222,
      "loss": 1.2484,
      "theoretical_loss": 3.3228798242866677,
      "tokens_seen": 2975268864
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001002020202020202,
      "loss": 1.3999,
      "theoretical_loss": 3.3228741873522223,
      "tokens_seen": 2975334400
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010018181818181818,
      "loss": 1.4958,
      "theoretical_loss": 3.3228685505767013,
      "tokens_seen": 2975399936
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010016161616161616,
      "loss": 1.3415,
      "theoretical_loss": 3.3228629139600967,
      "tokens_seen": 2975465472
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010014141414141414,
      "loss": 1.2126,
      "theoretical_loss": 3.322857277502401,
      "tokens_seen": 2975531008
    },
    {
      "epoch": 0.8,
      "objective/train/aligned_token_count": 0.3375488817691803,
      "objective/train/avg_token_score": 0.016468120738863945,
      "objective/train/docs_used": 1674243,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1135029792785645,
      "objective/train/original_loss": 4.60894250869751,
      "objective/train/theoretical_loss": 3.3228516412036058,
      "objective/train/tokens_used": 1334455776,
      "objective/train/unlikelihood_loss": 0.05124470219016075,
      "theoretical_loss": 3.3228516412036058,
      "tokens_seen": 2975596544
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010012121212121213,
      "loss": 1.3643,
      "theoretical_loss": 3.3228516412036058,
      "tokens_seen": 2975596544
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010010101010101011,
      "loss": 1.3859,
      "theoretical_loss": 3.3228460050637034,
      "tokens_seen": 2975662080
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010008080808080808,
      "loss": 1.4666,
      "theoretical_loss": 3.3228403690826855,
      "tokens_seen": 2975727616
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010006060606060607,
      "loss": 1.4887,
      "theoretical_loss": 3.322834733260544,
      "tokens_seen": 2975793152
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010004040404040404,
      "loss": 1.3028,
      "theoretical_loss": 3.3228290975972716,
      "tokens_seen": 2975858688
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00010002020202020202,
      "loss": 1.6003,
      "theoretical_loss": 3.32282346209286,
      "tokens_seen": 2975924224
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0001,
      "loss": 1.6369,
      "theoretical_loss": 3.322817826747301,
      "tokens_seen": 2975989760
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.997979797979798e-05,
      "loss": 1.2994,
      "theoretical_loss": 3.3228121915605864,
      "tokens_seen": 2976055296
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.995959595959596e-05,
      "loss": 1.4557,
      "theoretical_loss": 3.3228065565327087,
      "tokens_seen": 2976120832
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.993939393939394e-05,
      "loss": 1.4437,
      "theoretical_loss": 3.3228009216636605,
      "tokens_seen": 2976186368
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.991919191919193e-05,
      "loss": 1.3011,
      "theoretical_loss": 3.3227952869534323,
      "tokens_seen": 2976251904
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.989898989898991e-05,
      "loss": 1.2466,
      "theoretical_loss": 3.3227896524020175,
      "tokens_seen": 2976317440
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.987878787878787e-05,
      "loss": 1.2667,
      "theoretical_loss": 3.3227840180094073,
      "tokens_seen": 2976382976
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.985858585858585e-05,
      "loss": 1.5697,
      "theoretical_loss": 3.322778383775594,
      "tokens_seen": 2976448512
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.983838383838384e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.3227727497005697,
      "tokens_seen": 2976514048
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.981818181818182e-05,
      "loss": 1.2422,
      "theoretical_loss": 3.3227671157843264,
      "tokens_seen": 2976579584
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.97979797979798e-05,
      "loss": 1.1332,
      "theoretical_loss": 3.3227614820268565,
      "tokens_seen": 2976645120
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.977777777777779e-05,
      "loss": 1.4467,
      "theoretical_loss": 3.3227558484281507,
      "tokens_seen": 2976710656
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.975757575757576e-05,
      "loss": 1.4133,
      "theoretical_loss": 3.322750214988203,
      "tokens_seen": 2976776192
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.973737373737374e-05,
      "loss": 1.4253,
      "theoretical_loss": 3.3227445817070036,
      "tokens_seen": 2976841728
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.971717171717173e-05,
      "loss": 1.1697,
      "theoretical_loss": 3.3227389485845458,
      "tokens_seen": 2976907264
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.96969696969697e-05,
      "loss": 1.4238,
      "theoretical_loss": 3.3227333156208214,
      "tokens_seen": 2976972800
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.967676767676768e-05,
      "loss": 1.5195,
      "theoretical_loss": 3.3227276828158216,
      "tokens_seen": 2977038336
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.965656565656565e-05,
      "loss": 1.4375,
      "theoretical_loss": 3.3227220501695394,
      "tokens_seen": 2977103872
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.963636363636363e-05,
      "loss": 1.3094,
      "theoretical_loss": 3.322716417681966,
      "tokens_seen": 2977169408
    },
    {
      "epoch": 0.8,
      "objective/train/aligned_token_count": 0.4989613890647888,
      "objective/train/avg_token_score": 0.014074224978685379,
      "objective/train/docs_used": 1674993,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.415122389793396,
      "objective/train/original_loss": 4.279930114746094,
      "objective/train/theoretical_loss": 3.3227107853530944,
      "objective/train/tokens_used": 1336094176,
      "objective/train/unlikelihood_loss": 0.060086704790592194,
      "theoretical_loss": 3.3227107853530944,
      "tokens_seen": 2977234944
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.961616161616162e-05,
      "loss": 1.3196,
      "theoretical_loss": 3.3227107853530944,
      "tokens_seen": 2977234944
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.95959595959596e-05,
      "loss": 1.332,
      "theoretical_loss": 3.3227051531829157,
      "tokens_seen": 2977300480
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.957575757575759e-05,
      "loss": 1.1421,
      "theoretical_loss": 3.3226995211714225,
      "tokens_seen": 2977366016
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.955555555555556e-05,
      "loss": 1.3276,
      "theoretical_loss": 3.322693889318607,
      "tokens_seen": 2977431552
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.953535353535354e-05,
      "loss": 1.3401,
      "theoretical_loss": 3.3226882576244607,
      "tokens_seen": 2977497088
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.951515151515151e-05,
      "loss": 1.1831,
      "theoretical_loss": 3.322682626088976,
      "tokens_seen": 2977562624
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.94949494949495e-05,
      "loss": 1.2471,
      "theoretical_loss": 3.3226769947121446,
      "tokens_seen": 2977628160
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.947474747474748e-05,
      "loss": 1.1966,
      "theoretical_loss": 3.322671363493959,
      "tokens_seen": 2977693696
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.945454545454545e-05,
      "loss": 1.4578,
      "theoretical_loss": 3.322665732434411,
      "tokens_seen": 2977759232
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.943434343434343e-05,
      "loss": 1.5964,
      "theoretical_loss": 3.3226601015334927,
      "tokens_seen": 2977824768
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.941414141414142e-05,
      "loss": 1.3901,
      "theoretical_loss": 3.322654470791196,
      "tokens_seen": 2977890304
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.93939393939394e-05,
      "loss": 1.4775,
      "theoretical_loss": 3.3226488402075125,
      "tokens_seen": 2977955840
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.937373737373739e-05,
      "loss": 1.6638,
      "theoretical_loss": 3.322643209782435,
      "tokens_seen": 2978021376
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.935353535353534e-05,
      "loss": 1.5699,
      "theoretical_loss": 3.322637579515956,
      "tokens_seen": 2978086912
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.933333333333333e-05,
      "loss": 1.3606,
      "theoretical_loss": 3.3226319494080663,
      "tokens_seen": 2978152448
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.931313131313131e-05,
      "loss": 1.457,
      "theoretical_loss": 3.3226263194587586,
      "tokens_seen": 2978217984
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.92929292929293e-05,
      "loss": 1.5054,
      "theoretical_loss": 3.3226206896680246,
      "tokens_seen": 2978283520
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.927272727272728e-05,
      "loss": 1.2356,
      "theoretical_loss": 3.3226150600358566,
      "tokens_seen": 2978349056
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.925252525252526e-05,
      "loss": 1.3699,
      "theoretical_loss": 3.3226094305622467,
      "tokens_seen": 2978414592
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.923232323232323e-05,
      "loss": 1.5353,
      "theoretical_loss": 3.322603801247187,
      "tokens_seen": 2978480128
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.921212121212122e-05,
      "loss": 1.5344,
      "theoretical_loss": 3.3225981720906694,
      "tokens_seen": 2978545664
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.91919191919192e-05,
      "loss": 1.4583,
      "theoretical_loss": 3.3225925430926857,
      "tokens_seen": 2978611200
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.917171717171717e-05,
      "loss": 1.3888,
      "theoretical_loss": 3.3225869142532285,
      "tokens_seen": 2978676736
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.915151515151515e-05,
      "loss": 1.4682,
      "theoretical_loss": 3.3225812855722894,
      "tokens_seen": 2978742272
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.913131313131313e-05,
      "loss": 1.1313,
      "theoretical_loss": 3.3225756570498604,
      "tokens_seen": 2978807808
    },
    {
      "epoch": 0.8,
      "objective/train/aligned_token_count": 0.5460044145584106,
      "objective/train/avg_token_score": 0.006976807024329901,
      "objective/train/docs_used": 1676283,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7783664464950562,
      "objective/train/original_loss": 4.085821628570557,
      "objective/train/theoretical_loss": 3.3225700286859343,
      "objective/train/tokens_used": 1337732576,
      "objective/train/unlikelihood_loss": 0.046547070145606995,
      "theoretical_loss": 3.3225700286859343,
      "tokens_seen": 2978873344
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.911111111111111e-05,
      "loss": 1.2883,
      "theoretical_loss": 3.3225700286859343,
      "tokens_seen": 2978873344
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.90909090909091e-05,
      "loss": 1.2359,
      "theoretical_loss": 3.322564400480502,
      "tokens_seen": 2978938880
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.907070707070708e-05,
      "loss": 1.2832,
      "theoretical_loss": 3.3225587724335566,
      "tokens_seen": 2979004416
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.905050505050506e-05,
      "loss": 1.5942,
      "theoretical_loss": 3.3225531445450893,
      "tokens_seen": 2979069952
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.903030303030303e-05,
      "loss": 1.253,
      "theoretical_loss": 3.322547516815093,
      "tokens_seen": 2979135488
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.9010101010101e-05,
      "loss": 1.445,
      "theoretical_loss": 3.322541889243559,
      "tokens_seen": 2979201024
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.898989898989899e-05,
      "loss": 1.1377,
      "theoretical_loss": 3.3225362618304795,
      "tokens_seen": 2979266560
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.896969696969697e-05,
      "loss": 1.3209,
      "theoretical_loss": 3.322530634575847,
      "tokens_seen": 2979332096
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.894949494949495e-05,
      "loss": 1.2735,
      "theoretical_loss": 3.322525007479653,
      "tokens_seen": 2979397632
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.892929292929292e-05,
      "loss": 1.3332,
      "theoretical_loss": 3.32251938054189,
      "tokens_seen": 2979463168
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.890909090909091e-05,
      "loss": 1.1821,
      "theoretical_loss": 3.32251375376255,
      "tokens_seen": 2979528704
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.888888888888889e-05,
      "loss": 1.2155,
      "theoretical_loss": 3.322508127141625,
      "tokens_seen": 2979594240
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.886868686868688e-05,
      "loss": 1.2107,
      "theoretical_loss": 3.3225025006791067,
      "tokens_seen": 2979659776
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.884848484848486e-05,
      "loss": 1.2081,
      "theoretical_loss": 3.3224968743749876,
      "tokens_seen": 2979725312
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.882828282828283e-05,
      "loss": 1.2942,
      "theoretical_loss": 3.3224912482292597,
      "tokens_seen": 2979790848
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.88080808080808e-05,
      "loss": 1.2609,
      "theoretical_loss": 3.3224856222419143,
      "tokens_seen": 2979856384
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.878787878787878e-05,
      "loss": 1.4362,
      "theoretical_loss": 3.3224799964129446,
      "tokens_seen": 2979921920
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.876767676767677e-05,
      "loss": 1.2326,
      "theoretical_loss": 3.3224743707423423,
      "tokens_seen": 2979987456
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.874747474747475e-05,
      "loss": 1.3776,
      "theoretical_loss": 3.322468745230099,
      "tokens_seen": 2980052992
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.872727272727274e-05,
      "loss": 1.2825,
      "theoretical_loss": 3.3224631198762076,
      "tokens_seen": 2980118528
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.870707070707071e-05,
      "loss": 1.1904,
      "theoretical_loss": 3.322457494680659,
      "tokens_seen": 2980184064
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.868686868686869e-05,
      "loss": 1.4676,
      "theoretical_loss": 3.322451869643446,
      "tokens_seen": 2980249600
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.866666666666668e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.3224462447645613,
      "tokens_seen": 2980315136
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.864646464646465e-05,
      "loss": 1.367,
      "theoretical_loss": 3.3224406200439955,
      "tokens_seen": 2980380672
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.862626262626263e-05,
      "loss": 1.1934,
      "theoretical_loss": 3.322434995481742,
      "tokens_seen": 2980446208
    },
    {
      "epoch": 0.8,
      "objective/train/aligned_token_count": 0.2837243378162384,
      "objective/train/avg_token_score": 0.011892415583133698,
      "objective/train/docs_used": 1676817,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8564134240150452,
      "objective/train/original_loss": 4.280081748962402,
      "objective/train/theoretical_loss": 3.3224293710777917,
      "objective/train/tokens_used": 1339370976,
      "objective/train/unlikelihood_loss": 0.07432522624731064,
      "theoretical_loss": 3.3224293710777917,
      "tokens_seen": 2980511744
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.86060606060606e-05,
      "loss": 1.11,
      "theoretical_loss": 3.3224293710777917,
      "tokens_seen": 2980511744
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.858585858585858e-05,
      "loss": 1.481,
      "theoretical_loss": 3.322423746832137,
      "tokens_seen": 2980577280
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.856565656565657e-05,
      "loss": 1.2878,
      "theoretical_loss": 3.322418122744771,
      "tokens_seen": 2980642816
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.854545454545455e-05,
      "loss": 1.3979,
      "theoretical_loss": 3.3224124988156847,
      "tokens_seen": 2980708352
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.852525252525254e-05,
      "loss": 1.3079,
      "theoretical_loss": 3.3224068750448703,
      "tokens_seen": 2980773888
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.85050505050505e-05,
      "loss": 1.1612,
      "theoretical_loss": 3.3224012514323196,
      "tokens_seen": 2980839424
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.848484848484848e-05,
      "loss": 1.2538,
      "theoretical_loss": 3.3223956279780253,
      "tokens_seen": 2980904960
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.846464646464646e-05,
      "loss": 1.1947,
      "theoretical_loss": 3.3223900046819796,
      "tokens_seen": 2980970496
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.844444444444444e-05,
      "loss": 1.5316,
      "theoretical_loss": 3.3223843815441736,
      "tokens_seen": 2981036032
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.842424242424243e-05,
      "loss": 1.4408,
      "theoretical_loss": 3.3223787585646005,
      "tokens_seen": 2981101568
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.840404040404041e-05,
      "loss": 1.228,
      "theoretical_loss": 3.3223731357432515,
      "tokens_seen": 2981167104
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.838383838383838e-05,
      "loss": 1.2597,
      "theoretical_loss": 3.322367513080119,
      "tokens_seen": 2981232640
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.836363636363637e-05,
      "loss": 1.3171,
      "theoretical_loss": 3.3223618905751953,
      "tokens_seen": 2981298176
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.834343434343435e-05,
      "loss": 1.258,
      "theoretical_loss": 3.322356268228472,
      "tokens_seen": 2981363712
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.832323232323233e-05,
      "loss": 1.152,
      "theoretical_loss": 3.3223506460399417,
      "tokens_seen": 2981429248
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.83030303030303e-05,
      "loss": 1.1032,
      "theoretical_loss": 3.3223450240095955,
      "tokens_seen": 2981494784
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.828282828282828e-05,
      "loss": 1.2689,
      "theoretical_loss": 3.322339402137427,
      "tokens_seen": 2981560320
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.826262626262626e-05,
      "loss": 1.4103,
      "theoretical_loss": 3.3223337804234268,
      "tokens_seen": 2981625856
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.824242424242424e-05,
      "loss": 1.1915,
      "theoretical_loss": 3.322328158867588,
      "tokens_seen": 2981691392
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.822222222222223e-05,
      "loss": 1.1156,
      "theoretical_loss": 3.322322537469902,
      "tokens_seen": 2981756928
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.820202020202021e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.322316916230361,
      "tokens_seen": 2981822464
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.818181818181818e-05,
      "loss": 1.3923,
      "theoretical_loss": 3.3223112951489577,
      "tokens_seen": 2981888000
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.816161616161617e-05,
      "loss": 1.4599,
      "theoretical_loss": 3.3223056742256833,
      "tokens_seen": 2981953536
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.814141414141414e-05,
      "loss": 1.2641,
      "theoretical_loss": 3.3223000534605305,
      "tokens_seen": 2982019072
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.812121212121212e-05,
      "loss": 1.3664,
      "theoretical_loss": 3.322294432853491,
      "tokens_seen": 2982084608
    },
    {
      "epoch": 0.81,
      "objective/train/aligned_token_count": 0.3291178047657013,
      "objective/train/avg_token_score": 0.010801139287650585,
      "objective/train/docs_used": 1677869,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.010230541229248,
      "objective/train/original_loss": 4.815960884094238,
      "objective/train/theoretical_loss": 3.3222888124045573,
      "objective/train/tokens_used": 1341009376,
      "objective/train/unlikelihood_loss": 0.05664817988872528,
      "theoretical_loss": 3.3222888124045573,
      "tokens_seen": 2982150144
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.81010101010101e-05,
      "loss": 1.3605,
      "theoretical_loss": 3.3222888124045573,
      "tokens_seen": 2982150144
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.808080808080807e-05,
      "loss": 1.1703,
      "theoretical_loss": 3.322283192113721,
      "tokens_seen": 2982215680
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.806060606060606e-05,
      "loss": 1.4832,
      "theoretical_loss": 3.3222775719809743,
      "tokens_seen": 2982281216
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.804040404040404e-05,
      "loss": 1.0658,
      "theoretical_loss": 3.3222719520063095,
      "tokens_seen": 2982346752
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.802020202020203e-05,
      "loss": 1.2675,
      "theoretical_loss": 3.3222663321897183,
      "tokens_seen": 2982412288
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.800000000000001e-05,
      "loss": 1.4891,
      "theoretical_loss": 3.3222607125311936,
      "tokens_seen": 2982477824
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.7979797979798e-05,
      "loss": 1.4756,
      "theoretical_loss": 3.3222550930307264,
      "tokens_seen": 2982543360
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.795959595959595e-05,
      "loss": 1.2043,
      "theoretical_loss": 3.3222494736883093,
      "tokens_seen": 2982608896
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.793939393939394e-05,
      "loss": 1.3802,
      "theoretical_loss": 3.3222438545039346,
      "tokens_seen": 2982674432
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.791919191919192e-05,
      "loss": 1.3815,
      "theoretical_loss": 3.3222382354775943,
      "tokens_seen": 2982739968
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.78989898989899e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.3222326166092797,
      "tokens_seen": 2982805504
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.787878787878789e-05,
      "loss": 1.4386,
      "theoretical_loss": 3.322226997898984,
      "tokens_seen": 2982871040
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.785858585858586e-05,
      "loss": 1.3051,
      "theoretical_loss": 3.322221379346699,
      "tokens_seen": 2982936576
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.783838383838384e-05,
      "loss": 1.4576,
      "theoretical_loss": 3.322215760952416,
      "tokens_seen": 2983002112
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.781818181818183e-05,
      "loss": 1.3999,
      "theoretical_loss": 3.322210142716128,
      "tokens_seen": 2983067648
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.77979797979798e-05,
      "loss": 1.369,
      "theoretical_loss": 3.3222045246378267,
      "tokens_seen": 2983133184
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.777777777777778e-05,
      "loss": 1.377,
      "theoretical_loss": 3.3221989067175044,
      "tokens_seen": 2983198720
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.775757575757575e-05,
      "loss": 1.4052,
      "theoretical_loss": 3.3221932889551526,
      "tokens_seen": 2983264256
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.773737373737373e-05,
      "loss": 1.1995,
      "theoretical_loss": 3.3221876713507643,
      "tokens_seen": 2983329792
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.771717171717172e-05,
      "loss": 1.299,
      "theoretical_loss": 3.322182053904331,
      "tokens_seen": 2983395328
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.76969696969697e-05,
      "loss": 1.3534,
      "theoretical_loss": 3.3221764366158446,
      "tokens_seen": 2983460864
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.767676767676769e-05,
      "loss": 1.4401,
      "theoretical_loss": 3.322170819485298,
      "tokens_seen": 2983526400
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.765656565656566e-05,
      "loss": 1.4323,
      "theoretical_loss": 3.322165202512682,
      "tokens_seen": 2983591936
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.763636363636364e-05,
      "loss": 1.5388,
      "theoretical_loss": 3.3221595856979906,
      "tokens_seen": 2983657472
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.761616161616161e-05,
      "loss": 1.3503,
      "theoretical_loss": 3.322153969041214,
      "tokens_seen": 2983723008
    },
    {
      "epoch": 0.81,
      "objective/train/aligned_token_count": 0.455828458070755,
      "objective/train/avg_token_score": 0.013668995350599289,
      "objective/train/docs_used": 1678369,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4183601140975952,
      "objective/train/original_loss": 3.9421637058258057,
      "objective/train/theoretical_loss": 3.322148352542345,
      "objective/train/tokens_used": 1342647776,
      "objective/train/unlikelihood_loss": 0.08172054588794708,
      "theoretical_loss": 3.322148352542345,
      "tokens_seen": 2983788544
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.75959595959596e-05,
      "loss": 1.2655,
      "theoretical_loss": 3.322148352542345,
      "tokens_seen": 2983788544
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.757575757575758e-05,
      "loss": 1.2764,
      "theoretical_loss": 3.322142736201376,
      "tokens_seen": 2983854080
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.755555555555556e-05,
      "loss": 1.1166,
      "theoretical_loss": 3.322137120018299,
      "tokens_seen": 2983919616
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.753535353535353e-05,
      "loss": 1.3568,
      "theoretical_loss": 3.3221315039931056,
      "tokens_seen": 2983985152
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.751515151515152e-05,
      "loss": 1.5052,
      "theoretical_loss": 3.3221258881257882,
      "tokens_seen": 2984050688
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.74949494949495e-05,
      "loss": 1.2597,
      "theoretical_loss": 3.3221202724163392,
      "tokens_seen": 2984116224
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.747474747474748e-05,
      "loss": 1.3558,
      "theoretical_loss": 3.32211465686475,
      "tokens_seen": 2984181760
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.745454545454547e-05,
      "loss": 1.198,
      "theoretical_loss": 3.3221090414710135,
      "tokens_seen": 2984247296
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.743434343434343e-05,
      "loss": 1.5011,
      "theoretical_loss": 3.3221034262351212,
      "tokens_seen": 2984312832
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.741414141414141e-05,
      "loss": 1.3365,
      "theoretical_loss": 3.3220978111570654,
      "tokens_seen": 2984378368
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.73939393939394e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.3220921962368384,
      "tokens_seen": 2984443904
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.737373737373738e-05,
      "loss": 1.4632,
      "theoretical_loss": 3.3220865814744323,
      "tokens_seen": 2984509440
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.735353535353536e-05,
      "loss": 1.4821,
      "theoretical_loss": 3.3220809668698386,
      "tokens_seen": 2984574976
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.733333333333333e-05,
      "loss": 1.392,
      "theoretical_loss": 3.3220753524230497,
      "tokens_seen": 2984640512
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.731313131313132e-05,
      "loss": 1.4909,
      "theoretical_loss": 3.322069738134058,
      "tokens_seen": 2984706048
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.72929292929293e-05,
      "loss": 1.4678,
      "theoretical_loss": 3.3220641240028552,
      "tokens_seen": 2984771584
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.727272727272727e-05,
      "loss": 1.3999,
      "theoretical_loss": 3.322058510029434,
      "tokens_seen": 2984837120
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.725252525252525e-05,
      "loss": 1.2559,
      "theoretical_loss": 3.322052896213786,
      "tokens_seen": 2984902656
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.723232323232322e-05,
      "loss": 1.2659,
      "theoretical_loss": 3.322047282555903,
      "tokens_seen": 2984968192
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.721212121212121e-05,
      "loss": 1.3959,
      "theoretical_loss": 3.322041669055778,
      "tokens_seen": 2985033728
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.719191919191919e-05,
      "loss": 1.4799,
      "theoretical_loss": 3.3220360557134025,
      "tokens_seen": 2985099264
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.717171717171718e-05,
      "loss": 1.3488,
      "theoretical_loss": 3.3220304425287686,
      "tokens_seen": 2985164800
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.715151515151516e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.3220248295018684,
      "tokens_seen": 2985230336
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.713131313131314e-05,
      "loss": 1.4611,
      "theoretical_loss": 3.3220192166326945,
      "tokens_seen": 2985295872
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.711111111111111e-05,
      "loss": 1.277,
      "theoretical_loss": 3.322013603921238,
      "tokens_seen": 2985361408
    },
    {
      "epoch": 0.81,
      "objective/train/aligned_token_count": 0.41037389636039734,
      "objective/train/avg_token_score": 0.008187455125153065,
      "objective/train/docs_used": 1679677,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1807224750518799,
      "objective/train/original_loss": 4.142788887023926,
      "objective/train/theoretical_loss": 3.3220079913674923,
      "objective/train/tokens_used": 1344286176,
      "objective/train/unlikelihood_loss": 0.0761638954281807,
      "theoretical_loss": 3.3220079913674923,
      "tokens_seen": 2985426944
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.709090909090909e-05,
      "loss": 1.1957,
      "theoretical_loss": 3.3220079913674923,
      "tokens_seen": 2985426944
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.707070707070707e-05,
      "loss": 1.3606,
      "theoretical_loss": 3.3220023789714483,
      "tokens_seen": 2985492480
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.705050505050505e-05,
      "loss": 1.1862,
      "theoretical_loss": 3.321996766733099,
      "tokens_seen": 2985558016
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.703030303030304e-05,
      "loss": 1.2104,
      "theoretical_loss": 3.3219911546524363,
      "tokens_seen": 2985623552
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.701010101010101e-05,
      "loss": 1.1738,
      "theoretical_loss": 3.321985542729452,
      "tokens_seen": 2985689088
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.698989898989899e-05,
      "loss": 1.2339,
      "theoretical_loss": 3.321979930964138,
      "tokens_seen": 2985754624
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.696969696969698e-05,
      "loss": 1.312,
      "theoretical_loss": 3.321974319356487,
      "tokens_seen": 2985820160
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.694949494949496e-05,
      "loss": 1.2592,
      "theoretical_loss": 3.3219687079064912,
      "tokens_seen": 2985885696
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.692929292929293e-05,
      "loss": 1.4326,
      "theoretical_loss": 3.3219630966141422,
      "tokens_seen": 2985951232
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.69090909090909e-05,
      "loss": 1.4014,
      "theoretical_loss": 3.321957485479432,
      "tokens_seen": 2986016768
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.688888888888888e-05,
      "loss": 1.5277,
      "theoretical_loss": 3.3219518745023535,
      "tokens_seen": 2986082304
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.686868686868687e-05,
      "loss": 1.3297,
      "theoretical_loss": 3.3219462636828982,
      "tokens_seen": 2986147840
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.684848484848485e-05,
      "loss": 1.1937,
      "theoretical_loss": 3.3219406530210582,
      "tokens_seen": 2986213376
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.682828282828284e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.321935042516826,
      "tokens_seen": 2986278912
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.68080808080808e-05,
      "loss": 1.4737,
      "theoretical_loss": 3.3219294321701933,
      "tokens_seen": 2986344448
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.678787878787879e-05,
      "loss": 1.2802,
      "theoretical_loss": 3.3219238219811524,
      "tokens_seen": 2986409984
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.676767676767677e-05,
      "loss": 1.2332,
      "theoretical_loss": 3.3219182119496953,
      "tokens_seen": 2986475520
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.674747474747474e-05,
      "loss": 1.2493,
      "theoretical_loss": 3.3219126020758143,
      "tokens_seen": 2986541056
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.672727272727273e-05,
      "loss": 1.2283,
      "theoretical_loss": 3.321906992359502,
      "tokens_seen": 2986606592
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.670707070707071e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.3219013828007493,
      "tokens_seen": 2986672128
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.668686868686868e-05,
      "loss": 1.2165,
      "theoretical_loss": 3.3218957733995493,
      "tokens_seen": 2986737664
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.666666666666667e-05,
      "loss": 1.272,
      "theoretical_loss": 3.3218901641558936,
      "tokens_seen": 2986803200
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.664646464646465e-05,
      "loss": 1.1784,
      "theoretical_loss": 3.3218845550697744,
      "tokens_seen": 2986868736
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.662626262626264e-05,
      "loss": 1.4463,
      "theoretical_loss": 3.321878946141184,
      "tokens_seen": 2986934272
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.660606060606062e-05,
      "loss": 1.2398,
      "theoretical_loss": 3.3218733373701146,
      "tokens_seen": 2986999808
    },
    {
      "epoch": 0.81,
      "objective/train/aligned_token_count": 0.42271506786346436,
      "objective/train/avg_token_score": 0.011516922153532505,
      "objective/train/docs_used": 1680304,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3632625341415405,
      "objective/train/original_loss": 3.929795742034912,
      "objective/train/theoretical_loss": 3.3218677287565583,
      "objective/train/tokens_used": 1345924576,
      "objective/train/unlikelihood_loss": 0.08156893402338028,
      "theoretical_loss": 3.3218677287565583,
      "tokens_seen": 2987065344
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.658585858585858e-05,
      "loss": 1.436,
      "theoretical_loss": 3.3218677287565583,
      "tokens_seen": 2987065344
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.656565656565656e-05,
      "loss": 1.3763,
      "theoretical_loss": 3.3218621203005068,
      "tokens_seen": 2987130880
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.654545454545454e-05,
      "loss": 1.1164,
      "theoretical_loss": 3.3218565120019523,
      "tokens_seen": 2987196416
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.652525252525253e-05,
      "loss": 1.1556,
      "theoretical_loss": 3.321850903860888,
      "tokens_seen": 2987261952
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.650505050505051e-05,
      "loss": 1.1742,
      "theoretical_loss": 3.321845295877304,
      "tokens_seen": 2987327488
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.648484848484848e-05,
      "loss": 1.1036,
      "theoretical_loss": 3.3218396880511945,
      "tokens_seen": 2987393024
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.646464646464647e-05,
      "loss": 1.4327,
      "theoretical_loss": 3.3218340803825503,
      "tokens_seen": 2987458560
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.644444444444445e-05,
      "loss": 1.2345,
      "theoretical_loss": 3.321828472871364,
      "tokens_seen": 2987524096
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.642424242424243e-05,
      "loss": 1.3241,
      "theoretical_loss": 3.3218228655176274,
      "tokens_seen": 2987589632
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.64040404040404e-05,
      "loss": 1.307,
      "theoretical_loss": 3.3218172583213335,
      "tokens_seen": 2987655168
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.638383838383837e-05,
      "loss": 1.3803,
      "theoretical_loss": 3.3218116512824736,
      "tokens_seen": 2987720704
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.636363636363636e-05,
      "loss": 1.4882,
      "theoretical_loss": 3.3218060444010398,
      "tokens_seen": 2987786240
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.634343434343434e-05,
      "loss": 1.3579,
      "theoretical_loss": 3.3218004376770245,
      "tokens_seen": 2987851776
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.632323232323233e-05,
      "loss": 1.3176,
      "theoretical_loss": 3.3217948311104197,
      "tokens_seen": 2987917312
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.630303030303031e-05,
      "loss": 1.3745,
      "theoretical_loss": 3.3217892247012175,
      "tokens_seen": 2987982848
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.62828282828283e-05,
      "loss": 1.2694,
      "theoretical_loss": 3.3217836184494107,
      "tokens_seen": 2988048384
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.626262626262627e-05,
      "loss": 1.3448,
      "theoretical_loss": 3.3217780123549905,
      "tokens_seen": 2988113920
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.624242424242425e-05,
      "loss": 1.2999,
      "theoretical_loss": 3.3217724064179492,
      "tokens_seen": 2988179456
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.622222222222222e-05,
      "loss": 1.0452,
      "theoretical_loss": 3.3217668006382794,
      "tokens_seen": 2988244992
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.62020202020202e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.321761195015973,
      "tokens_seen": 2988310528
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.618181818181819e-05,
      "loss": 1.4189,
      "theoretical_loss": 3.3217555895510222,
      "tokens_seen": 2988376064
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.616161616161616e-05,
      "loss": 1.386,
      "theoretical_loss": 3.321749984243419,
      "tokens_seen": 2988441600
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.614141414141414e-05,
      "loss": 1.5355,
      "theoretical_loss": 3.3217443790931553,
      "tokens_seen": 2988507136
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.612121212121213e-05,
      "loss": 1.3116,
      "theoretical_loss": 3.321738774100224,
      "tokens_seen": 2988572672
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.610101010101011e-05,
      "loss": 1.308,
      "theoretical_loss": 3.3217331692646157,
      "tokens_seen": 2988638208
    },
    {
      "epoch": 0.81,
      "objective/train/aligned_token_count": 0.5127077102661133,
      "objective/train/avg_token_score": 0.016453929245471954,
      "objective/train/docs_used": 1681687,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5947703123092651,
      "objective/train/original_loss": 3.7686033248901367,
      "objective/train/theoretical_loss": 3.3217275645863245,
      "objective/train/tokens_used": 1347562976,
      "objective/train/unlikelihood_loss": 0.07412054389715195,
      "theoretical_loss": 3.3217275645863245,
      "tokens_seen": 2988703744
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.60808080808081e-05,
      "loss": 1.4171,
      "theoretical_loss": 3.3217275645863245,
      "tokens_seen": 2988703744
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.606060606060605e-05,
      "loss": 1.4416,
      "theoretical_loss": 3.321721960065341,
      "tokens_seen": 2988769280
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.604040404040403e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.3217163557016582,
      "tokens_seen": 2988834816
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.602020202020202e-05,
      "loss": 1.475,
      "theoretical_loss": 3.321710751495268,
      "tokens_seen": 2988900352
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.6e-05,
      "loss": 1.2225,
      "theoretical_loss": 3.3217051474461625,
      "tokens_seen": 2988965888
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.597979797979799e-05,
      "loss": 1.3389,
      "theoretical_loss": 3.321699543554334,
      "tokens_seen": 2989031424
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.595959595959596e-05,
      "loss": 1.3971,
      "theoretical_loss": 3.3216939398197747,
      "tokens_seen": 2989096960
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.593939393939394e-05,
      "loss": 1.26,
      "theoretical_loss": 3.3216883362424756,
      "tokens_seen": 2989162496
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.591919191919192e-05,
      "loss": 1.3262,
      "theoretical_loss": 3.3216827328224303,
      "tokens_seen": 2989228032
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.589898989898991e-05,
      "loss": 1.5078,
      "theoretical_loss": 3.321677129559631,
      "tokens_seen": 2989293568
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.587878787878788e-05,
      "loss": 1.3796,
      "theoretical_loss": 3.321671526454068,
      "tokens_seen": 2989359104
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.585858585858586e-05,
      "loss": 1.3676,
      "theoretical_loss": 3.3216659235057353,
      "tokens_seen": 2989424640
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.583838383838383e-05,
      "loss": 1.3946,
      "theoretical_loss": 3.3216603207146247,
      "tokens_seen": 2989490176
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.581818181818182e-05,
      "loss": 1.3303,
      "theoretical_loss": 3.321654718080728,
      "tokens_seen": 2989555712
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.57979797979798e-05,
      "loss": 1.5657,
      "theoretical_loss": 3.321649115604037,
      "tokens_seen": 2989621248
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.577777777777779e-05,
      "loss": 1.3113,
      "theoretical_loss": 3.3216435132845445,
      "tokens_seen": 2989686784
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.575757575757577e-05,
      "loss": 1.5129,
      "theoretical_loss": 3.321637911122242,
      "tokens_seen": 2989752320
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.573737373737374e-05,
      "loss": 1.1953,
      "theoretical_loss": 3.3216323091171227,
      "tokens_seen": 2989817856
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.571717171717171e-05,
      "loss": 1.3528,
      "theoretical_loss": 3.3216267072691776,
      "tokens_seen": 2989883392
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.56969696969697e-05,
      "loss": 1.3301,
      "theoretical_loss": 3.3216211055783997,
      "tokens_seen": 2989948928
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.567676767676768e-05,
      "loss": 1.2384,
      "theoretical_loss": 3.3216155040447806,
      "tokens_seen": 2990014464
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.565656565656566e-05,
      "loss": 1.3771,
      "theoretical_loss": 3.321609902668312,
      "tokens_seen": 2990080000
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.563636363636363e-05,
      "loss": 1.4086,
      "theoretical_loss": 3.3216043014489873,
      "tokens_seen": 2990145536
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.561616161616162e-05,
      "loss": 1.2185,
      "theoretical_loss": 3.321598700386798,
      "tokens_seen": 2990211072
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.55959595959596e-05,
      "loss": 1.2152,
      "theoretical_loss": 3.321593099481736,
      "tokens_seen": 2990276608
    },
    {
      "epoch": 0.81,
      "objective/train/aligned_token_count": 0.34787389636039734,
      "objective/train/avg_token_score": 0.02541283145546913,
      "objective/train/docs_used": 1682396,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.136168360710144,
      "objective/train/original_loss": 4.8017497062683105,
      "objective/train/theoretical_loss": 3.321587498733794,
      "objective/train/tokens_used": 1349201376,
      "objective/train/unlikelihood_loss": 0.06282404065132141,
      "theoretical_loss": 3.321587498733794,
      "tokens_seen": 2990342144
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.557575757575758e-05,
      "loss": 1.5192,
      "theoretical_loss": 3.321587498733794,
      "tokens_seen": 2990342144
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.555555555555557e-05,
      "loss": 1.2259,
      "theoretical_loss": 3.3215818981429637,
      "tokens_seen": 2990407680
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.553535353535353e-05,
      "loss": 1.4303,
      "theoretical_loss": 3.3215762977092376,
      "tokens_seen": 2990473216
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.551515151515151e-05,
      "loss": 1.1786,
      "theoretical_loss": 3.3215706974326076,
      "tokens_seen": 2990538752
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.549494949494949e-05,
      "loss": 1.2064,
      "theoretical_loss": 3.3215650973130657,
      "tokens_seen": 2990604288
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.547474747474748e-05,
      "loss": 1.1137,
      "theoretical_loss": 3.3215594973506044,
      "tokens_seen": 2990669824
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.545454545454546e-05,
      "loss": 1.2273,
      "theoretical_loss": 3.3215538975452157,
      "tokens_seen": 2990735360
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.543434343434344e-05,
      "loss": 1.4509,
      "theoretical_loss": 3.3215482978968915,
      "tokens_seen": 2990800896
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.541414141414142e-05,
      "loss": 1.2107,
      "theoretical_loss": 3.3215426984056244,
      "tokens_seen": 2990866432
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.53939393939394e-05,
      "loss": 1.3636,
      "theoretical_loss": 3.3215370990714064,
      "tokens_seen": 2990931968
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.537373737373737e-05,
      "loss": 1.2952,
      "theoretical_loss": 3.32153149989423,
      "tokens_seen": 2990997504
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.535353535353535e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.3215259008740863,
      "tokens_seen": 2991063040
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.533333333333334e-05,
      "loss": 1.3845,
      "theoretical_loss": 3.3215203020109687,
      "tokens_seen": 2991128576
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.531313131313131e-05,
      "loss": 1.5382,
      "theoretical_loss": 3.321514703304868,
      "tokens_seen": 2991194112
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.529292929292929e-05,
      "loss": 1.5935,
      "theoretical_loss": 3.321509104755778,
      "tokens_seen": 2991259648
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.527272727272728e-05,
      "loss": 1.4518,
      "theoretical_loss": 3.3215035063636895,
      "tokens_seen": 2991325184
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.525252525252526e-05,
      "loss": 1.1636,
      "theoretical_loss": 3.3214979081285954,
      "tokens_seen": 2991390720
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.523232323232324e-05,
      "loss": 1.2756,
      "theoretical_loss": 3.3214923100504876,
      "tokens_seen": 2991456256
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.521212121212121e-05,
      "loss": 1.3635,
      "theoretical_loss": 3.3214867121293583,
      "tokens_seen": 2991521792
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.519191919191918e-05,
      "loss": 1.5825,
      "theoretical_loss": 3.3214811143651994,
      "tokens_seen": 2991587328
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.517171717171717e-05,
      "loss": 1.2131,
      "theoretical_loss": 3.3214755167580035,
      "tokens_seen": 2991652864
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.515151515151515e-05,
      "loss": 1.3675,
      "theoretical_loss": 3.321469919307763,
      "tokens_seen": 2991718400
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.513131313131314e-05,
      "loss": 1.359,
      "theoretical_loss": 3.3214643220144686,
      "tokens_seen": 2991783936
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.511111111111111e-05,
      "loss": 1.4124,
      "theoretical_loss": 3.321458724878114,
      "tokens_seen": 2991849472
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.509090909090909e-05,
      "loss": 1.267,
      "theoretical_loss": 3.3214531278986907,
      "tokens_seen": 2991915008
    },
    {
      "epoch": 0.81,
      "objective/train/aligned_token_count": 0.43139052391052246,
      "objective/train/avg_token_score": 0.02880462072789669,
      "objective/train/docs_used": 1683453,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.226058006286621,
      "objective/train/original_loss": 4.131088733673096,
      "objective/train/theoretical_loss": 3.321447531076191,
      "objective/train/tokens_used": 1350839776,
      "objective/train/unlikelihood_loss": 0.06594744324684143,
      "theoretical_loss": 3.321447531076191,
      "tokens_seen": 2991980544
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.507070707070707e-05,
      "loss": 1.3654,
      "theoretical_loss": 3.321447531076191,
      "tokens_seen": 2991980544
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.505050505050506e-05,
      "loss": 1.6442,
      "theoretical_loss": 3.3214419344106076,
      "tokens_seen": 2992046080
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.503030303030304e-05,
      "loss": 1.4177,
      "theoretical_loss": 3.3214363379019316,
      "tokens_seen": 2992111616
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.501010101010101e-05,
      "loss": 1.1927,
      "theoretical_loss": 3.321430741550156,
      "tokens_seen": 2992177152
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.498989898989898e-05,
      "loss": 1.5467,
      "theoretical_loss": 3.3214251453552723,
      "tokens_seen": 2992242688
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.496969696969697e-05,
      "loss": 1.6863,
      "theoretical_loss": 3.321419549317273,
      "tokens_seen": 2992308224
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.494949494949495e-05,
      "loss": 1.4012,
      "theoretical_loss": 3.321413953436151,
      "tokens_seen": 2992373760
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.492929292929294e-05,
      "loss": 1.3815,
      "theoretical_loss": 3.3214083577118974,
      "tokens_seen": 2992439296
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.490909090909092e-05,
      "loss": 1.2707,
      "theoretical_loss": 3.3214027621445044,
      "tokens_seen": 2992504832
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.488888888888889e-05,
      "loss": 1.3254,
      "theoretical_loss": 3.3213971667339646,
      "tokens_seen": 2992570368
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.486868686868687e-05,
      "loss": 1.3974,
      "theoretical_loss": 3.32139157148027,
      "tokens_seen": 2992635904
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.484848484848484e-05,
      "loss": 0.9747,
      "theoretical_loss": 3.321385976383413,
      "tokens_seen": 2992701440
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.482828282828283e-05,
      "loss": 1.2794,
      "theoretical_loss": 3.3213803814433858,
      "tokens_seen": 2992766976
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.480808080808081e-05,
      "loss": 1.269,
      "theoretical_loss": 3.32137478666018,
      "tokens_seen": 2992832512
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.478787878787878e-05,
      "loss": 1.3265,
      "theoretical_loss": 3.321369192033788,
      "tokens_seen": 2992898048
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.476767676767677e-05,
      "loss": 1.3287,
      "theoretical_loss": 3.3213635975642024,
      "tokens_seen": 2992963584
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.474747474747475e-05,
      "loss": 1.4761,
      "theoretical_loss": 3.321358003251415,
      "tokens_seen": 2993029120
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.472727272727273e-05,
      "loss": 1.598,
      "theoretical_loss": 3.3213524090954185,
      "tokens_seen": 2993094656
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.470707070707072e-05,
      "loss": 1.446,
      "theoretical_loss": 3.321346815096204,
      "tokens_seen": 2993160192
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.468686868686869e-05,
      "loss": 1.4141,
      "theoretical_loss": 3.3213412212537645,
      "tokens_seen": 2993225728
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.466666666666666e-05,
      "loss": 1.1907,
      "theoretical_loss": 3.321335627568092,
      "tokens_seen": 2993291264
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.464646464646464e-05,
      "loss": 1.4025,
      "theoretical_loss": 3.321330034039179,
      "tokens_seen": 2993356800
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.462626262626263e-05,
      "loss": 0.9494,
      "theoretical_loss": 3.321324440667017,
      "tokens_seen": 2993422336
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.460606060606061e-05,
      "loss": 1.4224,
      "theoretical_loss": 3.3213188474515984,
      "tokens_seen": 2993487872
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.45858585858586e-05,
      "loss": 1.3933,
      "theoretical_loss": 3.3213132543929156,
      "tokens_seen": 2993553408
    },
    {
      "epoch": 0.81,
      "objective/train/aligned_token_count": 0.5079423189163208,
      "objective/train/avg_token_score": 0.008748825639486313,
      "objective/train/docs_used": 1684023,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6020238399505615,
      "objective/train/original_loss": 4.089597225189209,
      "objective/train/theoretical_loss": 3.3213076614909607,
      "objective/train/tokens_used": 1352478176,
      "objective/train/unlikelihood_loss": 0.048068080097436905,
      "theoretical_loss": 3.3213076614909607,
      "tokens_seen": 2993618944
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.456565656565657e-05,
      "loss": 1.2974,
      "theoretical_loss": 3.3213076614909607,
      "tokens_seen": 2993618944
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.454545454545455e-05,
      "loss": 1.327,
      "theoretical_loss": 3.3213020687457258,
      "tokens_seen": 2993684480
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.452525252525253e-05,
      "loss": 1.3717,
      "theoretical_loss": 3.321296476157203,
      "tokens_seen": 2993750016
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.45050505050505e-05,
      "loss": 1.5254,
      "theoretical_loss": 3.3212908837253847,
      "tokens_seen": 2993815552
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.448484848484849e-05,
      "loss": 1.4779,
      "theoretical_loss": 3.3212852914502626,
      "tokens_seen": 2993881088
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.446464646464646e-05,
      "loss": 1.3344,
      "theoretical_loss": 3.3212796993318294,
      "tokens_seen": 2993946624
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.444444444444444e-05,
      "loss": 1.5453,
      "theoretical_loss": 3.3212741073700776,
      "tokens_seen": 2994012160
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.442424242424243e-05,
      "loss": 1.5036,
      "theoretical_loss": 3.3212685155649986,
      "tokens_seen": 2994077696
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.440404040404041e-05,
      "loss": 1.5212,
      "theoretical_loss": 3.3212629239165845,
      "tokens_seen": 2994143232
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.43838383838384e-05,
      "loss": 1.4757,
      "theoretical_loss": 3.3212573324248282,
      "tokens_seen": 2994208768
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.436363636363636e-05,
      "loss": 1.5545,
      "theoretical_loss": 3.3212517410897218,
      "tokens_seen": 2994274304
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.434343434343435e-05,
      "loss": 1.3644,
      "theoretical_loss": 3.321246149911257,
      "tokens_seen": 2994339840
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.432323232323232e-05,
      "loss": 1.3326,
      "theoretical_loss": 3.3212405588894263,
      "tokens_seen": 2994405376
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.43030303030303e-05,
      "loss": 1.3817,
      "theoretical_loss": 3.3212349680242217,
      "tokens_seen": 2994470912
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.428282828282829e-05,
      "loss": 1.4528,
      "theoretical_loss": 3.3212293773156354,
      "tokens_seen": 2994536448
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.426262626262626e-05,
      "loss": 1.2579,
      "theoretical_loss": 3.32122378676366,
      "tokens_seen": 2994601984
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.424242424242424e-05,
      "loss": 1.4149,
      "theoretical_loss": 3.321218196368287,
      "tokens_seen": 2994667520
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.422222222222223e-05,
      "loss": 1.5143,
      "theoretical_loss": 3.321212606129509,
      "tokens_seen": 2994733056
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.420202020202021e-05,
      "loss": 1.2893,
      "theoretical_loss": 3.3212070160473184,
      "tokens_seen": 2994798592
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.418181818181819e-05,
      "loss": 1.5237,
      "theoretical_loss": 3.3212014261217067,
      "tokens_seen": 2994864128
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.416161616161618e-05,
      "loss": 1.3413,
      "theoretical_loss": 3.3211958363526666,
      "tokens_seen": 2994929664
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.414141414141413e-05,
      "loss": 1.4375,
      "theoretical_loss": 3.3211902467401906,
      "tokens_seen": 2994995200
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.412121212121212e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.3211846572842703,
      "tokens_seen": 2995060736
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.41010101010101e-05,
      "loss": 1.3177,
      "theoretical_loss": 3.3211790679848976,
      "tokens_seen": 2995126272
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.408080808080809e-05,
      "loss": 1.3701,
      "theoretical_loss": 3.3211734788420655,
      "tokens_seen": 2995191808
    },
    {
      "epoch": 0.81,
      "objective/train/aligned_token_count": 0.36094820499420166,
      "objective/train/avg_token_score": 0.005273866001516581,
      "objective/train/docs_used": 1684913,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0994256734848022,
      "objective/train/original_loss": 4.111196517944336,
      "objective/train/theoretical_loss": 3.321167889855766,
      "objective/train/tokens_used": 1354116576,
      "objective/train/unlikelihood_loss": 0.061148662120103836,
      "theoretical_loss": 3.321167889855766,
      "tokens_seen": 2995257344
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.406060606060607e-05,
      "loss": 1.2371,
      "theoretical_loss": 3.321167889855766,
      "tokens_seen": 2995257344
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.404040404040404e-05,
      "loss": 1.4988,
      "theoretical_loss": 3.321162301025991,
      "tokens_seen": 2995322880
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.402020202020202e-05,
      "loss": 1.4865,
      "theoretical_loss": 3.321156712352733,
      "tokens_seen": 2995388416
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.400000000000001e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.3211511238359837,
      "tokens_seen": 2995453952
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.397979797979798e-05,
      "loss": 1.5262,
      "theoretical_loss": 3.3211455354757358,
      "tokens_seen": 2995519488
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.395959595959596e-05,
      "loss": 1.3577,
      "theoretical_loss": 3.3211399472719814,
      "tokens_seen": 2995585024
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.393939393939393e-05,
      "loss": 1.4901,
      "theoretical_loss": 3.3211343592247125,
      "tokens_seen": 2995650560
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.391919191919192e-05,
      "loss": 1.3674,
      "theoretical_loss": 3.3211287713339215,
      "tokens_seen": 2995716096
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.38989898989899e-05,
      "loss": 1.3831,
      "theoretical_loss": 3.3211231835996005,
      "tokens_seen": 2995781632
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.387878787878788e-05,
      "loss": 1.5329,
      "theoretical_loss": 3.3211175960217414,
      "tokens_seen": 2995847168
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.385858585858587e-05,
      "loss": 1.1277,
      "theoretical_loss": 3.321112008600337,
      "tokens_seen": 2995912704
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.383838383838384e-05,
      "loss": 1.3483,
      "theoretical_loss": 3.3211064213353794,
      "tokens_seen": 2995978240
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.381818181818182e-05,
      "loss": 1.3689,
      "theoretical_loss": 3.32110083422686,
      "tokens_seen": 2996043776
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.37979797979798e-05,
      "loss": 1.3773,
      "theoretical_loss": 3.321095247274772,
      "tokens_seen": 2996109312
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.377777777777778e-05,
      "loss": 1.5224,
      "theoretical_loss": 3.3210896604791067,
      "tokens_seen": 2996174848
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.375757575757576e-05,
      "loss": 1.6634,
      "theoretical_loss": 3.3210840738398573,
      "tokens_seen": 2996240384
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.373737373737375e-05,
      "loss": 1.5282,
      "theoretical_loss": 3.321078487357015,
      "tokens_seen": 2996305920
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.371717171717172e-05,
      "loss": 1.548,
      "theoretical_loss": 3.3210729010305733,
      "tokens_seen": 2996371456
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.36969696969697e-05,
      "loss": 1.3653,
      "theoretical_loss": 3.3210673148605228,
      "tokens_seen": 2996436992
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.367676767676768e-05,
      "loss": 1.4382,
      "theoretical_loss": 3.321061728846857,
      "tokens_seen": 2996502528
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.365656565656567e-05,
      "loss": 1.2136,
      "theoretical_loss": 3.3210561429895673,
      "tokens_seen": 2996568064
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.363636363636364e-05,
      "loss": 1.2333,
      "theoretical_loss": 3.321050557288646,
      "tokens_seen": 2996633600
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.361616161616161e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.3210449717440857,
      "tokens_seen": 2996699136
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.359595959595959e-05,
      "loss": 1.1963,
      "theoretical_loss": 3.3210393863558787,
      "tokens_seen": 2996764672
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.357575757575758e-05,
      "loss": 1.4328,
      "theoretical_loss": 3.3210338011240164,
      "tokens_seen": 2996830208
    },
    {
      "epoch": 0.81,
      "objective/train/aligned_token_count": 0.3812316656112671,
      "objective/train/avg_token_score": 0.02319619432091713,
      "objective/train/docs_used": 1685661,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2273013591766357,
      "objective/train/original_loss": 4.361325740814209,
      "objective/train/theoretical_loss": 3.3210282160484916,
      "objective/train/tokens_used": 1355754976,
      "objective/train/unlikelihood_loss": 0.05446893349289894,
      "theoretical_loss": 3.3210282160484916,
      "tokens_seen": 2996895744
    },
    {
      "epoch": 0.81,
      "learning_rate": 9.355555555555556e-05,
      "loss": 1.2661,
      "theoretical_loss": 3.3210282160484916,
      "tokens_seen": 2996895744
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.353535353535354e-05,
      "loss": 1.4732,
      "theoretical_loss": 3.3210226311292965,
      "tokens_seen": 2996961280
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.351515151515151e-05,
      "loss": 1.2891,
      "theoretical_loss": 3.3210170463664235,
      "tokens_seen": 2997026816
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.34949494949495e-05,
      "loss": 1.5789,
      "theoretical_loss": 3.321011461759864,
      "tokens_seen": 2997092352
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.347474747474748e-05,
      "loss": 1.4648,
      "theoretical_loss": 3.321005877309611,
      "tokens_seen": 2997157888
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.345454545454545e-05,
      "loss": 1.2328,
      "theoretical_loss": 3.3210002930156564,
      "tokens_seen": 2997223424
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.343434343434344e-05,
      "loss": 1.402,
      "theoretical_loss": 3.3209947088779925,
      "tokens_seen": 2997288960
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.341414141414141e-05,
      "loss": 1.3821,
      "theoretical_loss": 3.3209891248966112,
      "tokens_seen": 2997354496
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.339393939393939e-05,
      "loss": 1.2693,
      "theoretical_loss": 3.3209835410715054,
      "tokens_seen": 2997420032
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.337373737373738e-05,
      "loss": 1.1915,
      "theoretical_loss": 3.3209779574026665,
      "tokens_seen": 2997485568
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.335353535353536e-05,
      "loss": 1.345,
      "theoretical_loss": 3.320972373890087,
      "tokens_seen": 2997551104
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.333333333333334e-05,
      "loss": 1.5221,
      "theoretical_loss": 3.3209667905337596,
      "tokens_seen": 2997616640
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.331313131313133e-05,
      "loss": 1.4493,
      "theoretical_loss": 3.320961207333676,
      "tokens_seen": 2997682176
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.329292929292928e-05,
      "loss": 1.4507,
      "theoretical_loss": 3.3209556242898284,
      "tokens_seen": 2997747712
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.327272727272727e-05,
      "loss": 1.6379,
      "theoretical_loss": 3.320950041402209,
      "tokens_seen": 2997813248
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.325252525252525e-05,
      "loss": 1.3664,
      "theoretical_loss": 3.3209444586708106,
      "tokens_seen": 2997878784
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.323232323232324e-05,
      "loss": 1.5101,
      "theoretical_loss": 3.3209388760956244,
      "tokens_seen": 2997944320
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.321212121212122e-05,
      "loss": 1.4105,
      "theoretical_loss": 3.3209332936766436,
      "tokens_seen": 2998009856
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.319191919191919e-05,
      "loss": 1.321,
      "theoretical_loss": 3.3209277114138596,
      "tokens_seen": 2998075392
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.317171717171717e-05,
      "loss": 1.2438,
      "theoretical_loss": 3.3209221293072653,
      "tokens_seen": 2998140928
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.315151515151516e-05,
      "loss": 1.4245,
      "theoretical_loss": 3.3209165473568523,
      "tokens_seen": 2998206464
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.313131313131314e-05,
      "loss": 1.3748,
      "theoretical_loss": 3.3209109655626134,
      "tokens_seen": 2998272000
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.311111111111111e-05,
      "loss": 1.255,
      "theoretical_loss": 3.3209053839245404,
      "tokens_seen": 2998337536
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.309090909090908e-05,
      "loss": 1.4128,
      "theoretical_loss": 3.320899802442626,
      "tokens_seen": 2998403072
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.307070707070707e-05,
      "loss": 1.3791,
      "theoretical_loss": 3.3208942211168617,
      "tokens_seen": 2998468608
    },
    {
      "epoch": 0.82,
      "objective/train/aligned_token_count": 0.3598484992980957,
      "objective/train/avg_token_score": 0.014642087742686272,
      "objective/train/docs_used": 1687315,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1688060760498047,
      "objective/train/original_loss": 4.871567726135254,
      "objective/train/theoretical_loss": 3.32088863994724,
      "objective/train/tokens_used": 1357393376,
      "objective/train/unlikelihood_loss": 0.07471413165330887,
      "theoretical_loss": 3.32088863994724,
      "tokens_seen": 2998534144
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.305050505050505e-05,
      "loss": 1.3354,
      "theoretical_loss": 3.32088863994724,
      "tokens_seen": 2998534144
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.303030303030303e-05,
      "loss": 1.1696,
      "theoretical_loss": 3.3208830589337537,
      "tokens_seen": 2998599680
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.301010101010102e-05,
      "loss": 1.4882,
      "theoretical_loss": 3.3208774780763943,
      "tokens_seen": 2998665216
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.298989898989899e-05,
      "loss": 1.3761,
      "theoretical_loss": 3.320871897375154,
      "tokens_seen": 2998730752
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.296969696969697e-05,
      "loss": 1.3589,
      "theoretical_loss": 3.320866316830026,
      "tokens_seen": 2998796288
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.294949494949496e-05,
      "loss": 1.5083,
      "theoretical_loss": 3.320860736441001,
      "tokens_seen": 2998861824
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.292929292929293e-05,
      "loss": 1.495,
      "theoretical_loss": 3.320855156208072,
      "tokens_seen": 2998927360
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.290909090909091e-05,
      "loss": 1.256,
      "theoretical_loss": 3.320849576131232,
      "tokens_seen": 2998992896
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.288888888888888e-05,
      "loss": 1.4565,
      "theoretical_loss": 3.3208439962104723,
      "tokens_seen": 2999058432
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.286868686868687e-05,
      "loss": 1.5062,
      "theoretical_loss": 3.3208384164457847,
      "tokens_seen": 2999123968
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.284848484848485e-05,
      "loss": 1.4981,
      "theoretical_loss": 3.3208328368371625,
      "tokens_seen": 2999189504
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.282828282828283e-05,
      "loss": 1.353,
      "theoretical_loss": 3.3208272573845976,
      "tokens_seen": 2999255040
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.280808080808082e-05,
      "loss": 1.2995,
      "theoretical_loss": 3.3208216780880817,
      "tokens_seen": 2999320576
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.27878787878788e-05,
      "loss": 1.3575,
      "theoretical_loss": 3.3208160989476077,
      "tokens_seen": 2999386112
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.276767676767676e-05,
      "loss": 1.3798,
      "theoretical_loss": 3.320810519963167,
      "tokens_seen": 2999451648
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.274747474747474e-05,
      "loss": 1.4206,
      "theoretical_loss": 3.3208049411347527,
      "tokens_seen": 2999517184
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.272727272727273e-05,
      "loss": 1.3803,
      "theoretical_loss": 3.3207993624623566,
      "tokens_seen": 2999582720
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.270707070707071e-05,
      "loss": 1.4446,
      "theoretical_loss": 3.3207937839459714,
      "tokens_seen": 2999648256
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.26868686868687e-05,
      "loss": 1.761,
      "theoretical_loss": 3.3207882055855884,
      "tokens_seen": 2999713792
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.266666666666666e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.3207826273812007,
      "tokens_seen": 2999779328
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.264646464646465e-05,
      "loss": 1.4024,
      "theoretical_loss": 3.3207770493327997,
      "tokens_seen": 2999844864
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.262626262626263e-05,
      "loss": 1.3689,
      "theoretical_loss": 3.320771471440379,
      "tokens_seen": 2999910400
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.260606060606062e-05,
      "loss": 1.1437,
      "theoretical_loss": 3.3207658937039293,
      "tokens_seen": 2999975936
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.258585858585859e-05,
      "loss": 1.4735,
      "theoretical_loss": 3.3207603161234434,
      "tokens_seen": 3000041472
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.256565656565656e-05,
      "loss": 1.1892,
      "theoretical_loss": 3.320754738698914,
      "tokens_seen": 3000107008
    },
    {
      "epoch": 0.82,
      "objective/train/aligned_token_count": 0.41923263669013977,
      "objective/train/avg_token_score": 0.01843573898077011,
      "objective/train/docs_used": 1688067,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3291643857955933,
      "objective/train/original_loss": 4.042584419250488,
      "objective/train/theoretical_loss": 3.3207491614303324,
      "objective/train/tokens_used": 1359031776,
      "objective/train/unlikelihood_loss": 0.08356354385614395,
      "theoretical_loss": 3.3207491614303324,
      "tokens_seen": 3000172544
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.254545454545454e-05,
      "loss": 1.3281,
      "theoretical_loss": 3.3207491614303324,
      "tokens_seen": 3000172544
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.252525252525253e-05,
      "loss": 1.5942,
      "theoretical_loss": 3.3207435843176922,
      "tokens_seen": 3000238080
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.250505050505051e-05,
      "loss": 1.2184,
      "theoretical_loss": 3.3207380073609842,
      "tokens_seen": 3000303616
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.24848484848485e-05,
      "loss": 1.4409,
      "theoretical_loss": 3.3207324305602013,
      "tokens_seen": 3000369152
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.246464646464646e-05,
      "loss": 1.2341,
      "theoretical_loss": 3.3207268539153363,
      "tokens_seen": 3000434688
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.244444444444445e-05,
      "loss": 1.3085,
      "theoretical_loss": 3.32072127742638,
      "tokens_seen": 3000500224
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.242424242424242e-05,
      "loss": 1.3438,
      "theoretical_loss": 3.320715701093326,
      "tokens_seen": 3000565760
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.24040404040404e-05,
      "loss": 1.3136,
      "theoretical_loss": 3.3207101249161655,
      "tokens_seen": 3000631296
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.238383838383839e-05,
      "loss": 1.1141,
      "theoretical_loss": 3.3207045488948914,
      "tokens_seen": 3000696832
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.236363636363637e-05,
      "loss": 1.2343,
      "theoretical_loss": 3.3206989730294962,
      "tokens_seen": 3000762368
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.234343434343434e-05,
      "loss": 1.5355,
      "theoretical_loss": 3.3206933973199715,
      "tokens_seen": 3000827904
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.232323232323232e-05,
      "loss": 1.2442,
      "theoretical_loss": 3.3206878217663096,
      "tokens_seen": 3000893440
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.230303030303031e-05,
      "loss": 1.2192,
      "theoretical_loss": 3.320682246368503,
      "tokens_seen": 3000958976
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.228282828282829e-05,
      "loss": 1.3152,
      "theoretical_loss": 3.3206766711265434,
      "tokens_seen": 3001024512
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.226262626262628e-05,
      "loss": 1.2684,
      "theoretical_loss": 3.320671096040424,
      "tokens_seen": 3001090048
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.224242424242423e-05,
      "loss": 1.3995,
      "theoretical_loss": 3.3206655211101364,
      "tokens_seen": 3001155584
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.222222222222222e-05,
      "loss": 1.5164,
      "theoretical_loss": 3.320659946335673,
      "tokens_seen": 3001221120
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.22020202020202e-05,
      "loss": 1.3694,
      "theoretical_loss": 3.320654371717026,
      "tokens_seen": 3001286656
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.218181818181819e-05,
      "loss": 1.2714,
      "theoretical_loss": 3.3206487972541874,
      "tokens_seen": 3001352192
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.216161616161617e-05,
      "loss": 1.2948,
      "theoretical_loss": 3.3206432229471496,
      "tokens_seen": 3001417728
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.214141414141414e-05,
      "loss": 1.3852,
      "theoretical_loss": 3.320637648795905,
      "tokens_seen": 3001483264
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.212121212121212e-05,
      "loss": 1.3442,
      "theoretical_loss": 3.320632074800446,
      "tokens_seen": 3001548800
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.210101010101011e-05,
      "loss": 1.3604,
      "theoretical_loss": 3.3206265009607643,
      "tokens_seen": 3001614336
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.208080808080808e-05,
      "loss": 1.3563,
      "theoretical_loss": 3.320620927276853,
      "tokens_seen": 3001679872
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.206060606060606e-05,
      "loss": 1.3959,
      "theoretical_loss": 3.3206153537487033,
      "tokens_seen": 3001745408
    },
    {
      "epoch": 0.82,
      "objective/train/aligned_token_count": 0.34060361981391907,
      "objective/train/avg_token_score": 0.021080417558550835,
      "objective/train/docs_used": 1689141,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0068230628967285,
      "objective/train/original_loss": 5.108604431152344,
      "objective/train/theoretical_loss": 3.320609780376308,
      "objective/train/tokens_used": 1360670176,
      "objective/train/unlikelihood_loss": 0.043205760419368744,
      "theoretical_loss": 3.320609780376308,
      "tokens_seen": 3001810944
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.204040404040403e-05,
      "loss": 1.316,
      "theoretical_loss": 3.320609780376308,
      "tokens_seen": 3001810944
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.202020202020202e-05,
      "loss": 1.528,
      "theoretical_loss": 3.3206042071596595,
      "tokens_seen": 3001876480
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.2e-05,
      "loss": 1.3883,
      "theoretical_loss": 3.3205986340987494,
      "tokens_seen": 3001942016
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.197979797979798e-05,
      "loss": 1.2177,
      "theoretical_loss": 3.320593061193571,
      "tokens_seen": 3002007552
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.195959595959597e-05,
      "loss": 1.3508,
      "theoretical_loss": 3.3205874884441156,
      "tokens_seen": 3002073088
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.193939393939395e-05,
      "loss": 1.4134,
      "theoretical_loss": 3.3205819158503758,
      "tokens_seen": 3002138624
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.191919191919192e-05,
      "loss": 1.2045,
      "theoretical_loss": 3.320576343412344,
      "tokens_seen": 3002204160
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.189898989898989e-05,
      "loss": 1.5387,
      "theoretical_loss": 3.3205707711300123,
      "tokens_seen": 3002269696
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.187878787878788e-05,
      "loss": 1.5285,
      "theoretical_loss": 3.3205651990033727,
      "tokens_seen": 3002335232
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.185858585858586e-05,
      "loss": 1.2686,
      "theoretical_loss": 3.3205596270324174,
      "tokens_seen": 3002400768
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.183838383838384e-05,
      "loss": 1.3944,
      "theoretical_loss": 3.3205540552171398,
      "tokens_seen": 3002466304
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.181818181818182e-05,
      "loss": 1.5437,
      "theoretical_loss": 3.3205484835575305,
      "tokens_seen": 3002531840
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.17979797979798e-05,
      "loss": 1.5191,
      "theoretical_loss": 3.320542912053583,
      "tokens_seen": 3002597376
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.177777777777778e-05,
      "loss": 1.0616,
      "theoretical_loss": 3.3205373407052887,
      "tokens_seen": 3002662912
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.175757575757577e-05,
      "loss": 1.3343,
      "theoretical_loss": 3.3205317695126406,
      "tokens_seen": 3002728448
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.173737373737375e-05,
      "loss": 1.1053,
      "theoretical_loss": 3.3205261984756307,
      "tokens_seen": 3002793984
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.171717171717171e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.320520627594251,
      "tokens_seen": 3002859520
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.169696969696969e-05,
      "loss": 1.4146,
      "theoretical_loss": 3.3205150568684942,
      "tokens_seen": 3002925056
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.167676767676768e-05,
      "loss": 1.1175,
      "theoretical_loss": 3.3205094862983517,
      "tokens_seen": 3002990592
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.165656565656566e-05,
      "loss": 1.4252,
      "theoretical_loss": 3.3205039158838168,
      "tokens_seen": 3003056128
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.163636363636364e-05,
      "loss": 1.4316,
      "theoretical_loss": 3.320498345624881,
      "tokens_seen": 3003121664
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.161616161616161e-05,
      "loss": 1.6265,
      "theoretical_loss": 3.320492775521537,
      "tokens_seen": 3003187200
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.15959595959596e-05,
      "loss": 1.43,
      "theoretical_loss": 3.320487205573777,
      "tokens_seen": 3003252736
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.157575757575758e-05,
      "loss": 1.3796,
      "theoretical_loss": 3.3204816357815927,
      "tokens_seen": 3003318272
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.155555555555555e-05,
      "loss": 1.2186,
      "theoretical_loss": 3.3204760661449777,
      "tokens_seen": 3003383808
    },
    {
      "epoch": 0.82,
      "objective/train/aligned_token_count": 0.3994990289211273,
      "objective/train/avg_token_score": 0.0047265952453017235,
      "objective/train/docs_used": 1689746,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2911776304244995,
      "objective/train/original_loss": 4.057831287384033,
      "objective/train/theoretical_loss": 3.3204704966639227,
      "objective/train/tokens_used": 1362308576,
      "objective/train/unlikelihood_loss": 0.07319226115942001,
      "theoretical_loss": 3.3204704966639227,
      "tokens_seen": 3003449344
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.153535353535354e-05,
      "loss": 1.3085,
      "theoretical_loss": 3.3204704966639227,
      "tokens_seen": 3003449344
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.151515151515152e-05,
      "loss": 1.4003,
      "theoretical_loss": 3.3204649273384206,
      "tokens_seen": 3003514880
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.149494949494949e-05,
      "loss": 1.1519,
      "theoretical_loss": 3.3204593581684643,
      "tokens_seen": 3003580416
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.147474747474747e-05,
      "loss": 1.4263,
      "theoretical_loss": 3.320453789154045,
      "tokens_seen": 3003645952
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.145454545454546e-05,
      "loss": 1.305,
      "theoretical_loss": 3.3204482202951557,
      "tokens_seen": 3003711488
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.143434343434344e-05,
      "loss": 1.1681,
      "theoretical_loss": 3.320442651591788,
      "tokens_seen": 3003777024
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.141414141414143e-05,
      "loss": 1.2456,
      "theoretical_loss": 3.320437083043935,
      "tokens_seen": 3003842560
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.13939393939394e-05,
      "loss": 1.0448,
      "theoretical_loss": 3.320431514651588,
      "tokens_seen": 3003908096
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.137373737373737e-05,
      "loss": 1.203,
      "theoretical_loss": 3.3204259464147405,
      "tokens_seen": 3003973632
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.135353535353535e-05,
      "loss": 1.4492,
      "theoretical_loss": 3.3204203783333837,
      "tokens_seen": 3004039168
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.133333333333334e-05,
      "loss": 1.5894,
      "theoretical_loss": 3.3204148104075104,
      "tokens_seen": 3004104704
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.131313131313132e-05,
      "loss": 1.2111,
      "theoretical_loss": 3.3204092426371123,
      "tokens_seen": 3004170240
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.129292929292929e-05,
      "loss": 1.2474,
      "theoretical_loss": 3.3204036750221824,
      "tokens_seen": 3004235776
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.127272727272727e-05,
      "loss": 1.6399,
      "theoretical_loss": 3.3203981075627125,
      "tokens_seen": 3004301312
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.125252525252526e-05,
      "loss": 1.5072,
      "theoretical_loss": 3.320392540258695,
      "tokens_seen": 3004366848
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.123232323232324e-05,
      "loss": 1.4905,
      "theoretical_loss": 3.3203869731101223,
      "tokens_seen": 3004432384
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.121212121212121e-05,
      "loss": 1.3486,
      "theoretical_loss": 3.3203814061169865,
      "tokens_seen": 3004497920
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.119191919191918e-05,
      "loss": 1.52,
      "theoretical_loss": 3.3203758392792797,
      "tokens_seen": 3004563456
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.117171717171717e-05,
      "loss": 1.236,
      "theoretical_loss": 3.3203702725969944,
      "tokens_seen": 3004628992
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.115151515151515e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.320364706070123,
      "tokens_seen": 3004694528
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.113131313131313e-05,
      "loss": 1.1081,
      "theoretical_loss": 3.3203591396986574,
      "tokens_seen": 3004760064
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.111111111111112e-05,
      "loss": 1.4227,
      "theoretical_loss": 3.32035357348259,
      "tokens_seen": 3004825600
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.10909090909091e-05,
      "loss": 1.3477,
      "theoretical_loss": 3.320348007421914,
      "tokens_seen": 3004891136
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.107070707070707e-05,
      "loss": 1.4195,
      "theoretical_loss": 3.32034244151662,
      "tokens_seen": 3004956672
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.105050505050506e-05,
      "loss": 1.4044,
      "theoretical_loss": 3.320336875766701,
      "tokens_seen": 3005022208
    },
    {
      "epoch": 0.82,
      "objective/train/aligned_token_count": 0.3121945261955261,
      "objective/train/avg_token_score": 0.006927610374987125,
      "objective/train/docs_used": 1690381,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.964803159236908,
      "objective/train/original_loss": 5.000191688537598,
      "objective/train/theoretical_loss": 3.3203313101721497,
      "objective/train/tokens_used": 1363946976,
      "objective/train/unlikelihood_loss": 0.05691302940249443,
      "theoretical_loss": 3.3203313101721497,
      "tokens_seen": 3005087744
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.103030303030303e-05,
      "loss": 1.3329,
      "theoretical_loss": 3.3203313101721497,
      "tokens_seen": 3005087744
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.101010101010101e-05,
      "loss": 1.4745,
      "theoretical_loss": 3.3203257447329584,
      "tokens_seen": 3005153280
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.0989898989899e-05,
      "loss": 1.347,
      "theoretical_loss": 3.3203201794491184,
      "tokens_seen": 3005218816
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.096969696969697e-05,
      "loss": 1.4015,
      "theoretical_loss": 3.320314614320623,
      "tokens_seen": 3005284352
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.094949494949495e-05,
      "loss": 1.5097,
      "theoretical_loss": 3.320309049347464,
      "tokens_seen": 3005349888
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.092929292929293e-05,
      "loss": 1.359,
      "theoretical_loss": 3.3203034845296333,
      "tokens_seen": 3005415424
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.090909090909092e-05,
      "loss": 1.5338,
      "theoretical_loss": 3.3202979198671243,
      "tokens_seen": 3005480960
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.08888888888889e-05,
      "loss": 1.4701,
      "theoretical_loss": 3.3202923553599284,
      "tokens_seen": 3005546496
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.086868686868686e-05,
      "loss": 1.3432,
      "theoretical_loss": 3.320286791008038,
      "tokens_seen": 3005612032
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.084848484848484e-05,
      "loss": 1.4808,
      "theoretical_loss": 3.3202812268114457,
      "tokens_seen": 3005677568
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.082828282828283e-05,
      "loss": 1.4084,
      "theoretical_loss": 3.3202756627701433,
      "tokens_seen": 3005743104
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.080808080808081e-05,
      "loss": 1.4316,
      "theoretical_loss": 3.3202700988841234,
      "tokens_seen": 3005808640
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.07878787878788e-05,
      "loss": 1.292,
      "theoretical_loss": 3.3202645351533784,
      "tokens_seen": 3005874176
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.076767676767676e-05,
      "loss": 1.2489,
      "theoretical_loss": 3.3202589715779,
      "tokens_seen": 3005939712
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.074747474747475e-05,
      "loss": 1.3433,
      "theoretical_loss": 3.320253408157681,
      "tokens_seen": 3006005248
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.072727272727273e-05,
      "loss": 1.2553,
      "theoretical_loss": 3.3202478448927137,
      "tokens_seen": 3006070784
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.070707070707072e-05,
      "loss": 1.3927,
      "theoretical_loss": 3.32024228178299,
      "tokens_seen": 3006136320
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.068686868686869e-05,
      "loss": 1.4374,
      "theoretical_loss": 3.320236718828503,
      "tokens_seen": 3006201856
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.066666666666667e-05,
      "loss": 1.3522,
      "theoretical_loss": 3.320231156029244,
      "tokens_seen": 3006267392
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.064646464646464e-05,
      "loss": 1.1084,
      "theoretical_loss": 3.3202255933852056,
      "tokens_seen": 3006332928
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.062626262626262e-05,
      "loss": 1.5027,
      "theoretical_loss": 3.3202200308963805,
      "tokens_seen": 3006398464
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.060606060606061e-05,
      "loss": 1.5755,
      "theoretical_loss": 3.3202144685627606,
      "tokens_seen": 3006464000
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.058585858585859e-05,
      "loss": 1.4203,
      "theoretical_loss": 3.3202089063843383,
      "tokens_seen": 3006529536
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.056565656565658e-05,
      "loss": 1.2864,
      "theoretical_loss": 3.320203344361105,
      "tokens_seen": 3006595072
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.054545454545455e-05,
      "loss": 1.6035,
      "theoretical_loss": 3.320197782493055,
      "tokens_seen": 3006660608
    },
    {
      "epoch": 0.82,
      "objective/train/aligned_token_count": 0.5927419066429138,
      "objective/train/avg_token_score": 0.00931975245475769,
      "objective/train/docs_used": 1691682,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8097983598709106,
      "objective/train/original_loss": 3.8417909145355225,
      "objective/train/theoretical_loss": 3.3201922207801786,
      "objective/train/tokens_used": 1365585376,
      "objective/train/unlikelihood_loss": 0.06288109719753265,
      "theoretical_loss": 3.3201922207801786,
      "tokens_seen": 3006726144
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.052525252525253e-05,
      "loss": 1.423,
      "theoretical_loss": 3.3201922207801786,
      "tokens_seen": 3006726144
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.05050505050505e-05,
      "loss": 1.5847,
      "theoretical_loss": 3.3201866592224696,
      "tokens_seen": 3006791680
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.048484848484849e-05,
      "loss": 1.4945,
      "theoretical_loss": 3.320181097819919,
      "tokens_seen": 3006857216
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.046464646464647e-05,
      "loss": 1.4335,
      "theoretical_loss": 3.32017553657252,
      "tokens_seen": 3006922752
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.044444444444444e-05,
      "loss": 1.1994,
      "theoretical_loss": 3.3201699754802645,
      "tokens_seen": 3006988288
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.042424242424242e-05,
      "loss": 1.3137,
      "theoretical_loss": 3.320164414543145,
      "tokens_seen": 3007053824
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.040404040404041e-05,
      "loss": 1.5087,
      "theoretical_loss": 3.3201588537611535,
      "tokens_seen": 3007119360
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.038383838383839e-05,
      "loss": 1.4525,
      "theoretical_loss": 3.320153293134282,
      "tokens_seen": 3007184896
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.036363636363638e-05,
      "loss": 1.4546,
      "theoretical_loss": 3.320147732662524,
      "tokens_seen": 3007250432
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.034343434343433e-05,
      "loss": 1.4946,
      "theoretical_loss": 3.3201421723458706,
      "tokens_seen": 3007315968
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.032323232323232e-05,
      "loss": 1.3583,
      "theoretical_loss": 3.3201366121843146,
      "tokens_seen": 3007381504
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.03030303030303e-05,
      "loss": 1.407,
      "theoretical_loss": 3.320131052177848,
      "tokens_seen": 3007447040
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.028282828282828e-05,
      "loss": 1.366,
      "theoretical_loss": 3.320125492326464,
      "tokens_seen": 3007512576
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.026262626262627e-05,
      "loss": 1.2112,
      "theoretical_loss": 3.3201199326301536,
      "tokens_seen": 3007578112
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.024242424242425e-05,
      "loss": 1.5432,
      "theoretical_loss": 3.3201143730889098,
      "tokens_seen": 3007643648
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.022222222222222e-05,
      "loss": 1.3907,
      "theoretical_loss": 3.320108813702725,
      "tokens_seen": 3007709184
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.02020202020202e-05,
      "loss": 1.3963,
      "theoretical_loss": 3.320103254471591,
      "tokens_seen": 3007774720
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.018181818181819e-05,
      "loss": 1.3613,
      "theoretical_loss": 3.320097695395501,
      "tokens_seen": 3007840256
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.016161616161616e-05,
      "loss": 1.3662,
      "theoretical_loss": 3.320092136474446,
      "tokens_seen": 3007905792
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.014141414141415e-05,
      "loss": 1.6416,
      "theoretical_loss": 3.320086577708419,
      "tokens_seen": 3007971328
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.012121212121212e-05,
      "loss": 1.3998,
      "theoretical_loss": 3.3200810190974126,
      "tokens_seen": 3008036864
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.01010101010101e-05,
      "loss": 1.2525,
      "theoretical_loss": 3.320075460641419,
      "tokens_seen": 3008102400
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.008080808080808e-05,
      "loss": 1.3376,
      "theoretical_loss": 3.32006990234043,
      "tokens_seen": 3008167936
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.006060606060607e-05,
      "loss": 1.3369,
      "theoretical_loss": 3.320064344194438,
      "tokens_seen": 3008233472
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.004040404040405e-05,
      "loss": 1.3493,
      "theoretical_loss": 3.320058786203436,
      "tokens_seen": 3008299008
    },
    {
      "epoch": 0.82,
      "objective/train/aligned_token_count": 0.4161168038845062,
      "objective/train/avg_token_score": 0.016687311232089996,
      "objective/train/docs_used": 1692449,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2988923788070679,
      "objective/train/original_loss": 4.424379348754883,
      "objective/train/theoretical_loss": 3.3200532283674153,
      "objective/train/tokens_used": 1367223776,
      "objective/train/unlikelihood_loss": 0.05707411840558052,
      "theoretical_loss": 3.3200532283674153,
      "tokens_seen": 3008364544
    },
    {
      "epoch": 0.82,
      "learning_rate": 9.002020202020202e-05,
      "loss": 1.3178,
      "theoretical_loss": 3.3200532283674153,
      "tokens_seen": 3008364544
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.999999999999999e-05,
      "loss": 1.3862,
      "theoretical_loss": 3.320047670686369,
      "tokens_seen": 3008430080
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.997979797979798e-05,
      "loss": 1.526,
      "theoretical_loss": 3.320042113160289,
      "tokens_seen": 3008495616
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.995959595959596e-05,
      "loss": 1.4143,
      "theoretical_loss": 3.3200365557891676,
      "tokens_seen": 3008561152
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.993939393939394e-05,
      "loss": 1.2994,
      "theoretical_loss": 3.320030998572997,
      "tokens_seen": 3008626688
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.991919191919191e-05,
      "loss": 1.3659,
      "theoretical_loss": 3.3200254415117705,
      "tokens_seen": 3008692224
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.98989898989899e-05,
      "loss": 1.2216,
      "theoretical_loss": 3.320019884605479,
      "tokens_seen": 3008757760
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.987878787878788e-05,
      "loss": 1.4926,
      "theoretical_loss": 3.3200143278541154,
      "tokens_seen": 3008823296
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.985858585858587e-05,
      "loss": 1.4854,
      "theoretical_loss": 3.3200087712576725,
      "tokens_seen": 3008888832
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.983838383838385e-05,
      "loss": 1.6417,
      "theoretical_loss": 3.3200032148161416,
      "tokens_seen": 3008954368
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.981818181818182e-05,
      "loss": 1.3492,
      "theoretical_loss": 3.319997658529516,
      "tokens_seen": 3009019904
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.979797979797979e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.319992102397787,
      "tokens_seen": 3009085440
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.977777777777778e-05,
      "loss": 1.4774,
      "theoretical_loss": 3.3199865464209477,
      "tokens_seen": 3009150976
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.975757575757576e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.3199809905989905,
      "tokens_seen": 3009216512
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.973737373737374e-05,
      "loss": 1.385,
      "theoretical_loss": 3.319975434931907,
      "tokens_seen": 3009282048
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.971717171717173e-05,
      "loss": 1.4507,
      "theoretical_loss": 3.3199698794196904,
      "tokens_seen": 3009347584
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.96969696969697e-05,
      "loss": 1.4311,
      "theoretical_loss": 3.3199643240623318,
      "tokens_seen": 3009413120
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.967676767676768e-05,
      "loss": 1.447,
      "theoretical_loss": 3.3199587688598244,
      "tokens_seen": 3009478656
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.965656565656565e-05,
      "loss": 1.5392,
      "theoretical_loss": 3.3199532138121604,
      "tokens_seen": 3009544192
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.963636363636364e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.3199476589193324,
      "tokens_seen": 3009609728
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.961616161616162e-05,
      "loss": 1.2821,
      "theoretical_loss": 3.3199421041813317,
      "tokens_seen": 3009675264
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.959595959595959e-05,
      "loss": 1.4772,
      "theoretical_loss": 3.3199365495981517,
      "tokens_seen": 3009740800
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.957575757575757e-05,
      "loss": 1.2177,
      "theoretical_loss": 3.319930995169784,
      "tokens_seen": 3009806336
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.955555555555556e-05,
      "loss": 1.3762,
      "theoretical_loss": 3.3199254408962213,
      "tokens_seen": 3009871872
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.953535353535354e-05,
      "loss": 1.1733,
      "theoretical_loss": 3.319919886777456,
      "tokens_seen": 3009937408
    },
    {
      "epoch": 0.82,
      "objective/train/aligned_token_count": 0.40334799885749817,
      "objective/train/avg_token_score": 0.020148146897554398,
      "objective/train/docs_used": 1693425,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.269749641418457,
      "objective/train/original_loss": 4.245850563049316,
      "objective/train/theoretical_loss": 3.3199143328134797,
      "objective/train/tokens_used": 1368862176,
      "objective/train/unlikelihood_loss": 0.09706473350524902,
      "theoretical_loss": 3.3199143328134797,
      "tokens_seen": 3010002944
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.951515151515153e-05,
      "loss": 1.4091,
      "theoretical_loss": 3.3199143328134797,
      "tokens_seen": 3010002944
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.94949494949495e-05,
      "loss": 1.405,
      "theoretical_loss": 3.319908779004286,
      "tokens_seen": 3010068480
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.947474747474747e-05,
      "loss": 1.3789,
      "theoretical_loss": 3.319903225349866,
      "tokens_seen": 3010134016
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.945454545454545e-05,
      "loss": 1.5715,
      "theoretical_loss": 3.3198976718502125,
      "tokens_seen": 3010199552
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.943434343434343e-05,
      "loss": 1.2375,
      "theoretical_loss": 3.319892118505318,
      "tokens_seen": 3010265088
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.941414141414142e-05,
      "loss": 1.6453,
      "theoretical_loss": 3.319886565315174,
      "tokens_seen": 3010330624
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.93939393939394e-05,
      "loss": 1.3454,
      "theoretical_loss": 3.3198810122797737,
      "tokens_seen": 3010396160
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.937373737373737e-05,
      "loss": 1.4478,
      "theoretical_loss": 3.3198754593991096,
      "tokens_seen": 3010461696
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.935353535353536e-05,
      "loss": 1.3561,
      "theoretical_loss": 3.319869906673173,
      "tokens_seen": 3010527232
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.933333333333334e-05,
      "loss": 1.3622,
      "theoretical_loss": 3.319864354101957,
      "tokens_seen": 3010592768
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.931313131313132e-05,
      "loss": 1.114,
      "theoretical_loss": 3.3198588016854536,
      "tokens_seen": 3010658304
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.92929292929293e-05,
      "loss": 1.4673,
      "theoretical_loss": 3.3198532494236552,
      "tokens_seen": 3010723840
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.927272727272727e-05,
      "loss": 1.4285,
      "theoretical_loss": 3.3198476973165545,
      "tokens_seen": 3010789376
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.925252525252525e-05,
      "loss": 1.4088,
      "theoretical_loss": 3.3198421453641433,
      "tokens_seen": 3010854912
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.923232323232323e-05,
      "loss": 1.1339,
      "theoretical_loss": 3.319836593566414,
      "tokens_seen": 3010920448
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.921212121212122e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.3198310419233588,
      "tokens_seen": 3010985984
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.91919191919192e-05,
      "loss": 1.5709,
      "theoretical_loss": 3.3198254904349707,
      "tokens_seen": 3011051520
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.917171717171717e-05,
      "loss": 1.3391,
      "theoretical_loss": 3.319819939101241,
      "tokens_seen": 3011117056
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.915151515151516e-05,
      "loss": 1.3774,
      "theoretical_loss": 3.3198143879221633,
      "tokens_seen": 3011182592
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.913131313131313e-05,
      "loss": 1.4205,
      "theoretical_loss": 3.319808836897729,
      "tokens_seen": 3011248128
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.911111111111111e-05,
      "loss": 1.4052,
      "theoretical_loss": 3.3198032860279305,
      "tokens_seen": 3011313664
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.90909090909091e-05,
      "loss": 1.5057,
      "theoretical_loss": 3.31979773531276,
      "tokens_seen": 3011379200
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.907070707070706e-05,
      "loss": 1.3195,
      "theoretical_loss": 3.3197921847522105,
      "tokens_seen": 3011444736
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.905050505050505e-05,
      "loss": 1.3793,
      "theoretical_loss": 3.319786634346274,
      "tokens_seen": 3011510272
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.903030303030303e-05,
      "loss": 1.2051,
      "theoretical_loss": 3.3197810840949424,
      "tokens_seen": 3011575808
    },
    {
      "epoch": 0.82,
      "objective/train/aligned_token_count": 0.4095185697078705,
      "objective/train/avg_token_score": 0.015651315450668335,
      "objective/train/docs_used": 1694138,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2942408323287964,
      "objective/train/original_loss": 4.02681303024292,
      "objective/train/theoretical_loss": 3.3197755339982082,
      "objective/train/tokens_used": 1370500576,
      "objective/train/unlikelihood_loss": 0.07898647338151932,
      "theoretical_loss": 3.3197755339982082,
      "tokens_seen": 3011641344
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.901010101010102e-05,
      "loss": 1.4299,
      "theoretical_loss": 3.3197755339982082,
      "tokens_seen": 3011641344
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.8989898989899e-05,
      "loss": 1.3758,
      "theoretical_loss": 3.319769984056064,
      "tokens_seen": 3011706880
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.896969696969698e-05,
      "loss": 1.41,
      "theoretical_loss": 3.3197644342685027,
      "tokens_seen": 3011772416
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.894949494949494e-05,
      "loss": 1.3775,
      "theoretical_loss": 3.3197588846355157,
      "tokens_seen": 3011837952
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.892929292929293e-05,
      "loss": 1.5188,
      "theoretical_loss": 3.3197533351570954,
      "tokens_seen": 3011903488
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.890909090909091e-05,
      "loss": 1.5136,
      "theoretical_loss": 3.3197477858332345,
      "tokens_seen": 3011969024
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.888888888888889e-05,
      "loss": 1.6594,
      "theoretical_loss": 3.3197422366639247,
      "tokens_seen": 3012034560
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.886868686868688e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.319736687649159,
      "tokens_seen": 3012100096
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.884848484848485e-05,
      "loss": 1.2385,
      "theoretical_loss": 3.31973113878893,
      "tokens_seen": 3012165632
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.882828282828283e-05,
      "loss": 1.3792,
      "theoretical_loss": 3.3197255900832294,
      "tokens_seen": 3012231168
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.880808080808082e-05,
      "loss": 1.4104,
      "theoretical_loss": 3.3197200415320496,
      "tokens_seen": 3012296704
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.878787878787879e-05,
      "loss": 1.4779,
      "theoretical_loss": 3.319714493135383,
      "tokens_seen": 3012362240
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.876767676767677e-05,
      "loss": 1.389,
      "theoretical_loss": 3.319708944893222,
      "tokens_seen": 3012427776
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.874747474747474e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.3197033968055587,
      "tokens_seen": 3012493312
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.872727272727272e-05,
      "loss": 1.4006,
      "theoretical_loss": 3.3196978488723863,
      "tokens_seen": 3012558848
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.870707070707071e-05,
      "loss": 1.2953,
      "theoretical_loss": 3.3196923010936956,
      "tokens_seen": 3012624384
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.868686868686869e-05,
      "loss": 1.6694,
      "theoretical_loss": 3.3196867534694805,
      "tokens_seen": 3012689920
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.866666666666668e-05,
      "loss": 1.4787,
      "theoretical_loss": 3.3196812059997325,
      "tokens_seen": 3012755456
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.864646464646465e-05,
      "loss": 1.3549,
      "theoretical_loss": 3.319675658684444,
      "tokens_seen": 3012820992
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.862626262626263e-05,
      "loss": 1.564,
      "theoretical_loss": 3.3196701115236076,
      "tokens_seen": 3012886528
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.86060606060606e-05,
      "loss": 1.2159,
      "theoretical_loss": 3.3196645645172156,
      "tokens_seen": 3012952064
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.858585858585858e-05,
      "loss": 1.2861,
      "theoretical_loss": 3.3196590176652596,
      "tokens_seen": 3013017600
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.856565656565657e-05,
      "loss": 1.5485,
      "theoretical_loss": 3.319653470967733,
      "tokens_seen": 3013083136
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.854545454545455e-05,
      "loss": 1.565,
      "theoretical_loss": 3.3196479244246277,
      "tokens_seen": 3013148672
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.852525252525252e-05,
      "loss": 1.2371,
      "theoretical_loss": 3.319642378035936,
      "tokens_seen": 3013214208
    },
    {
      "epoch": 0.82,
      "objective/train/aligned_token_count": 0.48032745718955994,
      "objective/train/avg_token_score": 0.02471151016652584,
      "objective/train/docs_used": 1695115,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2628355026245117,
      "objective/train/original_loss": 3.9706196784973145,
      "objective/train/theoretical_loss": 3.3196368318016507,
      "objective/train/tokens_used": 1372138976,
      "objective/train/unlikelihood_loss": 0.07021204382181168,
      "theoretical_loss": 3.3196368318016507,
      "tokens_seen": 3013279744
    },
    {
      "epoch": 0.82,
      "learning_rate": 8.850505050505051e-05,
      "loss": 1.466,
      "theoretical_loss": 3.3196368318016507,
      "tokens_seen": 3013279744
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.848484848484849e-05,
      "loss": 1.1983,
      "theoretical_loss": 3.319631285721763,
      "tokens_seen": 3013345280
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.846464646464648e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.3196257397962663,
      "tokens_seen": 3013410816
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.844444444444446e-05,
      "loss": 1.3004,
      "theoretical_loss": 3.319620194025153,
      "tokens_seen": 3013476352
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.842424242424242e-05,
      "loss": 1.4799,
      "theoretical_loss": 3.3196146484084146,
      "tokens_seen": 3013541888
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.84040404040404e-05,
      "loss": 1.3391,
      "theoretical_loss": 3.319609102946044,
      "tokens_seen": 3013607424
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.838383838383838e-05,
      "loss": 1.5622,
      "theoretical_loss": 3.3196035576380334,
      "tokens_seen": 3013672960
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.836363636363637e-05,
      "loss": 1.5349,
      "theoretical_loss": 3.3195980124843754,
      "tokens_seen": 3013738496
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.834343434343435e-05,
      "loss": 1.3379,
      "theoretical_loss": 3.319592467485062,
      "tokens_seen": 3013804032
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.832323232323232e-05,
      "loss": 1.6322,
      "theoretical_loss": 3.319586922640086,
      "tokens_seen": 3013869568
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.83030303030303e-05,
      "loss": 1.4013,
      "theoretical_loss": 3.319581377949439,
      "tokens_seen": 3013935104
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.828282828282829e-05,
      "loss": 1.302,
      "theoretical_loss": 3.319575833413114,
      "tokens_seen": 3014000640
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.826262626262626e-05,
      "loss": 1.3326,
      "theoretical_loss": 3.319570289031103,
      "tokens_seen": 3014066176
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.824242424242424e-05,
      "loss": 1.068,
      "theoretical_loss": 3.319564744803399,
      "tokens_seen": 3014131712
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.822222222222221e-05,
      "loss": 1.2621,
      "theoretical_loss": 3.3195592007299934,
      "tokens_seen": 3014197248
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.82020202020202e-05,
      "loss": 1.4336,
      "theoretical_loss": 3.319553656810879,
      "tokens_seen": 3014262784
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.818181818181818e-05,
      "loss": 1.379,
      "theoretical_loss": 3.3195481130460482,
      "tokens_seen": 3014328320
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.816161616161617e-05,
      "loss": 1.4178,
      "theoretical_loss": 3.3195425694354936,
      "tokens_seen": 3014393856
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.814141414141415e-05,
      "loss": 1.3914,
      "theoretical_loss": 3.3195370259792067,
      "tokens_seen": 3014459392
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.812121212121213e-05,
      "loss": 1.3046,
      "theoretical_loss": 3.319531482677181,
      "tokens_seen": 3014524928
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.81010101010101e-05,
      "loss": 1.4415,
      "theoretical_loss": 3.319525939529408,
      "tokens_seen": 3014590464
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.808080808080808e-05,
      "loss": 1.246,
      "theoretical_loss": 3.3195203965358804,
      "tokens_seen": 3014656000
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.806060606060606e-05,
      "loss": 1.5064,
      "theoretical_loss": 3.3195148536965906,
      "tokens_seen": 3014721536
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.804040404040404e-05,
      "loss": 1.173,
      "theoretical_loss": 3.3195093110115304,
      "tokens_seen": 3014787072
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.802020202020203e-05,
      "loss": 1.238,
      "theoretical_loss": 3.319503768480693,
      "tokens_seen": 3014852608
    },
    {
      "epoch": 0.83,
      "objective/train/aligned_token_count": 0.3404814302921295,
      "objective/train/avg_token_score": 0.05260157957673073,
      "objective/train/docs_used": 1695676,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1712831258773804,
      "objective/train/original_loss": 5.194584846496582,
      "objective/train/theoretical_loss": 3.31949822610407,
      "objective/train/tokens_used": 1373777376,
      "objective/train/unlikelihood_loss": 0.04949592798948288,
      "theoretical_loss": 3.31949822610407,
      "tokens_seen": 3014918144
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.8e-05,
      "loss": 1.3566,
      "theoretical_loss": 3.31949822610407,
      "tokens_seen": 3014918144
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.797979797979798e-05,
      "loss": 1.2811,
      "theoretical_loss": 3.3194926838816543,
      "tokens_seen": 3014983680
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.795959595959597e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.319487141813438,
      "tokens_seen": 3015049216
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.793939393939395e-05,
      "loss": 1.3138,
      "theoretical_loss": 3.3194815998994134,
      "tokens_seen": 3015114752
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.791919191919192e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.319476058139573,
      "tokens_seen": 3015180288
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.789898989898989e-05,
      "loss": 1.3616,
      "theoretical_loss": 3.3194705165339093,
      "tokens_seen": 3015245824
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.787878787878787e-05,
      "loss": 1.5193,
      "theoretical_loss": 3.319464975082415,
      "tokens_seen": 3015311360
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.785858585858586e-05,
      "loss": 1.2838,
      "theoretical_loss": 3.319459433785081,
      "tokens_seen": 3015376896
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.783838383838384e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.319453892641901,
      "tokens_seen": 3015442432
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.781818181818183e-05,
      "loss": 1.3562,
      "theoretical_loss": 3.319448351652867,
      "tokens_seen": 3015507968
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.77979797979798e-05,
      "loss": 0.9958,
      "theoretical_loss": 3.3194428108179714,
      "tokens_seen": 3015573504
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.777777777777778e-05,
      "loss": 1.3305,
      "theoretical_loss": 3.319437270137206,
      "tokens_seen": 3015639040
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.775757575757576e-05,
      "loss": 1.2689,
      "theoretical_loss": 3.319431729610564,
      "tokens_seen": 3015704576
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.773737373737373e-05,
      "loss": 1.4115,
      "theoretical_loss": 3.3194261892380377,
      "tokens_seen": 3015770112
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.771717171717172e-05,
      "loss": 1.255,
      "theoretical_loss": 3.319420649019619,
      "tokens_seen": 3015835648
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.76969696969697e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.3194151089553006,
      "tokens_seen": 3015901184
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.767676767676767e-05,
      "loss": 1.3307,
      "theoretical_loss": 3.3194095690450744,
      "tokens_seen": 3015966720
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.765656565656566e-05,
      "loss": 1.3029,
      "theoretical_loss": 3.3194040292889335,
      "tokens_seen": 3016032256
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.763636363636364e-05,
      "loss": 1.3858,
      "theoretical_loss": 3.3193984896868693,
      "tokens_seen": 3016097792
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.761616161616163e-05,
      "loss": 1.4685,
      "theoretical_loss": 3.3193929502388753,
      "tokens_seen": 3016163328
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.759595959595961e-05,
      "loss": 1.3883,
      "theoretical_loss": 3.319387410944943,
      "tokens_seen": 3016228864
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.757575757575757e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.319381871805065,
      "tokens_seen": 3016294400
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.755555555555555e-05,
      "loss": 1.1439,
      "theoretical_loss": 3.3193763328192336,
      "tokens_seen": 3016359936
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.753535353535353e-05,
      "loss": 1.1077,
      "theoretical_loss": 3.3193707939874417,
      "tokens_seen": 3016425472
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.751515151515152e-05,
      "loss": 1.3721,
      "theoretical_loss": 3.319365255309681,
      "tokens_seen": 3016491008
    },
    {
      "epoch": 0.83,
      "objective/train/aligned_token_count": 0.3390151560306549,
      "objective/train/avg_token_score": 0.00862281396985054,
      "objective/train/docs_used": 1696888,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1023715734481812,
      "objective/train/original_loss": 4.481023788452148,
      "objective/train/theoretical_loss": 3.3193597167859443,
      "objective/train/tokens_used": 1375415776,
      "objective/train/unlikelihood_loss": 0.07725152373313904,
      "theoretical_loss": 3.3193597167859443,
      "tokens_seen": 3016556544
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.74949494949495e-05,
      "loss": 1.2227,
      "theoretical_loss": 3.3193597167859443,
      "tokens_seen": 3016556544
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.747474747474747e-05,
      "loss": 1.2054,
      "theoretical_loss": 3.3193541784162237,
      "tokens_seen": 3016622080
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.745454545454546e-05,
      "loss": 1.1218,
      "theoretical_loss": 3.3193486402005115,
      "tokens_seen": 3016687616
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.743434343434344e-05,
      "loss": 1.4822,
      "theoretical_loss": 3.3193431021388005,
      "tokens_seen": 3016753152
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.741414141414142e-05,
      "loss": 1.2304,
      "theoretical_loss": 3.3193375642310827,
      "tokens_seen": 3016818688
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.73939393939394e-05,
      "loss": 1.2015,
      "theoretical_loss": 3.3193320264773507,
      "tokens_seen": 3016884224
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.737373737373736e-05,
      "loss": 1.3142,
      "theoretical_loss": 3.3193264888775964,
      "tokens_seen": 3016949760
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.735353535353535e-05,
      "loss": 1.1745,
      "theoretical_loss": 3.3193209514318127,
      "tokens_seen": 3017015296
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.733333333333333e-05,
      "loss": 1.4824,
      "theoretical_loss": 3.319315414139992,
      "tokens_seen": 3017080832
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.731313131313132e-05,
      "loss": 1.3843,
      "theoretical_loss": 3.3193098770021265,
      "tokens_seen": 3017146368
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.72929292929293e-05,
      "loss": 1.3465,
      "theoretical_loss": 3.3193043400182085,
      "tokens_seen": 3017211904
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.727272727272728e-05,
      "loss": 1.2874,
      "theoretical_loss": 3.3192988031882305,
      "tokens_seen": 3017277440
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.725252525252526e-05,
      "loss": 1.2689,
      "theoretical_loss": 3.319293266512185,
      "tokens_seen": 3017342976
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.723232323232324e-05,
      "loss": 1.2333,
      "theoretical_loss": 3.3192877299900636,
      "tokens_seen": 3017408512
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.721212121212121e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.3192821936218597,
      "tokens_seen": 3017474048
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.71919191919192e-05,
      "loss": 1.1294,
      "theoretical_loss": 3.319276657407565,
      "tokens_seen": 3017539584
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.717171717171718e-05,
      "loss": 1.304,
      "theoretical_loss": 3.3192711213471724,
      "tokens_seen": 3017605120
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.715151515151515e-05,
      "loss": 1.4157,
      "theoretical_loss": 3.319265585440674,
      "tokens_seen": 3017670656
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.713131313131313e-05,
      "loss": 1.2732,
      "theoretical_loss": 3.319260049688062,
      "tokens_seen": 3017736192
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.711111111111112e-05,
      "loss": 1.422,
      "theoretical_loss": 3.3192545140893293,
      "tokens_seen": 3017801728
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.70909090909091e-05,
      "loss": 1.5486,
      "theoretical_loss": 3.3192489786444677,
      "tokens_seen": 3017867264
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.707070707070708e-05,
      "loss": 1.197,
      "theoretical_loss": 3.3192434433534697,
      "tokens_seen": 3017932800
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.705050505050504e-05,
      "loss": 1.4733,
      "theoretical_loss": 3.3192379082163277,
      "tokens_seen": 3017998336
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.703030303030302e-05,
      "loss": 1.4009,
      "theoretical_loss": 3.3192323732330347,
      "tokens_seen": 3018063872
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.701010101010101e-05,
      "loss": 1.3979,
      "theoretical_loss": 3.319226838403582,
      "tokens_seen": 3018129408
    },
    {
      "epoch": 0.83,
      "objective/train/aligned_token_count": 0.5178396701812744,
      "objective/train/avg_token_score": 0.009794346988201141,
      "objective/train/docs_used": 1697654,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5901719331741333,
      "objective/train/original_loss": 4.212164402008057,
      "objective/train/theoretical_loss": 3.319221303727963,
      "objective/train/tokens_used": 1377054176,
      "objective/train/unlikelihood_loss": 0.05368603393435478,
      "theoretical_loss": 3.319221303727963,
      "tokens_seen": 3018194944
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.698989898989899e-05,
      "loss": 1.3879,
      "theoretical_loss": 3.319221303727963,
      "tokens_seen": 3018194944
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.696969696969698e-05,
      "loss": 1.5148,
      "theoretical_loss": 3.3192157692061697,
      "tokens_seen": 3018260480
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.694949494949495e-05,
      "loss": 1.3506,
      "theoretical_loss": 3.319210234838194,
      "tokens_seen": 3018326016
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.692929292929293e-05,
      "loss": 1.3416,
      "theoretical_loss": 3.319204700624029,
      "tokens_seen": 3018391552
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.690909090909091e-05,
      "loss": 1.3948,
      "theoretical_loss": 3.3191991665636666,
      "tokens_seen": 3018457088
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.68888888888889e-05,
      "loss": 1.4254,
      "theoretical_loss": 3.3191936326571,
      "tokens_seen": 3018522624
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.686868686868687e-05,
      "loss": 1.3414,
      "theoretical_loss": 3.3191880989043203,
      "tokens_seen": 3018588160
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.684848484848485e-05,
      "loss": 1.4406,
      "theoretical_loss": 3.319182565305321,
      "tokens_seen": 3018653696
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.682828282828282e-05,
      "loss": 1.2639,
      "theoretical_loss": 3.319177031860094,
      "tokens_seen": 3018719232
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.680808080808081e-05,
      "loss": 1.4408,
      "theoretical_loss": 3.319171498568631,
      "tokens_seen": 3018784768
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.678787878787879e-05,
      "loss": 1.3192,
      "theoretical_loss": 3.3191659654309262,
      "tokens_seen": 3018850304
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.676767676767678e-05,
      "loss": 1.5392,
      "theoretical_loss": 3.3191604324469703,
      "tokens_seen": 3018915840
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.674747474747476e-05,
      "loss": 1.3212,
      "theoretical_loss": 3.3191548996167564,
      "tokens_seen": 3018981376
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.672727272727273e-05,
      "loss": 1.2253,
      "theoretical_loss": 3.3191493669402767,
      "tokens_seen": 3019046912
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.67070707070707e-05,
      "loss": 1.4734,
      "theoretical_loss": 3.319143834417524,
      "tokens_seen": 3019112448
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.668686868686868e-05,
      "loss": 1.1149,
      "theoretical_loss": 3.3191383020484904,
      "tokens_seen": 3019177984
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.666666666666667e-05,
      "loss": 1.3468,
      "theoretical_loss": 3.319132769833168,
      "tokens_seen": 3019243520
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.664646464646465e-05,
      "loss": 1.3712,
      "theoretical_loss": 3.3191272377715495,
      "tokens_seen": 3019309056
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.662626262626262e-05,
      "loss": 1.389,
      "theoretical_loss": 3.319121705863627,
      "tokens_seen": 3019374592
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.66060606060606e-05,
      "loss": 1.3696,
      "theoretical_loss": 3.3191161741093937,
      "tokens_seen": 3019440128
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.658585858585859e-05,
      "loss": 1.1651,
      "theoretical_loss": 3.3191106425088415,
      "tokens_seen": 3019505664
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.656565656565657e-05,
      "loss": 1.4324,
      "theoretical_loss": 3.3191051110619627,
      "tokens_seen": 3019571200
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.654545454545456e-05,
      "loss": 1.24,
      "theoretical_loss": 3.319099579768749,
      "tokens_seen": 3019636736
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.652525252525252e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.319094048629194,
      "tokens_seen": 3019702272
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.65050505050505e-05,
      "loss": 1.4085,
      "theoretical_loss": 3.31908851764329,
      "tokens_seen": 3019767808
    },
    {
      "epoch": 0.83,
      "objective/train/aligned_token_count": 0.4922409653663635,
      "objective/train/avg_token_score": 0.01194742415100336,
      "objective/train/docs_used": 1699020,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.494076132774353,
      "objective/train/original_loss": 3.7357516288757324,
      "objective/train/theoretical_loss": 3.3190829868110288,
      "objective/train/tokens_used": 1378692576,
      "objective/train/unlikelihood_loss": 0.07323186844587326,
      "theoretical_loss": 3.3190829868110288,
      "tokens_seen": 3019833344
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.648484848484848e-05,
      "loss": 1.5554,
      "theoretical_loss": 3.3190829868110288,
      "tokens_seen": 3019833344
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.646464646464647e-05,
      "loss": 1.1528,
      "theoretical_loss": 3.319077456132403,
      "tokens_seen": 3019898880
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.644444444444445e-05,
      "loss": 1.5563,
      "theoretical_loss": 3.319071925607405,
      "tokens_seen": 3019964416
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.642424242424242e-05,
      "loss": 1.376,
      "theoretical_loss": 3.3190663952360273,
      "tokens_seen": 3020029952
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.64040404040404e-05,
      "loss": 1.2958,
      "theoretical_loss": 3.319060865018262,
      "tokens_seen": 3020095488
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.638383838383839e-05,
      "loss": 1.2136,
      "theoretical_loss": 3.319055334954102,
      "tokens_seen": 3020161024
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.636363636363636e-05,
      "loss": 1.4196,
      "theoretical_loss": 3.319049805043539,
      "tokens_seen": 3020226560
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.634343434343434e-05,
      "loss": 1.21,
      "theoretical_loss": 3.3190442752865663,
      "tokens_seen": 3020292096
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.632323232323233e-05,
      "loss": 1.2498,
      "theoretical_loss": 3.3190387456831756,
      "tokens_seen": 3020357632
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.63030303030303e-05,
      "loss": 1.3153,
      "theoretical_loss": 3.31903321623336,
      "tokens_seen": 3020423168
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.628282828282828e-05,
      "loss": 1.4863,
      "theoretical_loss": 3.3190276869371105,
      "tokens_seen": 3020488704
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.626262626262627e-05,
      "loss": 1.381,
      "theoretical_loss": 3.319022157794421,
      "tokens_seen": 3020554240
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.624242424242425e-05,
      "loss": 1.4788,
      "theoretical_loss": 3.3190166288052834,
      "tokens_seen": 3020619776
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.622222222222223e-05,
      "loss": 1.4128,
      "theoretical_loss": 3.31901109996969,
      "tokens_seen": 3020685312
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.62020202020202e-05,
      "loss": 1.2635,
      "theoretical_loss": 3.3190055712876334,
      "tokens_seen": 3020750848
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.618181818181817e-05,
      "loss": 1.2862,
      "theoretical_loss": 3.3190000427591055,
      "tokens_seen": 3020816384
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.616161616161616e-05,
      "loss": 1.3656,
      "theoretical_loss": 3.3189945143840993,
      "tokens_seen": 3020881920
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.614141414141414e-05,
      "loss": 1.3649,
      "theoretical_loss": 3.3189889861626067,
      "tokens_seen": 3020947456
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.612121212121213e-05,
      "loss": 1.4766,
      "theoretical_loss": 3.3189834580946207,
      "tokens_seen": 3021012992
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.61010101010101e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.318977930180133,
      "tokens_seen": 3021078528
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.608080808080808e-05,
      "loss": 0.9308,
      "theoretical_loss": 3.3189724024191367,
      "tokens_seen": 3021144064
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.606060606060606e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.318966874811624,
      "tokens_seen": 3021209600
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.604040404040405e-05,
      "loss": 1.333,
      "theoretical_loss": 3.318961347357587,
      "tokens_seen": 3021275136
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.602020202020203e-05,
      "loss": 1.5977,
      "theoretical_loss": 3.3189558200570186,
      "tokens_seen": 3021340672
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.599999999999999e-05,
      "loss": 1.0661,
      "theoretical_loss": 3.3189502929099106,
      "tokens_seen": 3021406208
    },
    {
      "epoch": 0.83,
      "objective/train/aligned_token_count": 0.3960166275501251,
      "objective/train/avg_token_score": 0.009528541006147861,
      "objective/train/docs_used": 1699637,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0561459064483643,
      "objective/train/original_loss": 4.358443737030029,
      "objective/train/theoretical_loss": 3.318944765916256,
      "objective/train/tokens_used": 1380330976,
      "objective/train/unlikelihood_loss": 0.04811561107635498,
      "theoretical_loss": 3.318944765916256,
      "tokens_seen": 3021471744
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.597979797979797e-05,
      "loss": 1.318,
      "theoretical_loss": 3.318944765916256,
      "tokens_seen": 3021471744
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.595959595959596e-05,
      "loss": 1.2539,
      "theoretical_loss": 3.3189392390760464,
      "tokens_seen": 3021537280
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.593939393939394e-05,
      "loss": 1.1466,
      "theoretical_loss": 3.318933712389275,
      "tokens_seen": 3021602816
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.591919191919193e-05,
      "loss": 1.252,
      "theoretical_loss": 3.3189281858559343,
      "tokens_seen": 3021668352
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.589898989898991e-05,
      "loss": 1.2704,
      "theoretical_loss": 3.318922659476016,
      "tokens_seen": 3021733888
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.587878787878788e-05,
      "loss": 1.1407,
      "theoretical_loss": 3.3189171332495135,
      "tokens_seen": 3021799424
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.585858585858586e-05,
      "loss": 1.1901,
      "theoretical_loss": 3.3189116071764184,
      "tokens_seen": 3021864960
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.583838383838383e-05,
      "loss": 1.3291,
      "theoretical_loss": 3.318906081256723,
      "tokens_seen": 3021930496
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.581818181818182e-05,
      "loss": 1.2247,
      "theoretical_loss": 3.31890055549042,
      "tokens_seen": 3021996032
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.57979797979798e-05,
      "loss": 1.3063,
      "theoretical_loss": 3.318895029877502,
      "tokens_seen": 3022061568
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.577777777777777e-05,
      "loss": 1.4427,
      "theoretical_loss": 3.3188895044179616,
      "tokens_seen": 3022127104
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.575757575757576e-05,
      "loss": 1.5116,
      "theoretical_loss": 3.3188839791117903,
      "tokens_seen": 3022192640
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.573737373737374e-05,
      "loss": 1.3321,
      "theoretical_loss": 3.3188784539589817,
      "tokens_seen": 3022258176
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.571717171717172e-05,
      "loss": 1.3693,
      "theoretical_loss": 3.3188729289595273,
      "tokens_seen": 3022323712
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.569696969696971e-05,
      "loss": 1.2792,
      "theoretical_loss": 3.31886740411342,
      "tokens_seen": 3022389248
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.567676767676768e-05,
      "loss": 1.4025,
      "theoretical_loss": 3.318861879420652,
      "tokens_seen": 3022454784
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.565656565656565e-05,
      "loss": 1.453,
      "theoretical_loss": 3.3188563548812158,
      "tokens_seen": 3022520320
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.563636363636363e-05,
      "loss": 1.6693,
      "theoretical_loss": 3.3188508304951037,
      "tokens_seen": 3022585856
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.561616161616162e-05,
      "loss": 1.4096,
      "theoretical_loss": 3.3188453062623084,
      "tokens_seen": 3022651392
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.55959595959596e-05,
      "loss": 1.3179,
      "theoretical_loss": 3.318839782182822,
      "tokens_seen": 3022716928
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.557575757575757e-05,
      "loss": 1.1187,
      "theoretical_loss": 3.3188342582566372,
      "tokens_seen": 3022782464
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.555555555555556e-05,
      "loss": 1.2515,
      "theoretical_loss": 3.318828734483746,
      "tokens_seen": 3022848000
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.553535353535354e-05,
      "loss": 1.4039,
      "theoretical_loss": 3.3188232108641413,
      "tokens_seen": 3022913536
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.551515151515152e-05,
      "loss": 1.461,
      "theoretical_loss": 3.3188176873978152,
      "tokens_seen": 3022979072
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.54949494949495e-05,
      "loss": 1.4092,
      "theoretical_loss": 3.3188121640847608,
      "tokens_seen": 3023044608
    },
    {
      "epoch": 0.83,
      "objective/train/aligned_token_count": 0.38624143600463867,
      "objective/train/avg_token_score": 0.014024905860424042,
      "objective/train/docs_used": 1700711,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1858272552490234,
      "objective/train/original_loss": 4.949784755706787,
      "objective/train/theoretical_loss": 3.3188066409249695,
      "objective/train/tokens_used": 1381969376,
      "objective/train/unlikelihood_loss": 0.0482003279030323,
      "theoretical_loss": 3.3188066409249695,
      "tokens_seen": 3023110144
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.547474747474748e-05,
      "loss": 1.3863,
      "theoretical_loss": 3.3188066409249695,
      "tokens_seen": 3023110144
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.545454545454545e-05,
      "loss": 1.393,
      "theoretical_loss": 3.3188011179184342,
      "tokens_seen": 3023175680
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.543434343434343e-05,
      "loss": 1.3262,
      "theoretical_loss": 3.3187955950651475,
      "tokens_seen": 3023241216
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.541414141414142e-05,
      "loss": 1.3647,
      "theoretical_loss": 3.3187900723651014,
      "tokens_seen": 3023306752
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.53939393939394e-05,
      "loss": 1.4637,
      "theoretical_loss": 3.318784549818289,
      "tokens_seen": 3023372288
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.537373737373738e-05,
      "loss": 1.1559,
      "theoretical_loss": 3.318779027424702,
      "tokens_seen": 3023437824
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.535353535353535e-05,
      "loss": 1.4948,
      "theoretical_loss": 3.3187735051843332,
      "tokens_seen": 3023503360
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.533333333333334e-05,
      "loss": 1.4662,
      "theoretical_loss": 3.318767983097175,
      "tokens_seen": 3023568896
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.531313131313131e-05,
      "loss": 1.1945,
      "theoretical_loss": 3.31876246116322,
      "tokens_seen": 3023634432
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.529292929292929e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.3187569393824603,
      "tokens_seen": 3023699968
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.527272727272728e-05,
      "loss": 1.3121,
      "theoretical_loss": 3.3187514177548882,
      "tokens_seen": 3023765504
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.525252525252525e-05,
      "loss": 1.4647,
      "theoretical_loss": 3.3187458962804968,
      "tokens_seen": 3023831040
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.523232323232323e-05,
      "loss": 1.3489,
      "theoretical_loss": 3.318740374959278,
      "tokens_seen": 3023896576
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.521212121212122e-05,
      "loss": 1.3621,
      "theoretical_loss": 3.318734853791224,
      "tokens_seen": 3023962112
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.51919191919192e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.318729332776328,
      "tokens_seen": 3024027648
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.517171717171718e-05,
      "loss": 1.4373,
      "theoretical_loss": 3.318723811914582,
      "tokens_seen": 3024093184
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.515151515151514e-05,
      "loss": 1.301,
      "theoretical_loss": 3.318718291205978,
      "tokens_seen": 3024158720
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.513131313131312e-05,
      "loss": 1.3053,
      "theoretical_loss": 3.3187127706505093,
      "tokens_seen": 3024224256
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.511111111111111e-05,
      "loss": 1.3185,
      "theoretical_loss": 3.318707250248168,
      "tokens_seen": 3024289792
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.509090909090909e-05,
      "loss": 1.4106,
      "theoretical_loss": 3.318701729998946,
      "tokens_seen": 3024355328
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.507070707070708e-05,
      "loss": 1.358,
      "theoretical_loss": 3.3186962099028365,
      "tokens_seen": 3024420864
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.505050505050506e-05,
      "loss": 1.055,
      "theoretical_loss": 3.3186906899598316,
      "tokens_seen": 3024486400
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.503030303030303e-05,
      "loss": 1.2886,
      "theoretical_loss": 3.3186851701699234,
      "tokens_seen": 3024551936
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.501010101010101e-05,
      "loss": 1.1468,
      "theoretical_loss": 3.318679650533105,
      "tokens_seen": 3024617472
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.4989898989899e-05,
      "loss": 1.1193,
      "theoretical_loss": 3.3186741310493684,
      "tokens_seen": 3024683008
    },
    {
      "epoch": 0.83,
      "objective/train/aligned_token_count": 0.3581989109516144,
      "objective/train/avg_token_score": 0.01199006475508213,
      "objective/train/docs_used": 1701440,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8649300932884216,
      "objective/train/original_loss": 4.489315509796143,
      "objective/train/theoretical_loss": 3.3186686117187065,
      "objective/train/tokens_used": 1383607776,
      "objective/train/unlikelihood_loss": 0.0568234920501709,
      "theoretical_loss": 3.3186686117187065,
      "tokens_seen": 3024748544
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.496969696969697e-05,
      "loss": 1.1441,
      "theoretical_loss": 3.3186686117187065,
      "tokens_seen": 3024748544
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.494949494949495e-05,
      "loss": 1.3097,
      "theoretical_loss": 3.318663092541111,
      "tokens_seen": 3024814080
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.492929292929292e-05,
      "loss": 1.5099,
      "theoretical_loss": 3.318657573516575,
      "tokens_seen": 3024879616
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.490909090909091e-05,
      "loss": 1.505,
      "theoretical_loss": 3.318652054645091,
      "tokens_seen": 3024945152
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.488888888888889e-05,
      "loss": 1.3422,
      "theoretical_loss": 3.3186465359266504,
      "tokens_seen": 3025010688
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.486868686868687e-05,
      "loss": 1.1287,
      "theoretical_loss": 3.3186410173612466,
      "tokens_seen": 3025076224
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.484848484848486e-05,
      "loss": 1.4571,
      "theoretical_loss": 3.318635498948872,
      "tokens_seen": 3025141760
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.482828282828283e-05,
      "loss": 1.4235,
      "theoretical_loss": 3.3186299806895185,
      "tokens_seen": 3025207296
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.480808080808081e-05,
      "loss": 1.5327,
      "theoretical_loss": 3.3186244625831796,
      "tokens_seen": 3025272832
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.478787878787878e-05,
      "loss": 1.4313,
      "theoretical_loss": 3.318618944629846,
      "tokens_seen": 3025338368
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.476767676767677e-05,
      "loss": 1.3167,
      "theoretical_loss": 3.318613426829512,
      "tokens_seen": 3025403904
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.474747474747475e-05,
      "loss": 1.445,
      "theoretical_loss": 3.318607909182169,
      "tokens_seen": 3025469440
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.472727272727272e-05,
      "loss": 1.3282,
      "theoretical_loss": 3.3186023916878096,
      "tokens_seen": 3025534976
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.47070707070707e-05,
      "loss": 1.2242,
      "theoretical_loss": 3.3185968743464263,
      "tokens_seen": 3025600512
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.468686868686869e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.3185913571580117,
      "tokens_seen": 3025666048
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.466666666666667e-05,
      "loss": 1.4263,
      "theoretical_loss": 3.318585840122558,
      "tokens_seen": 3025731584
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.464646464646466e-05,
      "loss": 1.0479,
      "theoretical_loss": 3.3185803232400577,
      "tokens_seen": 3025797120
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.462626262626263e-05,
      "loss": 1.341,
      "theoretical_loss": 3.318574806510503,
      "tokens_seen": 3025862656
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.46060606060606e-05,
      "loss": 1.2964,
      "theoretical_loss": 3.318569289933887,
      "tokens_seen": 3025928192
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.458585858585858e-05,
      "loss": 1.3201,
      "theoretical_loss": 3.318563773510202,
      "tokens_seen": 3025993728
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.456565656565657e-05,
      "loss": 1.3819,
      "theoretical_loss": 3.31855825723944,
      "tokens_seen": 3026059264
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.454545454545455e-05,
      "loss": 1.4464,
      "theoretical_loss": 3.3185527411215934,
      "tokens_seen": 3026124800
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.452525252525253e-05,
      "loss": 1.3321,
      "theoretical_loss": 3.3185472251566552,
      "tokens_seen": 3026190336
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.45050505050505e-05,
      "loss": 1.1501,
      "theoretical_loss": 3.3185417093446175,
      "tokens_seen": 3026255872
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.448484848484849e-05,
      "loss": 1.2806,
      "theoretical_loss": 3.318536193685473,
      "tokens_seen": 3026321408
    },
    {
      "epoch": 0.83,
      "objective/train/aligned_token_count": 0.19861926138401031,
      "objective/train/avg_token_score": 0.015986135229468346,
      "objective/train/docs_used": 1702703,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.747661828994751,
      "objective/train/original_loss": 5.9110941886901855,
      "objective/train/theoretical_loss": 3.3185306781792137,
      "objective/train/tokens_used": 1385246176,
      "objective/train/unlikelihood_loss": 0.03576838970184326,
      "theoretical_loss": 3.3185306781792137,
      "tokens_seen": 3026386944
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.446464646464647e-05,
      "loss": 1.251,
      "theoretical_loss": 3.3185306781792137,
      "tokens_seen": 3026386944
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.444444444444444e-05,
      "loss": 1.4103,
      "theoretical_loss": 3.3185251628258325,
      "tokens_seen": 3026452480
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.442424242424243e-05,
      "loss": 1.5084,
      "theoretical_loss": 3.3185196476253216,
      "tokens_seen": 3026518016
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.44040404040404e-05,
      "loss": 1.5282,
      "theoretical_loss": 3.3185141325776737,
      "tokens_seen": 3026583552
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.438383838383838e-05,
      "loss": 1.304,
      "theoretical_loss": 3.318508617682881,
      "tokens_seen": 3026649088
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.436363636363637e-05,
      "loss": 1.4644,
      "theoretical_loss": 3.318503102940936,
      "tokens_seen": 3026714624
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.434343434343435e-05,
      "loss": 1.3361,
      "theoretical_loss": 3.318497588351831,
      "tokens_seen": 3026780160
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.432323232323233e-05,
      "loss": 1.4712,
      "theoretical_loss": 3.318492073915559,
      "tokens_seen": 3026845696
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.43030303030303e-05,
      "loss": 1.2644,
      "theoretical_loss": 3.3184865596321123,
      "tokens_seen": 3026911232
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.428282828282827e-05,
      "loss": 1.2289,
      "theoretical_loss": 3.318481045501483,
      "tokens_seen": 3026976768
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.426262626262626e-05,
      "loss": 1.3668,
      "theoretical_loss": 3.3184755315236636,
      "tokens_seen": 3027042304
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.424242424242424e-05,
      "loss": 1.2863,
      "theoretical_loss": 3.3184700176986466,
      "tokens_seen": 3027107840
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.422222222222223e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.3184645040264247,
      "tokens_seen": 3027173376
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.420202020202021e-05,
      "loss": 1.3681,
      "theoretical_loss": 3.31845899050699,
      "tokens_seen": 3027238912
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.418181818181818e-05,
      "loss": 1.2461,
      "theoretical_loss": 3.318453477140335,
      "tokens_seen": 3027304448
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.416161616161616e-05,
      "loss": 1.088,
      "theoretical_loss": 3.318447963926453,
      "tokens_seen": 3027369984
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.414141414141415e-05,
      "loss": 1.4237,
      "theoretical_loss": 3.3184424508653354,
      "tokens_seen": 3027435520
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.412121212121213e-05,
      "loss": 1.4514,
      "theoretical_loss": 3.318436937956975,
      "tokens_seen": 3027501056
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.41010101010101e-05,
      "loss": 1.3054,
      "theoretical_loss": 3.3184314252013642,
      "tokens_seen": 3027566592
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.408080808080807e-05,
      "loss": 1.4611,
      "theoretical_loss": 3.318425912598496,
      "tokens_seen": 3027632128
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.406060606060606e-05,
      "loss": 1.3625,
      "theoretical_loss": 3.3184204001483617,
      "tokens_seen": 3027697664
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.404040404040404e-05,
      "loss": 1.4106,
      "theoretical_loss": 3.318414887850955,
      "tokens_seen": 3027763200
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.402020202020202e-05,
      "loss": 1.4374,
      "theoretical_loss": 3.318409375706268,
      "tokens_seen": 3027828736
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.400000000000001e-05,
      "loss": 1.1323,
      "theoretical_loss": 3.318403863714292,
      "tokens_seen": 3027894272
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.397979797979798e-05,
      "loss": 1.3119,
      "theoretical_loss": 3.3183983518750217,
      "tokens_seen": 3027959808
    },
    {
      "epoch": 0.83,
      "objective/train/aligned_token_count": 0.37781035900115967,
      "objective/train/avg_token_score": 0.010626877658069134,
      "objective/train/docs_used": 1703090,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2165758609771729,
      "objective/train/original_loss": 4.246699333190918,
      "objective/train/theoretical_loss": 3.318392840188448,
      "objective/train/tokens_used": 1386884576,
      "objective/train/unlikelihood_loss": 0.05951658636331558,
      "theoretical_loss": 3.318392840188448,
      "tokens_seen": 3028025344
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.395959595959596e-05,
      "loss": 1.393,
      "theoretical_loss": 3.318392840188448,
      "tokens_seen": 3028025344
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.393939393939393e-05,
      "loss": 1.449,
      "theoretical_loss": 3.3183873286545635,
      "tokens_seen": 3028090880
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.391919191919192e-05,
      "loss": 1.1054,
      "theoretical_loss": 3.3183818172733606,
      "tokens_seen": 3028156416
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.38989898989899e-05,
      "loss": 1.4772,
      "theoretical_loss": 3.3183763060448324,
      "tokens_seen": 3028221952
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.387878787878787e-05,
      "loss": 1.3749,
      "theoretical_loss": 3.318370794968971,
      "tokens_seen": 3028287488
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.385858585858586e-05,
      "loss": 1.2217,
      "theoretical_loss": 3.3183652840457687,
      "tokens_seen": 3028353024
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.383838383838384e-05,
      "loss": 1.4334,
      "theoretical_loss": 3.318359773275218,
      "tokens_seen": 3028418560
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.381818181818182e-05,
      "loss": 1.4773,
      "theoretical_loss": 3.318354262657312,
      "tokens_seen": 3028484096
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.379797979797981e-05,
      "loss": 1.2325,
      "theoretical_loss": 3.3183487521920423,
      "tokens_seen": 3028549632
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.377777777777779e-05,
      "loss": 1.3201,
      "theoretical_loss": 3.318343241879402,
      "tokens_seen": 3028615168
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.375757575757575e-05,
      "loss": 1.1976,
      "theoretical_loss": 3.318337731719383,
      "tokens_seen": 3028680704
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.373737373737373e-05,
      "loss": 1.315,
      "theoretical_loss": 3.3183322217119784,
      "tokens_seen": 3028746240
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.371717171717172e-05,
      "loss": 1.1383,
      "theoretical_loss": 3.31832671185718,
      "tokens_seen": 3028811776
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.36969696969697e-05,
      "loss": 1.2183,
      "theoretical_loss": 3.318321202154981,
      "tokens_seen": 3028877312
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.36969696969697e-05,
      "loss": 1.3481,
      "theoretical_loss": 3.318315692605373,
      "tokens_seen": 3028942848
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.367676767676768e-05,
      "loss": 1.3415,
      "theoretical_loss": 3.3183101832083497,
      "tokens_seen": 3029008384
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.365656565656565e-05,
      "loss": 1.322,
      "theoretical_loss": 3.3183046739639024,
      "tokens_seen": 3029073920
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.363636363636364e-05,
      "loss": 1.4704,
      "theoretical_loss": 3.318299164872024,
      "tokens_seen": 3029139456
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.361616161616162e-05,
      "loss": 1.3083,
      "theoretical_loss": 3.318293655932707,
      "tokens_seen": 3029204992
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.359595959595961e-05,
      "loss": 1.4431,
      "theoretical_loss": 3.318288147145944,
      "tokens_seen": 3029270528
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.357575757575758e-05,
      "loss": 1.2822,
      "theoretical_loss": 3.318282638511727,
      "tokens_seen": 3029336064
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.355555555555555e-05,
      "loss": 1.0288,
      "theoretical_loss": 3.318277130030049,
      "tokens_seen": 3029401600
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.353535353535353e-05,
      "loss": 1.4597,
      "theoretical_loss": 3.3182716217009025,
      "tokens_seen": 3029467136
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.351515151515152e-05,
      "loss": 1.2385,
      "theoretical_loss": 3.3182661135242797,
      "tokens_seen": 3029532672
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.34949494949495e-05,
      "loss": 1.27,
      "theoretical_loss": 3.318260605500173,
      "tokens_seen": 3029598208
    },
    {
      "epoch": 0.83,
      "objective/train/aligned_token_count": 0.4094574749469757,
      "objective/train/avg_token_score": 0.018313316628336906,
      "objective/train/docs_used": 1704525,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2836830615997314,
      "objective/train/original_loss": 4.789156913757324,
      "objective/train/theoretical_loss": 3.318255097628575,
      "objective/train/tokens_used": 1388522976,
      "objective/train/unlikelihood_loss": 0.0568431131541729,
      "theoretical_loss": 3.318255097628575,
      "tokens_seen": 3029663744
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.347474747474748e-05,
      "loss": 1.3234,
      "theoretical_loss": 3.318255097628575,
      "tokens_seen": 3029663744
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.345454545454545e-05,
      "loss": 1.3801,
      "theoretical_loss": 3.3182495899094784,
      "tokens_seen": 3029729280
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.343434343434344e-05,
      "loss": 1.3481,
      "theoretical_loss": 3.318244082342875,
      "tokens_seen": 3029794816
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.341414141414141e-05,
      "loss": 1.3028,
      "theoretical_loss": 3.3182385749287584,
      "tokens_seen": 3029860352
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.339393939393939e-05,
      "loss": 1.1922,
      "theoretical_loss": 3.31823306766712,
      "tokens_seen": 3029925888
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.337373737373738e-05,
      "loss": 1.3627,
      "theoretical_loss": 3.318227560557953,
      "tokens_seen": 3029991424
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.335353535353536e-05,
      "loss": 1.3846,
      "theoretical_loss": 3.318222053601249,
      "tokens_seen": 3030056960
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.333333333333333e-05,
      "loss": 1.1893,
      "theoretical_loss": 3.318216546797002,
      "tokens_seen": 3030122496
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.331313131313131e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.3182110401452034,
      "tokens_seen": 3030188032
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.32929292929293e-05,
      "loss": 1.2319,
      "theoretical_loss": 3.318205533645845,
      "tokens_seen": 3030253568
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.327272727272728e-05,
      "loss": 1.1708,
      "theoretical_loss": 3.318200027298921,
      "tokens_seen": 3030319104
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.325252525252527e-05,
      "loss": 1.2461,
      "theoretical_loss": 3.318194521104423,
      "tokens_seen": 3030384640
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.323232323232322e-05,
      "loss": 1.119,
      "theoretical_loss": 3.3181890150623428,
      "tokens_seen": 3030450176
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.321212121212121e-05,
      "loss": 1.5673,
      "theoretical_loss": 3.318183509172674,
      "tokens_seen": 3030515712
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.319191919191919e-05,
      "loss": 1.4288,
      "theoretical_loss": 3.318178003435409,
      "tokens_seen": 3030581248
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.317171717171718e-05,
      "loss": 1.1829,
      "theoretical_loss": 3.318172497850539,
      "tokens_seen": 3030646784
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.315151515151516e-05,
      "loss": 0.9225,
      "theoretical_loss": 3.3181669924180586,
      "tokens_seen": 3030712320
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.313131313131313e-05,
      "loss": 1.3789,
      "theoretical_loss": 3.3181614871379583,
      "tokens_seen": 3030777856
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.311111111111111e-05,
      "loss": 1.3013,
      "theoretical_loss": 3.3181559820102318,
      "tokens_seen": 3030843392
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.30909090909091e-05,
      "loss": 1.2708,
      "theoretical_loss": 3.3181504770348713,
      "tokens_seen": 3030908928
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.307070707070707e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.318144972211869,
      "tokens_seen": 3030974464
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.305050505050505e-05,
      "loss": 1.1661,
      "theoretical_loss": 3.3181394675412177,
      "tokens_seen": 3031040000
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.303030303030302e-05,
      "loss": 1.3877,
      "theoretical_loss": 3.3181339630229094,
      "tokens_seen": 3031105536
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.3010101010101e-05,
      "loss": 1.1102,
      "theoretical_loss": 3.3181284586569375,
      "tokens_seen": 3031171072
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.298989898989899e-05,
      "loss": 1.2461,
      "theoretical_loss": 3.3181229544432935,
      "tokens_seen": 3031236608
    },
    {
      "epoch": 0.84,
      "objective/train/aligned_token_count": 0.40139296650886536,
      "objective/train/avg_token_score": 0.011943846009671688,
      "objective/train/docs_used": 1705064,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1952029466629028,
      "objective/train/original_loss": 4.152256011962891,
      "objective/train/theoretical_loss": 3.3181174503819704,
      "objective/train/tokens_used": 1390161376,
      "objective/train/unlikelihood_loss": 0.06271233409643173,
      "theoretical_loss": 3.3181174503819704,
      "tokens_seen": 3031302144
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.296969696969697e-05,
      "loss": 1.2299,
      "theoretical_loss": 3.3181174503819704,
      "tokens_seen": 3031302144
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.294949494949496e-05,
      "loss": 1.4052,
      "theoretical_loss": 3.318111946472961,
      "tokens_seen": 3031367680
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.292929292929294e-05,
      "loss": 1.4934,
      "theoretical_loss": 3.3181064427162568,
      "tokens_seen": 3031433216
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.290909090909091e-05,
      "loss": 1.3391,
      "theoretical_loss": 3.3181009391118517,
      "tokens_seen": 3031498752
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.288888888888888e-05,
      "loss": 1.278,
      "theoretical_loss": 3.318095435659737,
      "tokens_seen": 3031564288
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.286868686868687e-05,
      "loss": 1.5196,
      "theoretical_loss": 3.3180899323599053,
      "tokens_seen": 3031629824
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.284848484848485e-05,
      "loss": 1.5137,
      "theoretical_loss": 3.31808442921235,
      "tokens_seen": 3031695360
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.282828282828283e-05,
      "loss": 1.2875,
      "theoretical_loss": 3.318078926217063,
      "tokens_seen": 3031760896
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.28080808080808e-05,
      "loss": 1.1576,
      "theoretical_loss": 3.3180734233740363,
      "tokens_seen": 3031826432
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.278787878787879e-05,
      "loss": 1.3544,
      "theoretical_loss": 3.318067920683263,
      "tokens_seen": 3031891968
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.276767676767677e-05,
      "loss": 1.3977,
      "theoretical_loss": 3.3180624181447356,
      "tokens_seen": 3031957504
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.274747474747476e-05,
      "loss": 1.3012,
      "theoretical_loss": 3.3180569157584463,
      "tokens_seen": 3032023040
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.272727272727274e-05,
      "loss": 1.1817,
      "theoretical_loss": 3.318051413524388,
      "tokens_seen": 3032088576
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.27070707070707e-05,
      "loss": 1.2681,
      "theoretical_loss": 3.318045911442553,
      "tokens_seen": 3032154112
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.268686868686868e-05,
      "loss": 1.2585,
      "theoretical_loss": 3.3180404095129337,
      "tokens_seen": 3032219648
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.266666666666667e-05,
      "loss": 1.233,
      "theoretical_loss": 3.3180349077355227,
      "tokens_seen": 3032285184
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.264646464646465e-05,
      "loss": 1.1415,
      "theoretical_loss": 3.3180294061103126,
      "tokens_seen": 3032350720
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.262626262626263e-05,
      "loss": 1.2484,
      "theoretical_loss": 3.3180239046372955,
      "tokens_seen": 3032416256
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.26060606060606e-05,
      "loss": 1.203,
      "theoretical_loss": 3.318018403316464,
      "tokens_seen": 3032481792
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.258585858585859e-05,
      "loss": 1.2933,
      "theoretical_loss": 3.3180129021478115,
      "tokens_seen": 3032547328
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.256565656565657e-05,
      "loss": 1.1177,
      "theoretical_loss": 3.318007401131329,
      "tokens_seen": 3032612864
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.254545454545454e-05,
      "loss": 1.3799,
      "theoretical_loss": 3.3180019002670105,
      "tokens_seen": 3032678400
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.252525252525253e-05,
      "loss": 1.2746,
      "theoretical_loss": 3.3179963995548474,
      "tokens_seen": 3032743936
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.250505050505051e-05,
      "loss": 1.4203,
      "theoretical_loss": 3.3179908989948323,
      "tokens_seen": 3032809472
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.248484848484848e-05,
      "loss": 1.5437,
      "theoretical_loss": 3.3179853985869587,
      "tokens_seen": 3032875008
    },
    {
      "epoch": 0.84,
      "objective/train/aligned_token_count": 0.5162512063980103,
      "objective/train/avg_token_score": 0.013542130589485168,
      "objective/train/docs_used": 1706237,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.8202773332595825,
      "objective/train/original_loss": 3.918135166168213,
      "objective/train/theoretical_loss": 3.317979898331218,
      "objective/train/tokens_used": 1391799776,
      "objective/train/unlikelihood_loss": 0.07113587856292725,
      "theoretical_loss": 3.317979898331218,
      "tokens_seen": 3032940544
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.246464646464646e-05,
      "loss": 1.3239,
      "theoretical_loss": 3.317979898331218,
      "tokens_seen": 3032940544
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.244444444444445e-05,
      "loss": 1.283,
      "theoretical_loss": 3.3179743982276033,
      "tokens_seen": 3033006080
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.242424242424243e-05,
      "loss": 1.1834,
      "theoretical_loss": 3.3179688982761064,
      "tokens_seen": 3033071616
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.240404040404042e-05,
      "loss": 1.1404,
      "theoretical_loss": 3.3179633984767207,
      "tokens_seen": 3033137152
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.238383838383839e-05,
      "loss": 1.4685,
      "theoretical_loss": 3.3179578988294383,
      "tokens_seen": 3033202688
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.236363636363636e-05,
      "loss": 1.4712,
      "theoretical_loss": 3.3179523993342515,
      "tokens_seen": 3033268224
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.234343434343434e-05,
      "loss": 1.1987,
      "theoretical_loss": 3.317946899991153,
      "tokens_seen": 3033333760
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.232323232323233e-05,
      "loss": 1.2785,
      "theoretical_loss": 3.317941400800136,
      "tokens_seen": 3033399296
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.230303030303031e-05,
      "loss": 1.4162,
      "theoretical_loss": 3.317935901761192,
      "tokens_seen": 3033464832
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.228282828282828e-05,
      "loss": 1.3787,
      "theoretical_loss": 3.3179304028743135,
      "tokens_seen": 3033530368
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.226262626262626e-05,
      "loss": 1.5114,
      "theoretical_loss": 3.3179249041394936,
      "tokens_seen": 3033595904
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.224242424242425e-05,
      "loss": 1.4781,
      "theoretical_loss": 3.317919405556725,
      "tokens_seen": 3033661440
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.222222222222223e-05,
      "loss": 1.395,
      "theoretical_loss": 3.317913907125999,
      "tokens_seen": 3033726976
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.22020202020202e-05,
      "loss": 1.3335,
      "theoretical_loss": 3.31790840884731,
      "tokens_seen": 3033792512
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.218181818181817e-05,
      "loss": 1.4014,
      "theoretical_loss": 3.3179029107206484,
      "tokens_seen": 3033858048
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.216161616161616e-05,
      "loss": 1.4508,
      "theoretical_loss": 3.317897412746008,
      "tokens_seen": 3033923584
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.214141414141414e-05,
      "loss": 1.3538,
      "theoretical_loss": 3.317891914923381,
      "tokens_seen": 3033989120
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.212121212121212e-05,
      "loss": 1.3931,
      "theoretical_loss": 3.31788641725276,
      "tokens_seen": 3034054656
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.210101010101011e-05,
      "loss": 1.2847,
      "theoretical_loss": 3.3178809197341375,
      "tokens_seen": 3034120192
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.208080808080809e-05,
      "loss": 1.2195,
      "theoretical_loss": 3.317875422367506,
      "tokens_seen": 3034185728
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.206060606060606e-05,
      "loss": 1.4484,
      "theoretical_loss": 3.3178699251528583,
      "tokens_seen": 3034251264
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.204040404040405e-05,
      "loss": 1.4001,
      "theoretical_loss": 3.3178644280901866,
      "tokens_seen": 3034316800
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.202020202020202e-05,
      "loss": 1.366,
      "theoretical_loss": 3.3178589311794826,
      "tokens_seen": 3034382336
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.2e-05,
      "loss": 1.5461,
      "theoretical_loss": 3.3178534344207407,
      "tokens_seen": 3034447872
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.197979797979798e-05,
      "loss": 1.3647,
      "theoretical_loss": 3.3178479378139514,
      "tokens_seen": 3034513408
    },
    {
      "epoch": 0.84,
      "objective/train/aligned_token_count": 0.5421554446220398,
      "objective/train/avg_token_score": 0.008277331478893757,
      "objective/train/docs_used": 1706768,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.736639380455017,
      "objective/train/original_loss": 4.100271701812744,
      "objective/train/theoretical_loss": 3.317842441359109,
      "objective/train/tokens_used": 1393438176,
      "objective/train/unlikelihood_loss": 0.03881938010454178,
      "theoretical_loss": 3.317842441359109,
      "tokens_seen": 3034578944
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.195959595959596e-05,
      "loss": 1.3871,
      "theoretical_loss": 3.317842441359109,
      "tokens_seen": 3034578944
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.193939393939394e-05,
      "loss": 1.3346,
      "theoretical_loss": 3.3178369450562046,
      "tokens_seen": 3034644480
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.191919191919192e-05,
      "loss": 1.2978,
      "theoretical_loss": 3.3178314489052316,
      "tokens_seen": 3034710016
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.189898989898991e-05,
      "loss": 1.5175,
      "theoretical_loss": 3.3178259529061824,
      "tokens_seen": 3034775552
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.187878787878789e-05,
      "loss": 1.6673,
      "theoretical_loss": 3.317820457059049,
      "tokens_seen": 3034841088
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.185858585858585e-05,
      "loss": 1.2754,
      "theoretical_loss": 3.3178149613638244,
      "tokens_seen": 3034906624
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.183838383838383e-05,
      "loss": 1.3461,
      "theoretical_loss": 3.3178094658205013,
      "tokens_seen": 3034972160
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.181818181818182e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.3178039704290714,
      "tokens_seen": 3035037696
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.17979797979798e-05,
      "loss": 1.5586,
      "theoretical_loss": 3.3177984751895284,
      "tokens_seen": 3035103232
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.177777777777778e-05,
      "loss": 1.1932,
      "theoretical_loss": 3.3177929801018635,
      "tokens_seen": 3035168768
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.175757575757575e-05,
      "loss": 1.5544,
      "theoretical_loss": 3.31778748516607,
      "tokens_seen": 3035234304
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.173737373737374e-05,
      "loss": 1.432,
      "theoretical_loss": 3.317781990382141,
      "tokens_seen": 3035299840
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.171717171717172e-05,
      "loss": 1.5407,
      "theoretical_loss": 3.3177764957500675,
      "tokens_seen": 3035365376
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.16969696969697e-05,
      "loss": 1.218,
      "theoretical_loss": 3.3177710012698434,
      "tokens_seen": 3035430912
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.167676767676768e-05,
      "loss": 1.2804,
      "theoretical_loss": 3.31776550694146,
      "tokens_seen": 3035496448
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.165656565656566e-05,
      "loss": 1.6973,
      "theoretical_loss": 3.317760012764911,
      "tokens_seen": 3035561984
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.163636363636363e-05,
      "loss": 1.4782,
      "theoretical_loss": 3.3177545187401885,
      "tokens_seen": 3035627520
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.161616161616161e-05,
      "loss": 1.4346,
      "theoretical_loss": 3.317749024867285,
      "tokens_seen": 3035693056
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.15959595959596e-05,
      "loss": 1.3392,
      "theoretical_loss": 3.317743531146193,
      "tokens_seen": 3035758592
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.157575757575758e-05,
      "loss": 1.647,
      "theoretical_loss": 3.317738037576905,
      "tokens_seen": 3035824128
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.155555555555557e-05,
      "loss": 1.5835,
      "theoretical_loss": 3.3177325441594134,
      "tokens_seen": 3035889664
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.153535353535354e-05,
      "loss": 1.3273,
      "theoretical_loss": 3.317727050893711,
      "tokens_seen": 3035955200
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.151515151515152e-05,
      "loss": 1.2775,
      "theoretical_loss": 3.31772155777979,
      "tokens_seen": 3036020736
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.149494949494949e-05,
      "loss": 1.4419,
      "theoretical_loss": 3.317716064817643,
      "tokens_seen": 3036086272
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.147474747474748e-05,
      "loss": 1.299,
      "theoretical_loss": 3.3177105720072633,
      "tokens_seen": 3036151808
    },
    {
      "epoch": 0.84,
      "objective/train/aligned_token_count": 0.48918622732162476,
      "objective/train/avg_token_score": 0.00501604937016964,
      "objective/train/docs_used": 1707993,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4733476638793945,
      "objective/train/original_loss": 4.019478797912598,
      "objective/train/theoretical_loss": 3.3177050793486424,
      "objective/train/tokens_used": 1395076576,
      "objective/train/unlikelihood_loss": 0.05594341456890106,
      "theoretical_loss": 3.3177050793486424,
      "tokens_seen": 3036217344
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.145454545454546e-05,
      "loss": 1.4319,
      "theoretical_loss": 3.3177050793486424,
      "tokens_seen": 3036217344
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.143434343434343e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.317699586841773,
      "tokens_seen": 3036282880
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.141414141414141e-05,
      "loss": 1.3501,
      "theoretical_loss": 3.317694094486648,
      "tokens_seen": 3036348416
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.13939393939394e-05,
      "loss": 1.3997,
      "theoretical_loss": 3.31768860228326,
      "tokens_seen": 3036413952
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.137373737373738e-05,
      "loss": 1.386,
      "theoretical_loss": 3.3176831102316013,
      "tokens_seen": 3036479488
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.135353535353537e-05,
      "loss": 1.2998,
      "theoretical_loss": 3.3176776183316647,
      "tokens_seen": 3036545024
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.133333333333332e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.3176721265834423,
      "tokens_seen": 3036610560
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.13131313131313e-05,
      "loss": 1.3625,
      "theoretical_loss": 3.3176666349869266,
      "tokens_seen": 3036676096
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.129292929292929e-05,
      "loss": 1.5149,
      "theoretical_loss": 3.3176611435421104,
      "tokens_seen": 3036741632
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.127272727272727e-05,
      "loss": 1.5094,
      "theoretical_loss": 3.317655652248986,
      "tokens_seen": 3036807168
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.125252525252526e-05,
      "loss": 1.274,
      "theoretical_loss": 3.317650161107547,
      "tokens_seen": 3036872704
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.123232323232324e-05,
      "loss": 1.376,
      "theoretical_loss": 3.3176446701177844,
      "tokens_seen": 3036938240
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.121212121212121e-05,
      "loss": 1.303,
      "theoretical_loss": 3.3176391792796913,
      "tokens_seen": 3037003776
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.11919191919192e-05,
      "loss": 1.4651,
      "theoretical_loss": 3.3176336885932605,
      "tokens_seen": 3037069312
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.117171717171718e-05,
      "loss": 1.4232,
      "theoretical_loss": 3.3176281980584847,
      "tokens_seen": 3037134848
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.115151515151515e-05,
      "loss": 1.2718,
      "theoretical_loss": 3.3176227076753557,
      "tokens_seen": 3037200384
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.113131313131314e-05,
      "loss": 1.1057,
      "theoretical_loss": 3.3176172174438667,
      "tokens_seen": 3037265920
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.11111111111111e-05,
      "loss": 1.4547,
      "theoretical_loss": 3.31761172736401,
      "tokens_seen": 3037331456
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.109090909090909e-05,
      "loss": 1.4156,
      "theoretical_loss": 3.317606237435778,
      "tokens_seen": 3037396992
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.107070707070707e-05,
      "loss": 1.2341,
      "theoretical_loss": 3.3176007476591636,
      "tokens_seen": 3037462528
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.105050505050506e-05,
      "loss": 1.2717,
      "theoretical_loss": 3.3175952580341592,
      "tokens_seen": 3037528064
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.103030303030304e-05,
      "loss": 1.2367,
      "theoretical_loss": 3.317589768560757,
      "tokens_seen": 3037593600
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.101010101010101e-05,
      "loss": 1.1897,
      "theoretical_loss": 3.3175842792389503,
      "tokens_seen": 3037659136
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.098989898989898e-05,
      "loss": 1.1888,
      "theoretical_loss": 3.3175787900687306,
      "tokens_seen": 3037724672
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.096969696969697e-05,
      "loss": 1.4251,
      "theoretical_loss": 3.317573301050091,
      "tokens_seen": 3037790208
    },
    {
      "epoch": 0.84,
      "objective/train/aligned_token_count": 0.3750610947608948,
      "objective/train/avg_token_score": 0.02307865023612976,
      "objective/train/docs_used": 1708712,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0759658813476562,
      "objective/train/original_loss": 4.243686676025391,
      "objective/train/theoretical_loss": 3.3175678121830243,
      "objective/train/tokens_used": 1396714976,
      "objective/train/unlikelihood_loss": 0.08723418414592743,
      "theoretical_loss": 3.3175678121830243,
      "tokens_seen": 3037855744
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.094949494949495e-05,
      "loss": 1.3733,
      "theoretical_loss": 3.3175678121830243,
      "tokens_seen": 3037855744
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.092929292929293e-05,
      "loss": 1.1927,
      "theoretical_loss": 3.3175623234675227,
      "tokens_seen": 3037921280
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.09090909090909e-05,
      "loss": 1.424,
      "theoretical_loss": 3.317556834903579,
      "tokens_seen": 3037986816
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.088888888888889e-05,
      "loss": 1.3884,
      "theoretical_loss": 3.3175513464911854,
      "tokens_seen": 3038052352
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.086868686868687e-05,
      "loss": 1.317,
      "theoretical_loss": 3.3175458582303348,
      "tokens_seen": 3038117888
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.084848484848486e-05,
      "loss": 1.254,
      "theoretical_loss": 3.3175403701210193,
      "tokens_seen": 3038183424
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.082828282828284e-05,
      "loss": 1.2213,
      "theoretical_loss": 3.317534882163232,
      "tokens_seen": 3038248960
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.080808080808081e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.317529394356965,
      "tokens_seen": 3038314496
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.078787878787878e-05,
      "loss": 1.6312,
      "theoretical_loss": 3.317523906702211,
      "tokens_seen": 3038380032
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.076767676767677e-05,
      "loss": 1.301,
      "theoretical_loss": 3.3175184191989624,
      "tokens_seen": 3038445568
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.074747474747475e-05,
      "loss": 1.3727,
      "theoretical_loss": 3.3175129318472125,
      "tokens_seen": 3038511104
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.072727272727273e-05,
      "loss": 1.349,
      "theoretical_loss": 3.3175074446469526,
      "tokens_seen": 3038576640
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.070707070707072e-05,
      "loss": 1.2291,
      "theoretical_loss": 3.3175019575981763,
      "tokens_seen": 3038642176
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.068686868686869e-05,
      "loss": 1.4142,
      "theoretical_loss": 3.317496470700876,
      "tokens_seen": 3038707712
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.066666666666667e-05,
      "loss": 1.5067,
      "theoretical_loss": 3.3174909839550435,
      "tokens_seen": 3038773248
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.064646464646464e-05,
      "loss": 1.2141,
      "theoretical_loss": 3.317485497360672,
      "tokens_seen": 3038838784
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.062626262626263e-05,
      "loss": 1.5122,
      "theoretical_loss": 3.317480010917754,
      "tokens_seen": 3038904320
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.060606060606061e-05,
      "loss": 1.5106,
      "theoretical_loss": 3.3174745246262822,
      "tokens_seen": 3038969856
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.058585858585858e-05,
      "loss": 1.1245,
      "theoretical_loss": 3.3174690384862484,
      "tokens_seen": 3039035392
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.056565656565656e-05,
      "loss": 1.6288,
      "theoretical_loss": 3.317463552497646,
      "tokens_seen": 3039100928
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.054545454545455e-05,
      "loss": 1.3189,
      "theoretical_loss": 3.3174580666604676,
      "tokens_seen": 3039166464
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.052525252525253e-05,
      "loss": 1.3973,
      "theoretical_loss": 3.317452580974705,
      "tokens_seen": 3039232000
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.050505050505052e-05,
      "loss": 1.3677,
      "theoretical_loss": 3.317447095440351,
      "tokens_seen": 3039297536
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.048484848484849e-05,
      "loss": 1.313,
      "theoretical_loss": 3.3174416100573985,
      "tokens_seen": 3039363072
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.046464646464646e-05,
      "loss": 1.4393,
      "theoretical_loss": 3.31743612482584,
      "tokens_seen": 3039428608
    },
    {
      "epoch": 0.84,
      "objective/train/aligned_token_count": 0.2927052676677704,
      "objective/train/avg_token_score": 0.04868728294968605,
      "objective/train/docs_used": 1709972,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8954604864120483,
      "objective/train/original_loss": 5.026909828186035,
      "objective/train/theoretical_loss": 3.317430639745668,
      "objective/train/tokens_used": 1398353376,
      "objective/train/unlikelihood_loss": 0.05441543459892273,
      "theoretical_loss": 3.317430639745668,
      "tokens_seen": 3039494144
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.044444444444444e-05,
      "loss": 1.3243,
      "theoretical_loss": 3.317430639745668,
      "tokens_seen": 3039494144
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.042424242424242e-05,
      "loss": 1.382,
      "theoretical_loss": 3.3174251548168745,
      "tokens_seen": 3039559680
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.040404040404041e-05,
      "loss": 1.126,
      "theoretical_loss": 3.317419670039453,
      "tokens_seen": 3039625216
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.038383838383839e-05,
      "loss": 1.2981,
      "theoretical_loss": 3.317414185413395,
      "tokens_seen": 3039690752
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.036363636363636e-05,
      "loss": 1.4811,
      "theoretical_loss": 3.317408700938694,
      "tokens_seen": 3039756288
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.034343434343435e-05,
      "loss": 1.2479,
      "theoretical_loss": 3.3174032166153427,
      "tokens_seen": 3039821824
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.032323232323233e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.317397732443333,
      "tokens_seen": 3039887360
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.030303030303031e-05,
      "loss": 1.3703,
      "theoretical_loss": 3.317392248422657,
      "tokens_seen": 3039952896
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.028282828282829e-05,
      "loss": 1.3631,
      "theoretical_loss": 3.317386764553308,
      "tokens_seen": 3040018432
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.026262626262626e-05,
      "loss": 1.4081,
      "theoretical_loss": 3.317381280835279,
      "tokens_seen": 3040083968
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.024242424242424e-05,
      "loss": 1.2965,
      "theoretical_loss": 3.317375797268562,
      "tokens_seen": 3040149504
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.022222222222222e-05,
      "loss": 1.4724,
      "theoretical_loss": 3.317370313853149,
      "tokens_seen": 3040215040
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.020202020202021e-05,
      "loss": 1.3121,
      "theoretical_loss": 3.317364830589034,
      "tokens_seen": 3040280576
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.018181818181819e-05,
      "loss": 1.3115,
      "theoretical_loss": 3.317359347476208,
      "tokens_seen": 3040346112
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.016161616161616e-05,
      "loss": 1.3808,
      "theoretical_loss": 3.3173538645146645,
      "tokens_seen": 3040411648
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.014141414141415e-05,
      "loss": 1.4078,
      "theoretical_loss": 3.317348381704396,
      "tokens_seen": 3040477184
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.012121212121212e-05,
      "loss": 1.2136,
      "theoretical_loss": 3.3173428990453946,
      "tokens_seen": 3040542720
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.01010101010101e-05,
      "loss": 1.3383,
      "theoretical_loss": 3.3173374165376535,
      "tokens_seen": 3040608256
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.008080808080808e-05,
      "loss": 1.445,
      "theoretical_loss": 3.3173319341811647,
      "tokens_seen": 3040673792
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.006060606060605e-05,
      "loss": 1.2799,
      "theoretical_loss": 3.317326451975921,
      "tokens_seen": 3040739328
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.004040404040404e-05,
      "loss": 1.3623,
      "theoretical_loss": 3.3173209699219153,
      "tokens_seen": 3040804864
    },
    {
      "epoch": 0.84,
      "learning_rate": 8.002020202020202e-05,
      "loss": 1.4015,
      "theoretical_loss": 3.3173154880191396,
      "tokens_seen": 3040870400
    },
    {
      "epoch": 0.84,
      "learning_rate": 8e-05,
      "loss": 1.2398,
      "theoretical_loss": 3.317310006267587,
      "tokens_seen": 3040935936
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.997979797979799e-05,
      "loss": 1.1236,
      "theoretical_loss": 3.3173045246672492,
      "tokens_seen": 3041001472
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.995959595959596e-05,
      "loss": 1.5168,
      "theoretical_loss": 3.31729904321812,
      "tokens_seen": 3041067008
    },
    {
      "epoch": 0.84,
      "objective/train/aligned_token_count": 0.35728248953819275,
      "objective/train/avg_token_score": 0.052419263869524,
      "objective/train/docs_used": 1710687,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1120922565460205,
      "objective/train/original_loss": 4.65952730178833,
      "objective/train/theoretical_loss": 3.317293561920191,
      "objective/train/tokens_used": 1399991776,
      "objective/train/unlikelihood_loss": 0.05755076929926872,
      "theoretical_loss": 3.317293561920191,
      "tokens_seen": 3041132544
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.993939393939393e-05,
      "loss": 1.204,
      "theoretical_loss": 3.317293561920191,
      "tokens_seen": 3041132544
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.991919191919192e-05,
      "loss": 1.6411,
      "theoretical_loss": 3.317288080773455,
      "tokens_seen": 3041198080
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.98989898989899e-05,
      "loss": 1.3083,
      "theoretical_loss": 3.317282599777905,
      "tokens_seen": 3041263616
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.987878787878788e-05,
      "loss": 1.4588,
      "theoretical_loss": 3.317277118933533,
      "tokens_seen": 3041329152
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.985858585858587e-05,
      "loss": 1.4528,
      "theoretical_loss": 3.317271638240332,
      "tokens_seen": 3041394688
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.983838383838384e-05,
      "loss": 1.2884,
      "theoretical_loss": 3.317266157698294,
      "tokens_seen": 3041460224
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.981818181818182e-05,
      "loss": 1.1416,
      "theoretical_loss": 3.3172606773074125,
      "tokens_seen": 3041525760
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.97979797979798e-05,
      "loss": 1.2402,
      "theoretical_loss": 3.317255197067679,
      "tokens_seen": 3041591296
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.977777777777778e-05,
      "loss": 1.2001,
      "theoretical_loss": 3.317249716979087,
      "tokens_seen": 3041656832
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.975757575757576e-05,
      "loss": 1.3442,
      "theoretical_loss": 3.3172442370416286,
      "tokens_seen": 3041722368
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.973737373737373e-05,
      "loss": 1.2205,
      "theoretical_loss": 3.3172387572552964,
      "tokens_seen": 3041787904
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.971717171717171e-05,
      "loss": 1.1834,
      "theoretical_loss": 3.3172332776200832,
      "tokens_seen": 3041853440
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.96969696969697e-05,
      "loss": 1.3921,
      "theoretical_loss": 3.3172277981359812,
      "tokens_seen": 3041918976
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.967676767676768e-05,
      "loss": 1.3414,
      "theoretical_loss": 3.3172223188029832,
      "tokens_seen": 3041984512
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.965656565656567e-05,
      "loss": 1.2131,
      "theoretical_loss": 3.317216839621082,
      "tokens_seen": 3042050048
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.963636363636364e-05,
      "loss": 1.3065,
      "theoretical_loss": 3.31721136059027,
      "tokens_seen": 3042115584
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.961616161616162e-05,
      "loss": 1.2045,
      "theoretical_loss": 3.317205881710539,
      "tokens_seen": 3042181120
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.959595959595959e-05,
      "loss": 1.1959,
      "theoretical_loss": 3.317200402981883,
      "tokens_seen": 3042246656
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.957575757575757e-05,
      "loss": 1.1286,
      "theoretical_loss": 3.3171949244042938,
      "tokens_seen": 3042312192
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.955555555555556e-05,
      "loss": 1.3535,
      "theoretical_loss": 3.317189445977764,
      "tokens_seen": 3042377728
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.953535353535353e-05,
      "loss": 1.3665,
      "theoretical_loss": 3.3171839677022863,
      "tokens_seen": 3042443264
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.951515151515151e-05,
      "loss": 1.344,
      "theoretical_loss": 3.317178489577853,
      "tokens_seen": 3042508800
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.94949494949495e-05,
      "loss": 1.1376,
      "theoretical_loss": 3.3171730116044573,
      "tokens_seen": 3042574336
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.947474747474748e-05,
      "loss": 1.4181,
      "theoretical_loss": 3.317167533782091,
      "tokens_seen": 3042639872
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.945454545454547e-05,
      "loss": 1.4564,
      "theoretical_loss": 3.3171620561107473,
      "tokens_seen": 3042705408
    },
    {
      "epoch": 0.84,
      "objective/train/aligned_token_count": 0.4305351972579956,
      "objective/train/avg_token_score": 0.013207322917878628,
      "objective/train/docs_used": 1712243,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2599613666534424,
      "objective/train/original_loss": 4.030418872833252,
      "objective/train/theoretical_loss": 3.3171565785904185,
      "objective/train/tokens_used": 1401630176,
      "objective/train/unlikelihood_loss": 0.07559382915496826,
      "theoretical_loss": 3.3171565785904185,
      "tokens_seen": 3042770944
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.943434343434345e-05,
      "loss": 1.2427,
      "theoretical_loss": 3.3171565785904185,
      "tokens_seen": 3042770944
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.94141414141414e-05,
      "loss": 1.3003,
      "theoretical_loss": 3.3171511012210972,
      "tokens_seen": 3042836480
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.939393939393939e-05,
      "loss": 1.4669,
      "theoretical_loss": 3.317145624002776,
      "tokens_seen": 3042902016
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.937373737373737e-05,
      "loss": 1.1483,
      "theoretical_loss": 3.317140146935448,
      "tokens_seen": 3042967552
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.935353535353536e-05,
      "loss": 1.3768,
      "theoretical_loss": 3.317134670019105,
      "tokens_seen": 3043033088
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.933333333333334e-05,
      "loss": 1.2737,
      "theoretical_loss": 3.3171291932537397,
      "tokens_seen": 3043098624
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.931313131313131e-05,
      "loss": 1.4096,
      "theoretical_loss": 3.317123716639345,
      "tokens_seen": 3043164160
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.92929292929293e-05,
      "loss": 1.3553,
      "theoretical_loss": 3.317118240175913,
      "tokens_seen": 3043229696
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.927272727272728e-05,
      "loss": 1.3059,
      "theoretical_loss": 3.3171127638634372,
      "tokens_seen": 3043295232
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.925252525252525e-05,
      "loss": 1.2469,
      "theoretical_loss": 3.3171072877019094,
      "tokens_seen": 3043360768
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.923232323232323e-05,
      "loss": 1.4722,
      "theoretical_loss": 3.3171018116913222,
      "tokens_seen": 3043426304
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.92121212121212e-05,
      "loss": 1.3132,
      "theoretical_loss": 3.317096335831669,
      "tokens_seen": 3043491840
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.919191919191919e-05,
      "loss": 1.3454,
      "theoretical_loss": 3.317090860122941,
      "tokens_seen": 3043557376
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.917171717171717e-05,
      "loss": 1.5216,
      "theoretical_loss": 3.3170853845651322,
      "tokens_seen": 3043622912
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.915151515151516e-05,
      "loss": 1.3251,
      "theoretical_loss": 3.3170799091582346,
      "tokens_seen": 3043688448
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.913131313131314e-05,
      "loss": 1.4909,
      "theoretical_loss": 3.3170744339022407,
      "tokens_seen": 3043753984
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.911111111111111e-05,
      "loss": 1.3206,
      "theoretical_loss": 3.317068958797143,
      "tokens_seen": 3043819520
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.90909090909091e-05,
      "loss": 1.3665,
      "theoretical_loss": 3.3170634838429347,
      "tokens_seen": 3043885056
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.907070707070707e-05,
      "loss": 1.483,
      "theoretical_loss": 3.3170580090396076,
      "tokens_seen": 3043950592
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.905050505050505e-05,
      "loss": 1.4397,
      "theoretical_loss": 3.3170525343871544,
      "tokens_seen": 3044016128
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.903030303030303e-05,
      "loss": 1.4179,
      "theoretical_loss": 3.317047059885568,
      "tokens_seen": 3044081664
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.901010101010102e-05,
      "loss": 1.3873,
      "theoretical_loss": 3.3170415855348416,
      "tokens_seen": 3044147200
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.898989898989899e-05,
      "loss": 1.4287,
      "theoretical_loss": 3.317036111334967,
      "tokens_seen": 3044212736
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.896969696969697e-05,
      "loss": 1.251,
      "theoretical_loss": 3.3170306372859364,
      "tokens_seen": 3044278272
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.894949494949496e-05,
      "loss": 1.219,
      "theoretical_loss": 3.317025163387743,
      "tokens_seen": 3044343808
    },
    {
      "epoch": 0.84,
      "objective/train/aligned_token_count": 0.388868510723114,
      "objective/train/avg_token_score": 0.024917105212807655,
      "objective/train/docs_used": 1712674,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.204413652420044,
      "objective/train/original_loss": 4.581943988800049,
      "objective/train/theoretical_loss": 3.3170196896403796,
      "objective/train/tokens_used": 1403268576,
      "objective/train/unlikelihood_loss": 0.06992856413125992,
      "theoretical_loss": 3.3170196896403796,
      "tokens_seen": 3044409344
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.892929292929294e-05,
      "loss": 1.3982,
      "theoretical_loss": 3.3170196896403796,
      "tokens_seen": 3044409344
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.890909090909091e-05,
      "loss": 1.3995,
      "theoretical_loss": 3.3170142160438383,
      "tokens_seen": 3044474880
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.888888888888888e-05,
      "loss": 1.3401,
      "theoretical_loss": 3.3170087425981123,
      "tokens_seen": 3044540416
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.886868686868686e-05,
      "loss": 1.3178,
      "theoretical_loss": 3.3170032693031937,
      "tokens_seen": 3044605952
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.884848484848485e-05,
      "loss": 1.4528,
      "theoretical_loss": 3.316997796159075,
      "tokens_seen": 3044671488
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.882828282828283e-05,
      "loss": 1.3605,
      "theoretical_loss": 3.3169923231657488,
      "tokens_seen": 3044737024
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.880808080808082e-05,
      "loss": 1.2419,
      "theoretical_loss": 3.3169868503232083,
      "tokens_seen": 3044802560
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.878787878787879e-05,
      "loss": 1.0038,
      "theoretical_loss": 3.3169813776314454,
      "tokens_seen": 3044868096
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.876767676767677e-05,
      "loss": 1.1549,
      "theoretical_loss": 3.3169759050904535,
      "tokens_seen": 3044933632
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.874747474747475e-05,
      "loss": 1.2665,
      "theoretical_loss": 3.3169704327002245,
      "tokens_seen": 3044999168
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.872727272727273e-05,
      "loss": 1.3479,
      "theoretical_loss": 3.316964960460751,
      "tokens_seen": 3045064704
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.870707070707071e-05,
      "loss": 1.1941,
      "theoretical_loss": 3.316959488372026,
      "tokens_seen": 3045130240
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.868686868686868e-05,
      "loss": 1.2203,
      "theoretical_loss": 3.3169540164340416,
      "tokens_seen": 3045195776
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.866666666666666e-05,
      "loss": 1.1677,
      "theoretical_loss": 3.3169485446467912,
      "tokens_seen": 3045261312
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.864646464646465e-05,
      "loss": 1.4794,
      "theoretical_loss": 3.316943073010267,
      "tokens_seen": 3045326848
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.862626262626263e-05,
      "loss": 1.4053,
      "theoretical_loss": 3.316937601524461,
      "tokens_seen": 3045392384
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.860606060606062e-05,
      "loss": 1.3158,
      "theoretical_loss": 3.316932130189367,
      "tokens_seen": 3045457920
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.85858585858586e-05,
      "loss": 1.393,
      "theoretical_loss": 3.3169266590049764,
      "tokens_seen": 3045523456
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.856565656565656e-05,
      "loss": 1.3666,
      "theoretical_loss": 3.3169211879712828,
      "tokens_seen": 3045588992
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.854545454545454e-05,
      "loss": 1.2959,
      "theoretical_loss": 3.316915717088278,
      "tokens_seen": 3045654528
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.852525252525252e-05,
      "loss": 1.4835,
      "theoretical_loss": 3.316910246355955,
      "tokens_seen": 3045720064
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.850505050505051e-05,
      "loss": 1.3528,
      "theoretical_loss": 3.316904775774306,
      "tokens_seen": 3045785600
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.848484848484849e-05,
      "loss": 1.449,
      "theoretical_loss": 3.316899305343325,
      "tokens_seen": 3045851136
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.846464646464646e-05,
      "loss": 1.6076,
      "theoretical_loss": 3.316893835063003,
      "tokens_seen": 3045916672
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.844444444444445e-05,
      "loss": 1.3625,
      "theoretical_loss": 3.316888364933333,
      "tokens_seen": 3045982208
    },
    {
      "epoch": 0.84,
      "objective/train/aligned_token_count": 0.48747557401657104,
      "objective/train/avg_token_score": 0.013353968970477581,
      "objective/train/docs_used": 1713719,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6127694845199585,
      "objective/train/original_loss": 4.13460636138916,
      "objective/train/theoretical_loss": 3.3168828949543085,
      "objective/train/tokens_used": 1404906976,
      "objective/train/unlikelihood_loss": 0.0662284716963768,
      "theoretical_loss": 3.3168828949543085,
      "tokens_seen": 3046047744
    },
    {
      "epoch": 0.84,
      "learning_rate": 7.842424242424243e-05,
      "loss": 1.4257,
      "theoretical_loss": 3.3168828949543085,
      "tokens_seen": 3046047744
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.840404040404041e-05,
      "loss": 1.3492,
      "theoretical_loss": 3.316877425125921,
      "tokens_seen": 3046113280
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.838383838383838e-05,
      "loss": 1.594,
      "theoretical_loss": 3.316871955448163,
      "tokens_seen": 3046178816
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.836363636363636e-05,
      "loss": 1.495,
      "theoretical_loss": 3.3168664859210284,
      "tokens_seen": 3046244352
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.834343434343434e-05,
      "loss": 1.4053,
      "theoretical_loss": 3.316861016544509,
      "tokens_seen": 3046309888
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.832323232323232e-05,
      "loss": 1.4702,
      "theoretical_loss": 3.316855547318597,
      "tokens_seen": 3046375424
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.830303030303031e-05,
      "loss": 1.5495,
      "theoretical_loss": 3.316850078243286,
      "tokens_seen": 3046440960
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.828282828282829e-05,
      "loss": 1.0164,
      "theoretical_loss": 3.3168446093185677,
      "tokens_seen": 3046506496
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.826262626262626e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.3168391405444355,
      "tokens_seen": 3046572032
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.824242424242425e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.316833671920881,
      "tokens_seen": 3046637568
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.822222222222222e-05,
      "loss": 1.2868,
      "theoretical_loss": 3.3168282034478977,
      "tokens_seen": 3046703104
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.82020202020202e-05,
      "loss": 1.2339,
      "theoretical_loss": 3.3168227351254784,
      "tokens_seen": 3046768640
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.818181818181818e-05,
      "loss": 1.5859,
      "theoretical_loss": 3.316817266953615,
      "tokens_seen": 3046834176
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.816161616161617e-05,
      "loss": 1.3441,
      "theoretical_loss": 3.3168117989323003,
      "tokens_seen": 3046899712
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.814141414141414e-05,
      "loss": 1.5056,
      "theoretical_loss": 3.316806331061527,
      "tokens_seen": 3046965248
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.812121212121212e-05,
      "loss": 1.3559,
      "theoretical_loss": 3.316800863341288,
      "tokens_seen": 3047030784
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.81010101010101e-05,
      "loss": 1.4567,
      "theoretical_loss": 3.316795395771575,
      "tokens_seen": 3047096320
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.808080808080809e-05,
      "loss": 1.3159,
      "theoretical_loss": 3.3167899283523816,
      "tokens_seen": 3047161856
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.806060606060607e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.3167844610837003,
      "tokens_seen": 3047227392
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.804040404040403e-05,
      "loss": 1.4661,
      "theoretical_loss": 3.3167789939655234,
      "tokens_seen": 3047292928
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.802020202020201e-05,
      "loss": 1.3961,
      "theoretical_loss": 3.3167735269978436,
      "tokens_seen": 3047358464
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.8e-05,
      "loss": 1.5238,
      "theoretical_loss": 3.3167680601806535,
      "tokens_seen": 3047424000
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.797979797979798e-05,
      "loss": 1.5886,
      "theoretical_loss": 3.3167625935139458,
      "tokens_seen": 3047489536
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.795959595959597e-05,
      "loss": 1.403,
      "theoretical_loss": 3.3167571269977127,
      "tokens_seen": 3047555072
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.793939393939394e-05,
      "loss": 1.4965,
      "theoretical_loss": 3.316751660631948,
      "tokens_seen": 3047620608
    },
    {
      "epoch": 0.85,
      "objective/train/aligned_token_count": 0.3328445851802826,
      "objective/train/avg_token_score": 0.05854041874408722,
      "objective/train/docs_used": 1714459,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.095767617225647,
      "objective/train/original_loss": 5.026232719421387,
      "objective/train/theoretical_loss": 3.316746194416643,
      "objective/train/tokens_used": 1406545376,
      "objective/train/unlikelihood_loss": 0.04737464711070061,
      "theoretical_loss": 3.316746194416643,
      "tokens_seen": 3047686144
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.791919191919192e-05,
      "loss": 1.256,
      "theoretical_loss": 3.316746194416643,
      "tokens_seen": 3047686144
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.78989898989899e-05,
      "loss": 1.2613,
      "theoretical_loss": 3.3167407283517907,
      "tokens_seen": 3047751680
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.787878787878789e-05,
      "loss": 1.3093,
      "theoretical_loss": 3.3167352624373843,
      "tokens_seen": 3047817216
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.785858585858586e-05,
      "loss": 1.179,
      "theoretical_loss": 3.3167297966734157,
      "tokens_seen": 3047882752
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.783838383838383e-05,
      "loss": 1.4246,
      "theoretical_loss": 3.316724331059878,
      "tokens_seen": 3047948288
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.781818181818181e-05,
      "loss": 1.6001,
      "theoretical_loss": 3.316718865596763,
      "tokens_seen": 3048013824
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.77979797979798e-05,
      "loss": 1.3492,
      "theoretical_loss": 3.3167134002840646,
      "tokens_seen": 3048079360
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.777777777777778e-05,
      "loss": 1.3235,
      "theoretical_loss": 3.3167079351217748,
      "tokens_seen": 3048144896
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.775757575757577e-05,
      "loss": 1.4269,
      "theoretical_loss": 3.3167024701098864,
      "tokens_seen": 3048210432
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.773737373737375e-05,
      "loss": 1.3738,
      "theoretical_loss": 3.3166970052483915,
      "tokens_seen": 3048275968
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.771717171717172e-05,
      "loss": 1.3822,
      "theoretical_loss": 3.316691540537283,
      "tokens_seen": 3048341504
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.769696969696969e-05,
      "loss": 1.5278,
      "theoretical_loss": 3.3166860759765537,
      "tokens_seen": 3048407040
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.767676767676767e-05,
      "loss": 1.4047,
      "theoretical_loss": 3.3166806115661966,
      "tokens_seen": 3048472576
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.765656565656566e-05,
      "loss": 1.1002,
      "theoretical_loss": 3.3166751473062033,
      "tokens_seen": 3048538112
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.763636363636364e-05,
      "loss": 1.6757,
      "theoretical_loss": 3.3166696831965674,
      "tokens_seen": 3048603648
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.761616161616161e-05,
      "loss": 1.2331,
      "theoretical_loss": 3.316664219237281,
      "tokens_seen": 3048669184
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.75959595959596e-05,
      "loss": 1.1565,
      "theoretical_loss": 3.3166587554283367,
      "tokens_seen": 3048734720
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.757575757575758e-05,
      "loss": 1.298,
      "theoretical_loss": 3.316653291769727,
      "tokens_seen": 3048800256
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.755555555555556e-05,
      "loss": 1.3658,
      "theoretical_loss": 3.3166478282614458,
      "tokens_seen": 3048865792
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.753535353535355e-05,
      "loss": 1.295,
      "theoretical_loss": 3.3166423649034837,
      "tokens_seen": 3048931328
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.75151515151515e-05,
      "loss": 1.4648,
      "theoretical_loss": 3.3166369016958352,
      "tokens_seen": 3048996864
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.749494949494949e-05,
      "loss": 1.3266,
      "theoretical_loss": 3.316631438638492,
      "tokens_seen": 3049062400
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.747474747474747e-05,
      "loss": 1.1863,
      "theoretical_loss": 3.316625975731446,
      "tokens_seen": 3049127936
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.745454545454546e-05,
      "loss": 1.2849,
      "theoretical_loss": 3.3166205129746915,
      "tokens_seen": 3049193472
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.743434343434344e-05,
      "loss": 1.1892,
      "theoretical_loss": 3.3166150503682204,
      "tokens_seen": 3049259008
    },
    {
      "epoch": 0.85,
      "objective/train/aligned_token_count": 0.3598484992980957,
      "objective/train/avg_token_score": 0.01566394418478012,
      "objective/train/docs_used": 1714997,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0930602550506592,
      "objective/train/original_loss": 4.263375282287598,
      "objective/train/theoretical_loss": 3.316609587912025,
      "objective/train/tokens_used": 1408183776,
      "objective/train/unlikelihood_loss": 0.06456009298563004,
      "theoretical_loss": 3.316609587912025,
      "tokens_seen": 3049324544
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.741414141414141e-05,
      "loss": 1.359,
      "theoretical_loss": 3.316609587912025,
      "tokens_seen": 3049324544
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.73939393939394e-05,
      "loss": 1.4098,
      "theoretical_loss": 3.3166041256060983,
      "tokens_seen": 3049390080
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.737373737373738e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.3165986634504327,
      "tokens_seen": 3049455616
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.735353535353535e-05,
      "loss": 1.4309,
      "theoretical_loss": 3.316593201445021,
      "tokens_seen": 3049521152
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.733333333333333e-05,
      "loss": 1.4826,
      "theoretical_loss": 3.3165877395898558,
      "tokens_seen": 3049586688
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.731313131313132e-05,
      "loss": 1.4674,
      "theoretical_loss": 3.3165822778849297,
      "tokens_seen": 3049652224
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.729292929292929e-05,
      "loss": 1.5576,
      "theoretical_loss": 3.316576816330236,
      "tokens_seen": 3049717760
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.727272727272727e-05,
      "loss": 1.3777,
      "theoretical_loss": 3.316571354925766,
      "tokens_seen": 3049783296
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.725252525252526e-05,
      "loss": 1.3671,
      "theoretical_loss": 3.3165658936715134,
      "tokens_seen": 3049848832
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.723232323232324e-05,
      "loss": 1.4957,
      "theoretical_loss": 3.3165604325674702,
      "tokens_seen": 3049914368
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.721212121212122e-05,
      "loss": 1.1804,
      "theoretical_loss": 3.3165549716136296,
      "tokens_seen": 3049979904
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.71919191919192e-05,
      "loss": 1.3757,
      "theoretical_loss": 3.316549510809984,
      "tokens_seen": 3050045440
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.717171717171716e-05,
      "loss": 1.1465,
      "theoretical_loss": 3.316544050156526,
      "tokens_seen": 3050110976
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.715151515151515e-05,
      "loss": 1.3698,
      "theoretical_loss": 3.3165385896532484,
      "tokens_seen": 3050176512
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.713131313131313e-05,
      "loss": 1.417,
      "theoretical_loss": 3.3165331293001437,
      "tokens_seen": 3050242048
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.711111111111112e-05,
      "loss": 1.1906,
      "theoretical_loss": 3.316527669097204,
      "tokens_seen": 3050307584
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.709090909090909e-05,
      "loss": 1.4999,
      "theoretical_loss": 3.3165222090444235,
      "tokens_seen": 3050373120
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.707070707070707e-05,
      "loss": 1.4178,
      "theoretical_loss": 3.316516749141793,
      "tokens_seen": 3050438656
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.705050505050506e-05,
      "loss": 1.2671,
      "theoretical_loss": 3.3165112893893065,
      "tokens_seen": 3050504192
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.703030303030304e-05,
      "loss": 1.2642,
      "theoretical_loss": 3.316505829786956,
      "tokens_seen": 3050569728
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.701010101010102e-05,
      "loss": 1.1909,
      "theoretical_loss": 3.316500370334734,
      "tokens_seen": 3050635264
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.698989898989898e-05,
      "loss": 1.417,
      "theoretical_loss": 3.316494911032634,
      "tokens_seen": 3050700800
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.696969696969696e-05,
      "loss": 1.1565,
      "theoretical_loss": 3.3164894518806474,
      "tokens_seen": 3050766336
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.694949494949495e-05,
      "loss": 1.3088,
      "theoretical_loss": 3.316483992878768,
      "tokens_seen": 3050831872
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.692929292929293e-05,
      "loss": 1.1588,
      "theoretical_loss": 3.316478534026988,
      "tokens_seen": 3050897408
    },
    {
      "epoch": 0.85,
      "objective/train/aligned_token_count": 0.45369014143943787,
      "objective/train/avg_token_score": 0.011538133956491947,
      "objective/train/docs_used": 1716070,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.427859902381897,
      "objective/train/original_loss": 3.9827561378479004,
      "objective/train/theoretical_loss": 3.3164730753253,
      "objective/train/tokens_used": 1409822176,
      "objective/train/unlikelihood_loss": 0.05690734088420868,
      "theoretical_loss": 3.3164730753253,
      "tokens_seen": 3050962944
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.690909090909092e-05,
      "loss": 1.3522,
      "theoretical_loss": 3.3164730753253,
      "tokens_seen": 3050962944
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.68888888888889e-05,
      "loss": 1.4908,
      "theoretical_loss": 3.3164676167736964,
      "tokens_seen": 3051028480
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.686868686868687e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.3164621583721705,
      "tokens_seen": 3051094016
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.684848484848485e-05,
      "loss": 1.3483,
      "theoretical_loss": 3.3164567001207144,
      "tokens_seen": 3051159552
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.682828282828282e-05,
      "loss": 1.3322,
      "theoretical_loss": 3.316451242019321,
      "tokens_seen": 3051225088
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.680808080808081e-05,
      "loss": 1.4378,
      "theoretical_loss": 3.316445784067983,
      "tokens_seen": 3051290624
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.678787878787879e-05,
      "loss": 1.5975,
      "theoretical_loss": 3.3164403262666924,
      "tokens_seen": 3051356160
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.676767676767676e-05,
      "loss": 1.294,
      "theoretical_loss": 3.316434868615443,
      "tokens_seen": 3051421696
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.674747474747475e-05,
      "loss": 1.2195,
      "theoretical_loss": 3.3164294111142265,
      "tokens_seen": 3051487232
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.672727272727273e-05,
      "loss": 1.3944,
      "theoretical_loss": 3.316423953763036,
      "tokens_seen": 3051552768
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.670707070707071e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.3164184965618637,
      "tokens_seen": 3051618304
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.66868686868687e-05,
      "loss": 1.2418,
      "theoretical_loss": 3.316413039510703,
      "tokens_seen": 3051683840
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.666666666666667e-05,
      "loss": 1.441,
      "theoretical_loss": 3.3164075826095463,
      "tokens_seen": 3051749376
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.664646464646464e-05,
      "loss": 1.4323,
      "theoretical_loss": 3.3164021258583856,
      "tokens_seen": 3051814912
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.662626262626262e-05,
      "loss": 1.3388,
      "theoretical_loss": 3.316396669257214,
      "tokens_seen": 3051880448
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.660606060606061e-05,
      "loss": 1.3699,
      "theoretical_loss": 3.3163912128060247,
      "tokens_seen": 3051945984
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.658585858585859e-05,
      "loss": 1.3906,
      "theoretical_loss": 3.31638575650481,
      "tokens_seen": 3052011520
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.656565656565656e-05,
      "loss": 1.5267,
      "theoretical_loss": 3.316380300353562,
      "tokens_seen": 3052077056
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.654545454545455e-05,
      "loss": 1.2665,
      "theoretical_loss": 3.316374844352274,
      "tokens_seen": 3052142592
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.652525252525253e-05,
      "loss": 1.2071,
      "theoretical_loss": 3.3163693885009384,
      "tokens_seen": 3052208128
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.650505050505051e-05,
      "loss": 1.3283,
      "theoretical_loss": 3.316363932799548,
      "tokens_seen": 3052273664
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.648484848484848e-05,
      "loss": 1.4059,
      "theoretical_loss": 3.3163584772480954,
      "tokens_seen": 3052339200
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.646464646464647e-05,
      "loss": 1.2657,
      "theoretical_loss": 3.3163530218465733,
      "tokens_seen": 3052404736
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.644444444444444e-05,
      "loss": 1.4175,
      "theoretical_loss": 3.316347566594974,
      "tokens_seen": 3052470272
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.642424242424242e-05,
      "loss": 1.1945,
      "theoretical_loss": 3.3163421114932907,
      "tokens_seen": 3052535808
    },
    {
      "epoch": 0.85,
      "objective/train/aligned_token_count": 0.45032989978790283,
      "objective/train/avg_token_score": 0.007256459444761276,
      "objective/train/docs_used": 1717617,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3248347043991089,
      "objective/train/original_loss": 3.9605352878570557,
      "objective/train/theoretical_loss": 3.3163366565415155,
      "objective/train/tokens_used": 1411460576,
      "objective/train/unlikelihood_loss": 0.05996003746986389,
      "theoretical_loss": 3.3163366565415155,
      "tokens_seen": 3052601344
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.64040404040404e-05,
      "loss": 1.4368,
      "theoretical_loss": 3.3163366565415155,
      "tokens_seen": 3052601344
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.638383838383839e-05,
      "loss": 1.2557,
      "theoretical_loss": 3.316331201739642,
      "tokens_seen": 3052666880
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.636363636363637e-05,
      "loss": 1.3179,
      "theoretical_loss": 3.3163257470876615,
      "tokens_seen": 3052732416
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.634343434343434e-05,
      "loss": 1.5317,
      "theoretical_loss": 3.3163202925855675,
      "tokens_seen": 3052797952
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.632323232323233e-05,
      "loss": 1.4161,
      "theoretical_loss": 3.316314838233353,
      "tokens_seen": 3052863488
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.63030303030303e-05,
      "loss": 1.5334,
      "theoretical_loss": 3.31630938403101,
      "tokens_seen": 3052929024
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.628282828282828e-05,
      "loss": 1.4999,
      "theoretical_loss": 3.3163039299785315,
      "tokens_seen": 3052994560
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.626262626262627e-05,
      "loss": 1.4281,
      "theoretical_loss": 3.31629847607591,
      "tokens_seen": 3053060096
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.624242424242424e-05,
      "loss": 1.6151,
      "theoretical_loss": 3.316293022323138,
      "tokens_seen": 3053125632
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.622222222222222e-05,
      "loss": 1.247,
      "theoretical_loss": 3.3162875687202087,
      "tokens_seen": 3053191168
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.62020202020202e-05,
      "loss": 1.3772,
      "theoretical_loss": 3.3162821152671142,
      "tokens_seen": 3053256704
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.618181818181819e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.316276661963848,
      "tokens_seen": 3053322240
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.616161616161617e-05,
      "loss": 1.2433,
      "theoretical_loss": 3.3162712088104014,
      "tokens_seen": 3053387776
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.614141414141413e-05,
      "loss": 1.3958,
      "theoretical_loss": 3.3162657558067687,
      "tokens_seen": 3053453312
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.612121212121211e-05,
      "loss": 1.433,
      "theoretical_loss": 3.316260302952941,
      "tokens_seen": 3053518848
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.61010101010101e-05,
      "loss": 1.4175,
      "theoretical_loss": 3.316254850248912,
      "tokens_seen": 3053584384
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.608080808080808e-05,
      "loss": 1.3959,
      "theoretical_loss": 3.3162493976946745,
      "tokens_seen": 3053649920
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.606060606060607e-05,
      "loss": 1.5275,
      "theoretical_loss": 3.3162439452902204,
      "tokens_seen": 3053715456
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.604040404040405e-05,
      "loss": 1.39,
      "theoretical_loss": 3.3162384930355424,
      "tokens_seen": 3053780992
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.602020202020202e-05,
      "loss": 1.3809,
      "theoretical_loss": 3.3162330409306335,
      "tokens_seen": 3053846528
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.6e-05,
      "loss": 1.4809,
      "theoretical_loss": 3.3162275889754866,
      "tokens_seen": 3053912064
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.597979797979799e-05,
      "loss": 1.6539,
      "theoretical_loss": 3.316222137170094,
      "tokens_seen": 3053977600
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.595959595959596e-05,
      "loss": 1.2021,
      "theoretical_loss": 3.316216685514449,
      "tokens_seen": 3054043136
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.593939393939394e-05,
      "loss": 1.5383,
      "theoretical_loss": 3.3162112340085432,
      "tokens_seen": 3054108672
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.591919191919191e-05,
      "loss": 1.274,
      "theoretical_loss": 3.31620578265237,
      "tokens_seen": 3054174208
    },
    {
      "epoch": 0.85,
      "objective/train/aligned_token_count": 0.4120234549045563,
      "objective/train/avg_token_score": 0.011010357178747654,
      "objective/train/docs_used": 1718284,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.203764796257019,
      "objective/train/original_loss": 4.426965713500977,
      "objective/train/theoretical_loss": 3.316200331445922,
      "objective/train/tokens_used": 1413098976,
      "objective/train/unlikelihood_loss": 0.05123637989163399,
      "theoretical_loss": 3.316200331445922,
      "tokens_seen": 3054239744
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.58989898989899e-05,
      "loss": 1.0877,
      "theoretical_loss": 3.316200331445922,
      "tokens_seen": 3054239744
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.587878787878788e-05,
      "loss": 1.2295,
      "theoretical_loss": 3.316194880389192,
      "tokens_seen": 3054305280
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.585858585858586e-05,
      "loss": 1.3373,
      "theoretical_loss": 3.316189429482172,
      "tokens_seen": 3054370816
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.583838383838385e-05,
      "loss": 1.4979,
      "theoretical_loss": 3.3161839787248555,
      "tokens_seen": 3054436352
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.581818181818182e-05,
      "loss": 1.4326,
      "theoretical_loss": 3.3161785281172347,
      "tokens_seen": 3054501888
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.57979797979798e-05,
      "loss": 1.3158,
      "theoretical_loss": 3.3161730776593026,
      "tokens_seen": 3054567424
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.577777777777777e-05,
      "loss": 1.1381,
      "theoretical_loss": 3.316167627351051,
      "tokens_seen": 3054632960
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.575757575757576e-05,
      "loss": 1.2262,
      "theoretical_loss": 3.3161621771924743,
      "tokens_seen": 3054698496
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.573737373737374e-05,
      "loss": 1.4811,
      "theoretical_loss": 3.3161567271835635,
      "tokens_seen": 3054764032
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.571717171717171e-05,
      "loss": 1.4074,
      "theoretical_loss": 3.316151277324312,
      "tokens_seen": 3054829568
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.56969696969697e-05,
      "loss": 1.3274,
      "theoretical_loss": 3.3161458276147124,
      "tokens_seen": 3054895104
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.567676767676768e-05,
      "loss": 1.4935,
      "theoretical_loss": 3.3161403780547576,
      "tokens_seen": 3054960640
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.565656565656566e-05,
      "loss": 1.4317,
      "theoretical_loss": 3.31613492864444,
      "tokens_seen": 3055026176
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.563636363636365e-05,
      "loss": 1.2565,
      "theoretical_loss": 3.3161294793837524,
      "tokens_seen": 3055091712
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.561616161616162e-05,
      "loss": 1.5756,
      "theoretical_loss": 3.316124030272687,
      "tokens_seen": 3055157248
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.559595959595959e-05,
      "loss": 1.4035,
      "theoretical_loss": 3.3161185813112373,
      "tokens_seen": 3055222784
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.557575757575757e-05,
      "loss": 1.3625,
      "theoretical_loss": 3.3161131324993955,
      "tokens_seen": 3055288320
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.555555555555556e-05,
      "loss": 1.2635,
      "theoretical_loss": 3.316107683837154,
      "tokens_seen": 3055353856
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.553535353535354e-05,
      "loss": 1.3965,
      "theoretical_loss": 3.3161022353245064,
      "tokens_seen": 3055419392
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.551515151515152e-05,
      "loss": 1.4185,
      "theoretical_loss": 3.316096786961445,
      "tokens_seen": 3055484928
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.54949494949495e-05,
      "loss": 1.2991,
      "theoretical_loss": 3.316091338747962,
      "tokens_seen": 3055550464
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.547474747474748e-05,
      "loss": 1.4024,
      "theoretical_loss": 3.3160858906840502,
      "tokens_seen": 3055616000
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.545454545454546e-05,
      "loss": 1.3345,
      "theoretical_loss": 3.3160804427697026,
      "tokens_seen": 3055681536
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.543434343434343e-05,
      "loss": 1.3519,
      "theoretical_loss": 3.3160749950049118,
      "tokens_seen": 3055747072
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.541414141414142e-05,
      "loss": 1.5342,
      "theoretical_loss": 3.3160695473896706,
      "tokens_seen": 3055812608
    },
    {
      "epoch": 0.85,
      "objective/train/aligned_token_count": 0.461815744638443,
      "objective/train/avg_token_score": 0.004791140556335449,
      "objective/train/docs_used": 1719055,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.344924807548523,
      "objective/train/original_loss": 3.9360427856445312,
      "objective/train/theoretical_loss": 3.316064099923971,
      "objective/train/tokens_used": 1414737376,
      "objective/train/unlikelihood_loss": 0.07483018934726715,
      "theoretical_loss": 3.316064099923971,
      "tokens_seen": 3055878144
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.539393939393939e-05,
      "loss": 1.4709,
      "theoretical_loss": 3.316064099923971,
      "tokens_seen": 3055878144
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.537373737373737e-05,
      "loss": 1.3528,
      "theoretical_loss": 3.3160586526078073,
      "tokens_seen": 3055943680
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.535353535353536e-05,
      "loss": 1.2628,
      "theoretical_loss": 3.3160532054411704,
      "tokens_seen": 3056009216
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.533333333333334e-05,
      "loss": 1.4203,
      "theoretical_loss": 3.316047758424054,
      "tokens_seen": 3056074752
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.531313131313132e-05,
      "loss": 1.5342,
      "theoretical_loss": 3.3160423115564504,
      "tokens_seen": 3056140288
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.52929292929293e-05,
      "loss": 1.4956,
      "theoretical_loss": 3.3160368648383525,
      "tokens_seen": 3056205824
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.527272727272726e-05,
      "loss": 1.3081,
      "theoretical_loss": 3.316031418269753,
      "tokens_seen": 3056271360
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.525252525252525e-05,
      "loss": 1.3079,
      "theoretical_loss": 3.316025971850644,
      "tokens_seen": 3056336896
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.523232323232323e-05,
      "loss": 1.4541,
      "theoretical_loss": 3.316020525581019,
      "tokens_seen": 3056402432
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.521212121212122e-05,
      "loss": 1.5971,
      "theoretical_loss": 3.3160150794608705,
      "tokens_seen": 3056467968
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.51919191919192e-05,
      "loss": 1.4899,
      "theoretical_loss": 3.3160096334901907,
      "tokens_seen": 3056533504
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.517171717171717e-05,
      "loss": 1.2005,
      "theoretical_loss": 3.3160041876689728,
      "tokens_seen": 3056599040
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.515151515151515e-05,
      "loss": 1.3329,
      "theoretical_loss": 3.3159987419972095,
      "tokens_seen": 3056664576
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.513131313131314e-05,
      "loss": 1.3925,
      "theoretical_loss": 3.3159932964748933,
      "tokens_seen": 3056730112
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.511111111111112e-05,
      "loss": 1.2043,
      "theoretical_loss": 3.315987851102017,
      "tokens_seen": 3056795648
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.509090909090909e-05,
      "loss": 1.4459,
      "theoretical_loss": 3.3159824058785734,
      "tokens_seen": 3056861184
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.507070707070706e-05,
      "loss": 1.486,
      "theoretical_loss": 3.3159769608045546,
      "tokens_seen": 3056926720
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.505050505050505e-05,
      "loss": 1.3803,
      "theoretical_loss": 3.3159715158799536,
      "tokens_seen": 3056992256
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.503030303030303e-05,
      "loss": 1.0578,
      "theoretical_loss": 3.3159660711047634,
      "tokens_seen": 3057057792
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.501010101010102e-05,
      "loss": 1.286,
      "theoretical_loss": 3.3159606264789767,
      "tokens_seen": 3057123328
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.4989898989899e-05,
      "loss": 1.3385,
      "theoretical_loss": 3.315955182002586,
      "tokens_seen": 3057188864
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.496969696969697e-05,
      "loss": 1.5712,
      "theoretical_loss": 3.315949737675584,
      "tokens_seen": 3057254400
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.494949494949495e-05,
      "loss": 1.165,
      "theoretical_loss": 3.3159442934979637,
      "tokens_seen": 3057319936
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.492929292929292e-05,
      "loss": 1.2575,
      "theoretical_loss": 3.315938849469717,
      "tokens_seen": 3057385472
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.490909090909091e-05,
      "loss": 1.2593,
      "theoretical_loss": 3.315933405590837,
      "tokens_seen": 3057451008
    },
    {
      "epoch": 0.85,
      "objective/train/aligned_token_count": 0.3159213066101074,
      "objective/train/avg_token_score": 0.008669550530612469,
      "objective/train/docs_used": 1720294,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9925927519798279,
      "objective/train/original_loss": 4.755394458770752,
      "objective/train/theoretical_loss": 3.315927961861317,
      "objective/train/tokens_used": 1416375776,
      "objective/train/unlikelihood_loss": 0.05916186049580574,
      "theoretical_loss": 3.315927961861317,
      "tokens_seen": 3057516544
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.488888888888889e-05,
      "loss": 1.2421,
      "theoretical_loss": 3.315927961861317,
      "tokens_seen": 3057516544
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.486868686868686e-05,
      "loss": 1.1809,
      "theoretical_loss": 3.315922518281149,
      "tokens_seen": 3057582080
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.484848484848485e-05,
      "loss": 1.2664,
      "theoretical_loss": 3.315917074850326,
      "tokens_seen": 3057647616
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.482828282828283e-05,
      "loss": 1.3916,
      "theoretical_loss": 3.3159116315688406,
      "tokens_seen": 3057713152
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.480808080808081e-05,
      "loss": 1.4754,
      "theoretical_loss": 3.3159061884366854,
      "tokens_seen": 3057778688
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.47878787878788e-05,
      "loss": 1.3794,
      "theoretical_loss": 3.3159007454538534,
      "tokens_seen": 3057844224
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.476767676767678e-05,
      "loss": 1.221,
      "theoretical_loss": 3.315895302620337,
      "tokens_seen": 3057909760
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.474747474747474e-05,
      "loss": 1.1564,
      "theoretical_loss": 3.315889859936129,
      "tokens_seen": 3057975296
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.472727272727272e-05,
      "loss": 1.2173,
      "theoretical_loss": 3.3158844174012225,
      "tokens_seen": 3058040832
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.47070707070707e-05,
      "loss": 1.4781,
      "theoretical_loss": 3.3158789750156092,
      "tokens_seen": 3058106368
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.468686868686869e-05,
      "loss": 1.248,
      "theoretical_loss": 3.3158735327792828,
      "tokens_seen": 3058171904
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.466666666666667e-05,
      "loss": 1.4457,
      "theoretical_loss": 3.315868090692236,
      "tokens_seen": 3058237440
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.464646464646464e-05,
      "loss": 1.4567,
      "theoretical_loss": 3.3158626487544605,
      "tokens_seen": 3058302976
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.462626262626263e-05,
      "loss": 1.5109,
      "theoretical_loss": 3.31585720696595,
      "tokens_seen": 3058368512
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.460606060606061e-05,
      "loss": 1.5267,
      "theoretical_loss": 3.315851765326697,
      "tokens_seen": 3058434048
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.45858585858586e-05,
      "loss": 1.5208,
      "theoretical_loss": 3.315846323836694,
      "tokens_seen": 3058499584
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.456565656565657e-05,
      "loss": 1.3631,
      "theoretical_loss": 3.3158408824959333,
      "tokens_seen": 3058565120
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.454545454545454e-05,
      "loss": 1.3035,
      "theoretical_loss": 3.3158354413044084,
      "tokens_seen": 3058630656
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.452525252525252e-05,
      "loss": 1.2957,
      "theoretical_loss": 3.315830000262112,
      "tokens_seen": 3058696192
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.45050505050505e-05,
      "loss": 1.5217,
      "theoretical_loss": 3.3158245593690365,
      "tokens_seen": 3058761728
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.448484848484849e-05,
      "loss": 1.0596,
      "theoretical_loss": 3.315819118625174,
      "tokens_seen": 3058827264
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.446464646464647e-05,
      "loss": 1.4045,
      "theoretical_loss": 3.3158136780305183,
      "tokens_seen": 3058892800
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.444444444444444e-05,
      "loss": 1.2969,
      "theoretical_loss": 3.315808237585062,
      "tokens_seen": 3058958336
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.442424242424243e-05,
      "loss": 1.397,
      "theoretical_loss": 3.315802797288797,
      "tokens_seen": 3059023872
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.44040404040404e-05,
      "loss": 1.368,
      "theoretical_loss": 3.3157973571417165,
      "tokens_seen": 3059089408
    },
    {
      "epoch": 0.85,
      "objective/train/aligned_token_count": 0.5259652733802795,
      "objective/train/avg_token_score": 0.003761256579309702,
      "objective/train/docs_used": 1721360,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6076905727386475,
      "objective/train/original_loss": 3.7786715030670166,
      "objective/train/theoretical_loss": 3.3157919171438133,
      "objective/train/tokens_used": 1418014176,
      "objective/train/unlikelihood_loss": 0.07750105112791061,
      "theoretical_loss": 3.3157919171438133,
      "tokens_seen": 3059154944
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.438383838383838e-05,
      "loss": 1.3539,
      "theoretical_loss": 3.3157919171438133,
      "tokens_seen": 3059154944
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.436363636363637e-05,
      "loss": 1.3629,
      "theoretical_loss": 3.31578647729508,
      "tokens_seen": 3059220480
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.434343434343435e-05,
      "loss": 1.3595,
      "theoretical_loss": 3.3157810375955092,
      "tokens_seen": 3059286016
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.432323232323232e-05,
      "loss": 1.464,
      "theoretical_loss": 3.315775598045094,
      "tokens_seen": 3059351552
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.43030303030303e-05,
      "loss": 1.4739,
      "theoretical_loss": 3.3157701586438266,
      "tokens_seen": 3059417088
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.428282828282829e-05,
      "loss": 1.3998,
      "theoretical_loss": 3.3157647193916997,
      "tokens_seen": 3059482624
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.426262626262627e-05,
      "loss": 1.3928,
      "theoretical_loss": 3.315759280288707,
      "tokens_seen": 3059548160
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.424242424242426e-05,
      "loss": 1.331,
      "theoretical_loss": 3.31575384133484,
      "tokens_seen": 3059613696
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.422222222222221e-05,
      "loss": 1.4505,
      "theoretical_loss": 3.315748402530092,
      "tokens_seen": 3059679232
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.42020202020202e-05,
      "loss": 1.3458,
      "theoretical_loss": 3.3157429638744556,
      "tokens_seen": 3059744768
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.418181818181818e-05,
      "loss": 1.5187,
      "theoretical_loss": 3.3157375253679233,
      "tokens_seen": 3059810304
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.416161616161617e-05,
      "loss": 1.3095,
      "theoretical_loss": 3.3157320870104883,
      "tokens_seen": 3059875840
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.414141414141415e-05,
      "loss": 1.3621,
      "theoretical_loss": 3.3157266488021433,
      "tokens_seen": 3059941376
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.412121212121212e-05,
      "loss": 1.4685,
      "theoretical_loss": 3.3157212107428804,
      "tokens_seen": 3060006912
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.41010101010101e-05,
      "loss": 1.3578,
      "theoretical_loss": 3.315715772832693,
      "tokens_seen": 3060072448
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.408080808080809e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.3157103350715738,
      "tokens_seen": 3060137984
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.406060606060606e-05,
      "loss": 1.3858,
      "theoretical_loss": 3.315704897459515,
      "tokens_seen": 3060203520
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.404040404040404e-05,
      "loss": 1.417,
      "theoretical_loss": 3.3156994599965093,
      "tokens_seen": 3060269056
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.402020202020201e-05,
      "loss": 1.4637,
      "theoretical_loss": 3.31569402268255,
      "tokens_seen": 3060334592
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.4e-05,
      "loss": 1.5247,
      "theoretical_loss": 3.3156885855176297,
      "tokens_seen": 3060400128
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.397979797979798e-05,
      "loss": 1.439,
      "theoretical_loss": 3.3156831485017406,
      "tokens_seen": 3060465664
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.395959595959596e-05,
      "loss": 1.6807,
      "theoretical_loss": 3.315677711634876,
      "tokens_seen": 3060531200
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.393939393939395e-05,
      "loss": 1.4701,
      "theoretical_loss": 3.3156722749170284,
      "tokens_seen": 3060596736
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.391919191919193e-05,
      "loss": 1.4029,
      "theoretical_loss": 3.31566683834819,
      "tokens_seen": 3060662272
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.38989898989899e-05,
      "loss": 1.416,
      "theoretical_loss": 3.315661401928355,
      "tokens_seen": 3060727808
    },
    {
      "epoch": 0.85,
      "objective/train/aligned_token_count": 0.4665200412273407,
      "objective/train/avg_token_score": 0.017265059053897858,
      "objective/train/docs_used": 1721958,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5267146825790405,
      "objective/train/original_loss": 4.405810832977295,
      "objective/train/theoretical_loss": 3.315655965657515,
      "objective/train/tokens_used": 1419652576,
      "objective/train/unlikelihood_loss": 0.05424108728766441,
      "theoretical_loss": 3.315655965657515,
      "tokens_seen": 3060793344
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.387878787878787e-05,
      "loss": 1.453,
      "theoretical_loss": 3.315655965657515,
      "tokens_seen": 3060793344
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.385858585858586e-05,
      "loss": 1.5157,
      "theoretical_loss": 3.315650529535662,
      "tokens_seen": 3060858880
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.383838383838384e-05,
      "loss": 1.4058,
      "theoretical_loss": 3.315645093562791,
      "tokens_seen": 3060924416
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.381818181818182e-05,
      "loss": 1.3928,
      "theoretical_loss": 3.3156396577388922,
      "tokens_seen": 3060989952
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.37979797979798e-05,
      "loss": 1.4059,
      "theoretical_loss": 3.31563422206396,
      "tokens_seen": 3061055488
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.377777777777778e-05,
      "loss": 1.4418,
      "theoretical_loss": 3.3156287865379865,
      "tokens_seen": 3061121024
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.375757575757576e-05,
      "loss": 1.2672,
      "theoretical_loss": 3.3156233511609647,
      "tokens_seen": 3061186560
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.373737373737375e-05,
      "loss": 1.3842,
      "theoretical_loss": 3.315617915932887,
      "tokens_seen": 3061252096
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.371717171717173e-05,
      "loss": 1.237,
      "theoretical_loss": 3.3156124808537464,
      "tokens_seen": 3061317632
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.369696969696969e-05,
      "loss": 1.4803,
      "theoretical_loss": 3.3156070459235356,
      "tokens_seen": 3061383168
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.367676767676767e-05,
      "loss": 1.2226,
      "theoretical_loss": 3.315601611142247,
      "tokens_seen": 3061448704
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.365656565656566e-05,
      "loss": 1.3145,
      "theoretical_loss": 3.315596176509874,
      "tokens_seen": 3061514240
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.363636363636364e-05,
      "loss": 1.616,
      "theoretical_loss": 3.3155907420264086,
      "tokens_seen": 3061579776
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.361616161616162e-05,
      "loss": 1.2391,
      "theoretical_loss": 3.315585307691844,
      "tokens_seen": 3061645312
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.35959595959596e-05,
      "loss": 1.4481,
      "theoretical_loss": 3.315579873506173,
      "tokens_seen": 3061710848
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.357575757575758e-05,
      "loss": 1.3956,
      "theoretical_loss": 3.315574439469388,
      "tokens_seen": 3061776384
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.355555555555556e-05,
      "loss": 1.4672,
      "theoretical_loss": 3.3155690055814815,
      "tokens_seen": 3061841920
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.353535353535353e-05,
      "loss": 1.6098,
      "theoretical_loss": 3.315563571842447,
      "tokens_seen": 3061907456
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.351515151515152e-05,
      "loss": 1.7046,
      "theoretical_loss": 3.3155581382522765,
      "tokens_seen": 3061972992
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.349494949494949e-05,
      "loss": 1.2315,
      "theoretical_loss": 3.3155527048109636,
      "tokens_seen": 3062038528
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.347474747474747e-05,
      "loss": 1.2781,
      "theoretical_loss": 3.3155472715185,
      "tokens_seen": 3062104064
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.345454545454545e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.3155418383748794,
      "tokens_seen": 3062169600
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.343434343434344e-05,
      "loss": 1.4254,
      "theoretical_loss": 3.3155364053800938,
      "tokens_seen": 3062235136
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.341414141414142e-05,
      "loss": 1.3764,
      "theoretical_loss": 3.315530972534136,
      "tokens_seen": 3062300672
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.33939393939394e-05,
      "loss": 1.3552,
      "theoretical_loss": 3.3155255398369996,
      "tokens_seen": 3062366208
    },
    {
      "epoch": 0.85,
      "objective/train/aligned_token_count": 0.36375856399536133,
      "objective/train/avg_token_score": 0.0089182173833251,
      "objective/train/docs_used": 1723218,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.122045874595642,
      "objective/train/original_loss": 4.700098037719727,
      "objective/train/theoretical_loss": 3.315520107288676,
      "objective/train/tokens_used": 1421290976,
      "objective/train/unlikelihood_loss": 0.046707578003406525,
      "theoretical_loss": 3.315520107288676,
      "tokens_seen": 3062431744
    },
    {
      "epoch": 0.85,
      "learning_rate": 7.337373737373738e-05,
      "loss": 1.4543,
      "theoretical_loss": 3.315520107288676,
      "tokens_seen": 3062431744
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.335353535353535e-05,
      "loss": 1.443,
      "theoretical_loss": 3.315514674889159,
      "tokens_seen": 3062497280
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.333333333333333e-05,
      "loss": 1.4651,
      "theoretical_loss": 3.315509242638441,
      "tokens_seen": 3062562816
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.331313131313132e-05,
      "loss": 1.4187,
      "theoretical_loss": 3.315503810536515,
      "tokens_seen": 3062628352
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.32929292929293e-05,
      "loss": 1.3925,
      "theoretical_loss": 3.315498378583373,
      "tokens_seen": 3062693888
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.327272727272727e-05,
      "loss": 1.5248,
      "theoretical_loss": 3.315492946779008,
      "tokens_seen": 3062759424
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.325252525252525e-05,
      "loss": 1.3932,
      "theoretical_loss": 3.3154875151234133,
      "tokens_seen": 3062824960
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.323232323232324e-05,
      "loss": 1.3935,
      "theoretical_loss": 3.3154820836165815,
      "tokens_seen": 3062890496
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.321212121212122e-05,
      "loss": 1.4148,
      "theoretical_loss": 3.315476652258505,
      "tokens_seen": 3062956032
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.319191919191919e-05,
      "loss": 1.5117,
      "theoretical_loss": 3.3154712210491764,
      "tokens_seen": 3063021568
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.317171717171716e-05,
      "loss": 1.6429,
      "theoretical_loss": 3.315465789988589,
      "tokens_seen": 3063087104
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.315151515151515e-05,
      "loss": 1.3423,
      "theoretical_loss": 3.315460359076735,
      "tokens_seen": 3063152640
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.313131313131313e-05,
      "loss": 1.7213,
      "theoretical_loss": 3.315454928313607,
      "tokens_seen": 3063218176
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.311111111111111e-05,
      "loss": 1.2746,
      "theoretical_loss": 3.315449497699199,
      "tokens_seen": 3063283712
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.30909090909091e-05,
      "loss": 1.5557,
      "theoretical_loss": 3.3154440672335026,
      "tokens_seen": 3063349248
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.307070707070707e-05,
      "loss": 1.3945,
      "theoretical_loss": 3.3154386369165105,
      "tokens_seen": 3063414784
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.305050505050505e-05,
      "loss": 1.3176,
      "theoretical_loss": 3.3154332067482164,
      "tokens_seen": 3063480320
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.303030303030304e-05,
      "loss": 1.4777,
      "theoretical_loss": 3.3154277767286118,
      "tokens_seen": 3063545856
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.301010101010101e-05,
      "loss": 1.2598,
      "theoretical_loss": 3.3154223468576904,
      "tokens_seen": 3063611392
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.298989898989899e-05,
      "loss": 1.6485,
      "theoretical_loss": 3.3154169171354444,
      "tokens_seen": 3063676928
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.296969696969697e-05,
      "loss": 1.5131,
      "theoretical_loss": 3.3154114875618674,
      "tokens_seen": 3063742464
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.294949494949495e-05,
      "loss": 1.4559,
      "theoretical_loss": 3.3154060581369507,
      "tokens_seen": 3063808000
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.292929292929293e-05,
      "loss": 1.4916,
      "theoretical_loss": 3.3154006288606883,
      "tokens_seen": 3063873536
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.290909090909091e-05,
      "loss": 1.3427,
      "theoretical_loss": 3.315395199733073,
      "tokens_seen": 3063939072
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.28888888888889e-05,
      "loss": 1.3463,
      "theoretical_loss": 3.3153897707540962,
      "tokens_seen": 3064004608
    },
    {
      "epoch": 0.86,
      "objective/train/aligned_token_count": 0.41477271914482117,
      "objective/train/avg_token_score": 0.025422554463148117,
      "objective/train/docs_used": 1723856,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1921305656433105,
      "objective/train/original_loss": 4.320864677429199,
      "objective/train/theoretical_loss": 3.315384341923752,
      "objective/train/tokens_used": 1422929376,
      "objective/train/unlikelihood_loss": 0.07576357573270798,
      "theoretical_loss": 3.315384341923752,
      "tokens_seen": 3064070144
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.286868686868688e-05,
      "loss": 1.2482,
      "theoretical_loss": 3.315384341923752,
      "tokens_seen": 3064070144
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.284848484848484e-05,
      "loss": 1.4414,
      "theoretical_loss": 3.3153789132420326,
      "tokens_seen": 3064135680
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.282828282828282e-05,
      "loss": 1.3391,
      "theoretical_loss": 3.315373484708931,
      "tokens_seen": 3064201216
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.28080808080808e-05,
      "loss": 1.4476,
      "theoretical_loss": 3.3153680563244396,
      "tokens_seen": 3064266752
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.278787878787879e-05,
      "loss": 1.4514,
      "theoretical_loss": 3.3153626280885513,
      "tokens_seen": 3064332288
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.276767676767677e-05,
      "loss": 1.3854,
      "theoretical_loss": 3.3153572000012588,
      "tokens_seen": 3064397824
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.274747474747474e-05,
      "loss": 1.2982,
      "theoretical_loss": 3.315351772062555,
      "tokens_seen": 3064463360
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.272727272727273e-05,
      "loss": 1.3678,
      "theoretical_loss": 3.3153463442724327,
      "tokens_seen": 3064528896
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.270707070707071e-05,
      "loss": 1.335,
      "theoretical_loss": 3.315340916630885,
      "tokens_seen": 3064594432
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.26868686868687e-05,
      "loss": 1.3714,
      "theoretical_loss": 3.3153354891379037,
      "tokens_seen": 3064659968
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.266666666666667e-05,
      "loss": 1.1445,
      "theoretical_loss": 3.3153300617934818,
      "tokens_seen": 3064725504
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.264646464646464e-05,
      "loss": 1.4886,
      "theoretical_loss": 3.315324634597613,
      "tokens_seen": 3064791040
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.262626262626262e-05,
      "loss": 1.5413,
      "theoretical_loss": 3.3153192075502886,
      "tokens_seen": 3064856576
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.26060606060606e-05,
      "loss": 1.3506,
      "theoretical_loss": 3.3153137806515027,
      "tokens_seen": 3064922112
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.258585858585859e-05,
      "loss": 1.3554,
      "theoretical_loss": 3.315308353901248,
      "tokens_seen": 3064987648
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.256565656565657e-05,
      "loss": 1.4363,
      "theoretical_loss": 3.315302927299516,
      "tokens_seen": 3065053184
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.254545454545456e-05,
      "loss": 1.6298,
      "theoretical_loss": 3.3152975008463,
      "tokens_seen": 3065118720
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.252525252525253e-05,
      "loss": 1.2158,
      "theoretical_loss": 3.3152920745415937,
      "tokens_seen": 3065184256
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.25050505050505e-05,
      "loss": 1.4648,
      "theoretical_loss": 3.3152866483853884,
      "tokens_seen": 3065249792
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.248484848484848e-05,
      "loss": 1.4577,
      "theoretical_loss": 3.315281222377678,
      "tokens_seen": 3065315328
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.246464646464647e-05,
      "loss": 1.4406,
      "theoretical_loss": 3.3152757965184554,
      "tokens_seen": 3065380864
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.244444444444445e-05,
      "loss": 1.3681,
      "theoretical_loss": 3.315270370807712,
      "tokens_seen": 3065446400
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.242424242424242e-05,
      "loss": 1.3984,
      "theoretical_loss": 3.315264945245442,
      "tokens_seen": 3065511936
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.24040404040404e-05,
      "loss": 1.5859,
      "theoretical_loss": 3.315259519831637,
      "tokens_seen": 3065577472
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.238383838383839e-05,
      "loss": 1.5356,
      "theoretical_loss": 3.3152540945662907,
      "tokens_seen": 3065643008
    },
    {
      "epoch": 0.86,
      "objective/train/aligned_token_count": 0.4813660681247711,
      "objective/train/avg_token_score": 0.0057468898594379425,
      "objective/train/docs_used": 1724430,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.495861291885376,
      "objective/train/original_loss": 4.0261430740356445,
      "objective/train/theoretical_loss": 3.3152486694493954,
      "objective/train/tokens_used": 1424567776,
      "objective/train/unlikelihood_loss": 0.0772835910320282,
      "theoretical_loss": 3.3152486694493954,
      "tokens_seen": 3065708544
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.236363636363637e-05,
      "loss": 1.6163,
      "theoretical_loss": 3.3152486694493954,
      "tokens_seen": 3065708544
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.234343434343436e-05,
      "loss": 1.4573,
      "theoretical_loss": 3.3152432444809437,
      "tokens_seen": 3065774080
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.232323232323231e-05,
      "loss": 1.5374,
      "theoretical_loss": 3.315237819660929,
      "tokens_seen": 3065839616
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.23030303030303e-05,
      "loss": 1.4284,
      "theoretical_loss": 3.3152323949893434,
      "tokens_seen": 3065905152
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.228282828282828e-05,
      "loss": 1.3551,
      "theoretical_loss": 3.3152269704661803,
      "tokens_seen": 3065970688
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.226262626262626e-05,
      "loss": 1.312,
      "theoretical_loss": 3.315221546091432,
      "tokens_seen": 3066036224
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.224242424242425e-05,
      "loss": 1.3828,
      "theoretical_loss": 3.315216121865091,
      "tokens_seen": 3066101760
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.222222222222222e-05,
      "loss": 1.4881,
      "theoretical_loss": 3.3152106977871503,
      "tokens_seen": 3066167296
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.22020202020202e-05,
      "loss": 1.5809,
      "theoretical_loss": 3.3152052738576034,
      "tokens_seen": 3066232832
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.218181818181819e-05,
      "loss": 1.4477,
      "theoretical_loss": 3.3151998500764424,
      "tokens_seen": 3066298368
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.216161616161617e-05,
      "loss": 1.6285,
      "theoretical_loss": 3.31519442644366,
      "tokens_seen": 3066363904
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.214141414141414e-05,
      "loss": 1.397,
      "theoretical_loss": 3.315189002959249,
      "tokens_seen": 3066429440
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.212121212121213e-05,
      "loss": 1.2662,
      "theoretical_loss": 3.3151835796232025,
      "tokens_seen": 3066494976
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.21010101010101e-05,
      "loss": 1.4119,
      "theoretical_loss": 3.315178156435513,
      "tokens_seen": 3066560512
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.208080808080808e-05,
      "loss": 1.4786,
      "theoretical_loss": 3.315172733396173,
      "tokens_seen": 3066626048
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.206060606060606e-05,
      "loss": 1.4402,
      "theoretical_loss": 3.315167310505176,
      "tokens_seen": 3066691584
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.204040404040405e-05,
      "loss": 1.2768,
      "theoretical_loss": 3.3151618877625144,
      "tokens_seen": 3066757120
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.202020202020203e-05,
      "loss": 1.1968,
      "theoretical_loss": 3.315156465168181,
      "tokens_seen": 3066822656
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.2e-05,
      "loss": 1.3482,
      "theoretical_loss": 3.315151042722168,
      "tokens_seen": 3066888192
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.197979797979797e-05,
      "loss": 1.3922,
      "theoretical_loss": 3.3151456204244694,
      "tokens_seen": 3066953728
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.195959595959596e-05,
      "loss": 1.5269,
      "theoretical_loss": 3.3151401982750768,
      "tokens_seen": 3067019264
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.193939393939394e-05,
      "loss": 1.5155,
      "theoretical_loss": 3.3151347762739833,
      "tokens_seen": 3067084800
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.191919191919192e-05,
      "loss": 1.4943,
      "theoretical_loss": 3.315129354421182,
      "tokens_seen": 3067150336
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.18989898989899e-05,
      "loss": 1.2752,
      "theoretical_loss": 3.3151239327166655,
      "tokens_seen": 3067215872
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.187878787878788e-05,
      "loss": 1.5437,
      "theoretical_loss": 3.315118511160427,
      "tokens_seen": 3067281408
    },
    {
      "epoch": 0.86,
      "objective/train/aligned_token_count": 0.4523460268974304,
      "objective/train/avg_token_score": 0.00838436372578144,
      "objective/train/docs_used": 1725702,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5420401096343994,
      "objective/train/original_loss": 4.444885730743408,
      "objective/train/theoretical_loss": 3.315113089752458,
      "objective/train/tokens_used": 1426206176,
      "objective/train/unlikelihood_loss": 0.06120666489005089,
      "theoretical_loss": 3.315113089752458,
      "tokens_seen": 3067346944
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.185858585858586e-05,
      "loss": 1.3876,
      "theoretical_loss": 3.315113089752458,
      "tokens_seen": 3067346944
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.183838383838385e-05,
      "loss": 1.3544,
      "theoretical_loss": 3.315107668492753,
      "tokens_seen": 3067412480
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.181818181818183e-05,
      "loss": 1.4592,
      "theoretical_loss": 3.3151022473813034,
      "tokens_seen": 3067478016
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.179797979797979e-05,
      "loss": 1.4325,
      "theoretical_loss": 3.3150968264181024,
      "tokens_seen": 3067543552
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.177777777777777e-05,
      "loss": 1.4463,
      "theoretical_loss": 3.3150914056031433,
      "tokens_seen": 3067609088
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.175757575757576e-05,
      "loss": 1.2678,
      "theoretical_loss": 3.3150859849364185,
      "tokens_seen": 3067674624
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.173737373737374e-05,
      "loss": 1.4749,
      "theoretical_loss": 3.31508056441792,
      "tokens_seen": 3067740160
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.171717171717172e-05,
      "loss": 1.2526,
      "theoretical_loss": 3.3150751440476416,
      "tokens_seen": 3067805696
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.169696969696971e-05,
      "loss": 1.465,
      "theoretical_loss": 3.315069723825576,
      "tokens_seen": 3067871232
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.167676767676768e-05,
      "loss": 1.2327,
      "theoretical_loss": 3.315064303751716,
      "tokens_seen": 3067936768
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.165656565656566e-05,
      "loss": 1.2401,
      "theoretical_loss": 3.3150588838260537,
      "tokens_seen": 3068002304
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.163636363636363e-05,
      "loss": 1.3486,
      "theoretical_loss": 3.3150534640485825,
      "tokens_seen": 3068067840
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.161616161616162e-05,
      "loss": 1.5371,
      "theoretical_loss": 3.315048044419295,
      "tokens_seen": 3068133376
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.15959595959596e-05,
      "loss": 1.5687,
      "theoretical_loss": 3.315042624938184,
      "tokens_seen": 3068198912
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.157575757575757e-05,
      "loss": 1.4048,
      "theoretical_loss": 3.315037205605242,
      "tokens_seen": 3068264448
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.155555555555555e-05,
      "loss": 1.3342,
      "theoretical_loss": 3.315031786420462,
      "tokens_seen": 3068329984
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.153535353535354e-05,
      "loss": 1.6783,
      "theoretical_loss": 3.315026367383837,
      "tokens_seen": 3068395520
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.151515151515152e-05,
      "loss": 1.3224,
      "theoretical_loss": 3.31502094849536,
      "tokens_seen": 3068461056
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.14949494949495e-05,
      "loss": 1.4066,
      "theoretical_loss": 3.315015529755023,
      "tokens_seen": 3068526592
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.147474747474748e-05,
      "loss": 1.4647,
      "theoretical_loss": 3.3150101111628194,
      "tokens_seen": 3068592128
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.145454545454545e-05,
      "loss": 1.4386,
      "theoretical_loss": 3.315004692718742,
      "tokens_seen": 3068657664
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.143434343434343e-05,
      "loss": 1.3212,
      "theoretical_loss": 3.314999274422783,
      "tokens_seen": 3068723200
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.141414141414141e-05,
      "loss": 1.3783,
      "theoretical_loss": 3.3149938562749357,
      "tokens_seen": 3068788736
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.13939393939394e-05,
      "loss": 1.6329,
      "theoretical_loss": 3.3149884382751926,
      "tokens_seen": 3068854272
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.137373737373737e-05,
      "loss": 1.3226,
      "theoretical_loss": 3.3149830204235466,
      "tokens_seen": 3068919808
    },
    {
      "epoch": 0.86,
      "objective/train/aligned_token_count": 0.43841642141342163,
      "objective/train/avg_token_score": 0.007751962635666132,
      "objective/train/docs_used": 1726520,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.369861125946045,
      "objective/train/original_loss": 4.560239315032959,
      "objective/train/theoretical_loss": 3.314977602719991,
      "objective/train/tokens_used": 1427844576,
      "objective/train/unlikelihood_loss": 0.05186886712908745,
      "theoretical_loss": 3.314977602719991,
      "tokens_seen": 3068985344
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.135353535353535e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.314977602719991,
      "tokens_seen": 3068985344
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.133333333333334e-05,
      "loss": 1.483,
      "theoretical_loss": 3.314972185164518,
      "tokens_seen": 3069050880
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.131313131313132e-05,
      "loss": 1.4016,
      "theoretical_loss": 3.31496676775712,
      "tokens_seen": 3069116416
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.12929292929293e-05,
      "loss": 1.3196,
      "theoretical_loss": 3.314961350497791,
      "tokens_seen": 3069181952
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.127272727272728e-05,
      "loss": 1.3645,
      "theoretical_loss": 3.3149559333865226,
      "tokens_seen": 3069247488
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.125252525252525e-05,
      "loss": 1.5155,
      "theoretical_loss": 3.314950516423308,
      "tokens_seen": 3069313024
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.123232323232323e-05,
      "loss": 1.3538,
      "theoretical_loss": 3.3149450996081407,
      "tokens_seen": 3069378560
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.121212121212121e-05,
      "loss": 1.4889,
      "theoretical_loss": 3.314939682941012,
      "tokens_seen": 3069444096
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.11919191919192e-05,
      "loss": 1.344,
      "theoretical_loss": 3.3149342664219166,
      "tokens_seen": 3069509632
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.117171717171718e-05,
      "loss": 1.3315,
      "theoretical_loss": 3.3149288500508454,
      "tokens_seen": 3069575168
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.115151515151515e-05,
      "loss": 1.4438,
      "theoretical_loss": 3.3149234338277926,
      "tokens_seen": 3069640704
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.113131313131314e-05,
      "loss": 1.5404,
      "theoretical_loss": 3.31491801775275,
      "tokens_seen": 3069706240
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.11111111111111e-05,
      "loss": 1.306,
      "theoretical_loss": 3.314912601825711,
      "tokens_seen": 3069771776
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.109090909090909e-05,
      "loss": 1.5253,
      "theoretical_loss": 3.3149071860466686,
      "tokens_seen": 3069837312
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.107070707070707e-05,
      "loss": 1.1309,
      "theoretical_loss": 3.3149017704156147,
      "tokens_seen": 3069902848
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.105050505050504e-05,
      "loss": 1.2576,
      "theoretical_loss": 3.314896354932543,
      "tokens_seen": 3069968384
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.103030303030303e-05,
      "loss": 1.2016,
      "theoretical_loss": 3.3148909395974457,
      "tokens_seen": 3070033920
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.101010101010101e-05,
      "loss": 1.5008,
      "theoretical_loss": 3.314885524410316,
      "tokens_seen": 3070099456
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.0989898989899e-05,
      "loss": 1.3128,
      "theoretical_loss": 3.3148801093711464,
      "tokens_seen": 3070164992
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.096969696969698e-05,
      "loss": 1.4259,
      "theoretical_loss": 3.31487469447993,
      "tokens_seen": 3070230528
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.094949494949495e-05,
      "loss": 1.3104,
      "theoretical_loss": 3.314869279736659,
      "tokens_seen": 3070296064
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.092929292929292e-05,
      "loss": 1.3752,
      "theoretical_loss": 3.3148638651413274,
      "tokens_seen": 3070361600
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.09090909090909e-05,
      "loss": 1.245,
      "theoretical_loss": 3.3148584506939267,
      "tokens_seen": 3070427136
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.088888888888889e-05,
      "loss": 1.2412,
      "theoretical_loss": 3.31485303639445,
      "tokens_seen": 3070492672
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.086868686868687e-05,
      "loss": 1.0315,
      "theoretical_loss": 3.3148476222428904,
      "tokens_seen": 3070558208
    },
    {
      "epoch": 0.86,
      "objective/train/aligned_token_count": 0.41550585627555847,
      "objective/train/avg_token_score": 0.009799261577427387,
      "objective/train/docs_used": 1727712,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2563297748565674,
      "objective/train/original_loss": 4.768157005310059,
      "objective/train/theoretical_loss": 3.314842208239241,
      "objective/train/tokens_used": 1429482976,
      "objective/train/unlikelihood_loss": 0.053655531257390976,
      "theoretical_loss": 3.314842208239241,
      "tokens_seen": 3070623744
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.084848484848486e-05,
      "loss": 1.427,
      "theoretical_loss": 3.314842208239241,
      "tokens_seen": 3070623744
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.082828282828283e-05,
      "loss": 1.4034,
      "theoretical_loss": 3.314836794383494,
      "tokens_seen": 3070689280
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.080808080808081e-05,
      "loss": 1.5111,
      "theoretical_loss": 3.314831380675643,
      "tokens_seen": 3070754816
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.07878787878788e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.3148259671156795,
      "tokens_seen": 3070820352
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.076767676767677e-05,
      "loss": 1.3484,
      "theoretical_loss": 3.3148205537035973,
      "tokens_seen": 3070885888
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.074747474747475e-05,
      "loss": 1.4397,
      "theoretical_loss": 3.314815140439389,
      "tokens_seen": 3070951424
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.072727272727272e-05,
      "loss": 1.3979,
      "theoretical_loss": 3.3148097273230475,
      "tokens_seen": 3071016960
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.07070707070707e-05,
      "loss": 1.4189,
      "theoretical_loss": 3.314804314354565,
      "tokens_seen": 3071082496
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.068686868686869e-05,
      "loss": 1.3719,
      "theoretical_loss": 3.314798901533935,
      "tokens_seen": 3071148032
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.066666666666667e-05,
      "loss": 1.4247,
      "theoretical_loss": 3.31479348886115,
      "tokens_seen": 3071213568
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.064646464646466e-05,
      "loss": 1.4554,
      "theoretical_loss": 3.314788076336203,
      "tokens_seen": 3071279104
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.062626262626263e-05,
      "loss": 1.1624,
      "theoretical_loss": 3.314782663959087,
      "tokens_seen": 3071344640
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.060606060606061e-05,
      "loss": 1.415,
      "theoretical_loss": 3.314777251729794,
      "tokens_seen": 3071410176
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.058585858585858e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.3147718396483175,
      "tokens_seen": 3071475712
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.056565656565656e-05,
      "loss": 1.452,
      "theoretical_loss": 3.31476642771465,
      "tokens_seen": 3071541248
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.054545454545455e-05,
      "loss": 1.4954,
      "theoretical_loss": 3.314761015928785,
      "tokens_seen": 3071606784
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.052525252525252e-05,
      "loss": 1.3659,
      "theoretical_loss": 3.314755604290714,
      "tokens_seen": 3071672320
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.05050505050505e-05,
      "loss": 1.5167,
      "theoretical_loss": 3.314750192800431,
      "tokens_seen": 3071737856
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.048484848484849e-05,
      "loss": 1.4567,
      "theoretical_loss": 3.314744781457928,
      "tokens_seen": 3071803392
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.046464646464647e-05,
      "loss": 1.4587,
      "theoretical_loss": 3.314739370263198,
      "tokens_seen": 3071868928
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.044444444444446e-05,
      "loss": 1.375,
      "theoretical_loss": 3.3147339592162344,
      "tokens_seen": 3071934464
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.042424242424244e-05,
      "loss": 1.3796,
      "theoretical_loss": 3.3147285483170297,
      "tokens_seen": 3072000000
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.04040404040404e-05,
      "loss": 1.5532,
      "theoretical_loss": 3.314723137565576,
      "tokens_seen": 3072065536
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.038383838383838e-05,
      "loss": 1.4022,
      "theoretical_loss": 3.3147177269618675,
      "tokens_seen": 3072131072
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.036363636363636e-05,
      "loss": 1.4005,
      "theoretical_loss": 3.3147123165058954,
      "tokens_seen": 3072196608
    },
    {
      "epoch": 0.86,
      "objective/train/aligned_token_count": 0.4577834904193878,
      "objective/train/avg_token_score": 0.019130781292915344,
      "objective/train/docs_used": 1728325,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5656864643096924,
      "objective/train/original_loss": 4.695202827453613,
      "objective/train/theoretical_loss": 3.314706906197654,
      "objective/train/tokens_used": 1431121376,
      "objective/train/unlikelihood_loss": 0.05910246819257736,
      "theoretical_loss": 3.314706906197654,
      "tokens_seen": 3072262144
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.034343434343435e-05,
      "loss": 1.5616,
      "theoretical_loss": 3.314706906197654,
      "tokens_seen": 3072262144
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.032323232323233e-05,
      "loss": 1.3401,
      "theoretical_loss": 3.3147014960371353,
      "tokens_seen": 3072327680
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.03030303030303e-05,
      "loss": 1.392,
      "theoretical_loss": 3.314696086024332,
      "tokens_seen": 3072393216
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.028282828282829e-05,
      "loss": 1.6484,
      "theoretical_loss": 3.3146906761592376,
      "tokens_seen": 3072458752
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.026262626262627e-05,
      "loss": 1.4434,
      "theoretical_loss": 3.3146852664418445,
      "tokens_seen": 3072524288
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.024242424242424e-05,
      "loss": 1.4214,
      "theoretical_loss": 3.314679856872145,
      "tokens_seen": 3072589824
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.022222222222222e-05,
      "loss": 1.476,
      "theoretical_loss": 3.314674447450133,
      "tokens_seen": 3072655360
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.02020202020202e-05,
      "loss": 1.2694,
      "theoretical_loss": 3.3146690381758006,
      "tokens_seen": 3072720896
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.018181818181818e-05,
      "loss": 1.4589,
      "theoretical_loss": 3.3146636290491407,
      "tokens_seen": 3072786432
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.016161616161616e-05,
      "loss": 1.2871,
      "theoretical_loss": 3.314658220070146,
      "tokens_seen": 3072851968
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.014141414141415e-05,
      "loss": 1.4561,
      "theoretical_loss": 3.3146528112388096,
      "tokens_seen": 3072917504
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.012121212121213e-05,
      "loss": 1.3653,
      "theoretical_loss": 3.3146474025551247,
      "tokens_seen": 3072983040
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.01010101010101e-05,
      "loss": 1.4305,
      "theoretical_loss": 3.3146419940190834,
      "tokens_seen": 3073048576
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.008080808080807e-05,
      "loss": 1.441,
      "theoretical_loss": 3.3146365856306783,
      "tokens_seen": 3073114112
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.006060606060606e-05,
      "loss": 1.3638,
      "theoretical_loss": 3.3146311773899035,
      "tokens_seen": 3073179648
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.004040404040404e-05,
      "loss": 1.242,
      "theoretical_loss": 3.3146257692967507,
      "tokens_seen": 3073245184
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.002020202020202e-05,
      "loss": 1.2142,
      "theoretical_loss": 3.314620361351213,
      "tokens_seen": 3073310720
    },
    {
      "epoch": 0.86,
      "learning_rate": 7.000000000000001e-05,
      "loss": 1.5144,
      "theoretical_loss": 3.314614953553283,
      "tokens_seen": 3073376256
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.997979797979798e-05,
      "loss": 1.334,
      "theoretical_loss": 3.3146095459029543,
      "tokens_seen": 3073441792
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.995959595959596e-05,
      "loss": 1.4634,
      "theoretical_loss": 3.314604138400219,
      "tokens_seen": 3073507328
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.993939393939395e-05,
      "loss": 1.2414,
      "theoretical_loss": 3.31459873104507,
      "tokens_seen": 3073572864
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.991919191919193e-05,
      "loss": 1.414,
      "theoretical_loss": 3.3145933238375003,
      "tokens_seen": 3073638400
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.98989898989899e-05,
      "loss": 1.5413,
      "theoretical_loss": 3.314587916777503,
      "tokens_seen": 3073703936
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.987878787878787e-05,
      "loss": 1.2133,
      "theoretical_loss": 3.3145825098650703,
      "tokens_seen": 3073769472
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.985858585858585e-05,
      "loss": 1.1406,
      "theoretical_loss": 3.314577103100196,
      "tokens_seen": 3073835008
    },
    {
      "epoch": 0.86,
      "objective/train/aligned_token_count": 0.3983382284641266,
      "objective/train/avg_token_score": 0.04411346837878227,
      "objective/train/docs_used": 1729584,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2973413467407227,
      "objective/train/original_loss": 4.0482258796691895,
      "objective/train/theoretical_loss": 3.3145716964828713,
      "objective/train/tokens_used": 1432759776,
      "objective/train/unlikelihood_loss": 0.07682612538337708,
      "theoretical_loss": 3.3145716964828713,
      "tokens_seen": 3073900544
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.983838383838384e-05,
      "loss": 1.3875,
      "theoretical_loss": 3.3145716964828713,
      "tokens_seen": 3073900544
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.981818181818182e-05,
      "loss": 1.2302,
      "theoretical_loss": 3.3145662900130906,
      "tokens_seen": 3073966080
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.97979797979798e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.314560883690846,
      "tokens_seen": 3074031616
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.977777777777778e-05,
      "loss": 1.2938,
      "theoretical_loss": 3.31455547751613,
      "tokens_seen": 3074097152
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.975757575757576e-05,
      "loss": 1.4039,
      "theoretical_loss": 3.3145500714889367,
      "tokens_seen": 3074162688
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.973737373737374e-05,
      "loss": 1.407,
      "theoretical_loss": 3.3145446656092576,
      "tokens_seen": 3074228224
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.971717171717172e-05,
      "loss": 1.2301,
      "theoretical_loss": 3.3145392598770864,
      "tokens_seen": 3074293760
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.96969696969697e-05,
      "loss": 1.1716,
      "theoretical_loss": 3.3145338542924154,
      "tokens_seen": 3074359296
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.967676767676767e-05,
      "loss": 1.3228,
      "theoretical_loss": 3.3145284488552376,
      "tokens_seen": 3074424832
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.965656565656565e-05,
      "loss": 1.3472,
      "theoretical_loss": 3.3145230435655457,
      "tokens_seen": 3074490368
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.963636363636364e-05,
      "loss": 1.2666,
      "theoretical_loss": 3.314517638423333,
      "tokens_seen": 3074555904
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.961616161616162e-05,
      "loss": 1.3634,
      "theoretical_loss": 3.3145122334285917,
      "tokens_seen": 3074621440
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.95959595959596e-05,
      "loss": 1.3774,
      "theoretical_loss": 3.314506828581315,
      "tokens_seen": 3074686976
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.957575757575759e-05,
      "loss": 1.3118,
      "theoretical_loss": 3.3145014238814956,
      "tokens_seen": 3074752512
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.955555555555555e-05,
      "loss": 1.1013,
      "theoretical_loss": 3.3144960193291264,
      "tokens_seen": 3074818048
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.953535353535353e-05,
      "loss": 1.4038,
      "theoretical_loss": 3.3144906149242006,
      "tokens_seen": 3074883584
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.951515151515151e-05,
      "loss": 1.1062,
      "theoretical_loss": 3.31448521066671,
      "tokens_seen": 3074949120
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.94949494949495e-05,
      "loss": 1.4441,
      "theoretical_loss": 3.314479806556649,
      "tokens_seen": 3075014656
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.947474747474748e-05,
      "loss": 1.4506,
      "theoretical_loss": 3.314474402594009,
      "tokens_seen": 3075080192
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.945454545454545e-05,
      "loss": 1.2797,
      "theoretical_loss": 3.3144689987787834,
      "tokens_seen": 3075145728
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.943434343434344e-05,
      "loss": 1.3682,
      "theoretical_loss": 3.3144635951109653,
      "tokens_seen": 3075211264
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.941414141414142e-05,
      "loss": 1.4238,
      "theoretical_loss": 3.314458191590547,
      "tokens_seen": 3075276800
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.93939393939394e-05,
      "loss": 1.3745,
      "theoretical_loss": 3.3144527882175217,
      "tokens_seen": 3075342336
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.937373737373737e-05,
      "loss": 1.5208,
      "theoretical_loss": 3.3144473849918823,
      "tokens_seen": 3075407872
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.935353535353535e-05,
      "loss": 1.2347,
      "theoretical_loss": 3.314441981913621,
      "tokens_seen": 3075473408
    },
    {
      "epoch": 0.86,
      "objective/train/aligned_token_count": 0.29869258403778076,
      "objective/train/avg_token_score": 0.01598837599158287,
      "objective/train/docs_used": 1730199,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8469123244285583,
      "objective/train/original_loss": 5.0171074867248535,
      "objective/train/theoretical_loss": 3.3144365789827317,
      "objective/train/tokens_used": 1434398176,
      "objective/train/unlikelihood_loss": 0.0552532896399498,
      "theoretical_loss": 3.3144365789827317,
      "tokens_seen": 3075538944
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.933333333333333e-05,
      "loss": 1.3266,
      "theoretical_loss": 3.3144365789827317,
      "tokens_seen": 3075538944
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.931313131313131e-05,
      "loss": 1.2142,
      "theoretical_loss": 3.3144311761992062,
      "tokens_seen": 3075604480
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.92929292929293e-05,
      "loss": 1.2104,
      "theoretical_loss": 3.314425773563038,
      "tokens_seen": 3075670016
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.927272727272728e-05,
      "loss": 1.187,
      "theoretical_loss": 3.3144203710742195,
      "tokens_seen": 3075735552
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.925252525252525e-05,
      "loss": 1.4196,
      "theoretical_loss": 3.314414968732744,
      "tokens_seen": 3075801088
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.923232323232324e-05,
      "loss": 1.4276,
      "theoretical_loss": 3.3144095665386044,
      "tokens_seen": 3075866624
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.92121212121212e-05,
      "loss": 1.5895,
      "theoretical_loss": 3.3144041644917928,
      "tokens_seen": 3075932160
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.919191919191919e-05,
      "loss": 1.375,
      "theoretical_loss": 3.314398762592303,
      "tokens_seen": 3075997696
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.917171717171717e-05,
      "loss": 1.4517,
      "theoretical_loss": 3.3143933608401266,
      "tokens_seen": 3076063232
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.915151515151516e-05,
      "loss": 1.2864,
      "theoretical_loss": 3.314387959235258,
      "tokens_seen": 3076128768
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.913131313131313e-05,
      "loss": 1.2755,
      "theoretical_loss": 3.3143825577776886,
      "tokens_seen": 3076194304
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.911111111111111e-05,
      "loss": 1.2188,
      "theoretical_loss": 3.3143771564674123,
      "tokens_seen": 3076259840
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.90909090909091e-05,
      "loss": 1.4839,
      "theoretical_loss": 3.314371755304421,
      "tokens_seen": 3076325376
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.907070707070708e-05,
      "loss": 1.2591,
      "theoretical_loss": 3.3143663542887087,
      "tokens_seen": 3076390912
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.905050505050506e-05,
      "loss": 1.1743,
      "theoretical_loss": 3.3143609534202674,
      "tokens_seen": 3076456448
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.903030303030302e-05,
      "loss": 1.4535,
      "theoretical_loss": 3.31435555269909,
      "tokens_seen": 3076521984
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.9010101010101e-05,
      "loss": 1.1004,
      "theoretical_loss": 3.31435015212517,
      "tokens_seen": 3076587520
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.898989898989899e-05,
      "loss": 1.2695,
      "theoretical_loss": 3.314344751698499,
      "tokens_seen": 3076653056
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.896969696969697e-05,
      "loss": 1.0634,
      "theoretical_loss": 3.3143393514190715,
      "tokens_seen": 3076718592
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.894949494949496e-05,
      "loss": 1.3929,
      "theoretical_loss": 3.314333951286879,
      "tokens_seen": 3076784128
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.892929292929293e-05,
      "loss": 1.3321,
      "theoretical_loss": 3.314328551301915,
      "tokens_seen": 3076849664
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.890909090909091e-05,
      "loss": 1.1877,
      "theoretical_loss": 3.314323151464172,
      "tokens_seen": 3076915200
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.88888888888889e-05,
      "loss": 1.2458,
      "theoretical_loss": 3.314317751773643,
      "tokens_seen": 3076980736
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.886868686868688e-05,
      "loss": 1.3385,
      "theoretical_loss": 3.3143123522303206,
      "tokens_seen": 3077046272
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.884848484848485e-05,
      "loss": 1.2628,
      "theoretical_loss": 3.3143069528341984,
      "tokens_seen": 3077111808
    },
    {
      "epoch": 0.86,
      "objective/train/aligned_token_count": 0.40890762209892273,
      "objective/train/avg_token_score": 0.016667621210217476,
      "objective/train/docs_used": 1730859,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1481164693832397,
      "objective/train/original_loss": 4.386240005493164,
      "objective/train/theoretical_loss": 3.3143015535852687,
      "objective/train/tokens_used": 1436036576,
      "objective/train/unlikelihood_loss": 0.06950560212135315,
      "theoretical_loss": 3.3143015535852687,
      "tokens_seen": 3077177344
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.882828282828282e-05,
      "loss": 1.2698,
      "theoretical_loss": 3.3143015535852687,
      "tokens_seen": 3077177344
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.88080808080808e-05,
      "loss": 1.4712,
      "theoretical_loss": 3.3142961544835243,
      "tokens_seen": 3077242880
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.878787878787879e-05,
      "loss": 1.3051,
      "theoretical_loss": 3.3142907555289582,
      "tokens_seen": 3077308416
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.876767676767677e-05,
      "loss": 1.2116,
      "theoretical_loss": 3.3142853567215633,
      "tokens_seen": 3077373952
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.874747474747476e-05,
      "loss": 1.3424,
      "theoretical_loss": 3.3142799580613325,
      "tokens_seen": 3077439488
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.872727272727274e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.314274559548258,
      "tokens_seen": 3077505024
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.870707070707071e-05,
      "loss": 1.2807,
      "theoretical_loss": 3.3142691611823336,
      "tokens_seen": 3077570560
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.868686868686868e-05,
      "loss": 1.201,
      "theoretical_loss": 3.314263762963552,
      "tokens_seen": 3077636096
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.866666666666666e-05,
      "loss": 1.4608,
      "theoretical_loss": 3.3142583648919053,
      "tokens_seen": 3077701632
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.864646464646465e-05,
      "loss": 1.5125,
      "theoretical_loss": 3.3142529669673872,
      "tokens_seen": 3077767168
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.862626262626263e-05,
      "loss": 1.2963,
      "theoretical_loss": 3.31424756918999,
      "tokens_seen": 3077832704
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.86060606060606e-05,
      "loss": 1.2936,
      "theoretical_loss": 3.314242171559707,
      "tokens_seen": 3077898240
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.858585858585859e-05,
      "loss": 1.4952,
      "theoretical_loss": 3.3142367740765306,
      "tokens_seen": 3077963776
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.856565656565657e-05,
      "loss": 1.3396,
      "theoretical_loss": 3.3142313767404543,
      "tokens_seen": 3078029312
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.854545454545455e-05,
      "loss": 1.2697,
      "theoretical_loss": 3.31422597955147,
      "tokens_seen": 3078094848
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.852525252525254e-05,
      "loss": 1.3205,
      "theoretical_loss": 3.3142205825095714,
      "tokens_seen": 3078160384
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.85050505050505e-05,
      "loss": 1.3603,
      "theoretical_loss": 3.314215185614751,
      "tokens_seen": 3078225920
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.848484848484848e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.314209788867002,
      "tokens_seen": 3078291456
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.846464646464646e-05,
      "loss": 1.1999,
      "theoretical_loss": 3.3142043922663165,
      "tokens_seen": 3078356992
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.844444444444445e-05,
      "loss": 1.3218,
      "theoretical_loss": 3.3141989958126885,
      "tokens_seen": 3078422528
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.842424242424243e-05,
      "loss": 1.385,
      "theoretical_loss": 3.3141935995061096,
      "tokens_seen": 3078488064
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.84040404040404e-05,
      "loss": 1.456,
      "theoretical_loss": 3.314188203346574,
      "tokens_seen": 3078553600
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.838383838383839e-05,
      "loss": 1.4607,
      "theoretical_loss": 3.314182807334073,
      "tokens_seen": 3078619136
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.836363636363637e-05,
      "loss": 1.1676,
      "theoretical_loss": 3.314177411468601,
      "tokens_seen": 3078684672
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.834343434343434e-05,
      "loss": 1.1492,
      "theoretical_loss": 3.3141720157501497,
      "tokens_seen": 3078750208
    },
    {
      "epoch": 0.86,
      "objective/train/aligned_token_count": 0.3105449676513672,
      "objective/train/avg_token_score": 0.039454683661460876,
      "objective/train/docs_used": 1732124,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8731678128242493,
      "objective/train/original_loss": 4.892634391784668,
      "objective/train/theoretical_loss": 3.3141666201787126,
      "objective/train/tokens_used": 1437674976,
      "objective/train/unlikelihood_loss": 0.06003899499773979,
      "theoretical_loss": 3.3141666201787126,
      "tokens_seen": 3078815744
    },
    {
      "epoch": 0.86,
      "learning_rate": 6.832323232323232e-05,
      "loss": 1.185,
      "theoretical_loss": 3.3141666201787126,
      "tokens_seen": 3078815744
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.830303030303031e-05,
      "loss": 1.5023,
      "theoretical_loss": 3.314161224754282,
      "tokens_seen": 3078881280
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.828282828282828e-05,
      "loss": 1.2948,
      "theoretical_loss": 3.314155829476852,
      "tokens_seen": 3078946816
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.826262626262626e-05,
      "loss": 1.2632,
      "theoretical_loss": 3.314150434346414,
      "tokens_seen": 3079012352
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.824242424242425e-05,
      "loss": 1.405,
      "theoretical_loss": 3.314145039362962,
      "tokens_seen": 3079077888
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.822222222222223e-05,
      "loss": 1.1738,
      "theoretical_loss": 3.314139644526488,
      "tokens_seen": 3079143424
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.820202020202021e-05,
      "loss": 1.3184,
      "theoretical_loss": 3.3141342498369855,
      "tokens_seen": 3079208960
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.818181818181818e-05,
      "loss": 1.4369,
      "theoretical_loss": 3.314128855294447,
      "tokens_seen": 3079274496
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.816161616161615e-05,
      "loss": 1.2476,
      "theoretical_loss": 3.3141234608988652,
      "tokens_seen": 3079340032
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.814141414141414e-05,
      "loss": 1.3454,
      "theoretical_loss": 3.3141180666502335,
      "tokens_seen": 3079405568
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.812121212121212e-05,
      "loss": 1.4211,
      "theoretical_loss": 3.3141126725485446,
      "tokens_seen": 3079471104
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.81010101010101e-05,
      "loss": 1.2513,
      "theoretical_loss": 3.314107278593791,
      "tokens_seen": 3079536640
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.808080808080808e-05,
      "loss": 1.3808,
      "theoretical_loss": 3.3141018847859662,
      "tokens_seen": 3079602176
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.806060606060606e-05,
      "loss": 1.4326,
      "theoretical_loss": 3.3140964911250625,
      "tokens_seen": 3079667712
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.804040404040405e-05,
      "loss": 1.1961,
      "theoretical_loss": 3.3140910976110733,
      "tokens_seen": 3079733248
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.802020202020203e-05,
      "loss": 1.2603,
      "theoretical_loss": 3.3140857042439906,
      "tokens_seen": 3079798784
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.800000000000001e-05,
      "loss": 1.1839,
      "theoretical_loss": 3.314080311023808,
      "tokens_seen": 3079864320
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.797979797979797e-05,
      "loss": 1.3516,
      "theoretical_loss": 3.314074917950519,
      "tokens_seen": 3079929856
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.795959595959595e-05,
      "loss": 1.4675,
      "theoretical_loss": 3.314069525024115,
      "tokens_seen": 3079995392
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.793939393939394e-05,
      "loss": 1.322,
      "theoretical_loss": 3.314064132244589,
      "tokens_seen": 3080060928
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.791919191919192e-05,
      "loss": 1.2932,
      "theoretical_loss": 3.3140587396119354,
      "tokens_seen": 3080126464
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.78989898989899e-05,
      "loss": 1.501,
      "theoretical_loss": 3.314053347126146,
      "tokens_seen": 3080192000
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.787878787878789e-05,
      "loss": 1.1144,
      "theoretical_loss": 3.3140479547872133,
      "tokens_seen": 3080257536
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.785858585858586e-05,
      "loss": 1.0588,
      "theoretical_loss": 3.314042562595131,
      "tokens_seen": 3080323072
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.783838383838384e-05,
      "loss": 1.2552,
      "theoretical_loss": 3.3140371705498914,
      "tokens_seen": 3080388608
    },
    {
      "epoch": 0.87,
      "objective/train/aligned_token_count": 0.5535801649093628,
      "objective/train/avg_token_score": 0.03102460317313671,
      "objective/train/docs_used": 1732703,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.9124468564987183,
      "objective/train/original_loss": 4.536565780639648,
      "objective/train/theoretical_loss": 3.314031778651488,
      "objective/train/tokens_used": 1439313376,
      "objective/train/unlikelihood_loss": 0.03262428566813469,
      "theoretical_loss": 3.314031778651488,
      "tokens_seen": 3080454144
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.781818181818181e-05,
      "loss": 1.2698,
      "theoretical_loss": 3.314031778651488,
      "tokens_seen": 3080454144
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.77979797979798e-05,
      "loss": 1.1198,
      "theoretical_loss": 3.314026386899913,
      "tokens_seen": 3080519680
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.777777777777778e-05,
      "loss": 1.2281,
      "theoretical_loss": 3.3140209952951594,
      "tokens_seen": 3080585216
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.775757575757575e-05,
      "loss": 1.3869,
      "theoretical_loss": 3.3140156038372206,
      "tokens_seen": 3080650752
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.773737373737374e-05,
      "loss": 1.3668,
      "theoretical_loss": 3.314010212526089,
      "tokens_seen": 3080716288
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.771717171717172e-05,
      "loss": 1.371,
      "theoretical_loss": 3.314004821361758,
      "tokens_seen": 3080781824
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.76969696969697e-05,
      "loss": 1.4135,
      "theoretical_loss": 3.3139994303442197,
      "tokens_seen": 3080847360
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.767676767676769e-05,
      "loss": 1.3838,
      "theoretical_loss": 3.3139940394734673,
      "tokens_seen": 3080912896
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.765656565656566e-05,
      "loss": 1.3312,
      "theoretical_loss": 3.313988648749494,
      "tokens_seen": 3080978432
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.763636363636363e-05,
      "loss": 1.4022,
      "theoretical_loss": 3.313983258172292,
      "tokens_seen": 3081043968
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.761616161616161e-05,
      "loss": 1.4535,
      "theoretical_loss": 3.3139778677418548,
      "tokens_seen": 3081109504
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.75959595959596e-05,
      "loss": 1.4342,
      "theoretical_loss": 3.3139724774581754,
      "tokens_seen": 3081175040
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.757575757575758e-05,
      "loss": 1.3223,
      "theoretical_loss": 3.313967087321246,
      "tokens_seen": 3081240576
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.755555555555555e-05,
      "loss": 1.44,
      "theoretical_loss": 3.31396169733106,
      "tokens_seen": 3081306112
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.753535353535354e-05,
      "loss": 1.1792,
      "theoretical_loss": 3.3139563074876106,
      "tokens_seen": 3081371648
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.751515151515152e-05,
      "loss": 1.4776,
      "theoretical_loss": 3.3139509177908897,
      "tokens_seen": 3081437184
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.74949494949495e-05,
      "loss": 1.2696,
      "theoretical_loss": 3.3139455282408905,
      "tokens_seen": 3081502720
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.747474747474747e-05,
      "loss": 1.1486,
      "theoretical_loss": 3.313940138837607,
      "tokens_seen": 3081568256
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.745454545454546e-05,
      "loss": 1.2741,
      "theoretical_loss": 3.3139347495810303,
      "tokens_seen": 3081633792
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.743434343434343e-05,
      "loss": 1.3641,
      "theoretical_loss": 3.3139293604711546,
      "tokens_seen": 3081699328
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.741414141414141e-05,
      "loss": 1.2633,
      "theoretical_loss": 3.3139239715079722,
      "tokens_seen": 3081764864
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.73939393939394e-05,
      "loss": 1.2176,
      "theoretical_loss": 3.313918582691476,
      "tokens_seen": 3081830400
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.737373737373738e-05,
      "loss": 1.4008,
      "theoretical_loss": 3.3139131940216595,
      "tokens_seen": 3081895936
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.735353535353536e-05,
      "loss": 1.3412,
      "theoretical_loss": 3.313907805498515,
      "tokens_seen": 3081961472
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.733333333333333e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.3139024171220353,
      "tokens_seen": 3082027008
    },
    {
      "epoch": 0.87,
      "objective/train/aligned_token_count": 0.4754398763179779,
      "objective/train/avg_token_score": 0.005251714494079351,
      "objective/train/docs_used": 1733946,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3580626249313354,
      "objective/train/original_loss": 3.6132917404174805,
      "objective/train/theoretical_loss": 3.313897028892214,
      "objective/train/tokens_used": 1440951776,
      "objective/train/unlikelihood_loss": 0.07232429832220078,
      "theoretical_loss": 3.313897028892214,
      "tokens_seen": 3082092544
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.731313131313132e-05,
      "loss": 1.4771,
      "theoretical_loss": 3.313897028892214,
      "tokens_seen": 3082092544
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.729292929292929e-05,
      "loss": 1.2277,
      "theoretical_loss": 3.3138916408090426,
      "tokens_seen": 3082158080
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.727272727272727e-05,
      "loss": 1.4038,
      "theoretical_loss": 3.3138862528725155,
      "tokens_seen": 3082223616
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.725252525252526e-05,
      "loss": 1.4223,
      "theoretical_loss": 3.3138808650826252,
      "tokens_seen": 3082289152
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.723232323232323e-05,
      "loss": 1.1524,
      "theoretical_loss": 3.313875477439364,
      "tokens_seen": 3082354688
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.721212121212121e-05,
      "loss": 1.3266,
      "theoretical_loss": 3.313870089942725,
      "tokens_seen": 3082420224
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.71919191919192e-05,
      "loss": 1.1426,
      "theoretical_loss": 3.3138647025927015,
      "tokens_seen": 3082485760
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.717171717171718e-05,
      "loss": 1.355,
      "theoretical_loss": 3.313859315389286,
      "tokens_seen": 3082551296
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.715151515151516e-05,
      "loss": 1.2617,
      "theoretical_loss": 3.313853928332472,
      "tokens_seen": 3082616832
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.713131313131312e-05,
      "loss": 1.3782,
      "theoretical_loss": 3.313848541422251,
      "tokens_seen": 3082682368
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.71111111111111e-05,
      "loss": 1.5823,
      "theoretical_loss": 3.313843154658618,
      "tokens_seen": 3082747904
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.709090909090909e-05,
      "loss": 1.6497,
      "theoretical_loss": 3.313837768041564,
      "tokens_seen": 3082813440
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.707070707070707e-05,
      "loss": 1.3105,
      "theoretical_loss": 3.313832381571083,
      "tokens_seen": 3082878976
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.705050505050506e-05,
      "loss": 1.4009,
      "theoretical_loss": 3.313826995247167,
      "tokens_seen": 3082944512
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.703030303030303e-05,
      "loss": 1.1269,
      "theoretical_loss": 3.3138216090698096,
      "tokens_seen": 3083010048
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.701010101010101e-05,
      "loss": 1.1212,
      "theoretical_loss": 3.313816223039004,
      "tokens_seen": 3083075584
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.6989898989899e-05,
      "loss": 1.2397,
      "theoretical_loss": 3.313810837154742,
      "tokens_seen": 3083141120
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.696969696969698e-05,
      "loss": 1.3077,
      "theoretical_loss": 3.3138054514170174,
      "tokens_seen": 3083206656
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.694949494949495e-05,
      "loss": 1.1759,
      "theoretical_loss": 3.3138000658258226,
      "tokens_seen": 3083272192
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.692929292929293e-05,
      "loss": 1.3765,
      "theoretical_loss": 3.313794680381151,
      "tokens_seen": 3083337728
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.69090909090909e-05,
      "loss": 1.3914,
      "theoretical_loss": 3.313789295082995,
      "tokens_seen": 3083403264
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.688888888888889e-05,
      "loss": 1.0139,
      "theoretical_loss": 3.3137839099313475,
      "tokens_seen": 3083468800
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.686868686868687e-05,
      "loss": 1.3719,
      "theoretical_loss": 3.313778524926202,
      "tokens_seen": 3083534336
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.684848484848485e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.3137731400675507,
      "tokens_seen": 3083599872
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.682828282828284e-05,
      "loss": 1.4655,
      "theoretical_loss": 3.313767755355387,
      "tokens_seen": 3083665408
    },
    {
      "epoch": 0.87,
      "objective/train/aligned_token_count": 0.5432551503181458,
      "objective/train/avg_token_score": 0.005819818936288357,
      "objective/train/docs_used": 1734728,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4318212270736694,
      "objective/train/original_loss": 3.8242740631103516,
      "objective/train/theoretical_loss": 3.3137623707897035,
      "objective/train/tokens_used": 1442590176,
      "objective/train/unlikelihood_loss": 0.04773322120308876,
      "theoretical_loss": 3.3137623707897035,
      "tokens_seen": 3083730944
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.680808080808081e-05,
      "loss": 1.4088,
      "theoretical_loss": 3.3137623707897035,
      "tokens_seen": 3083730944
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.678787878787878e-05,
      "loss": 1.3782,
      "theoretical_loss": 3.3137569863704934,
      "tokens_seen": 3083796480
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.676767676767676e-05,
      "loss": 1.29,
      "theoretical_loss": 3.3137516020977493,
      "tokens_seen": 3083862016
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.674747474747475e-05,
      "loss": 1.1659,
      "theoretical_loss": 3.313746217971464,
      "tokens_seen": 3083927552
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.672727272727273e-05,
      "loss": 1.3222,
      "theoretical_loss": 3.3137408339916306,
      "tokens_seen": 3083993088
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.67070707070707e-05,
      "loss": 1.4403,
      "theoretical_loss": 3.3137354501582426,
      "tokens_seen": 3084058624
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.668686868686869e-05,
      "loss": 1.5043,
      "theoretical_loss": 3.3137300664712916,
      "tokens_seen": 3084124160
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.666666666666667e-05,
      "loss": 1.3578,
      "theoretical_loss": 3.3137246829307716,
      "tokens_seen": 3084189696
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.664646464646465e-05,
      "loss": 1.3811,
      "theoretical_loss": 3.3137192995366753,
      "tokens_seen": 3084255232
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.662626262626264e-05,
      "loss": 1.2305,
      "theoretical_loss": 3.3137139162889953,
      "tokens_seen": 3084320768
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.66060606060606e-05,
      "loss": 1.3807,
      "theoretical_loss": 3.3137085331877243,
      "tokens_seen": 3084386304
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.658585858585858e-05,
      "loss": 1.375,
      "theoretical_loss": 3.313703150232856,
      "tokens_seen": 3084451840
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.656565656565656e-05,
      "loss": 1.246,
      "theoretical_loss": 3.313697767424383,
      "tokens_seen": 3084517376
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.654545454545455e-05,
      "loss": 1.2293,
      "theoretical_loss": 3.3136923847622977,
      "tokens_seen": 3084582912
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.652525252525253e-05,
      "loss": 1.4878,
      "theoretical_loss": 3.3136870022465934,
      "tokens_seen": 3084648448
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.650505050505051e-05,
      "loss": 1.3745,
      "theoretical_loss": 3.313681619877263,
      "tokens_seen": 3084713984
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.648484848484848e-05,
      "loss": 1.2708,
      "theoretical_loss": 3.3136762376543,
      "tokens_seen": 3084779520
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.646464646464647e-05,
      "loss": 1.0005,
      "theoretical_loss": 3.313670855577696,
      "tokens_seen": 3084845056
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.644444444444445e-05,
      "loss": 1.5441,
      "theoretical_loss": 3.3136654736474447,
      "tokens_seen": 3084910592
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.642424242424242e-05,
      "loss": 1.261,
      "theoretical_loss": 3.313660091863539,
      "tokens_seen": 3084976128
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.640404040404041e-05,
      "loss": 1.5191,
      "theoretical_loss": 3.313654710225972,
      "tokens_seen": 3085041664
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.638383838383838e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.313649328734736,
      "tokens_seen": 3085107200
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.636363636363636e-05,
      "loss": 1.6617,
      "theoretical_loss": 3.3136439473898243,
      "tokens_seen": 3085172736
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.634343434343435e-05,
      "loss": 1.3024,
      "theoretical_loss": 3.31363856619123,
      "tokens_seen": 3085238272
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.632323232323233e-05,
      "loss": 1.5326,
      "theoretical_loss": 3.313633185138946,
      "tokens_seen": 3085303808
    },
    {
      "epoch": 0.87,
      "objective/train/aligned_token_count": 0.32869014143943787,
      "objective/train/avg_token_score": 0.011612819507718086,
      "objective/train/docs_used": 1736039,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.9300379157066345,
      "objective/train/original_loss": 4.554388999938965,
      "objective/train/theoretical_loss": 3.3136278042329645,
      "objective/train/tokens_used": 1444228576,
      "objective/train/unlikelihood_loss": 0.048059023916721344,
      "theoretical_loss": 3.3136278042329645,
      "tokens_seen": 3085369344
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.630303030303031e-05,
      "loss": 1.2502,
      "theoretical_loss": 3.3136278042329645,
      "tokens_seen": 3085369344
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.628282828282828e-05,
      "loss": 1.1855,
      "theoretical_loss": 3.313622423473279,
      "tokens_seen": 3085434880
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.626262626262625e-05,
      "loss": 1.484,
      "theoretical_loss": 3.3136170428598826,
      "tokens_seen": 3085500416
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.624242424242424e-05,
      "loss": 1.3578,
      "theoretical_loss": 3.313611662392768,
      "tokens_seen": 3085565952
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.622222222222222e-05,
      "loss": 1.4431,
      "theoretical_loss": 3.313606282071928,
      "tokens_seen": 3085631488
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.62020202020202e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.3136009018973556,
      "tokens_seen": 3085697024
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.618181818181818e-05,
      "loss": 1.1029,
      "theoretical_loss": 3.3135955218690434,
      "tokens_seen": 3085762560
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.616161616161616e-05,
      "loss": 1.2131,
      "theoretical_loss": 3.313590141986985,
      "tokens_seen": 3085828096
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.614141414141414e-05,
      "loss": 1.4511,
      "theoretical_loss": 3.313584762251173,
      "tokens_seen": 3085893632
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.612121212121213e-05,
      "loss": 1.1919,
      "theoretical_loss": 3.3135793826616,
      "tokens_seen": 3085959168
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.610101010101011e-05,
      "loss": 1.1742,
      "theoretical_loss": 3.3135740032182595,
      "tokens_seen": 3086024704
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.608080808080808e-05,
      "loss": 1.0671,
      "theoretical_loss": 3.3135686239211437,
      "tokens_seen": 3086090240
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.606060606060605e-05,
      "loss": 1.1685,
      "theoretical_loss": 3.313563244770246,
      "tokens_seen": 3086155776
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.604040404040404e-05,
      "loss": 1.3549,
      "theoretical_loss": 3.3135578657655596,
      "tokens_seen": 3086221312
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.602020202020202e-05,
      "loss": 1.412,
      "theoretical_loss": 3.3135524869070765,
      "tokens_seen": 3086286848
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.6e-05,
      "loss": 1.1501,
      "theoretical_loss": 3.3135471081947907,
      "tokens_seen": 3086352384
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.597979797979799e-05,
      "loss": 1.0131,
      "theoretical_loss": 3.313541729628694,
      "tokens_seen": 3086417920
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.595959595959596e-05,
      "loss": 1.1781,
      "theoretical_loss": 3.3135363512087808,
      "tokens_seen": 3086483456
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.593939393939394e-05,
      "loss": 1.1003,
      "theoretical_loss": 3.3135309729350424,
      "tokens_seen": 3086548992
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.591919191919191e-05,
      "loss": 1.3631,
      "theoretical_loss": 3.313525594807473,
      "tokens_seen": 3086614528
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.58989898989899e-05,
      "loss": 1.3777,
      "theoretical_loss": 3.3135202168260647,
      "tokens_seen": 3086680064
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.587878787878788e-05,
      "loss": 1.3435,
      "theoretical_loss": 3.3135148389908107,
      "tokens_seen": 3086745600
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.585858585858585e-05,
      "loss": 1.1922,
      "theoretical_loss": 3.313509461301704,
      "tokens_seen": 3086811136
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.583838383838384e-05,
      "loss": 1.2651,
      "theoretical_loss": 3.3135040837587377,
      "tokens_seen": 3086876672
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.581818181818182e-05,
      "loss": 1.3956,
      "theoretical_loss": 3.313498706361904,
      "tokens_seen": 3086942208
    },
    {
      "epoch": 0.87,
      "objective/train/aligned_token_count": 0.4340175986289978,
      "objective/train/avg_token_score": 0.008154028095304966,
      "objective/train/docs_used": 1736664,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5367728471755981,
      "objective/train/original_loss": 4.431552410125732,
      "objective/train/theoretical_loss": 3.313493329111197,
      "objective/train/tokens_used": 1445866976,
      "objective/train/unlikelihood_loss": 0.05472097918391228,
      "theoretical_loss": 3.313493329111197,
      "tokens_seen": 3087007744
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.57979797979798e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.313493329111197,
      "tokens_seen": 3087007744
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.577777777777779e-05,
      "loss": 1.1845,
      "theoretical_loss": 3.313487952006609,
      "tokens_seen": 3087073280
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.575757575757576e-05,
      "loss": 1.1247,
      "theoretical_loss": 3.3134825750481323,
      "tokens_seen": 3087138816
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.573737373737373e-05,
      "loss": 1.2686,
      "theoretical_loss": 3.3134771982357605,
      "tokens_seen": 3087204352
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.571717171717171e-05,
      "loss": 1.4866,
      "theoretical_loss": 3.3134718215694865,
      "tokens_seen": 3087269888
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.56969696969697e-05,
      "loss": 1.1561,
      "theoretical_loss": 3.313466445049303,
      "tokens_seen": 3087335424
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.567676767676768e-05,
      "loss": 1.3678,
      "theoretical_loss": 3.3134610686752035,
      "tokens_seen": 3087400960
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.565656565656566e-05,
      "loss": 1.3402,
      "theoretical_loss": 3.3134556924471807,
      "tokens_seen": 3087466496
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.563636363636364e-05,
      "loss": 1.4461,
      "theoretical_loss": 3.313450316365227,
      "tokens_seen": 3087532032
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.561616161616162e-05,
      "loss": 1.5468,
      "theoretical_loss": 3.313444940429336,
      "tokens_seen": 3087597568
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.55959595959596e-05,
      "loss": 1.4144,
      "theoretical_loss": 3.3134395646394994,
      "tokens_seen": 3087663104
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.557575757575759e-05,
      "loss": 1.3429,
      "theoretical_loss": 3.313434188995712,
      "tokens_seen": 3087728640
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.555555555555556e-05,
      "loss": 1.2863,
      "theoretical_loss": 3.3134288134979655,
      "tokens_seen": 3087794176
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.553535353535353e-05,
      "loss": 1.4411,
      "theoretical_loss": 3.313423438146253,
      "tokens_seen": 3087859712
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.551515151515151e-05,
      "loss": 1.3055,
      "theoretical_loss": 3.313418062940568,
      "tokens_seen": 3087925248
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.54949494949495e-05,
      "loss": 1.5197,
      "theoretical_loss": 3.3134126878809025,
      "tokens_seen": 3087990784
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.547474747474748e-05,
      "loss": 1.4101,
      "theoretical_loss": 3.31340731296725,
      "tokens_seen": 3088056320
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.545454545454546e-05,
      "loss": 1.1874,
      "theoretical_loss": 3.313401938199603,
      "tokens_seen": 3088121856
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.543434343434343e-05,
      "loss": 1.2853,
      "theoretical_loss": 3.3133965635779554,
      "tokens_seen": 3088187392
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.541414141414142e-05,
      "loss": 1.2956,
      "theoretical_loss": 3.3133911891022994,
      "tokens_seen": 3088252928
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.539393939393939e-05,
      "loss": 1.3688,
      "theoretical_loss": 3.3133858147726283,
      "tokens_seen": 3088318464
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.537373737373737e-05,
      "loss": 1.2563,
      "theoretical_loss": 3.313380440588934,
      "tokens_seen": 3088384000
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.535353535353536e-05,
      "loss": 1.4649,
      "theoretical_loss": 3.3133750665512114,
      "tokens_seen": 3088449536
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.533333333333333e-05,
      "loss": 1.2728,
      "theoretical_loss": 3.3133696926594514,
      "tokens_seen": 3088515072
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.531313131313131e-05,
      "loss": 1.1024,
      "theoretical_loss": 3.3133643189136484,
      "tokens_seen": 3088580608
    },
    {
      "epoch": 0.87,
      "objective/train/aligned_token_count": 0.4456256031990051,
      "objective/train/avg_token_score": 0.030121799558401108,
      "objective/train/docs_used": 1737965,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3233076333999634,
      "objective/train/original_loss": 4.245978355407715,
      "objective/train/theoretical_loss": 3.313358945313794,
      "objective/train/tokens_used": 1447505376,
      "objective/train/unlikelihood_loss": 0.07466845214366913,
      "theoretical_loss": 3.313358945313794,
      "tokens_seen": 3088646144
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.52929292929293e-05,
      "loss": 1.4985,
      "theoretical_loss": 3.313358945313794,
      "tokens_seen": 3088646144
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.527272727272728e-05,
      "loss": 1.2903,
      "theoretical_loss": 3.3133535718598823,
      "tokens_seen": 3088711680
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.525252525252526e-05,
      "loss": 1.568,
      "theoretical_loss": 3.313348198551906,
      "tokens_seen": 3088777216
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.523232323232325e-05,
      "loss": 1.4521,
      "theoretical_loss": 3.313342825389858,
      "tokens_seen": 3088842752
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.52121212121212e-05,
      "loss": 1.3645,
      "theoretical_loss": 3.3133374523737307,
      "tokens_seen": 3088908288
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.519191919191919e-05,
      "loss": 1.3231,
      "theoretical_loss": 3.3133320795035175,
      "tokens_seen": 3088973824
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.517171717171717e-05,
      "loss": 1.4127,
      "theoretical_loss": 3.313326706779211,
      "tokens_seen": 3089039360
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.515151515151516e-05,
      "loss": 1.389,
      "theoretical_loss": 3.313321334200805,
      "tokens_seen": 3089104896
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.513131313131314e-05,
      "loss": 1.2505,
      "theoretical_loss": 3.313315961768292,
      "tokens_seen": 3089170432
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.511111111111111e-05,
      "loss": 1.1537,
      "theoretical_loss": 3.3133105894816643,
      "tokens_seen": 3089235968
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.50909090909091e-05,
      "loss": 1.3158,
      "theoretical_loss": 3.3133052173409157,
      "tokens_seen": 3089301504
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.507070707070708e-05,
      "loss": 1.2138,
      "theoretical_loss": 3.3132998453460387,
      "tokens_seen": 3089367040
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.505050505050505e-05,
      "loss": 1.3894,
      "theoretical_loss": 3.3132944734970264,
      "tokens_seen": 3089432576
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.503030303030303e-05,
      "loss": 1.2004,
      "theoretical_loss": 3.3132891017938717,
      "tokens_seen": 3089498112
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.5010101010101e-05,
      "loss": 1.2954,
      "theoretical_loss": 3.3132837302365674,
      "tokens_seen": 3089563648
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.498989898989899e-05,
      "loss": 1.4837,
      "theoretical_loss": 3.313278358825107,
      "tokens_seen": 3089629184
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.496969696969697e-05,
      "loss": 1.6036,
      "theoretical_loss": 3.3132729875594826,
      "tokens_seen": 3089694720
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.494949494949495e-05,
      "loss": 1.248,
      "theoretical_loss": 3.313267616439688,
      "tokens_seen": 3089760256
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.492929292929294e-05,
      "loss": 1.3209,
      "theoretical_loss": 3.3132622454657152,
      "tokens_seen": 3089825792
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.490909090909091e-05,
      "loss": 1.2851,
      "theoretical_loss": 3.3132568746375584,
      "tokens_seen": 3089891328
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.488888888888889e-05,
      "loss": 1.5061,
      "theoretical_loss": 3.3132515039552093,
      "tokens_seen": 3089956864
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.486868686868686e-05,
      "loss": 1.6495,
      "theoretical_loss": 3.3132461334186614,
      "tokens_seen": 3090022400
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.484848484848485e-05,
      "loss": 1.2643,
      "theoretical_loss": 3.313240763027908,
      "tokens_seen": 3090087936
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.482828282828283e-05,
      "loss": 1.3697,
      "theoretical_loss": 3.3132353927829414,
      "tokens_seen": 3090153472
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.480808080808081e-05,
      "loss": 1.3105,
      "theoretical_loss": 3.313230022683755,
      "tokens_seen": 3090219008
    },
    {
      "epoch": 0.87,
      "objective/train/aligned_token_count": 0.4040811359882355,
      "objective/train/avg_token_score": 0.06727245450019836,
      "objective/train/docs_used": 1738543,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2791539430618286,
      "objective/train/original_loss": 5.033825874328613,
      "objective/train/theoretical_loss": 3.3132246527303413,
      "objective/train/tokens_used": 1449143776,
      "objective/train/unlikelihood_loss": 0.03325916826725006,
      "theoretical_loss": 3.3132246527303413,
      "tokens_seen": 3090284544
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.478787878787879e-05,
      "loss": 1.3697,
      "theoretical_loss": 3.3132246527303413,
      "tokens_seen": 3090284544
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.476767676767677e-05,
      "loss": 1.2789,
      "theoretical_loss": 3.313219282922694,
      "tokens_seen": 3090350080
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.474747474747475e-05,
      "loss": 1.2965,
      "theoretical_loss": 3.313213913260805,
      "tokens_seen": 3090415616
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.472727272727274e-05,
      "loss": 1.1388,
      "theoretical_loss": 3.3132085437446683,
      "tokens_seen": 3090481152
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.470707070707071e-05,
      "loss": 1.2582,
      "theoretical_loss": 3.3132031743742765,
      "tokens_seen": 3090546688
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.468686868686868e-05,
      "loss": 1.4221,
      "theoretical_loss": 3.313197805149622,
      "tokens_seen": 3090612224
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.466666666666666e-05,
      "loss": 1.4107,
      "theoretical_loss": 3.3131924360706986,
      "tokens_seen": 3090677760
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.464646464646465e-05,
      "loss": 1.3346,
      "theoretical_loss": 3.3131870671374983,
      "tokens_seen": 3090743296
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.462626262626263e-05,
      "loss": 1.4176,
      "theoretical_loss": 3.3131816983500153,
      "tokens_seen": 3090808832
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.460606060606061e-05,
      "loss": 1.3185,
      "theoretical_loss": 3.3131763297082415,
      "tokens_seen": 3090874368
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.458585858585858e-05,
      "loss": 1.4078,
      "theoretical_loss": 3.3131709612121703,
      "tokens_seen": 3090939904
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.456565656565657e-05,
      "loss": 1.138,
      "theoretical_loss": 3.3131655928617945,
      "tokens_seen": 3091005440
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.454545454545455e-05,
      "loss": 1.3933,
      "theoretical_loss": 3.313160224657107,
      "tokens_seen": 3091070976
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.452525252525252e-05,
      "loss": 1.4597,
      "theoretical_loss": 3.3131548565981013,
      "tokens_seen": 3091136512
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.45050505050505e-05,
      "loss": 1.2942,
      "theoretical_loss": 3.3131494886847697,
      "tokens_seen": 3091202048
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.448484848484848e-05,
      "loss": 1.2366,
      "theoretical_loss": 3.3131441209171055,
      "tokens_seen": 3091267584
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.446464646464646e-05,
      "loss": 1.32,
      "theoretical_loss": 3.3131387532951013,
      "tokens_seen": 3091333120
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.444444444444444e-05,
      "loss": 1.3026,
      "theoretical_loss": 3.313133385818751,
      "tokens_seen": 3091398656
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.442424242424243e-05,
      "loss": 1.1847,
      "theoretical_loss": 3.313128018488046,
      "tokens_seen": 3091464192
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.440404040404041e-05,
      "loss": 1.3937,
      "theoretical_loss": 3.313122651302981,
      "tokens_seen": 3091529728
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.43838383838384e-05,
      "loss": 1.2737,
      "theoretical_loss": 3.3131172842635475,
      "tokens_seen": 3091595264
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.436363636363635e-05,
      "loss": 1.2427,
      "theoretical_loss": 3.3131119173697394,
      "tokens_seen": 3091660800
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.434343434343434e-05,
      "loss": 1.3898,
      "theoretical_loss": 3.313106550621549,
      "tokens_seen": 3091726336
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.432323232323232e-05,
      "loss": 1.3752,
      "theoretical_loss": 3.31310118401897,
      "tokens_seen": 3091791872
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.43030303030303e-05,
      "loss": 1.3283,
      "theoretical_loss": 3.3130958175619947,
      "tokens_seen": 3091857408
    },
    {
      "epoch": 0.87,
      "objective/train/aligned_token_count": 0.48765885829925537,
      "objective/train/avg_token_score": 0.005869174841791391,
      "objective/train/docs_used": 1739219,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5323609113693237,
      "objective/train/original_loss": 4.0369133949279785,
      "objective/train/theoretical_loss": 3.3130904512506163,
      "objective/train/tokens_used": 1450782176,
      "objective/train/unlikelihood_loss": 0.06851207464933395,
      "theoretical_loss": 3.3130904512506163,
      "tokens_seen": 3091922944
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.428282828282829e-05,
      "loss": 1.4105,
      "theoretical_loss": 3.3130904512506163,
      "tokens_seen": 3091922944
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.426262626262626e-05,
      "loss": 1.3387,
      "theoretical_loss": 3.313085085084828,
      "tokens_seen": 3091988480
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.424242424242424e-05,
      "loss": 1.2392,
      "theoretical_loss": 3.3130797190646226,
      "tokens_seen": 3092054016
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.422222222222223e-05,
      "loss": 1.3566,
      "theoretical_loss": 3.3130743531899927,
      "tokens_seen": 3092119552
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.420202020202021e-05,
      "loss": 1.1502,
      "theoretical_loss": 3.313068987460932,
      "tokens_seen": 3092185088
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.418181818181818e-05,
      "loss": 1.2457,
      "theoretical_loss": 3.3130636218774328,
      "tokens_seen": 3092250624
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.416161616161615e-05,
      "loss": 1.6624,
      "theoretical_loss": 3.313058256439488,
      "tokens_seen": 3092316160
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.414141414141414e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.3130528911470916,
      "tokens_seen": 3092381696
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.412121212121212e-05,
      "loss": 1.2795,
      "theoretical_loss": 3.3130475260002354,
      "tokens_seen": 3092447232
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.41010101010101e-05,
      "loss": 1.3019,
      "theoretical_loss": 3.3130421609989127,
      "tokens_seen": 3092512768
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.408080808080809e-05,
      "loss": 1.3868,
      "theoretical_loss": 3.313036796143117,
      "tokens_seen": 3092578304
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.406060606060606e-05,
      "loss": 1.3622,
      "theoretical_loss": 3.3130314314328406,
      "tokens_seen": 3092643840
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.404040404040404e-05,
      "loss": 1.4785,
      "theoretical_loss": 3.3130260668680767,
      "tokens_seen": 3092709376
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.402020202020203e-05,
      "loss": 1.2719,
      "theoretical_loss": 3.313020702448818,
      "tokens_seen": 3092774912
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.4e-05,
      "loss": 1.2824,
      "theoretical_loss": 3.313015338175058,
      "tokens_seen": 3092840448
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.397979797979798e-05,
      "loss": 1.2195,
      "theoretical_loss": 3.31300997404679,
      "tokens_seen": 3092905984
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.395959595959597e-05,
      "loss": 1.4483,
      "theoretical_loss": 3.313004610064006,
      "tokens_seen": 3092971520
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.393939393939394e-05,
      "loss": 1.3174,
      "theoretical_loss": 3.312999246226699,
      "tokens_seen": 3093037056
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.391919191919192e-05,
      "loss": 1.5936,
      "theoretical_loss": 3.3129938825348626,
      "tokens_seen": 3093102592
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.38989898989899e-05,
      "loss": 1.4124,
      "theoretical_loss": 3.31298851898849,
      "tokens_seen": 3093168128
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.387878787878789e-05,
      "loss": 1.183,
      "theoretical_loss": 3.312983155587573,
      "tokens_seen": 3093233664
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.385858585858587e-05,
      "loss": 1.3305,
      "theoretical_loss": 3.3129777923321058,
      "tokens_seen": 3093299200
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.383838383838383e-05,
      "loss": 1.3607,
      "theoretical_loss": 3.3129724292220804,
      "tokens_seen": 3093364736
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.381818181818181e-05,
      "loss": 1.4171,
      "theoretical_loss": 3.3129670662574906,
      "tokens_seen": 3093430272
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.37979797979798e-05,
      "loss": 1.4099,
      "theoretical_loss": 3.3129617034383285,
      "tokens_seen": 3093495808
    },
    {
      "epoch": 0.87,
      "objective/train/aligned_token_count": 0.5401393175125122,
      "objective/train/avg_token_score": 0.02374294400215149,
      "objective/train/docs_used": 1739859,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6016216278076172,
      "objective/train/original_loss": 4.0241379737854,
      "objective/train/theoretical_loss": 3.312956340764588,
      "objective/train/tokens_used": 1452420576,
      "objective/train/unlikelihood_loss": 0.07347038388252258,
      "theoretical_loss": 3.312956340764588,
      "tokens_seen": 3093561344
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.377777777777778e-05,
      "loss": 1.3202,
      "theoretical_loss": 3.312956340764588,
      "tokens_seen": 3093561344
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.375757575757576e-05,
      "loss": 1.5176,
      "theoretical_loss": 3.3129509782362616,
      "tokens_seen": 3093626880
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.373737373737373e-05,
      "loss": 1.3551,
      "theoretical_loss": 3.3129456158533426,
      "tokens_seen": 3093692416
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.371717171717172e-05,
      "loss": 1.3003,
      "theoretical_loss": 3.3129402536158232,
      "tokens_seen": 3093757952
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.36969696969697e-05,
      "loss": 1.5038,
      "theoretical_loss": 3.312934891523697,
      "tokens_seen": 3093823488
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.367676767676769e-05,
      "loss": 1.4098,
      "theoretical_loss": 3.312929529576957,
      "tokens_seen": 3093889024
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.365656565656566e-05,
      "loss": 1.4045,
      "theoretical_loss": 3.312924167775596,
      "tokens_seen": 3093954560
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.363636363636363e-05,
      "loss": 1.2815,
      "theoretical_loss": 3.3129188061196064,
      "tokens_seen": 3094020096
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.361616161616161e-05,
      "loss": 1.349,
      "theoretical_loss": 3.3129134446089825,
      "tokens_seen": 3094085632
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.35959595959596e-05,
      "loss": 1.3258,
      "theoretical_loss": 3.3129080832437166,
      "tokens_seen": 3094151168
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.357575757575758e-05,
      "loss": 1.5199,
      "theoretical_loss": 3.3129027220238014,
      "tokens_seen": 3094216704
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.355555555555556e-05,
      "loss": 1.4055,
      "theoretical_loss": 3.31289736094923,
      "tokens_seen": 3094282240
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.353535353535355e-05,
      "loss": 1.2162,
      "theoretical_loss": 3.312892000019996,
      "tokens_seen": 3094347776
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.351515151515152e-05,
      "loss": 1.2784,
      "theoretical_loss": 3.3128866392360914,
      "tokens_seen": 3094413312
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.349494949494949e-05,
      "loss": 1.2339,
      "theoretical_loss": 3.31288127859751,
      "tokens_seen": 3094478848
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.347474747474747e-05,
      "loss": 1.538,
      "theoretical_loss": 3.312875918104244,
      "tokens_seen": 3094544384
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.345454545454546e-05,
      "loss": 1.3007,
      "theoretical_loss": 3.3128705577562876,
      "tokens_seen": 3094609920
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.343434343434344e-05,
      "loss": 1.3201,
      "theoretical_loss": 3.3128651975536325,
      "tokens_seen": 3094675456
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.341414141414141e-05,
      "loss": 1.4166,
      "theoretical_loss": 3.3128598374962723,
      "tokens_seen": 3094740992
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.33939393939394e-05,
      "loss": 1.2934,
      "theoretical_loss": 3.3128544775842,
      "tokens_seen": 3094806528
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.337373737373738e-05,
      "loss": 1.4108,
      "theoretical_loss": 3.312849117817408,
      "tokens_seen": 3094872064
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.335353535353536e-05,
      "loss": 1.4265,
      "theoretical_loss": 3.3128437581958905,
      "tokens_seen": 3094937600
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.333333333333335e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.3128383987196397,
      "tokens_seen": 3095003136
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.33131313131313e-05,
      "loss": 1.317,
      "theoretical_loss": 3.3128330393886483,
      "tokens_seen": 3095068672
    },
    {
      "epoch": 0.87,
      "learning_rate": 6.329292929292929e-05,
      "loss": 1.2531,
      "theoretical_loss": 3.3128276802029095,
      "tokens_seen": 3095134208
    },
    {
      "epoch": 0.87,
      "objective/train/aligned_token_count": 0.28164711594581604,
      "objective/train/avg_token_score": 0.017628099769353867,
      "objective/train/docs_used": 1741115,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8385324478149414,
      "objective/train/original_loss": 5.506045818328857,
      "objective/train/theoretical_loss": 3.312822321162417,
      "objective/train/tokens_used": 1454058976,
      "objective/train/unlikelihood_loss": 0.03306974843144417,
      "theoretical_loss": 3.312822321162417,
      "tokens_seen": 3095199744
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.327272727272727e-05,
      "loss": 1.1216,
      "theoretical_loss": 3.312822321162417,
      "tokens_seen": 3095199744
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.325252525252525e-05,
      "loss": 1.3939,
      "theoretical_loss": 3.3128169622671626,
      "tokens_seen": 3095265280
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.323232323232324e-05,
      "loss": 1.3857,
      "theoretical_loss": 3.31281160351714,
      "tokens_seen": 3095330816
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.321212121212121e-05,
      "loss": 1.3275,
      "theoretical_loss": 3.312806244912342,
      "tokens_seen": 3095396352
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.319191919191919e-05,
      "loss": 1.3468,
      "theoretical_loss": 3.312800886452762,
      "tokens_seen": 3095461888
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.317171717171718e-05,
      "loss": 1.3306,
      "theoretical_loss": 3.312795528138393,
      "tokens_seen": 3095527424
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.315151515151516e-05,
      "loss": 1.5015,
      "theoretical_loss": 3.312790169969227,
      "tokens_seen": 3095592960
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.313131313131313e-05,
      "loss": 1.2854,
      "theoretical_loss": 3.312784811945258,
      "tokens_seen": 3095658496
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.311111111111112e-05,
      "loss": 1.3773,
      "theoretical_loss": 3.3127794540664786,
      "tokens_seen": 3095724032
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.309090909090909e-05,
      "loss": 1.335,
      "theoretical_loss": 3.3127740963328813,
      "tokens_seen": 3095789568
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.307070707070707e-05,
      "loss": 1.3948,
      "theoretical_loss": 3.31276873874446,
      "tokens_seen": 3095855104
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.305050505050505e-05,
      "loss": 1.2892,
      "theoretical_loss": 3.3127633813012074,
      "tokens_seen": 3095920640
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.303030303030304e-05,
      "loss": 1.3646,
      "theoretical_loss": 3.3127580240031165,
      "tokens_seen": 3095986176
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.301010101010102e-05,
      "loss": 1.3697,
      "theoretical_loss": 3.31275266685018,
      "tokens_seen": 3096051712
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.298989898989899e-05,
      "loss": 1.2079,
      "theoretical_loss": 3.312747309842391,
      "tokens_seen": 3096117248
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.296969696969696e-05,
      "loss": 1.2,
      "theoretical_loss": 3.312741952979743,
      "tokens_seen": 3096182784
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.294949494949495e-05,
      "loss": 1.3862,
      "theoretical_loss": 3.312736596262228,
      "tokens_seen": 3096248320
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.292929292929293e-05,
      "loss": 1.2966,
      "theoretical_loss": 3.3127312396898403,
      "tokens_seen": 3096313856
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.290909090909091e-05,
      "loss": 1.3254,
      "theoretical_loss": 3.312725883262572,
      "tokens_seen": 3096379392
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.288888888888888e-05,
      "loss": 1.3332,
      "theoretical_loss": 3.312720526980416,
      "tokens_seen": 3096444928
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.286868686868687e-05,
      "loss": 1.1528,
      "theoretical_loss": 3.3127151708433655,
      "tokens_seen": 3096510464
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.284848484848485e-05,
      "loss": 1.6089,
      "theoretical_loss": 3.312709814851414,
      "tokens_seen": 3096576000
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.282828282828284e-05,
      "loss": 1.3146,
      "theoretical_loss": 3.3127044590045536,
      "tokens_seen": 3096641536
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.280808080808082e-05,
      "loss": 1.1954,
      "theoretical_loss": 3.312699103302778,
      "tokens_seen": 3096707072
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.278787878787878e-05,
      "loss": 1.307,
      "theoretical_loss": 3.31269374774608,
      "tokens_seen": 3096772608
    },
    {
      "epoch": 0.88,
      "objective/train/aligned_token_count": 0.42179861664772034,
      "objective/train/avg_token_score": 0.0100519098341465,
      "objective/train/docs_used": 1741744,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2280328273773193,
      "objective/train/original_loss": 4.144233226776123,
      "objective/train/theoretical_loss": 3.3126883923344526,
      "objective/train/tokens_used": 1455697376,
      "objective/train/unlikelihood_loss": 0.05300737917423248,
      "theoretical_loss": 3.3126883923344526,
      "tokens_seen": 3096838144
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.276767676767676e-05,
      "loss": 1.2182,
      "theoretical_loss": 3.3126883923344526,
      "tokens_seen": 3096838144
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.274747474747475e-05,
      "loss": 1.2815,
      "theoretical_loss": 3.312683037067889,
      "tokens_seen": 3096903680
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.272727272727273e-05,
      "loss": 1.5321,
      "theoretical_loss": 3.3126776819463815,
      "tokens_seen": 3096969216
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.270707070707071e-05,
      "loss": 1.5086,
      "theoretical_loss": 3.312672326969924,
      "tokens_seen": 3097034752
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.26868686868687e-05,
      "loss": 1.179,
      "theoretical_loss": 3.3126669721385085,
      "tokens_seen": 3097100288
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.266666666666667e-05,
      "loss": 1.4435,
      "theoretical_loss": 3.312661617452129,
      "tokens_seen": 3097165824
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.264646464646465e-05,
      "loss": 1.3746,
      "theoretical_loss": 3.312656262910778,
      "tokens_seen": 3097231360
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.262626262626262e-05,
      "loss": 1.4814,
      "theoretical_loss": 3.3126509085144487,
      "tokens_seen": 3097296896
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.26060606060606e-05,
      "loss": 1.1903,
      "theoretical_loss": 3.312645554263134,
      "tokens_seen": 3097362432
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.258585858585859e-05,
      "loss": 1.4002,
      "theoretical_loss": 3.3126402001568267,
      "tokens_seen": 3097427968
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.256565656565656e-05,
      "loss": 1.2242,
      "theoretical_loss": 3.31263484619552,
      "tokens_seen": 3097493504
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.254545454545454e-05,
      "loss": 1.5902,
      "theoretical_loss": 3.312629492379207,
      "tokens_seen": 3097559040
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.252525252525253e-05,
      "loss": 1.2162,
      "theoretical_loss": 3.3126241387078803,
      "tokens_seen": 3097624576
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.250505050505051e-05,
      "loss": 1.4788,
      "theoretical_loss": 3.3126187851815336,
      "tokens_seen": 3097690112
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.248484848484848e-05,
      "loss": 1.375,
      "theoretical_loss": 3.3126134318001594,
      "tokens_seen": 3097755648
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.246464646464647e-05,
      "loss": 1.3671,
      "theoretical_loss": 3.312608078563751,
      "tokens_seen": 3097821184
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.244444444444444e-05,
      "loss": 1.4424,
      "theoretical_loss": 3.312602725472301,
      "tokens_seen": 3097886720
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.242424242424242e-05,
      "loss": 1.4053,
      "theoretical_loss": 3.312597372525803,
      "tokens_seen": 3097952256
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.24040404040404e-05,
      "loss": 1.2669,
      "theoretical_loss": 3.3125920197242493,
      "tokens_seen": 3098017792
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.238383838383839e-05,
      "loss": 1.2876,
      "theoretical_loss": 3.3125866670676336,
      "tokens_seen": 3098083328
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.236363636363636e-05,
      "loss": 1.2891,
      "theoretical_loss": 3.312581314555948,
      "tokens_seen": 3098148864
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.234343434343434e-05,
      "loss": 1.36,
      "theoretical_loss": 3.3125759621891864,
      "tokens_seen": 3098214400
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.232323232323233e-05,
      "loss": 1.5504,
      "theoretical_loss": 3.3125706099673415,
      "tokens_seen": 3098279936
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.230303030303031e-05,
      "loss": 1.3752,
      "theoretical_loss": 3.3125652578904066,
      "tokens_seen": 3098345472
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.228282828282828e-05,
      "loss": 1.4896,
      "theoretical_loss": 3.312559905958374,
      "tokens_seen": 3098411008
    },
    {
      "epoch": 0.88,
      "objective/train/aligned_token_count": 0.5229105353355408,
      "objective/train/avg_token_score": 0.0064886873587965965,
      "objective/train/docs_used": 1742793,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6618590354919434,
      "objective/train/original_loss": 4.132230758666992,
      "objective/train/theoretical_loss": 3.3125545541712373,
      "objective/train/tokens_used": 1457335776,
      "objective/train/unlikelihood_loss": 0.06788662075996399,
      "theoretical_loss": 3.3125545541712373,
      "tokens_seen": 3098476544
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.226262626262627e-05,
      "loss": 1.4474,
      "theoretical_loss": 3.3125545541712373,
      "tokens_seen": 3098476544
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.224242424242425e-05,
      "loss": 1.2525,
      "theoretical_loss": 3.3125492025289893,
      "tokens_seen": 3098542080
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.222222222222222e-05,
      "loss": 1.341,
      "theoretical_loss": 3.3125438510316227,
      "tokens_seen": 3098607616
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.22020202020202e-05,
      "loss": 1.3133,
      "theoretical_loss": 3.3125384996791314,
      "tokens_seen": 3098673152
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.218181818181817e-05,
      "loss": 1.4901,
      "theoretical_loss": 3.3125331484715077,
      "tokens_seen": 3098738688
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.216161616161616e-05,
      "loss": 1.3892,
      "theoretical_loss": 3.3125277974087446,
      "tokens_seen": 3098804224
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.214141414141414e-05,
      "loss": 1.2266,
      "theoretical_loss": 3.312522446490836,
      "tokens_seen": 3098869760
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.212121212121213e-05,
      "loss": 1.4082,
      "theoretical_loss": 3.3125170957177734,
      "tokens_seen": 3098935296
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.21010101010101e-05,
      "loss": 1.3431,
      "theoretical_loss": 3.312511745089551,
      "tokens_seen": 3099000832
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.208080808080808e-05,
      "loss": 1.4298,
      "theoretical_loss": 3.3125063946061615,
      "tokens_seen": 3099066368
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.206060606060606e-05,
      "loss": 1.4251,
      "theoretical_loss": 3.3125010442675977,
      "tokens_seen": 3099131904
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.204040404040405e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.3124956940738532,
      "tokens_seen": 3099197440
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.202020202020202e-05,
      "loss": 1.5877,
      "theoretical_loss": 3.31249034402492,
      "tokens_seen": 3099262976
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.2e-05,
      "loss": 1.3166,
      "theoretical_loss": 3.3124849941207923,
      "tokens_seen": 3099328512
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.197979797979799e-05,
      "loss": 1.4087,
      "theoretical_loss": 3.312479644361462,
      "tokens_seen": 3099394048
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.195959595959596e-05,
      "loss": 1.3429,
      "theoretical_loss": 3.3124742947469232,
      "tokens_seen": 3099459584
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.193939393939394e-05,
      "loss": 1.6494,
      "theoretical_loss": 3.312468945277168,
      "tokens_seen": 3099525120
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.191919191919192e-05,
      "loss": 1.5285,
      "theoretical_loss": 3.31246359595219,
      "tokens_seen": 3099590656
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.18989898989899e-05,
      "loss": 1.4326,
      "theoretical_loss": 3.312458246771982,
      "tokens_seen": 3099656192
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.187878787878788e-05,
      "loss": 1.4332,
      "theoretical_loss": 3.312452897736537,
      "tokens_seen": 3099721728
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.185858585858586e-05,
      "loss": 1.286,
      "theoretical_loss": 3.3124475488458485,
      "tokens_seen": 3099787264
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.183838383838383e-05,
      "loss": 1.1417,
      "theoretical_loss": 3.3124422000999085,
      "tokens_seen": 3099852800
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.181818181818182e-05,
      "loss": 1.3933,
      "theoretical_loss": 3.3124368514987106,
      "tokens_seen": 3099918336
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.17979797979798e-05,
      "loss": 1.3648,
      "theoretical_loss": 3.3124315030422484,
      "tokens_seen": 3099983872
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.177777777777779e-05,
      "loss": 1.5313,
      "theoretical_loss": 3.312426154730514,
      "tokens_seen": 3100049408
    },
    {
      "epoch": 0.88,
      "objective/train/aligned_token_count": 0.507331371307373,
      "objective/train/avg_token_score": 0.01199525035917759,
      "objective/train/docs_used": 1743460,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5499135255813599,
      "objective/train/original_loss": 3.8499817848205566,
      "objective/train/theoretical_loss": 3.312420806563501,
      "objective/train/tokens_used": 1458974176,
      "objective/train/unlikelihood_loss": 0.0794258788228035,
      "theoretical_loss": 3.312420806563501,
      "tokens_seen": 3100114944
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.175757575757576e-05,
      "loss": 1.3756,
      "theoretical_loss": 3.312420806563501,
      "tokens_seen": 3100114944
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.173737373737374e-05,
      "loss": 1.3944,
      "theoretical_loss": 3.312415458541202,
      "tokens_seen": 3100180480
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.171717171717172e-05,
      "loss": 1.1893,
      "theoretical_loss": 3.31241011066361,
      "tokens_seen": 3100246016
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.16969696969697e-05,
      "loss": 1.2832,
      "theoretical_loss": 3.3124047629307185,
      "tokens_seen": 3100311552
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.167676767676768e-05,
      "loss": 1.2052,
      "theoretical_loss": 3.3123994153425205,
      "tokens_seen": 3100377088
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.165656565656566e-05,
      "loss": 1.5098,
      "theoretical_loss": 3.3123940678990085,
      "tokens_seen": 3100442624
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.163636363636363e-05,
      "loss": 1.5591,
      "theoretical_loss": 3.3123887206001763,
      "tokens_seen": 3100508160
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.161616161616162e-05,
      "loss": 1.1675,
      "theoretical_loss": 3.312383373446016,
      "tokens_seen": 3100573696
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.15959595959596e-05,
      "loss": 1.3316,
      "theoretical_loss": 3.3123780264365212,
      "tokens_seen": 3100639232
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.157575757575757e-05,
      "loss": 1.5088,
      "theoretical_loss": 3.312372679571685,
      "tokens_seen": 3100704768
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.155555555555555e-05,
      "loss": 1.4007,
      "theoretical_loss": 3.3123673328515,
      "tokens_seen": 3100770304
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.153535353535354e-05,
      "loss": 1.1634,
      "theoretical_loss": 3.3123619862759597,
      "tokens_seen": 3100835840
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.151515151515152e-05,
      "loss": 1.4042,
      "theoretical_loss": 3.3123566398450572,
      "tokens_seen": 3100901376
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.14949494949495e-05,
      "loss": 1.2623,
      "theoretical_loss": 3.312351293558785,
      "tokens_seen": 3100966912
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.147474747474748e-05,
      "loss": 1.3935,
      "theoretical_loss": 3.312345947417136,
      "tokens_seen": 3101032448
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.145454545454546e-05,
      "loss": 1.4805,
      "theoretical_loss": 3.312340601420104,
      "tokens_seen": 3101097984
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.143434343434343e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.3123352555676813,
      "tokens_seen": 3101163520
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.141414141414142e-05,
      "loss": 1.5287,
      "theoretical_loss": 3.3123299098598618,
      "tokens_seen": 3101229056
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.13939393939394e-05,
      "loss": 1.243,
      "theoretical_loss": 3.3123245642966377,
      "tokens_seen": 3101294592
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.137373737373737e-05,
      "loss": 1.5686,
      "theoretical_loss": 3.3123192188780024,
      "tokens_seen": 3101360128
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.135353535353535e-05,
      "loss": 1.6458,
      "theoretical_loss": 3.312313873603949,
      "tokens_seen": 3101425664
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.133333333333334e-05,
      "loss": 1.4375,
      "theoretical_loss": 3.3123085284744698,
      "tokens_seen": 3101491200
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.131313131313131e-05,
      "loss": 1.5726,
      "theoretical_loss": 3.312303183489559,
      "tokens_seen": 3101556736
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.129292929292929e-05,
      "loss": 1.4672,
      "theoretical_loss": 3.312297838649209,
      "tokens_seen": 3101622272
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.127272727272728e-05,
      "loss": 1.3391,
      "theoretical_loss": 3.3122924939534126,
      "tokens_seen": 3101687808
    },
    {
      "epoch": 0.88,
      "objective/train/aligned_token_count": 0.5336021780967712,
      "objective/train/avg_token_score": 0.007012638729065657,
      "objective/train/docs_used": 1744846,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.724253535270691,
      "objective/train/original_loss": 4.19031286239624,
      "objective/train/theoretical_loss": 3.3122871494021635,
      "objective/train/tokens_used": 1460612576,
      "objective/train/unlikelihood_loss": 0.04923609271645546,
      "theoretical_loss": 3.3122871494021635,
      "tokens_seen": 3101753344
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.125252525252526e-05,
      "loss": 1.5019,
      "theoretical_loss": 3.3122871494021635,
      "tokens_seen": 3101753344
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.123232323232323e-05,
      "loss": 1.3672,
      "theoretical_loss": 3.3122818049954543,
      "tokens_seen": 3101818880
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.121212121212121e-05,
      "loss": 1.3501,
      "theoretical_loss": 3.3122764607332784,
      "tokens_seen": 3101884416
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.11919191919192e-05,
      "loss": 1.4388,
      "theoretical_loss": 3.3122711166156282,
      "tokens_seen": 3101949952
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.117171717171717e-05,
      "loss": 1.4039,
      "theoretical_loss": 3.3122657726424976,
      "tokens_seen": 3102015488
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.115151515151515e-05,
      "loss": 1.4096,
      "theoretical_loss": 3.3122604288138784,
      "tokens_seen": 3102081024
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.113131313131314e-05,
      "loss": 1.2096,
      "theoretical_loss": 3.312255085129765,
      "tokens_seen": 3102146560
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.111111111111111e-05,
      "loss": 1.2736,
      "theoretical_loss": 3.3122497415901493,
      "tokens_seen": 3102212096
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.109090909090909e-05,
      "loss": 1.5768,
      "theoretical_loss": 3.312244398195025,
      "tokens_seen": 3102277632
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.107070707070708e-05,
      "loss": 1.1651,
      "theoretical_loss": 3.312239054944385,
      "tokens_seen": 3102343168
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.105050505050505e-05,
      "loss": 1.5251,
      "theoretical_loss": 3.3122337118382226,
      "tokens_seen": 3102408704
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.1030303030303036e-05,
      "loss": 1.3816,
      "theoretical_loss": 3.3122283688765304,
      "tokens_seen": 3102474240
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.101010101010101e-05,
      "loss": 1.4238,
      "theoretical_loss": 3.3122230260593017,
      "tokens_seen": 3102539776
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.098989898989899e-05,
      "loss": 1.3879,
      "theoretical_loss": 3.3122176833865296,
      "tokens_seen": 3102605312
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0969696969696975e-05,
      "loss": 1.4427,
      "theoretical_loss": 3.3122123408582067,
      "tokens_seen": 3102670848
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.094949494949495e-05,
      "loss": 1.351,
      "theoretical_loss": 3.3122069984743265,
      "tokens_seen": 3102736384
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.092929292929293e-05,
      "loss": 1.4589,
      "theoretical_loss": 3.312201656234882,
      "tokens_seen": 3102801920
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0909090909090906e-05,
      "loss": 1.3452,
      "theoretical_loss": 3.312196314139866,
      "tokens_seen": 3102867456
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.088888888888889e-05,
      "loss": 1.6494,
      "theoretical_loss": 3.3121909721892715,
      "tokens_seen": 3102932992
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0868686868686874e-05,
      "loss": 1.5178,
      "theoretical_loss": 3.312185630383092,
      "tokens_seen": 3102998528
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0848484848484845e-05,
      "loss": 1.2764,
      "theoretical_loss": 3.3121802887213203,
      "tokens_seen": 3103064064
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.082828282828283e-05,
      "loss": 1.4451,
      "theoretical_loss": 3.3121749472039492,
      "tokens_seen": 3103129600
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.080808080808081e-05,
      "loss": 1.2213,
      "theoretical_loss": 3.312169605830972,
      "tokens_seen": 3103195136
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.078787878787879e-05,
      "loss": 1.5284,
      "theoretical_loss": 3.312164264602382,
      "tokens_seen": 3103260672
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0767676767676774e-05,
      "loss": 1.3733,
      "theoretical_loss": 3.312158923518172,
      "tokens_seen": 3103326208
    },
    {
      "epoch": 0.88,
      "objective/train/aligned_token_count": 0.32374143600463867,
      "objective/train/avg_token_score": 0.016244947910308838,
      "objective/train/docs_used": 1745390,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.958683431148529,
      "objective/train/original_loss": 4.1633501052856445,
      "objective/train/theoretical_loss": 3.312153582578335,
      "objective/train/tokens_used": 1462250976,
      "objective/train/unlikelihood_loss": 0.07154559344053268,
      "theoretical_loss": 3.312153582578335,
      "tokens_seen": 3103391744
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0747474747474744e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.312153582578335,
      "tokens_seen": 3103391744
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.072727272727273e-05,
      "loss": 1.2052,
      "theoretical_loss": 3.3121482417828636,
      "tokens_seen": 3103457280
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.070707070707071e-05,
      "loss": 1.3873,
      "theoretical_loss": 3.3121429011317516,
      "tokens_seen": 3103522816
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.068686868686869e-05,
      "loss": 1.4631,
      "theoretical_loss": 3.3121375606249917,
      "tokens_seen": 3103588352
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0666666666666666e-05,
      "loss": 1.3053,
      "theoretical_loss": 3.312132220262577,
      "tokens_seen": 3103653888
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.064646464646465e-05,
      "loss": 1.1693,
      "theoretical_loss": 3.3121268800445005,
      "tokens_seen": 3103719424
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.062626262626263e-05,
      "loss": 1.2568,
      "theoretical_loss": 3.3121215399707555,
      "tokens_seen": 3103784960
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.060606060606061e-05,
      "loss": 1.5248,
      "theoretical_loss": 3.312116200041335,
      "tokens_seen": 3103850496
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.058585858585858e-05,
      "loss": 1.5025,
      "theoretical_loss": 3.3121108602562317,
      "tokens_seen": 3103916032
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0565656565656566e-05,
      "loss": 1.366,
      "theoretical_loss": 3.312105520615439,
      "tokens_seen": 3103981568
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.054545454545455e-05,
      "loss": 1.4963,
      "theoretical_loss": 3.3121001811189497,
      "tokens_seen": 3104047104
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.052525252525253e-05,
      "loss": 1.4923,
      "theoretical_loss": 3.312094841766757,
      "tokens_seen": 3104112640
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0505050505050504e-05,
      "loss": 1.306,
      "theoretical_loss": 3.312089502558854,
      "tokens_seen": 3104178176
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.048484848484848e-05,
      "loss": 1.4349,
      "theoretical_loss": 3.312084163495234,
      "tokens_seen": 3104243712
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0464646464646465e-05,
      "loss": 1.3279,
      "theoretical_loss": 3.312078824575889,
      "tokens_seen": 3104309248
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.044444444444445e-05,
      "loss": 1.4835,
      "theoretical_loss": 3.312073485800813,
      "tokens_seen": 3104374784
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0424242424242426e-05,
      "loss": 1.2368,
      "theoretical_loss": 3.3120681471699993,
      "tokens_seen": 3104440320
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0404040404040404e-05,
      "loss": 1.4605,
      "theoretical_loss": 3.3120628086834403,
      "tokens_seen": 3104505856
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.038383838383839e-05,
      "loss": 1.314,
      "theoretical_loss": 3.312057470341129,
      "tokens_seen": 3104571392
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0363636363636365e-05,
      "loss": 1.3692,
      "theoretical_loss": 3.312052132143059,
      "tokens_seen": 3104636928
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.034343434343435e-05,
      "loss": 1.3906,
      "theoretical_loss": 3.312046794089223,
      "tokens_seen": 3104702464
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.032323232323232e-05,
      "loss": 1.2749,
      "theoretical_loss": 3.3120414561796143,
      "tokens_seen": 3104768000
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.03030303030303e-05,
      "loss": 1.2587,
      "theoretical_loss": 3.3120361184142255,
      "tokens_seen": 3104833536
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.028282828282829e-05,
      "loss": 1.187,
      "theoretical_loss": 3.3120307807930502,
      "tokens_seen": 3104899072
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0262626262626264e-05,
      "loss": 1.2278,
      "theoretical_loss": 3.312025443316081,
      "tokens_seen": 3104964608
    },
    {
      "epoch": 0.88,
      "objective/train/aligned_token_count": 0.4050586521625519,
      "objective/train/avg_token_score": 0.01767726242542267,
      "objective/train/docs_used": 1746455,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3521039485931396,
      "objective/train/original_loss": 4.111509323120117,
      "objective/train/theoretical_loss": 3.3120201059833114,
      "objective/train/tokens_used": 1463889376,
      "objective/train/unlikelihood_loss": 0.07980647683143616,
      "theoretical_loss": 3.3120201059833114,
      "tokens_seen": 3105030144
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.024242424242424e-05,
      "loss": 1.3103,
      "theoretical_loss": 3.3120201059833114,
      "tokens_seen": 3105030144
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0222222222222225e-05,
      "loss": 1.467,
      "theoretical_loss": 3.312014768794734,
      "tokens_seen": 3105095680
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.02020202020202e-05,
      "loss": 1.3687,
      "theoretical_loss": 3.3120094317503423,
      "tokens_seen": 3105161216
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0181818181818187e-05,
      "loss": 1.1842,
      "theoretical_loss": 3.3120040948501295,
      "tokens_seen": 3105226752
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.016161616161616e-05,
      "loss": 1.2106,
      "theoretical_loss": 3.311998758094088,
      "tokens_seen": 3105292288
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.014141414141414e-05,
      "loss": 1.4751,
      "theoretical_loss": 3.311993421482211,
      "tokens_seen": 3105357824
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0121212121212125e-05,
      "loss": 1.4881,
      "theoretical_loss": 3.311988085014492,
      "tokens_seen": 3105423360
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.01010101010101e-05,
      "loss": 1.4214,
      "theoretical_loss": 3.3119827486909235,
      "tokens_seen": 3105488896
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0080808080808086e-05,
      "loss": 1.5286,
      "theoretical_loss": 3.3119774125114994,
      "tokens_seen": 3105554432
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0060606060606056e-05,
      "loss": 1.1518,
      "theoretical_loss": 3.3119720764762115,
      "tokens_seen": 3105619968
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.004040404040404e-05,
      "loss": 1.4517,
      "theoretical_loss": 3.311966740585054,
      "tokens_seen": 3105685504
    },
    {
      "epoch": 0.88,
      "learning_rate": 6.0020202020202024e-05,
      "loss": 1.5644,
      "theoretical_loss": 3.3119614048380197,
      "tokens_seen": 3105751040
    },
    {
      "epoch": 0.88,
      "learning_rate": 6e-05,
      "loss": 1.1861,
      "theoretical_loss": 3.3119560692351016,
      "tokens_seen": 3105816576
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.997979797979798e-05,
      "loss": 1.5046,
      "theoretical_loss": 3.3119507337762926,
      "tokens_seen": 3105882112
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.995959595959596e-05,
      "loss": 1.2961,
      "theoretical_loss": 3.3119453984615856,
      "tokens_seen": 3105947648
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.993939393939394e-05,
      "loss": 1.3282,
      "theoretical_loss": 3.3119400632909746,
      "tokens_seen": 3106013184
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9919191919191924e-05,
      "loss": 1.4722,
      "theoretical_loss": 3.3119347282644513,
      "tokens_seen": 3106078720
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9898989898989894e-05,
      "loss": 1.6443,
      "theoretical_loss": 3.3119293933820098,
      "tokens_seen": 3106144256
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.987878787878788e-05,
      "loss": 1.4145,
      "theoretical_loss": 3.311924058643643,
      "tokens_seen": 3106209792
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.985858585858586e-05,
      "loss": 1.1745,
      "theoretical_loss": 3.3119187240493435,
      "tokens_seen": 3106275328
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.983838383838384e-05,
      "loss": 1.0134,
      "theoretical_loss": 3.311913389599105,
      "tokens_seen": 3106340864
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.981818181818182e-05,
      "loss": 1.4093,
      "theoretical_loss": 3.31190805529292,
      "tokens_seen": 3106406400
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.97979797979798e-05,
      "loss": 1.513,
      "theoretical_loss": 3.311902721130782,
      "tokens_seen": 3106471936
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.977777777777778e-05,
      "loss": 1.3232,
      "theoretical_loss": 3.3118973871126838,
      "tokens_seen": 3106537472
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.975757575757576e-05,
      "loss": 1.2395,
      "theoretical_loss": 3.3118920532386187,
      "tokens_seen": 3106603008
    },
    {
      "epoch": 0.88,
      "objective/train/aligned_token_count": 0.42320382595062256,
      "objective/train/avg_token_score": 0.01227737870067358,
      "objective/train/docs_used": 1746871,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3407150506973267,
      "objective/train/original_loss": 4.456091403961182,
      "objective/train/theoretical_loss": 3.3118867195085793,
      "objective/train/tokens_used": 1465527776,
      "objective/train/unlikelihood_loss": 0.0496998131275177,
      "theoretical_loss": 3.3118867195085793,
      "tokens_seen": 3106668544
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.973737373737374e-05,
      "loss": 1.411,
      "theoretical_loss": 3.3118867195085793,
      "tokens_seen": 3106668544
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9717171717171716e-05,
      "loss": 1.1645,
      "theoretical_loss": 3.3118813859225593,
      "tokens_seen": 3106734080
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.96969696969697e-05,
      "loss": 1.4431,
      "theoretical_loss": 3.3118760524805517,
      "tokens_seen": 3106799616
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.967676767676768e-05,
      "loss": 1.2984,
      "theoretical_loss": 3.311870719182549,
      "tokens_seen": 3106865152
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.965656565656566e-05,
      "loss": 1.5106,
      "theoretical_loss": 3.3118653860285447,
      "tokens_seen": 3106930688
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.963636363636363e-05,
      "loss": 1.461,
      "theoretical_loss": 3.311860053018532,
      "tokens_seen": 3106996224
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9616161616161615e-05,
      "loss": 1.2255,
      "theoretical_loss": 3.3118547201525037,
      "tokens_seen": 3107061760
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.95959595959596e-05,
      "loss": 1.3862,
      "theoretical_loss": 3.3118493874304527,
      "tokens_seen": 3107127296
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.957575757575758e-05,
      "loss": 1.2251,
      "theoretical_loss": 3.3118440548523727,
      "tokens_seen": 3107192832
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.955555555555556e-05,
      "loss": 1.5898,
      "theoretical_loss": 3.311838722418256,
      "tokens_seen": 3107258368
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.953535353535354e-05,
      "loss": 1.4388,
      "theoretical_loss": 3.3118333901280965,
      "tokens_seen": 3107323904
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9515151515151515e-05,
      "loss": 1.3734,
      "theoretical_loss": 3.311828057981887,
      "tokens_seen": 3107389440
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.94949494949495e-05,
      "loss": 1.2283,
      "theoretical_loss": 3.3118227259796202,
      "tokens_seen": 3107454976
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9474747474747476e-05,
      "loss": 1.3179,
      "theoretical_loss": 3.3118173941212894,
      "tokens_seen": 3107520512
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.945454545454545e-05,
      "loss": 1.4341,
      "theoretical_loss": 3.3118120624068874,
      "tokens_seen": 3107586048
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.943434343434344e-05,
      "loss": 1.4837,
      "theoretical_loss": 3.311806730836408,
      "tokens_seen": 3107651584
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9414141414141414e-05,
      "loss": 1.5212,
      "theoretical_loss": 3.311801399409844,
      "tokens_seen": 3107717120
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.93939393939394e-05,
      "loss": 1.6531,
      "theoretical_loss": 3.3117960681271885,
      "tokens_seen": 3107782656
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9373737373737376e-05,
      "loss": 1.493,
      "theoretical_loss": 3.311790736988434,
      "tokens_seen": 3107848192
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.935353535353535e-05,
      "loss": 1.1839,
      "theoretical_loss": 3.311785405993574,
      "tokens_seen": 3107913728
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.933333333333334e-05,
      "loss": 1.3508,
      "theoretical_loss": 3.3117800751426016,
      "tokens_seen": 3107979264
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9313131313131314e-05,
      "loss": 1.4573,
      "theoretical_loss": 3.31177474443551,
      "tokens_seen": 3108044800
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.929292929292929e-05,
      "loss": 1.4694,
      "theoretical_loss": 3.311769413872292,
      "tokens_seen": 3108110336
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9272727272727275e-05,
      "loss": 1.4726,
      "theoretical_loss": 3.311764083452941,
      "tokens_seen": 3108175872
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.925252525252525e-05,
      "loss": 1.3428,
      "theoretical_loss": 3.31175875317745,
      "tokens_seen": 3108241408
    },
    {
      "epoch": 0.88,
      "objective/train/aligned_token_count": 0.36418622732162476,
      "objective/train/avg_token_score": 0.012625249102711678,
      "objective/train/docs_used": 1748010,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1213479042053223,
      "objective/train/original_loss": 3.9500253200531006,
      "objective/train/theoretical_loss": 3.3117534230458117,
      "objective/train/tokens_used": 1467166176,
      "objective/train/unlikelihood_loss": 0.07355272769927979,
      "theoretical_loss": 3.3117534230458117,
      "tokens_seen": 3108306944
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9232323232323236e-05,
      "loss": 1.3294,
      "theoretical_loss": 3.3117534230458117,
      "tokens_seen": 3108306944
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.921212121212121e-05,
      "loss": 1.1508,
      "theoretical_loss": 3.3117480930580196,
      "tokens_seen": 3108372480
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.919191919191919e-05,
      "loss": 1.3408,
      "theoretical_loss": 3.311742763214067,
      "tokens_seen": 3108438016
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9171717171717175e-05,
      "loss": 1.2688,
      "theoretical_loss": 3.3117374335139464,
      "tokens_seen": 3108503552
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.915151515151515e-05,
      "loss": 1.3787,
      "theoretical_loss": 3.3117321039576515,
      "tokens_seen": 3108569088
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9131313131313136e-05,
      "loss": 1.2246,
      "theoretical_loss": 3.3117267745451744,
      "tokens_seen": 3108634624
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.911111111111111e-05,
      "loss": 1.2576,
      "theoretical_loss": 3.3117214452765094,
      "tokens_seen": 3108700160
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.909090909090909e-05,
      "loss": 1.3976,
      "theoretical_loss": 3.311716116151649,
      "tokens_seen": 3108765696
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.9070707070707074e-05,
      "loss": 1.5346,
      "theoretical_loss": 3.3117107871705858,
      "tokens_seen": 3108831232
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.905050505050505e-05,
      "loss": 1.4903,
      "theoretical_loss": 3.3117054583333134,
      "tokens_seen": 3108896768
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.903030303030303e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.311700129639825,
      "tokens_seen": 3108962304
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.901010101010101e-05,
      "loss": 1.2526,
      "theoretical_loss": 3.311694801090114,
      "tokens_seen": 3109027840
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.898989898989899e-05,
      "loss": 1.4132,
      "theoretical_loss": 3.311689472684173,
      "tokens_seen": 3109093376
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8969696969696973e-05,
      "loss": 1.3665,
      "theoretical_loss": 3.311684144421995,
      "tokens_seen": 3109158912
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.894949494949495e-05,
      "loss": 1.3429,
      "theoretical_loss": 3.311678816303573,
      "tokens_seen": 3109224448
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.892929292929293e-05,
      "loss": 1.3868,
      "theoretical_loss": 3.3116734883289003,
      "tokens_seen": 3109289984
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.890909090909091e-05,
      "loss": 1.4173,
      "theoretical_loss": 3.3116681604979705,
      "tokens_seen": 3109355520
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.888888888888889e-05,
      "loss": 1.5127,
      "theoretical_loss": 3.311662832810776,
      "tokens_seen": 3109421056
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.886868686868687e-05,
      "loss": 1.3741,
      "theoretical_loss": 3.31165750526731,
      "tokens_seen": 3109486592
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.884848484848485e-05,
      "loss": 1.3569,
      "theoretical_loss": 3.3116521778675656,
      "tokens_seen": 3109552128
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.882828282828283e-05,
      "loss": 1.5854,
      "theoretical_loss": 3.311646850611536,
      "tokens_seen": 3109617664
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.880808080808081e-05,
      "loss": 1.3956,
      "theoretical_loss": 3.3116415234992145,
      "tokens_seen": 3109683200
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.878787878787879e-05,
      "loss": 1.2267,
      "theoretical_loss": 3.311636196530594,
      "tokens_seen": 3109748736
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8767676767676766e-05,
      "loss": 1.3367,
      "theoretical_loss": 3.3116308697056676,
      "tokens_seen": 3109814272
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.874747474747475e-05,
      "loss": 1.3499,
      "theoretical_loss": 3.3116255430244284,
      "tokens_seen": 3109879808
    },
    {
      "epoch": 0.88,
      "objective/train/aligned_token_count": 0.37829911708831787,
      "objective/train/avg_token_score": 0.007665807381272316,
      "objective/train/docs_used": 1748667,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.070912480354309,
      "objective/train/original_loss": 4.159041881561279,
      "objective/train/theoretical_loss": 3.311620216486869,
      "objective/train/tokens_used": 1468804576,
      "objective/train/unlikelihood_loss": 0.07531969249248505,
      "theoretical_loss": 3.311620216486869,
      "tokens_seen": 3109945344
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.872727272727273e-05,
      "loss": 1.4993,
      "theoretical_loss": 3.311620216486869,
      "tokens_seen": 3109945344
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.870707070707071e-05,
      "loss": 1.639,
      "theoretical_loss": 3.3116148900929834,
      "tokens_seen": 3110010880
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.868686868686869e-05,
      "loss": 1.5637,
      "theoretical_loss": 3.3116095638427643,
      "tokens_seen": 3110076416
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8666666666666665e-05,
      "loss": 1.2394,
      "theoretical_loss": 3.3116042377362045,
      "tokens_seen": 3110141952
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.864646464646465e-05,
      "loss": 1.512,
      "theoretical_loss": 3.311598911773298,
      "tokens_seen": 3110207488
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8626262626262626e-05,
      "loss": 1.3925,
      "theoretical_loss": 3.3115935859540366,
      "tokens_seen": 3110273024
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.860606060606061e-05,
      "loss": 1.3675,
      "theoretical_loss": 3.311588260278414,
      "tokens_seen": 3110338560
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.858585858585859e-05,
      "loss": 1.3267,
      "theoretical_loss": 3.311582934746424,
      "tokens_seen": 3110404096
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8565656565656565e-05,
      "loss": 1.4575,
      "theoretical_loss": 3.3115776093580584,
      "tokens_seen": 3110469632
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.854545454545455e-05,
      "loss": 1.5,
      "theoretical_loss": 3.3115722841133115,
      "tokens_seen": 3110535168
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8525252525252526e-05,
      "loss": 1.3791,
      "theoretical_loss": 3.3115669590121755,
      "tokens_seen": 3110600704
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.85050505050505e-05,
      "loss": 1.3977,
      "theoretical_loss": 3.3115616340546437,
      "tokens_seen": 3110666240
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.848484848484849e-05,
      "loss": 1.5925,
      "theoretical_loss": 3.31155630924071,
      "tokens_seen": 3110731776
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8464646464646464e-05,
      "loss": 1.4515,
      "theoretical_loss": 3.311550984570366,
      "tokens_seen": 3110797312
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.844444444444445e-05,
      "loss": 1.3583,
      "theoretical_loss": 3.311545660043606,
      "tokens_seen": 3110862848
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8424242424242425e-05,
      "loss": 1.3675,
      "theoretical_loss": 3.3115403356604234,
      "tokens_seen": 3110928384
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.84040404040404e-05,
      "loss": 1.4879,
      "theoretical_loss": 3.31153501142081,
      "tokens_seen": 3110993920
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8383838383838386e-05,
      "loss": 1.4612,
      "theoretical_loss": 3.31152968732476,
      "tokens_seen": 3111059456
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8363636363636364e-05,
      "loss": 1.5361,
      "theoretical_loss": 3.311524363372266,
      "tokens_seen": 3111124992
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.834343434343435e-05,
      "loss": 1.2305,
      "theoretical_loss": 3.311519039563321,
      "tokens_seen": 3111190528
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8323232323232325e-05,
      "loss": 1.4598,
      "theoretical_loss": 3.3115137158979184,
      "tokens_seen": 3111256064
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.83030303030303e-05,
      "loss": 1.4948,
      "theoretical_loss": 3.311508392376051,
      "tokens_seen": 3111321600
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.8282828282828286e-05,
      "loss": 1.5258,
      "theoretical_loss": 3.3115030689977125,
      "tokens_seen": 3111387136
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.826262626262627e-05,
      "loss": 1.4513,
      "theoretical_loss": 3.311497745762895,
      "tokens_seen": 3111452672
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.824242424242424e-05,
      "loss": 1.6172,
      "theoretical_loss": 3.311492422671593,
      "tokens_seen": 3111518208
    },
    {
      "epoch": 0.88,
      "objective/train/aligned_token_count": 0.47012463212013245,
      "objective/train/avg_token_score": 0.009874004870653152,
      "objective/train/docs_used": 1749977,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2909672260284424,
      "objective/train/original_loss": 4.002129554748535,
      "objective/train/theoretical_loss": 3.3114870997237986,
      "objective/train/tokens_used": 1470442976,
      "objective/train/unlikelihood_loss": 0.058124542236328125,
      "theoretical_loss": 3.3114870997237986,
      "tokens_seen": 3111583744
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.8222222222222224e-05,
      "loss": 1.2022,
      "theoretical_loss": 3.3114870997237986,
      "tokens_seen": 3111583744
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.82020202020202e-05,
      "loss": 1.371,
      "theoretical_loss": 3.311481776919505,
      "tokens_seen": 3111649280
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.8181818181818185e-05,
      "loss": 1.2838,
      "theoretical_loss": 3.3114764542587056,
      "tokens_seen": 3111714816
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.816161616161616e-05,
      "loss": 1.0399,
      "theoretical_loss": 3.311471131741393,
      "tokens_seen": 3111780352
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.814141414141414e-05,
      "loss": 1.6238,
      "theoretical_loss": 3.311465809367561,
      "tokens_seen": 3111845888
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.8121212121212124e-05,
      "loss": 1.5922,
      "theoretical_loss": 3.3114604871372024,
      "tokens_seen": 3111911424
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.81010101010101e-05,
      "loss": 1.4976,
      "theoretical_loss": 3.31145516505031,
      "tokens_seen": 3111976960
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.808080808080808e-05,
      "loss": 1.3408,
      "theoretical_loss": 3.3114498431068773,
      "tokens_seen": 3112042496
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.806060606060606e-05,
      "loss": 1.48,
      "theoretical_loss": 3.311444521306898,
      "tokens_seen": 3112108032
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.804040404040404e-05,
      "loss": 1.4279,
      "theoretical_loss": 3.311439199650364,
      "tokens_seen": 3112173568
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.802020202020202e-05,
      "loss": 1.4511,
      "theoretical_loss": 3.3114338781372687,
      "tokens_seen": 3112239104
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.800000000000001e-05,
      "loss": 1.2029,
      "theoretical_loss": 3.3114285567676056,
      "tokens_seen": 3112304640
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.797979797979798e-05,
      "loss": 1.4402,
      "theoretical_loss": 3.3114232355413673,
      "tokens_seen": 3112370176
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.795959595959596e-05,
      "loss": 1.3949,
      "theoretical_loss": 3.311417914458548,
      "tokens_seen": 3112435712
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.793939393939394e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.31141259351914,
      "tokens_seen": 3112501248
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.791919191919192e-05,
      "loss": 1.1917,
      "theoretical_loss": 3.311407272723136,
      "tokens_seen": 3112566784
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.78989898989899e-05,
      "loss": 1.2947,
      "theoretical_loss": 3.31140195207053,
      "tokens_seen": 3112632320
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.787878787878788e-05,
      "loss": 1.3641,
      "theoretical_loss": 3.3113966315613146,
      "tokens_seen": 3112697856
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.785858585858586e-05,
      "loss": 1.3135,
      "theoretical_loss": 3.311391311195483,
      "tokens_seen": 3112763392
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7838383838383845e-05,
      "loss": 1.4947,
      "theoretical_loss": 3.311385990973029,
      "tokens_seen": 3112828928
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7818181818181815e-05,
      "loss": 1.5261,
      "theoretical_loss": 3.3113806708939446,
      "tokens_seen": 3112894464
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.77979797979798e-05,
      "loss": 1.5012,
      "theoretical_loss": 3.3113753509582233,
      "tokens_seen": 3112960000
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7777777777777776e-05,
      "loss": 1.4398,
      "theoretical_loss": 3.3113700311658585,
      "tokens_seen": 3113025536
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.775757575757576e-05,
      "loss": 1.5164,
      "theoretical_loss": 3.3113647115168434,
      "tokens_seen": 3113091072
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7737373737373744e-05,
      "loss": 1.3224,
      "theoretical_loss": 3.3113593920111706,
      "tokens_seen": 3113156608
    },
    {
      "epoch": 0.89,
      "objective/train/aligned_token_count": 0.5130131840705872,
      "objective/train/avg_token_score": 0.014172990806400776,
      "objective/train/docs_used": 1750566,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5457127094268799,
      "objective/train/original_loss": 3.9439444541931152,
      "objective/train/theoretical_loss": 3.3113540726488333,
      "objective/train/tokens_used": 1472081376,
      "objective/train/unlikelihood_loss": 0.06887193769216537,
      "theoretical_loss": 3.3113540726488333,
      "tokens_seen": 3113222144
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7717171717171715e-05,
      "loss": 1.4441,
      "theoretical_loss": 3.3113540726488333,
      "tokens_seen": 3113222144
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.76969696969697e-05,
      "loss": 1.6312,
      "theoretical_loss": 3.3113487534298254,
      "tokens_seen": 3113287680
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7676767676767676e-05,
      "loss": 1.3795,
      "theoretical_loss": 3.3113434343541392,
      "tokens_seen": 3113353216
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.765656565656566e-05,
      "loss": 1.1514,
      "theoretical_loss": 3.311338115421768,
      "tokens_seen": 3113418752
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.763636363636364e-05,
      "loss": 1.3905,
      "theoretical_loss": 3.311332796632705,
      "tokens_seen": 3113484288
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7616161616161614e-05,
      "loss": 1.3279,
      "theoretical_loss": 3.311327477986943,
      "tokens_seen": 3113549824
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.75959595959596e-05,
      "loss": 1.1762,
      "theoretical_loss": 3.3113221594844755,
      "tokens_seen": 3113615360
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.757575757575758e-05,
      "loss": 1.2028,
      "theoretical_loss": 3.311316841125296,
      "tokens_seen": 3113680896
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.755555555555555e-05,
      "loss": 1.1871,
      "theoretical_loss": 3.3113115229093966,
      "tokens_seen": 3113746432
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7535353535353537e-05,
      "loss": 1.4469,
      "theoretical_loss": 3.311306204836771,
      "tokens_seen": 3113811968
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7515151515151514e-05,
      "loss": 1.5588,
      "theoretical_loss": 3.311300886907413,
      "tokens_seen": 3113877504
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.74949494949495e-05,
      "loss": 1.3827,
      "theoretical_loss": 3.311295569121315,
      "tokens_seen": 3113943040
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7474747474747475e-05,
      "loss": 1.3225,
      "theoretical_loss": 3.3112902514784697,
      "tokens_seen": 3114008576
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.745454545454545e-05,
      "loss": 1.5548,
      "theoretical_loss": 3.3112849339788704,
      "tokens_seen": 3114074112
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7434343434343436e-05,
      "loss": 1.2906,
      "theoretical_loss": 3.311279616622511,
      "tokens_seen": 3114139648
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.741414141414142e-05,
      "loss": 1.312,
      "theoretical_loss": 3.311274299409384,
      "tokens_seen": 3114205184
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.73939393939394e-05,
      "loss": 1.4564,
      "theoretical_loss": 3.3112689823394827,
      "tokens_seen": 3114270720
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7373737373737374e-05,
      "loss": 1.3319,
      "theoretical_loss": 3.3112636654128007,
      "tokens_seen": 3114336256
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.735353535353535e-05,
      "loss": 1.3859,
      "theoretical_loss": 3.31125834862933,
      "tokens_seen": 3114401792
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7333333333333336e-05,
      "loss": 1.3027,
      "theoretical_loss": 3.3112530319890645,
      "tokens_seen": 3114467328
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.731313131313132e-05,
      "loss": 1.3828,
      "theoretical_loss": 3.3112477154919975,
      "tokens_seen": 3114532864
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.729292929292929e-05,
      "loss": 1.232,
      "theoretical_loss": 3.3112423991381217,
      "tokens_seen": 3114598400
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7272727272727274e-05,
      "loss": 1.4758,
      "theoretical_loss": 3.31123708292743,
      "tokens_seen": 3114663936
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.725252525252525e-05,
      "loss": 1.2323,
      "theoretical_loss": 3.3112317668599163,
      "tokens_seen": 3114729472
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7232323232323235e-05,
      "loss": 1.3882,
      "theoretical_loss": 3.311226450935573,
      "tokens_seen": 3114795008
    },
    {
      "epoch": 0.89,
      "objective/train/aligned_token_count": 0.5113025307655334,
      "objective/train/avg_token_score": 0.008897893130779266,
      "objective/train/docs_used": 1751732,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4948898553848267,
      "objective/train/original_loss": 4.125837802886963,
      "objective/train/theoretical_loss": 3.311221135154394,
      "objective/train/tokens_used": 1473719776,
      "objective/train/unlikelihood_loss": 0.05963914841413498,
      "theoretical_loss": 3.311221135154394,
      "tokens_seen": 3114860544
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.721212121212121e-05,
      "loss": 1.3133,
      "theoretical_loss": 3.311221135154394,
      "tokens_seen": 3114860544
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.719191919191919e-05,
      "loss": 1.4539,
      "theoretical_loss": 3.3112158195163715,
      "tokens_seen": 3114926080
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.717171717171717e-05,
      "loss": 1.2063,
      "theoretical_loss": 3.311210504021499,
      "tokens_seen": 3114991616
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.715151515151516e-05,
      "loss": 1.131,
      "theoretical_loss": 3.3112051886697706,
      "tokens_seen": 3115057152
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.7131313131313134e-05,
      "loss": 1.3568,
      "theoretical_loss": 3.311199873461178,
      "tokens_seen": 3115122688
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.711111111111111e-05,
      "loss": 1.4391,
      "theoretical_loss": 3.311194558395715,
      "tokens_seen": 3115188224
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.709090909090909e-05,
      "loss": 1.4926,
      "theoretical_loss": 3.3111892434733745,
      "tokens_seen": 3115253760
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.707070707070707e-05,
      "loss": 1.4403,
      "theoretical_loss": 3.31118392869415,
      "tokens_seen": 3115319296
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.705050505050506e-05,
      "loss": 1.327,
      "theoretical_loss": 3.3111786140580346,
      "tokens_seen": 3115384832
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.703030303030303e-05,
      "loss": 1.5208,
      "theoretical_loss": 3.311173299565021,
      "tokens_seen": 3115450368
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.701010101010101e-05,
      "loss": 1.3262,
      "theoretical_loss": 3.3111679852151026,
      "tokens_seen": 3115515904
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6989898989898995e-05,
      "loss": 1.4485,
      "theoretical_loss": 3.3111626710082724,
      "tokens_seen": 3115581440
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.696969696969697e-05,
      "loss": 1.1387,
      "theoretical_loss": 3.311157356944524,
      "tokens_seen": 3115646976
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.694949494949495e-05,
      "loss": 1.3137,
      "theoretical_loss": 3.3111520430238497,
      "tokens_seen": 3115712512
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.692929292929293e-05,
      "loss": 1.4803,
      "theoretical_loss": 3.3111467292462438,
      "tokens_seen": 3115778048
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.690909090909091e-05,
      "loss": 1.4669,
      "theoretical_loss": 3.311141415611698,
      "tokens_seen": 3115843584
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6888888888888895e-05,
      "loss": 1.4483,
      "theoretical_loss": 3.311136102120207,
      "tokens_seen": 3115909120
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6868686868686865e-05,
      "loss": 1.3501,
      "theoretical_loss": 3.311130788771763,
      "tokens_seen": 3115974656
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.684848484848485e-05,
      "loss": 1.6005,
      "theoretical_loss": 3.311125475566359,
      "tokens_seen": 3116040192
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6828282828282826e-05,
      "loss": 1.1829,
      "theoretical_loss": 3.3111201625039888,
      "tokens_seen": 3116105728
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.680808080808081e-05,
      "loss": 1.6111,
      "theoretical_loss": 3.3111148495846447,
      "tokens_seen": 3116171264
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6787878787878794e-05,
      "loss": 1.4998,
      "theoretical_loss": 3.3111095368083205,
      "tokens_seen": 3116236800
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6767676767676764e-05,
      "loss": 1.3165,
      "theoretical_loss": 3.3111042241750095,
      "tokens_seen": 3116302336
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.674747474747475e-05,
      "loss": 1.2831,
      "theoretical_loss": 3.311098911684704,
      "tokens_seen": 3116367872
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.672727272727273e-05,
      "loss": 1.272,
      "theoretical_loss": 3.311093599337398,
      "tokens_seen": 3116433408
    },
    {
      "epoch": 0.89,
      "objective/train/aligned_token_count": 0.23503176867961884,
      "objective/train/avg_token_score": 0.01828402653336525,
      "objective/train/docs_used": 1752323,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.7561014890670776,
      "objective/train/original_loss": 4.8860249519348145,
      "objective/train/theoretical_loss": 3.3110882871330847,
      "objective/train/tokens_used": 1475358176,
      "objective/train/unlikelihood_loss": 0.059419021010398865,
      "theoretical_loss": 3.3110882871330847,
      "tokens_seen": 3116498944
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.670707070707071e-05,
      "loss": 1.303,
      "theoretical_loss": 3.3110882871330847,
      "tokens_seen": 3116498944
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.668686868686869e-05,
      "loss": 1.4659,
      "theoretical_loss": 3.3110829750717565,
      "tokens_seen": 3116564480
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6666666666666664e-05,
      "loss": 1.3769,
      "theoretical_loss": 3.311077663153407,
      "tokens_seen": 3116630016
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.664646464646465e-05,
      "loss": 1.2738,
      "theoretical_loss": 3.3110723513780287,
      "tokens_seen": 3116695552
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.662626262626263e-05,
      "loss": 1.2492,
      "theoretical_loss": 3.3110670397456157,
      "tokens_seen": 3116761088
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.66060606060606e-05,
      "loss": 1.3169,
      "theoretical_loss": 3.3110617282561607,
      "tokens_seen": 3116826624
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6585858585858586e-05,
      "loss": 1.358,
      "theoretical_loss": 3.311056416909657,
      "tokens_seen": 3116892160
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.656565656565657e-05,
      "loss": 1.2251,
      "theoretical_loss": 3.311051105706097,
      "tokens_seen": 3116957696
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.654545454545455e-05,
      "loss": 1.3774,
      "theoretical_loss": 3.311045794645475,
      "tokens_seen": 3117023232
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.652525252525253e-05,
      "loss": 1.3415,
      "theoretical_loss": 3.311040483727784,
      "tokens_seen": 3117088768
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.65050505050505e-05,
      "loss": 1.3694,
      "theoretical_loss": 3.311035172953016,
      "tokens_seen": 3117154304
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6484848484848486e-05,
      "loss": 1.3414,
      "theoretical_loss": 3.311029862321165,
      "tokens_seen": 3117219840
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.646464646464647e-05,
      "loss": 1.639,
      "theoretical_loss": 3.3110245518322245,
      "tokens_seen": 3117285376
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.644444444444445e-05,
      "loss": 1.3688,
      "theoretical_loss": 3.3110192414861874,
      "tokens_seen": 3117350912
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6424242424242424e-05,
      "loss": 1.4718,
      "theoretical_loss": 3.311013931283046,
      "tokens_seen": 3117416448
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.64040404040404e-05,
      "loss": 1.3085,
      "theoretical_loss": 3.3110086212227947,
      "tokens_seen": 3117481984
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6383838383838385e-05,
      "loss": 1.4319,
      "theoretical_loss": 3.311003311305426,
      "tokens_seen": 3117547520
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.636363636363637e-05,
      "loss": 1.5217,
      "theoretical_loss": 3.3109980015309324,
      "tokens_seen": 3117613056
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.634343434343434e-05,
      "loss": 1.0552,
      "theoretical_loss": 3.3109926918993087,
      "tokens_seen": 3117678592
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6323232323232324e-05,
      "loss": 1.5416,
      "theoretical_loss": 3.3109873824105467,
      "tokens_seen": 3117744128
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.630303030303031e-05,
      "loss": 1.3866,
      "theoretical_loss": 3.3109820730646398,
      "tokens_seen": 3117809664
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6282828282828285e-05,
      "loss": 1.2464,
      "theoretical_loss": 3.3109767638615817,
      "tokens_seen": 3117875200
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.626262626262627e-05,
      "loss": 1.3168,
      "theoretical_loss": 3.310971454801365,
      "tokens_seen": 3117940736
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.624242424242424e-05,
      "loss": 1.2862,
      "theoretical_loss": 3.3109661458839827,
      "tokens_seen": 3118006272
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.622222222222222e-05,
      "loss": 1.1519,
      "theoretical_loss": 3.3109608371094286,
      "tokens_seen": 3118071808
    },
    {
      "epoch": 0.89,
      "objective/train/aligned_token_count": 0.37463343143463135,
      "objective/train/avg_token_score": 0.009226646274328232,
      "objective/train/docs_used": 1753489,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.1963626146316528,
      "objective/train/original_loss": 4.749354839324951,
      "objective/train/theoretical_loss": 3.3109555284776957,
      "objective/train/tokens_used": 1476996576,
      "objective/train/unlikelihood_loss": 0.06636327505111694,
      "theoretical_loss": 3.3109555284776957,
      "tokens_seen": 3118137344
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.620202020202021e-05,
      "loss": 1.3399,
      "theoretical_loss": 3.3109555284776957,
      "tokens_seen": 3118137344
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6181818181818184e-05,
      "loss": 1.4453,
      "theoretical_loss": 3.310950219988777,
      "tokens_seen": 3118202880
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.616161616161616e-05,
      "loss": 1.254,
      "theoretical_loss": 3.3109449116426655,
      "tokens_seen": 3118268416
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6141414141414145e-05,
      "loss": 1.4918,
      "theoretical_loss": 3.310939603439355,
      "tokens_seen": 3118333952
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.612121212121212e-05,
      "loss": 1.4313,
      "theoretical_loss": 3.3109342953788374,
      "tokens_seen": 3118399488
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6101010101010106e-05,
      "loss": 1.45,
      "theoretical_loss": 3.310928987461107,
      "tokens_seen": 3118465024
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.608080808080808e-05,
      "loss": 1.5972,
      "theoretical_loss": 3.310923679686157,
      "tokens_seen": 3118530560
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.606060606060606e-05,
      "loss": 1.2961,
      "theoretical_loss": 3.3109183720539797,
      "tokens_seen": 3118596096
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6040404040404045e-05,
      "loss": 1.2039,
      "theoretical_loss": 3.3109130645645686,
      "tokens_seen": 3118661632
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.602020202020202e-05,
      "loss": 1.4931,
      "theoretical_loss": 3.310907757217917,
      "tokens_seen": 3118727168
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.6e-05,
      "loss": 1.548,
      "theoretical_loss": 3.3109024500140185,
      "tokens_seen": 3118792704
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5979797979797976e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.3108971429528653,
      "tokens_seen": 3118858240
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.595959595959596e-05,
      "loss": 1.5197,
      "theoretical_loss": 3.3108918360344513,
      "tokens_seen": 3118923776
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5939393939393944e-05,
      "loss": 1.4406,
      "theoretical_loss": 3.3108865292587693,
      "tokens_seen": 3118989312
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.591919191919192e-05,
      "loss": 1.1079,
      "theoretical_loss": 3.3108812226258126,
      "tokens_seen": 3119054848
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.58989898989899e-05,
      "loss": 1.4758,
      "theoretical_loss": 3.3108759161355743,
      "tokens_seen": 3119120384
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.587878787878788e-05,
      "loss": 1.3433,
      "theoretical_loss": 3.310870609788048,
      "tokens_seen": 3119185920
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.585858585858586e-05,
      "loss": 1.4788,
      "theoretical_loss": 3.310865303583226,
      "tokens_seen": 3119251456
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5838383838383844e-05,
      "loss": 1.2951,
      "theoretical_loss": 3.310859997521102,
      "tokens_seen": 3119316992
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5818181818181814e-05,
      "loss": 1.2756,
      "theoretical_loss": 3.3108546916016692,
      "tokens_seen": 3119382528
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.57979797979798e-05,
      "loss": 1.5766,
      "theoretical_loss": 3.3108493858249206,
      "tokens_seen": 3119448064
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.577777777777778e-05,
      "loss": 1.3625,
      "theoretical_loss": 3.3108440801908494,
      "tokens_seen": 3119513600
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.575757575757576e-05,
      "loss": 1.4995,
      "theoretical_loss": 3.310838774699449,
      "tokens_seen": 3119579136
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5737373737373736e-05,
      "loss": 1.3776,
      "theoretical_loss": 3.310833469350712,
      "tokens_seen": 3119644672
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5717171717171714e-05,
      "loss": 1.4954,
      "theoretical_loss": 3.3108281641446324,
      "tokens_seen": 3119710208
    },
    {
      "epoch": 0.89,
      "objective/train/aligned_token_count": 0.5104472041130066,
      "objective/train/avg_token_score": 0.01837521605193615,
      "objective/train/docs_used": 1754072,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.638211727142334,
      "objective/train/original_loss": 4.6138529777526855,
      "objective/train/theoretical_loss": 3.3108228590812026,
      "objective/train/tokens_used": 1478634976,
      "objective/train/unlikelihood_loss": 0.044399913400411606,
      "theoretical_loss": 3.3108228590812026,
      "tokens_seen": 3119775744
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.56969696969697e-05,
      "loss": 1.6568,
      "theoretical_loss": 3.3108228590812026,
      "tokens_seen": 3119775744
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.567676767676768e-05,
      "loss": 1.5454,
      "theoretical_loss": 3.3108175541604163,
      "tokens_seen": 3119841280
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.565656565656565e-05,
      "loss": 1.4197,
      "theoretical_loss": 3.3108122493822663,
      "tokens_seen": 3119906816
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5636363636363636e-05,
      "loss": 1.4668,
      "theoretical_loss": 3.3108069447467456,
      "tokens_seen": 3119972352
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.561616161616162e-05,
      "loss": 1.5944,
      "theoretical_loss": 3.310801640253848,
      "tokens_seen": 3120037888
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.55959595959596e-05,
      "loss": 1.1438,
      "theoretical_loss": 3.310796335903566,
      "tokens_seen": 3120103424
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.557575757575758e-05,
      "loss": 1.4806,
      "theoretical_loss": 3.3107910316958935,
      "tokens_seen": 3120168960
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.555555555555555e-05,
      "loss": 1.6026,
      "theoretical_loss": 3.3107857276308232,
      "tokens_seen": 3120234496
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5535353535353535e-05,
      "loss": 1.2833,
      "theoretical_loss": 3.3107804237083482,
      "tokens_seen": 3120300032
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.551515151515152e-05,
      "loss": 1.4889,
      "theoretical_loss": 3.310775119928462,
      "tokens_seen": 3120365568
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5494949494949497e-05,
      "loss": 1.5752,
      "theoretical_loss": 3.310769816291158,
      "tokens_seen": 3120431104
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5474747474747474e-05,
      "loss": 1.316,
      "theoretical_loss": 3.3107645127964282,
      "tokens_seen": 3120496640
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.545454545454546e-05,
      "loss": 1.3199,
      "theoretical_loss": 3.310759209444267,
      "tokens_seen": 3120562176
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5434343434343435e-05,
      "loss": 1.3505,
      "theoretical_loss": 3.3107539062346665,
      "tokens_seen": 3120627712
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.541414141414142e-05,
      "loss": 1.4844,
      "theoretical_loss": 3.310748603167621,
      "tokens_seen": 3120693248
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.539393939393939e-05,
      "loss": 1.2961,
      "theoretical_loss": 3.3107433002431232,
      "tokens_seen": 3120758784
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.537373737373737e-05,
      "loss": 1.2536,
      "theoretical_loss": 3.310737997461166,
      "tokens_seen": 3120824320
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.535353535353536e-05,
      "loss": 1.5921,
      "theoretical_loss": 3.310732694821743,
      "tokens_seen": 3120889856
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5333333333333334e-05,
      "loss": 1.3897,
      "theoretical_loss": 3.3107273923248473,
      "tokens_seen": 3120955392
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.531313131313132e-05,
      "loss": 1.3934,
      "theoretical_loss": 3.3107220899704717,
      "tokens_seen": 3121020928
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.529292929292929e-05,
      "loss": 1.4569,
      "theoretical_loss": 3.3107167877586097,
      "tokens_seen": 3121086464
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.527272727272727e-05,
      "loss": 1.4667,
      "theoretical_loss": 3.3107114856892546,
      "tokens_seen": 3121152000
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.525252525252526e-05,
      "loss": 1.4864,
      "theoretical_loss": 3.310706183762399,
      "tokens_seen": 3121217536
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5232323232323234e-05,
      "loss": 1.3311,
      "theoretical_loss": 3.310700881978037,
      "tokens_seen": 3121283072
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.521212121212121e-05,
      "loss": 1.5355,
      "theoretical_loss": 3.310695580336161,
      "tokens_seen": 3121348608
    },
    {
      "epoch": 0.89,
      "objective/train/aligned_token_count": 0.4471529722213745,
      "objective/train/avg_token_score": 0.005897354334592819,
      "objective/train/docs_used": 1755494,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4884589910507202,
      "objective/train/original_loss": 4.368992805480957,
      "objective/train/theoretical_loss": 3.310690278836764,
      "objective/train/tokens_used": 1480273376,
      "objective/train/unlikelihood_loss": 0.0505085252225399,
      "theoretical_loss": 3.310690278836764,
      "tokens_seen": 3121414144
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5191919191919195e-05,
      "loss": 1.4187,
      "theoretical_loss": 3.310690278836764,
      "tokens_seen": 3121414144
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.517171717171717e-05,
      "loss": 1.3622,
      "theoretical_loss": 3.3106849774798404,
      "tokens_seen": 3121479680
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5151515151515156e-05,
      "loss": 1.6132,
      "theoretical_loss": 3.310679676265382,
      "tokens_seen": 3121545216
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5131313131313127e-05,
      "loss": 1.2555,
      "theoretical_loss": 3.3106743751933827,
      "tokens_seen": 3121610752
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.511111111111111e-05,
      "loss": 1.2196,
      "theoretical_loss": 3.310669074263836,
      "tokens_seen": 3121676288
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5090909090909094e-05,
      "loss": 1.4866,
      "theoretical_loss": 3.310663773476734,
      "tokens_seen": 3121741824
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.507070707070707e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.310658472832071,
      "tokens_seen": 3121807360
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.5050505050505056e-05,
      "loss": 1.3995,
      "theoretical_loss": 3.3106531723298396,
      "tokens_seen": 3121872896
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.503030303030303e-05,
      "loss": 1.4123,
      "theoretical_loss": 3.3106478719700325,
      "tokens_seen": 3121938432
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.501010101010101e-05,
      "loss": 1.3364,
      "theoretical_loss": 3.310642571752644,
      "tokens_seen": 3122003968
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4989898989898994e-05,
      "loss": 1.3259,
      "theoretical_loss": 3.310637271677667,
      "tokens_seen": 3122069504
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.496969696969697e-05,
      "loss": 1.4962,
      "theoretical_loss": 3.310631971745094,
      "tokens_seen": 3122135040
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.494949494949495e-05,
      "loss": 1.4263,
      "theoretical_loss": 3.3106266719549184,
      "tokens_seen": 3122200576
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.492929292929293e-05,
      "loss": 1.3417,
      "theoretical_loss": 3.3106213723071343,
      "tokens_seen": 3122266112
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.490909090909091e-05,
      "loss": 1.35,
      "theoretical_loss": 3.310616072801734,
      "tokens_seen": 3122331648
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4888888888888893e-05,
      "loss": 1.349,
      "theoretical_loss": 3.3106107734387105,
      "tokens_seen": 3122397184
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4868686868686864e-05,
      "loss": 1.3102,
      "theoretical_loss": 3.3106054742180575,
      "tokens_seen": 3122462720
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.484848484848485e-05,
      "loss": 1.2217,
      "theoretical_loss": 3.3106001751397685,
      "tokens_seen": 3122528256
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.482828282828283e-05,
      "loss": 1.5152,
      "theoretical_loss": 3.3105948762038357,
      "tokens_seen": 3122593792
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.480808080808081e-05,
      "loss": 1.2986,
      "theoretical_loss": 3.3105895774102527,
      "tokens_seen": 3122659328
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4787878787878786e-05,
      "loss": 1.4923,
      "theoretical_loss": 3.310584278759013,
      "tokens_seen": 3122724864
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.476767676767677e-05,
      "loss": 1.3882,
      "theoretical_loss": 3.31057898025011,
      "tokens_seen": 3122790400
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.474747474747475e-05,
      "loss": 1.2038,
      "theoretical_loss": 3.310573681883536,
      "tokens_seen": 3122855936
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.472727272727273e-05,
      "loss": 1.2645,
      "theoretical_loss": 3.310568383659285,
      "tokens_seen": 3122921472
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.470707070707071e-05,
      "loss": 1.5393,
      "theoretical_loss": 3.31056308557735,
      "tokens_seen": 3122987008
    },
    {
      "epoch": 0.89,
      "objective/train/aligned_token_count": 0.5250488519668579,
      "objective/train/avg_token_score": 0.008977536112070084,
      "objective/train/docs_used": 1755986,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.7008429765701294,
      "objective/train/original_loss": 3.9189841747283936,
      "objective/train/theoretical_loss": 3.3105577876377237,
      "objective/train/tokens_used": 1481911776,
      "objective/train/unlikelihood_loss": 0.08025264739990234,
      "theoretical_loss": 3.3105577876377237,
      "tokens_seen": 3123052544
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4686868686868686e-05,
      "loss": 1.5521,
      "theoretical_loss": 3.3105577876377237,
      "tokens_seen": 3123052544
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.466666666666667e-05,
      "loss": 1.637,
      "theoretical_loss": 3.3105524898403997,
      "tokens_seen": 3123118080
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.464646464646465e-05,
      "loss": 1.2455,
      "theoretical_loss": 3.3105471921853713,
      "tokens_seen": 3123183616
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.462626262626263e-05,
      "loss": 1.6947,
      "theoretical_loss": 3.3105418946726313,
      "tokens_seen": 3123249152
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.460606060606061e-05,
      "loss": 1.3524,
      "theoretical_loss": 3.3105365973021734,
      "tokens_seen": 3123314688
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4585858585858585e-05,
      "loss": 1.146,
      "theoretical_loss": 3.31053130007399,
      "tokens_seen": 3123380224
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.456565656565657e-05,
      "loss": 1.3404,
      "theoretical_loss": 3.310526002988076,
      "tokens_seen": 3123445760
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4545454545454546e-05,
      "loss": 1.423,
      "theoretical_loss": 3.3105207060444224,
      "tokens_seen": 3123511296
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4525252525252523e-05,
      "loss": 1.5747,
      "theoretical_loss": 3.3105154092430236,
      "tokens_seen": 3123576832
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.450505050505051e-05,
      "loss": 1.2816,
      "theoretical_loss": 3.3105101125838727,
      "tokens_seen": 3123642368
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4484848484848485e-05,
      "loss": 1.5482,
      "theoretical_loss": 3.3105048160669623,
      "tokens_seen": 3123707904
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.446464646464647e-05,
      "loss": 1.4614,
      "theoretical_loss": 3.3104995196922866,
      "tokens_seen": 3123773440
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.444444444444444e-05,
      "loss": 1.3009,
      "theoretical_loss": 3.3104942234598385,
      "tokens_seen": 3123838976
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.442424242424242e-05,
      "loss": 1.3782,
      "theoretical_loss": 3.3104889273696108,
      "tokens_seen": 3123904512
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.440404040404041e-05,
      "loss": 1.3777,
      "theoretical_loss": 3.310483631421597,
      "tokens_seen": 3123970048
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4383838383838384e-05,
      "loss": 1.4639,
      "theoretical_loss": 3.3104783356157896,
      "tokens_seen": 3124035584
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.436363636363637e-05,
      "loss": 1.4723,
      "theoretical_loss": 3.310473039952183,
      "tokens_seen": 3124101120
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4343434343434345e-05,
      "loss": 1.1635,
      "theoretical_loss": 3.3104677444307695,
      "tokens_seen": 3124166656
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.432323232323232e-05,
      "loss": 1.3735,
      "theoretical_loss": 3.3104624490515424,
      "tokens_seen": 3124232192
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4303030303030306e-05,
      "loss": 1.5016,
      "theoretical_loss": 3.3104571538144953,
      "tokens_seen": 3124297728
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4282828282828284e-05,
      "loss": 1.2836,
      "theoretical_loss": 3.310451858719621,
      "tokens_seen": 3124363264
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.426262626262626e-05,
      "loss": 1.3885,
      "theoretical_loss": 3.3104465637669134,
      "tokens_seen": 3124428800
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4242424242424245e-05,
      "loss": 1.3748,
      "theoretical_loss": 3.3104412689563647,
      "tokens_seen": 3124494336
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.422222222222222e-05,
      "loss": 1.3896,
      "theoretical_loss": 3.310435974287969,
      "tokens_seen": 3124559872
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4202020202020206e-05,
      "loss": 1.6036,
      "theoretical_loss": 3.3104306797617187,
      "tokens_seen": 3124625408
    },
    {
      "epoch": 0.89,
      "objective/train/aligned_token_count": 0.47831133008003235,
      "objective/train/avg_token_score": 0.015462594106793404,
      "objective/train/docs_used": 1757280,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5683847665786743,
      "objective/train/original_loss": 4.095974445343018,
      "objective/train/theoretical_loss": 3.3104253853776076,
      "objective/train/tokens_used": 1483550176,
      "objective/train/unlikelihood_loss": 0.0657786875963211,
      "theoretical_loss": 3.3104253853776076,
      "tokens_seen": 3124690944
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.418181818181818e-05,
      "loss": 1.3695,
      "theoretical_loss": 3.3104253853776076,
      "tokens_seen": 3124690944
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.416161616161616e-05,
      "loss": 1.3747,
      "theoretical_loss": 3.3104200911356285,
      "tokens_seen": 3124756480
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4141414141414144e-05,
      "loss": 1.3269,
      "theoretical_loss": 3.310414797035775,
      "tokens_seen": 3124822016
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.412121212121212e-05,
      "loss": 1.2731,
      "theoretical_loss": 3.31040950307804,
      "tokens_seen": 3124887552
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4101010101010105e-05,
      "loss": 1.2706,
      "theoretical_loss": 3.3104042092624164,
      "tokens_seen": 3124953088
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.408080808080808e-05,
      "loss": 1.3447,
      "theoretical_loss": 3.310398915588898,
      "tokens_seen": 3125018624
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.406060606060606e-05,
      "loss": 1.3556,
      "theoretical_loss": 3.310393622057479,
      "tokens_seen": 3125084160
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4040404040404044e-05,
      "loss": 1.3356,
      "theoretical_loss": 3.3103883286681497,
      "tokens_seen": 3125149696
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.402020202020202e-05,
      "loss": 1.3887,
      "theoretical_loss": 3.310383035420906,
      "tokens_seen": 3125215232
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.4e-05,
      "loss": 1.4257,
      "theoretical_loss": 3.31037774231574,
      "tokens_seen": 3125280768
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.397979797979798e-05,
      "loss": 1.447,
      "theoretical_loss": 3.310372449352645,
      "tokens_seen": 3125346304
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.395959595959596e-05,
      "loss": 1.3927,
      "theoretical_loss": 3.3103671565316137,
      "tokens_seen": 3125411840
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.393939393939394e-05,
      "loss": 1.2569,
      "theoretical_loss": 3.3103618638526404,
      "tokens_seen": 3125477376
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.391919191919192e-05,
      "loss": 1.3897,
      "theoretical_loss": 3.310356571315718,
      "tokens_seen": 3125542912
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.38989898989899e-05,
      "loss": 1.4368,
      "theoretical_loss": 3.310351278920839,
      "tokens_seen": 3125608448
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.387878787878788e-05,
      "loss": 1.6444,
      "theoretical_loss": 3.310345986667997,
      "tokens_seen": 3125673984
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.385858585858586e-05,
      "loss": 1.2706,
      "theoretical_loss": 3.310340694557186,
      "tokens_seen": 3125739520
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.383838383838384e-05,
      "loss": 1.1783,
      "theoretical_loss": 3.3103354025883975,
      "tokens_seen": 3125805056
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.381818181818182e-05,
      "loss": 1.3142,
      "theoretical_loss": 3.3103301107616265,
      "tokens_seen": 3125870592
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.37979797979798e-05,
      "loss": 1.3304,
      "theoretical_loss": 3.310324819076865,
      "tokens_seen": 3125936128
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.377777777777778e-05,
      "loss": 1.4506,
      "theoretical_loss": 3.310319527534107,
      "tokens_seen": 3126001664
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3757575757575765e-05,
      "loss": 1.5344,
      "theoretical_loss": 3.310314236133345,
      "tokens_seen": 3126067200
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3737373737373735e-05,
      "loss": 1.3056,
      "theoretical_loss": 3.310308944874573,
      "tokens_seen": 3126132736
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.371717171717172e-05,
      "loss": 1.7138,
      "theoretical_loss": 3.3103036537577832,
      "tokens_seen": 3126198272
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3696969696969696e-05,
      "loss": 1.3948,
      "theoretical_loss": 3.31029836278297,
      "tokens_seen": 3126263808
    },
    {
      "epoch": 0.89,
      "objective/train/aligned_token_count": 0.48497068881988525,
      "objective/train/avg_token_score": 0.04004567861557007,
      "objective/train/docs_used": 1757671,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3589171171188354,
      "objective/train/original_loss": 4.2097344398498535,
      "objective/train/theoretical_loss": 3.3102930719501256,
      "objective/train/tokens_used": 1485188576,
      "objective/train/unlikelihood_loss": 0.04839378222823143,
      "theoretical_loss": 3.3102930719501256,
      "tokens_seen": 3126329344
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.367676767676768e-05,
      "loss": 1.3758,
      "theoretical_loss": 3.3102930719501256,
      "tokens_seen": 3126329344
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.365656565656566e-05,
      "loss": 1.3502,
      "theoretical_loss": 3.3102877812592437,
      "tokens_seen": 3126394880
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3636363636363635e-05,
      "loss": 1.2629,
      "theoretical_loss": 3.310282490710317,
      "tokens_seen": 3126460416
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.361616161616162e-05,
      "loss": 1.7126,
      "theoretical_loss": 3.31027720030334,
      "tokens_seen": 3126525952
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3595959595959596e-05,
      "loss": 1.322,
      "theoretical_loss": 3.3102719100383045,
      "tokens_seen": 3126591488
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.357575757575757e-05,
      "loss": 1.4139,
      "theoretical_loss": 3.3102666199152044,
      "tokens_seen": 3126657024
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.355555555555556e-05,
      "loss": 1.3564,
      "theoretical_loss": 3.3102613299340327,
      "tokens_seen": 3126722560
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3535353535353534e-05,
      "loss": 1.3142,
      "theoretical_loss": 3.3102560400947834,
      "tokens_seen": 3126788096
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.351515151515152e-05,
      "loss": 1.5294,
      "theoretical_loss": 3.310250750397448,
      "tokens_seen": 3126853632
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.34949494949495e-05,
      "loss": 1.3711,
      "theoretical_loss": 3.3102454608420215,
      "tokens_seen": 3126919168
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.347474747474747e-05,
      "loss": 1.4383,
      "theoretical_loss": 3.3102401714284957,
      "tokens_seen": 3126984704
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3454545454545457e-05,
      "loss": 1.3929,
      "theoretical_loss": 3.3102348821568652,
      "tokens_seen": 3127050240
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3434343434343434e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.310229593027122,
      "tokens_seen": 3127115776
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.341414141414142e-05,
      "loss": 1.3426,
      "theoretical_loss": 3.3102243040392603,
      "tokens_seen": 3127181312
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3393939393939395e-05,
      "loss": 1.2359,
      "theoretical_loss": 3.3102190151932724,
      "tokens_seen": 3127246848
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.337373737373737e-05,
      "loss": 1.2562,
      "theoretical_loss": 3.3102137264891525,
      "tokens_seen": 3127312384
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3353535353535356e-05,
      "loss": 1.3702,
      "theoretical_loss": 3.310208437926893,
      "tokens_seen": 3127377920
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.333333333333334e-05,
      "loss": 1.2735,
      "theoretical_loss": 3.310203149506487,
      "tokens_seen": 3127443456
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.331313131313131e-05,
      "loss": 1.2253,
      "theoretical_loss": 3.3101978612279286,
      "tokens_seen": 3127508992
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3292929292929294e-05,
      "loss": 1.189,
      "theoretical_loss": 3.3101925730912107,
      "tokens_seen": 3127574528
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.327272727272727e-05,
      "loss": 1.4019,
      "theoretical_loss": 3.310187285096326,
      "tokens_seen": 3127640064
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3252525252525255e-05,
      "loss": 1.315,
      "theoretical_loss": 3.310181997243268,
      "tokens_seen": 3127705600
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.323232323232324e-05,
      "loss": 1.3051,
      "theoretical_loss": 3.3101767095320307,
      "tokens_seen": 3127771136
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.321212121212121e-05,
      "loss": 1.3747,
      "theoretical_loss": 3.310171421962606,
      "tokens_seen": 3127836672
    },
    {
      "epoch": 0.89,
      "learning_rate": 5.3191919191919194e-05,
      "loss": 1.4495,
      "theoretical_loss": 3.310166134534988,
      "tokens_seen": 3127902208
    },
    {
      "epoch": 0.89,
      "objective/train/aligned_token_count": 0.4962121248245239,
      "objective/train/avg_token_score": 0.005860667675733566,
      "objective/train/docs_used": 1758353,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.434056282043457,
      "objective/train/original_loss": 4.247581481933594,
      "objective/train/theoretical_loss": 3.31016084724917,
      "objective/train/tokens_used": 1486826976,
      "objective/train/unlikelihood_loss": 0.05410975217819214,
      "theoretical_loss": 3.31016084724917,
      "tokens_seen": 3127967744
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.317171717171717e-05,
      "loss": 1.1845,
      "theoretical_loss": 3.31016084724917,
      "tokens_seen": 3127967744
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.3151515151515155e-05,
      "loss": 1.3965,
      "theoretical_loss": 3.3101555601051444,
      "tokens_seen": 3128033280
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.313131313131313e-05,
      "loss": 1.2922,
      "theoretical_loss": 3.3101502731029053,
      "tokens_seen": 3128098816
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.311111111111111e-05,
      "loss": 1.1195,
      "theoretical_loss": 3.3101449862424457,
      "tokens_seen": 3128164352
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.309090909090909e-05,
      "loss": 1.349,
      "theoretical_loss": 3.3101396995237584,
      "tokens_seen": 3128229888
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.307070707070708e-05,
      "loss": 1.3562,
      "theoretical_loss": 3.3101344129468373,
      "tokens_seen": 3128295424
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.305050505050505e-05,
      "loss": 1.2627,
      "theoretical_loss": 3.3101291265116752,
      "tokens_seen": 3128360960
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.303030303030303e-05,
      "loss": 1.3676,
      "theoretical_loss": 3.310123840218265,
      "tokens_seen": 3128426496
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.301010101010101e-05,
      "loss": 1.3354,
      "theoretical_loss": 3.3101185540666007,
      "tokens_seen": 3128492032
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.298989898989899e-05,
      "loss": 1.605,
      "theoretical_loss": 3.3101132680566754,
      "tokens_seen": 3128557568
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.296969696969697e-05,
      "loss": 1.4929,
      "theoretical_loss": 3.3101079821884816,
      "tokens_seen": 3128623104
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.294949494949495e-05,
      "loss": 1.4245,
      "theoretical_loss": 3.310102696462013,
      "tokens_seen": 3128688640
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.292929292929293e-05,
      "loss": 1.2883,
      "theoretical_loss": 3.310097410877263,
      "tokens_seen": 3128754176
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2909090909090915e-05,
      "loss": 1.4851,
      "theoretical_loss": 3.310092125434225,
      "tokens_seen": 3128819712
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.288888888888889e-05,
      "loss": 1.3343,
      "theoretical_loss": 3.310086840132892,
      "tokens_seen": 3128885248
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.286868686868687e-05,
      "loss": 1.4588,
      "theoretical_loss": 3.3100815549732565,
      "tokens_seen": 3128950784
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2848484848484847e-05,
      "loss": 1.3604,
      "theoretical_loss": 3.310076269955313,
      "tokens_seen": 3129016320
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.282828282828283e-05,
      "loss": 1.4244,
      "theoretical_loss": 3.3100709850790535,
      "tokens_seen": 3129081856
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2808080808080815e-05,
      "loss": 1.4149,
      "theoretical_loss": 3.3100657003444725,
      "tokens_seen": 3129147392
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2787878787878785e-05,
      "loss": 1.4766,
      "theoretical_loss": 3.310060415751562,
      "tokens_seen": 3129212928
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.276767676767677e-05,
      "loss": 1.4378,
      "theoretical_loss": 3.310055131300316,
      "tokens_seen": 3129278464
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2747474747474746e-05,
      "loss": 1.2126,
      "theoretical_loss": 3.310049846990728,
      "tokens_seen": 3129344000
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.272727272727273e-05,
      "loss": 1.4361,
      "theoretical_loss": 3.3100445628227906,
      "tokens_seen": 3129409536
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.270707070707071e-05,
      "loss": 1.2247,
      "theoretical_loss": 3.310039278796497,
      "tokens_seen": 3129475072
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2686868686868684e-05,
      "loss": 1.4909,
      "theoretical_loss": 3.3100339949118407,
      "tokens_seen": 3129540608
    },
    {
      "epoch": 0.9,
      "objective/train/aligned_token_count": 0.49217987060546875,
      "objective/train/avg_token_score": 0.015217938460409641,
      "objective/train/docs_used": 1759339,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.6157103776931763,
      "objective/train/original_loss": 4.176531791687012,
      "objective/train/theoretical_loss": 3.3100287111688154,
      "objective/train/tokens_used": 1488465376,
      "objective/train/unlikelihood_loss": 0.05973109230399132,
      "theoretical_loss": 3.3100287111688154,
      "tokens_seen": 3129606144
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.266666666666667e-05,
      "loss": 1.56,
      "theoretical_loss": 3.3100287111688154,
      "tokens_seen": 3129606144
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.264646464646465e-05,
      "loss": 1.1816,
      "theoretical_loss": 3.310023427567413,
      "tokens_seen": 3129671680
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.262626262626263e-05,
      "loss": 1.383,
      "theoretical_loss": 3.3100181441076284,
      "tokens_seen": 3129737216
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.260606060606061e-05,
      "loss": 1.4676,
      "theoretical_loss": 3.3100128607894534,
      "tokens_seen": 3129802752
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2585858585858584e-05,
      "loss": 1.2084,
      "theoretical_loss": 3.310007577612882,
      "tokens_seen": 3129868288
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.256565656565657e-05,
      "loss": 1.3987,
      "theoretical_loss": 3.3100022945779077,
      "tokens_seen": 3129933824
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.254545454545455e-05,
      "loss": 1.2508,
      "theoretical_loss": 3.309997011684523,
      "tokens_seen": 3129999360
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.252525252525252e-05,
      "loss": 1.3514,
      "theoretical_loss": 3.3099917289327214,
      "tokens_seen": 3130064896
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2505050505050506e-05,
      "loss": 1.4349,
      "theoretical_loss": 3.3099864463224966,
      "tokens_seen": 3130130432
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.248484848484848e-05,
      "loss": 1.3063,
      "theoretical_loss": 3.3099811638538412,
      "tokens_seen": 3130195968
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.246464646464647e-05,
      "loss": 1.4106,
      "theoretical_loss": 3.309975881526749,
      "tokens_seen": 3130261504
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2444444444444445e-05,
      "loss": 1.1761,
      "theoretical_loss": 3.3099705993412125,
      "tokens_seen": 3130327040
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.242424242424242e-05,
      "loss": 1.219,
      "theoretical_loss": 3.309965317297226,
      "tokens_seen": 3130392576
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2404040404040406e-05,
      "loss": 1.5357,
      "theoretical_loss": 3.3099600353947816,
      "tokens_seen": 3130458112
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.238383838383839e-05,
      "loss": 1.2219,
      "theoretical_loss": 3.309954753633873,
      "tokens_seen": 3130523648
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.236363636363636e-05,
      "loss": 1.2418,
      "theoretical_loss": 3.3099494720144937,
      "tokens_seen": 3130589184
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2343434343434344e-05,
      "loss": 1.2373,
      "theoretical_loss": 3.309944190536637,
      "tokens_seen": 3130654720
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.232323232323232e-05,
      "loss": 1.5142,
      "theoretical_loss": 3.3099389092002958,
      "tokens_seen": 3130720256
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2303030303030305e-05,
      "loss": 1.4546,
      "theoretical_loss": 3.309933628005463,
      "tokens_seen": 3130785792
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.228282828282829e-05,
      "loss": 1.2087,
      "theoretical_loss": 3.309928346952133,
      "tokens_seen": 3130851328
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.226262626262626e-05,
      "loss": 1.3147,
      "theoretical_loss": 3.309923066040298,
      "tokens_seen": 3130916864
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2242424242424243e-05,
      "loss": 1.3925,
      "theoretical_loss": 3.3099177852699517,
      "tokens_seen": 3130982400
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.222222222222223e-05,
      "loss": 1.2409,
      "theoretical_loss": 3.3099125046410873,
      "tokens_seen": 3131047936
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2202020202020205e-05,
      "loss": 1.284,
      "theoretical_loss": 3.309907224153698,
      "tokens_seen": 3131113472
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.218181818181818e-05,
      "loss": 1.2427,
      "theoretical_loss": 3.309901943807777,
      "tokens_seen": 3131179008
    },
    {
      "epoch": 0.9,
      "objective/train/aligned_token_count": 0.4545454680919647,
      "objective/train/avg_token_score": 0.006941567175090313,
      "objective/train/docs_used": 1760143,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.5438612699508667,
      "objective/train/original_loss": 4.240050792694092,
      "objective/train/theoretical_loss": 3.3098966636033174,
      "objective/train/tokens_used": 1490103776,
      "objective/train/unlikelihood_loss": 0.060233987867832184,
      "theoretical_loss": 3.3098966636033174,
      "tokens_seen": 3131244544
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.216161616161616e-05,
      "loss": 1.3611,
      "theoretical_loss": 3.3098966636033174,
      "tokens_seen": 3131244544
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.214141414141414e-05,
      "loss": 1.0626,
      "theoretical_loss": 3.3098913835403128,
      "tokens_seen": 3131310080
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.212121212121213e-05,
      "loss": 1.3628,
      "theoretical_loss": 3.309886103618757,
      "tokens_seen": 3131375616
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.21010101010101e-05,
      "loss": 1.4059,
      "theoretical_loss": 3.3098808238386415,
      "tokens_seen": 3131441152
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.208080808080808e-05,
      "loss": 1.0782,
      "theoretical_loss": 3.309875544199961,
      "tokens_seen": 3131506688
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.206060606060606e-05,
      "loss": 1.2602,
      "theoretical_loss": 3.3098702647027083,
      "tokens_seen": 3131572224
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.204040404040404e-05,
      "loss": 1.196,
      "theoretical_loss": 3.3098649853468767,
      "tokens_seen": 3131637760
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2020202020202026e-05,
      "loss": 1.4812,
      "theoretical_loss": 3.3098597061324595,
      "tokens_seen": 3131703296
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.2e-05,
      "loss": 1.2278,
      "theoretical_loss": 3.3098544270594497,
      "tokens_seen": 3131768832
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.197979797979798e-05,
      "loss": 1.0198,
      "theoretical_loss": 3.3098491481278414,
      "tokens_seen": 3131834368
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1959595959595965e-05,
      "loss": 1.1706,
      "theoretical_loss": 3.3098438693376266,
      "tokens_seen": 3131899904
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.193939393939394e-05,
      "loss": 1.1906,
      "theoretical_loss": 3.3098385906887993,
      "tokens_seen": 3131965440
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.191919191919192e-05,
      "loss": 1.2985,
      "theoretical_loss": 3.3098333121813526,
      "tokens_seen": 3132030976
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1898989898989896e-05,
      "loss": 1.1692,
      "theoretical_loss": 3.30982803381528,
      "tokens_seen": 3132096512
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.187878787878788e-05,
      "loss": 1.3385,
      "theoretical_loss": 3.3098227555905746,
      "tokens_seen": 3132162048
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1858585858585864e-05,
      "loss": 1.5116,
      "theoretical_loss": 3.309817477507229,
      "tokens_seen": 3132227584
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1838383838383835e-05,
      "loss": 1.3424,
      "theoretical_loss": 3.3098121995652376,
      "tokens_seen": 3132293120
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.181818181818182e-05,
      "loss": 1.4025,
      "theoretical_loss": 3.3098069217645927,
      "tokens_seen": 3132358656
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.17979797979798e-05,
      "loss": 1.2905,
      "theoretical_loss": 3.3098016441052884,
      "tokens_seen": 3132424192
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.177777777777778e-05,
      "loss": 1.6446,
      "theoretical_loss": 3.3097963665873174,
      "tokens_seen": 3132489728
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.175757575757576e-05,
      "loss": 1.4794,
      "theoretical_loss": 3.3097910892106728,
      "tokens_seen": 3132555264
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1737373737373734e-05,
      "loss": 1.5895,
      "theoretical_loss": 3.3097858119753485,
      "tokens_seen": 3132620800
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.171717171717172e-05,
      "loss": 1.2607,
      "theoretical_loss": 3.309780534881337,
      "tokens_seen": 3132686336
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.16969696969697e-05,
      "loss": 1.4723,
      "theoretical_loss": 3.3097752579286324,
      "tokens_seen": 3132751872
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.167676767676768e-05,
      "loss": 1.3672,
      "theoretical_loss": 3.309769981117227,
      "tokens_seen": 3132817408
    },
    {
      "epoch": 0.9,
      "objective/train/aligned_token_count": 0.33449414372444153,
      "objective/train/avg_token_score": 0.013704377226531506,
      "objective/train/docs_used": 1761098,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0468885898590088,
      "objective/train/original_loss": 4.555112361907959,
      "objective/train/theoretical_loss": 3.3097647044471152,
      "objective/train/tokens_used": 1491742176,
      "objective/train/unlikelihood_loss": 0.0930151641368866,
      "theoretical_loss": 3.3097647044471152,
      "tokens_seen": 3132882944
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1656565656565656e-05,
      "loss": 1.1902,
      "theoretical_loss": 3.3097647044471152,
      "tokens_seen": 3132882944
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1636363636363634e-05,
      "loss": 1.4351,
      "theoretical_loss": 3.309759427918289,
      "tokens_seen": 3132948480
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.161616161616162e-05,
      "loss": 1.2899,
      "theoretical_loss": 3.309754151530743,
      "tokens_seen": 3133014016
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.15959595959596e-05,
      "loss": 1.5343,
      "theoretical_loss": 3.3097488752844693,
      "tokens_seen": 3133079552
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.157575757575757e-05,
      "loss": 1.1196,
      "theoretical_loss": 3.3097435991794617,
      "tokens_seen": 3133145088
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1555555555555556e-05,
      "loss": 1.1199,
      "theoretical_loss": 3.309738323215713,
      "tokens_seen": 3133210624
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.153535353535354e-05,
      "loss": 1.2333,
      "theoretical_loss": 3.3097330473932174,
      "tokens_seen": 3133276160
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.151515151515152e-05,
      "loss": 1.3023,
      "theoretical_loss": 3.3097277717119677,
      "tokens_seen": 3133341696
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1494949494949494e-05,
      "loss": 1.5387,
      "theoretical_loss": 3.3097224961719567,
      "tokens_seen": 3133407232
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.147474747474747e-05,
      "loss": 1.269,
      "theoretical_loss": 3.309717220773178,
      "tokens_seen": 3133472768
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1454545454545455e-05,
      "loss": 1.1692,
      "theoretical_loss": 3.309711945515625,
      "tokens_seen": 3133538304
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.143434343434344e-05,
      "loss": 1.3451,
      "theoretical_loss": 3.309706670399291,
      "tokens_seen": 3133603840
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1414141414141416e-05,
      "loss": 1.4027,
      "theoretical_loss": 3.3097013954241694,
      "tokens_seen": 3133669376
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1393939393939394e-05,
      "loss": 1.4904,
      "theoretical_loss": 3.3096961205902526,
      "tokens_seen": 3133734912
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.137373737373738e-05,
      "loss": 1.3688,
      "theoretical_loss": 3.309690845897535,
      "tokens_seen": 3133800448
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1353535353535355e-05,
      "loss": 1.204,
      "theoretical_loss": 3.309685571346009,
      "tokens_seen": 3133865984
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.133333333333334e-05,
      "loss": 1.4447,
      "theoretical_loss": 3.3096802969356682,
      "tokens_seen": 3133931520
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.131313131313131e-05,
      "loss": 1.281,
      "theoretical_loss": 3.3096750226665064,
      "tokens_seen": 3133997056
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.129292929292929e-05,
      "loss": 1.2511,
      "theoretical_loss": 3.3096697485385156,
      "tokens_seen": 3134062592
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.127272727272728e-05,
      "loss": 1.3567,
      "theoretical_loss": 3.30966447455169,
      "tokens_seen": 3134128128
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1252525252525254e-05,
      "loss": 1.5466,
      "theoretical_loss": 3.3096592007060233,
      "tokens_seen": 3134193664
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.123232323232323e-05,
      "loss": 1.1301,
      "theoretical_loss": 3.3096539270015075,
      "tokens_seen": 3134259200
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.121212121212121e-05,
      "loss": 1.5365,
      "theoretical_loss": 3.3096486534381366,
      "tokens_seen": 3134324736
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.119191919191919e-05,
      "loss": 1.2545,
      "theoretical_loss": 3.3096433800159044,
      "tokens_seen": 3134390272
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1171717171717177e-05,
      "loss": 1.1603,
      "theoretical_loss": 3.309638106734803,
      "tokens_seen": 3134455808
    },
    {
      "epoch": 0.9,
      "objective/train/aligned_token_count": 0.35184505581855774,
      "objective/train/avg_token_score": 0.02992575615644455,
      "objective/train/docs_used": 1761812,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0966085195541382,
      "objective/train/original_loss": 4.614572525024414,
      "objective/train/theoretical_loss": 3.309632833594826,
      "objective/train/tokens_used": 1493380576,
      "objective/train/unlikelihood_loss": 0.0913081243634224,
      "theoretical_loss": 3.309632833594826,
      "tokens_seen": 3134521344
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.115151515151515e-05,
      "loss": 1.4011,
      "theoretical_loss": 3.309632833594826,
      "tokens_seen": 3134521344
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.113131313131313e-05,
      "loss": 1.4124,
      "theoretical_loss": 3.3096275605959677,
      "tokens_seen": 3134586880
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1111111111111115e-05,
      "loss": 1.4246,
      "theoretical_loss": 3.30962228773822,
      "tokens_seen": 3134652416
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.109090909090909e-05,
      "loss": 1.5288,
      "theoretical_loss": 3.309617015021577,
      "tokens_seen": 3134717952
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1070707070707076e-05,
      "loss": 1.2626,
      "theoretical_loss": 3.3096117424460316,
      "tokens_seen": 3134783488
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1050505050505046e-05,
      "loss": 1.3836,
      "theoretical_loss": 3.3096064700115773,
      "tokens_seen": 3134849024
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.103030303030303e-05,
      "loss": 1.2801,
      "theoretical_loss": 3.3096011977182074,
      "tokens_seen": 3134914560
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.1010101010101014e-05,
      "loss": 1.1801,
      "theoretical_loss": 3.309595925565915,
      "tokens_seen": 3134980096
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.098989898989899e-05,
      "loss": 1.4501,
      "theoretical_loss": 3.309590653554693,
      "tokens_seen": 3135045632
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.096969696969697e-05,
      "loss": 1.388,
      "theoretical_loss": 3.3095853816845358,
      "tokens_seen": 3135111168
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.094949494949495e-05,
      "loss": 1.1735,
      "theoretical_loss": 3.309580109955436,
      "tokens_seen": 3135176704
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.092929292929293e-05,
      "loss": 1.3307,
      "theoretical_loss": 3.309574838367386,
      "tokens_seen": 3135242240
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0909090909090914e-05,
      "loss": 1.5011,
      "theoretical_loss": 3.3095695669203806,
      "tokens_seen": 3135307776
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0888888888888884e-05,
      "loss": 1.4223,
      "theoretical_loss": 3.3095642956144125,
      "tokens_seen": 3135373312
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.086868686868687e-05,
      "loss": 1.3046,
      "theoretical_loss": 3.3095590244494746,
      "tokens_seen": 3135438848
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.084848484848485e-05,
      "loss": 1.1874,
      "theoretical_loss": 3.309553753425561,
      "tokens_seen": 3135504384
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.082828282828283e-05,
      "loss": 1.7106,
      "theoretical_loss": 3.309548482542664,
      "tokens_seen": 3135569920
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.080808080808081e-05,
      "loss": 1.1269,
      "theoretical_loss": 3.3095432118007775,
      "tokens_seen": 3135635456
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0787878787878784e-05,
      "loss": 1.3608,
      "theoretical_loss": 3.309537941199894,
      "tokens_seen": 3135700992
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.076767676767677e-05,
      "loss": 1.439,
      "theoretical_loss": 3.3095326707400083,
      "tokens_seen": 3135766528
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.074747474747475e-05,
      "loss": 1.348,
      "theoretical_loss": 3.3095274004211124,
      "tokens_seen": 3135832064
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.072727272727273e-05,
      "loss": 1.524,
      "theoretical_loss": 3.3095221302432,
      "tokens_seen": 3135897600
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0707070707070706e-05,
      "loss": 1.3559,
      "theoretical_loss": 3.3095168602062643,
      "tokens_seen": 3135963136
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.068686868686869e-05,
      "loss": 1.2652,
      "theoretical_loss": 3.3095115903102985,
      "tokens_seen": 3136028672
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.066666666666667e-05,
      "loss": 1.2676,
      "theoretical_loss": 3.3095063205552964,
      "tokens_seen": 3136094208
    },
    {
      "debugging/Self-BLEU-5": 0.47506560066232434,
      "debugging/distinct-1-grams": 0.7038168750705822,
      "debugging/distinct-2-grams": 0.9567806381187414,
      "debugging/entropy-1-grams": 5.464426542413726,
      "debugging/entropy-2-grams": 6.431994342092753,
      "debugging/length": 564.0,
      "debugging/num_segments": 8,
      "epoch": 0.9,
      "objective/train/aligned_token_count": 0.42967987060546875,
      "objective/train/avg_token_score": 0.015942009165883064,
      "objective/train/docs_used": 1763087,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3303141593933105,
      "objective/train/original_loss": 4.268259048461914,
      "objective/train/theoretical_loss": 3.309501050941251,
      "objective/train/tokens_used": 1495018976,
      "objective/train/unlikelihood_loss": 0.06023964658379555,
      "theoretical_loss": 3.309501050941251,
      "tokens_seen": 3136159744
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.064646464646465e-05,
      "loss": 1.4975,
      "theoretical_loss": 3.309501050941251,
      "tokens_seen": 3136159744
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.062626262626262e-05,
      "loss": 1.3527,
      "theoretical_loss": 3.309495781468155,
      "tokens_seen": 3136225280
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0606060606060606e-05,
      "loss": 1.3235,
      "theoretical_loss": 3.3094905121360023,
      "tokens_seen": 3136290816
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.058585858585859e-05,
      "loss": 1.2674,
      "theoretical_loss": 3.309485242944786,
      "tokens_seen": 3136356352
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.056565656565657e-05,
      "loss": 1.1589,
      "theoretical_loss": 3.3094799738944998,
      "tokens_seen": 3136421888
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0545454545454544e-05,
      "loss": 1.4316,
      "theoretical_loss": 3.3094747049851363,
      "tokens_seen": 3136487424
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.052525252525253e-05,
      "loss": 1.2952,
      "theoretical_loss": 3.309469436216689,
      "tokens_seen": 3136552960
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0505050505050505e-05,
      "loss": 1.3994,
      "theoretical_loss": 3.3094641675891516,
      "tokens_seen": 3136618496
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.048484848484849e-05,
      "loss": 1.4528,
      "theoretical_loss": 3.309458899102517,
      "tokens_seen": 3136684032
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0464646464646466e-05,
      "loss": 1.1653,
      "theoretical_loss": 3.309453630756779,
      "tokens_seen": 3136749568
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.044444444444444e-05,
      "loss": 1.0603,
      "theoretical_loss": 3.3094483625519295,
      "tokens_seen": 3136815104
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.042424242424243e-05,
      "loss": 1.4922,
      "theoretical_loss": 3.309443094487963,
      "tokens_seen": 3136880640
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0404040404040404e-05,
      "loss": 1.1698,
      "theoretical_loss": 3.3094378265648734,
      "tokens_seen": 3136946176
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.038383838383839e-05,
      "loss": 1.409,
      "theoretical_loss": 3.3094325587826523,
      "tokens_seen": 3137011712
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.036363636363636e-05,
      "loss": 1.242,
      "theoretical_loss": 3.309427291141294,
      "tokens_seen": 3137077248
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.034343434343434e-05,
      "loss": 1.5453,
      "theoretical_loss": 3.3094220236407916,
      "tokens_seen": 3137142784
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.032323232323233e-05,
      "loss": 1.3947,
      "theoretical_loss": 3.3094167562811387,
      "tokens_seen": 3137208320
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0303030303030304e-05,
      "loss": 1.2354,
      "theoretical_loss": 3.309411489062328,
      "tokens_seen": 3137273856
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.028282828282828e-05,
      "loss": 1.0447,
      "theoretical_loss": 3.309406221984353,
      "tokens_seen": 3137339392
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0262626262626265e-05,
      "loss": 1.2999,
      "theoretical_loss": 3.309400955047207,
      "tokens_seen": 3137404928
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.024242424242424e-05,
      "loss": 1.4419,
      "theoretical_loss": 3.309395688250884,
      "tokens_seen": 3137470464
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0222222222222226e-05,
      "loss": 1.4156,
      "theoretical_loss": 3.309390421595376,
      "tokens_seen": 3137536000
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0202020202020203e-05,
      "loss": 1.2807,
      "theoretical_loss": 3.3093851550806774,
      "tokens_seen": 3137601536
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.018181818181818e-05,
      "loss": 1.4585,
      "theoretical_loss": 3.309379888706781,
      "tokens_seen": 3137667072
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0161616161616165e-05,
      "loss": 1.2158,
      "theoretical_loss": 3.3093746224736798,
      "tokens_seen": 3137732608
    },
    {
      "epoch": 0.9,
      "objective/train/aligned_token_count": 0.4782502353191376,
      "objective/train/avg_token_score": 0.03576279059052467,
      "objective/train/docs_used": 1763593,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4824802875518799,
      "objective/train/original_loss": 4.065350532531738,
      "objective/train/theoretical_loss": 3.3093693563813678,
      "objective/train/tokens_used": 1496657376,
      "objective/train/unlikelihood_loss": 0.06148296594619751,
      "theoretical_loss": 3.3093693563813678,
      "tokens_seen": 3137798144
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.014141414141414e-05,
      "loss": 1.4343,
      "theoretical_loss": 3.3093693563813678,
      "tokens_seen": 3137798144
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0121212121212126e-05,
      "loss": 1.2788,
      "theoretical_loss": 3.309364090429838,
      "tokens_seen": 3137863680
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.01010101010101e-05,
      "loss": 1.5339,
      "theoretical_loss": 3.3093588246190833,
      "tokens_seen": 3137929216
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.008080808080808e-05,
      "loss": 1.4281,
      "theoretical_loss": 3.3093535589490974,
      "tokens_seen": 3137994752
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.0060606060606064e-05,
      "loss": 1.278,
      "theoretical_loss": 3.3093482934198737,
      "tokens_seen": 3138060288
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.004040404040404e-05,
      "loss": 1.2475,
      "theoretical_loss": 3.3093430280314053,
      "tokens_seen": 3138125824
    },
    {
      "epoch": 0.9,
      "learning_rate": 5.002020202020202e-05,
      "loss": 1.226,
      "theoretical_loss": 3.3093377627836853,
      "tokens_seen": 3138191360
    },
    {
      "epoch": 0.9,
      "learning_rate": 5e-05,
      "loss": 1.3555,
      "theoretical_loss": 3.3093324976767073,
      "tokens_seen": 3138256896
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.997979797979798e-05,
      "loss": 1.4261,
      "theoretical_loss": 3.309327232710465,
      "tokens_seen": 3138322432
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9959595959595964e-05,
      "loss": 1.4955,
      "theoretical_loss": 3.3093219678849506,
      "tokens_seen": 3138387968
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9939393939393934e-05,
      "loss": 1.1746,
      "theoretical_loss": 3.3093167032001585,
      "tokens_seen": 3138453504
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.991919191919192e-05,
      "loss": 1.4459,
      "theoretical_loss": 3.3093114386560814,
      "tokens_seen": 3138519040
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.98989898989899e-05,
      "loss": 1.4388,
      "theoretical_loss": 3.309306174252712,
      "tokens_seen": 3138584576
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.987878787878788e-05,
      "loss": 1.2979,
      "theoretical_loss": 3.3093009099900454,
      "tokens_seen": 3138650112
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.985858585858586e-05,
      "loss": 1.4679,
      "theoretical_loss": 3.3092956458680733,
      "tokens_seen": 3138715648
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.983838383838384e-05,
      "loss": 1.5085,
      "theoretical_loss": 3.3092903818867896,
      "tokens_seen": 3138781184
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.981818181818182e-05,
      "loss": 1.2589,
      "theoretical_loss": 3.3092851180461875,
      "tokens_seen": 3138846720
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.97979797979798e-05,
      "loss": 1.4163,
      "theoretical_loss": 3.3092798543462605,
      "tokens_seen": 3138912256
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.977777777777778e-05,
      "loss": 1.3653,
      "theoretical_loss": 3.3092745907870014,
      "tokens_seen": 3138977792
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9757575757575756e-05,
      "loss": 1.4634,
      "theoretical_loss": 3.309269327368404,
      "tokens_seen": 3139043328
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.973737373737374e-05,
      "loss": 1.3405,
      "theoretical_loss": 3.3092640640904616,
      "tokens_seen": 3139108864
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.971717171717172e-05,
      "loss": 1.5039,
      "theoretical_loss": 3.309258800953167,
      "tokens_seen": 3139174400
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.96969696969697e-05,
      "loss": 1.3958,
      "theoretical_loss": 3.309253537956514,
      "tokens_seen": 3139239936
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.967676767676767e-05,
      "loss": 1.1133,
      "theoretical_loss": 3.309248275100496,
      "tokens_seen": 3139305472
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9656565656565655e-05,
      "loss": 1.2968,
      "theoretical_loss": 3.3092430123851058,
      "tokens_seen": 3139371008
    },
    {
      "epoch": 0.9,
      "objective/train/aligned_token_count": 0.5359237790107727,
      "objective/train/avg_token_score": 0.012179692275822163,
      "objective/train/docs_used": 1765019,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.724595308303833,
      "objective/train/original_loss": 4.283945083618164,
      "objective/train/theoretical_loss": 3.309237749810337,
      "objective/train/tokens_used": 1498295776,
      "objective/train/unlikelihood_loss": 0.048299092799425125,
      "theoretical_loss": 3.309237749810337,
      "tokens_seen": 3139436544
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.963636363636364e-05,
      "loss": 1.4353,
      "theoretical_loss": 3.309237749810337,
      "tokens_seen": 3139436544
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9616161616161616e-05,
      "loss": 1.1332,
      "theoretical_loss": 3.309232487376183,
      "tokens_seen": 3139502080
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.95959595959596e-05,
      "loss": 1.4117,
      "theoretical_loss": 3.309227225082637,
      "tokens_seen": 3139567616
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.957575757575758e-05,
      "loss": 1.2903,
      "theoretical_loss": 3.3092219629296924,
      "tokens_seen": 3139633152
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9555555555555555e-05,
      "loss": 1.3153,
      "theoretical_loss": 3.309216700917342,
      "tokens_seen": 3139698688
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.953535353535354e-05,
      "loss": 1.2189,
      "theoretical_loss": 3.30921143904558,
      "tokens_seen": 3139764224
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9515151515151516e-05,
      "loss": 1.485,
      "theoretical_loss": 3.309206177314399,
      "tokens_seen": 3139829760
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.949494949494949e-05,
      "loss": 1.5536,
      "theoretical_loss": 3.3092009157237925,
      "tokens_seen": 3139895296
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.947474747474748e-05,
      "loss": 1.1819,
      "theoretical_loss": 3.3091956542737537,
      "tokens_seen": 3139960832
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9454545454545454e-05,
      "loss": 1.4212,
      "theoretical_loss": 3.3091903929642763,
      "tokens_seen": 3140026368
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.943434343434344e-05,
      "loss": 1.2523,
      "theoretical_loss": 3.309185131795353,
      "tokens_seen": 3140091904
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9414141414141415e-05,
      "loss": 1.2987,
      "theoretical_loss": 3.309179870766978,
      "tokens_seen": 3140157440
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.939393939393939e-05,
      "loss": 1.4603,
      "theoretical_loss": 3.309174609879144,
      "tokens_seen": 3140222976
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9373737373737376e-05,
      "loss": 1.2354,
      "theoretical_loss": 3.309169349131844,
      "tokens_seen": 3140288512
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9353535353535354e-05,
      "loss": 1.396,
      "theoretical_loss": 3.309164088525072,
      "tokens_seen": 3140354048
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.933333333333334e-05,
      "loss": 1.3461,
      "theoretical_loss": 3.309158828058821,
      "tokens_seen": 3140419584
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9313131313131315e-05,
      "loss": 1.5078,
      "theoretical_loss": 3.3091535677330848,
      "tokens_seen": 3140485120
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.929292929292929e-05,
      "loss": 1.2455,
      "theoretical_loss": 3.3091483075478556,
      "tokens_seen": 3140550656
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9272727272727276e-05,
      "loss": 1.2507,
      "theoretical_loss": 3.3091430475031274,
      "tokens_seen": 3140616192
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.925252525252525e-05,
      "loss": 1.5041,
      "theoretical_loss": 3.309137787598894,
      "tokens_seen": 3140681728
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.923232323232323e-05,
      "loss": 1.1203,
      "theoretical_loss": 3.3091325278351476,
      "tokens_seen": 3140747264
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9212121212121214e-05,
      "loss": 1.4412,
      "theoretical_loss": 3.3091272682118826,
      "tokens_seen": 3140812800
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.919191919191919e-05,
      "loss": 1.3917,
      "theoretical_loss": 3.3091220087290916,
      "tokens_seen": 3140878336
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9171717171717175e-05,
      "loss": 1.2201,
      "theoretical_loss": 3.309116749386768,
      "tokens_seen": 3140943872
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.915151515151515e-05,
      "loss": 1.6356,
      "theoretical_loss": 3.3091114901849057,
      "tokens_seen": 3141009408
    },
    {
      "epoch": 0.9,
      "objective/train/aligned_token_count": 0.4044477045536041,
      "objective/train/avg_token_score": 0.008539588190615177,
      "objective/train/docs_used": 1765173,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.0937812328338623,
      "objective/train/original_loss": 4.380913257598877,
      "objective/train/theoretical_loss": 3.309106231123497,
      "objective/train/tokens_used": 1499934176,
      "objective/train/unlikelihood_loss": 0.05640077963471413,
      "theoretical_loss": 3.309106231123497,
      "tokens_seen": 3141074944
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.913131313131313e-05,
      "loss": 1.4926,
      "theoretical_loss": 3.309106231123497,
      "tokens_seen": 3141074944
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.9111111111111114e-05,
      "loss": 1.3253,
      "theoretical_loss": 3.3091009722025366,
      "tokens_seen": 3141140480
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.909090909090909e-05,
      "loss": 1.353,
      "theoretical_loss": 3.3090957134220167,
      "tokens_seen": 3141206016
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.907070707070707e-05,
      "loss": 1.4995,
      "theoretical_loss": 3.3090904547819306,
      "tokens_seen": 3141271552
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.905050505050505e-05,
      "loss": 1.0941,
      "theoretical_loss": 3.3090851962822723,
      "tokens_seen": 3141337088
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.903030303030303e-05,
      "loss": 1.4312,
      "theoretical_loss": 3.309079937923035,
      "tokens_seen": 3141402624
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.901010101010101e-05,
      "loss": 1.446,
      "theoretical_loss": 3.309074679704211,
      "tokens_seen": 3141468160
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8989898989899e-05,
      "loss": 1.3619,
      "theoretical_loss": 3.309069421625795,
      "tokens_seen": 3141533696
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.896969696969697e-05,
      "loss": 1.3197,
      "theoretical_loss": 3.30906416368778,
      "tokens_seen": 3141599232
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.894949494949495e-05,
      "loss": 1.2589,
      "theoretical_loss": 3.3090589058901587,
      "tokens_seen": 3141664768
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.892929292929293e-05,
      "loss": 1.3407,
      "theoretical_loss": 3.3090536482329247,
      "tokens_seen": 3141730304
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.890909090909091e-05,
      "loss": 1.2621,
      "theoretical_loss": 3.3090483907160717,
      "tokens_seen": 3141795840
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.888888888888889e-05,
      "loss": 1.3179,
      "theoretical_loss": 3.3090431333395927,
      "tokens_seen": 3141861376
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.886868686868687e-05,
      "loss": 1.4401,
      "theoretical_loss": 3.309037876103481,
      "tokens_seen": 3141926912
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.884848484848485e-05,
      "loss": 1.2562,
      "theoretical_loss": 3.30903261900773,
      "tokens_seen": 3141992448
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.882828282828283e-05,
      "loss": 1.3787,
      "theoretical_loss": 3.309027362052333,
      "tokens_seen": 3142057984
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8808080808080805e-05,
      "loss": 1.5056,
      "theoretical_loss": 3.3090221052372835,
      "tokens_seen": 3142123520
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.878787878787879e-05,
      "loss": 1.2381,
      "theoretical_loss": 3.3090168485625746,
      "tokens_seen": 3142189056
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8767676767676767e-05,
      "loss": 1.3998,
      "theoretical_loss": 3.3090115920281997,
      "tokens_seen": 3142254592
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.874747474747475e-05,
      "loss": 1.4112,
      "theoretical_loss": 3.3090063356341517,
      "tokens_seen": 3142320128
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8727272727272734e-05,
      "loss": 1.2542,
      "theoretical_loss": 3.309001079380425,
      "tokens_seen": 3142385664
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8707070707070705e-05,
      "loss": 1.3865,
      "theoretical_loss": 3.3089958232670122,
      "tokens_seen": 3142451200
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.868686868686869e-05,
      "loss": 1.3508,
      "theoretical_loss": 3.308990567293906,
      "tokens_seen": 3142516736
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8666666666666666e-05,
      "loss": 1.5007,
      "theoretical_loss": 3.308985311461101,
      "tokens_seen": 3142582272
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.864646464646465e-05,
      "loss": 1.4429,
      "theoretical_loss": 3.3089800557685902,
      "tokens_seen": 3142647808
    },
    {
      "epoch": 0.9,
      "objective/train/aligned_token_count": 0.42821359634399414,
      "objective/train/avg_token_score": 0.005702037364244461,
      "objective/train/docs_used": 1765556,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3094305992126465,
      "objective/train/original_loss": 4.545772552490234,
      "objective/train/theoretical_loss": 3.308974800216366,
      "objective/train/tokens_used": 1501572576,
      "objective/train/unlikelihood_loss": 0.061255261301994324,
      "theoretical_loss": 3.308974800216366,
      "tokens_seen": 3142713344
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.862626262626263e-05,
      "loss": 1.1727,
      "theoretical_loss": 3.308974800216366,
      "tokens_seen": 3142713344
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8606060606060604e-05,
      "loss": 1.3701,
      "theoretical_loss": 3.308969544804423,
      "tokens_seen": 3142778880
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.858585858585859e-05,
      "loss": 1.5203,
      "theoretical_loss": 3.308964289532754,
      "tokens_seen": 3142844416
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.856565656565657e-05,
      "loss": 1.4016,
      "theoretical_loss": 3.308959034401352,
      "tokens_seen": 3142909952
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.854545454545454e-05,
      "loss": 1.401,
      "theoretical_loss": 3.308953779410211,
      "tokens_seen": 3142975488
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.852525252525253e-05,
      "loss": 1.2136,
      "theoretical_loss": 3.308948524559323,
      "tokens_seen": 3143041024
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8505050505050504e-05,
      "loss": 1.3741,
      "theoretical_loss": 3.308943269848683,
      "tokens_seen": 3143106560
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.848484848484849e-05,
      "loss": 1.5369,
      "theoretical_loss": 3.308938015278284,
      "tokens_seen": 3143172096
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8464646464646465e-05,
      "loss": 1.3366,
      "theoretical_loss": 3.3089327608481183,
      "tokens_seen": 3143237632
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.844444444444444e-05,
      "loss": 1.2605,
      "theoretical_loss": 3.30892750655818,
      "tokens_seen": 3143303168
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8424242424242426e-05,
      "loss": 1.2428,
      "theoretical_loss": 3.3089222524084625,
      "tokens_seen": 3143368704
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.84040404040404e-05,
      "loss": 1.4488,
      "theoretical_loss": 3.3089169983989586,
      "tokens_seen": 3143434240
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.838383838383839e-05,
      "loss": 1.3491,
      "theoretical_loss": 3.308911744529662,
      "tokens_seen": 3143499776
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8363636363636364e-05,
      "loss": 1.1899,
      "theoretical_loss": 3.3089064908005663,
      "tokens_seen": 3143565312
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.834343434343434e-05,
      "loss": 1.3442,
      "theoretical_loss": 3.3089012372116646,
      "tokens_seen": 3143630848
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8323232323232326e-05,
      "loss": 1.4795,
      "theoretical_loss": 3.3088959837629504,
      "tokens_seen": 3143696384
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.830303030303031e-05,
      "loss": 1.2622,
      "theoretical_loss": 3.308890730454416,
      "tokens_seen": 3143761920
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.828282828282828e-05,
      "loss": 1.4604,
      "theoretical_loss": 3.3088854772860565,
      "tokens_seen": 3143827456
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8262626262626264e-05,
      "loss": 1.3068,
      "theoretical_loss": 3.3088802242578637,
      "tokens_seen": 3143892992
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.824242424242424e-05,
      "loss": 1.4981,
      "theoretical_loss": 3.3088749713698316,
      "tokens_seen": 3143958528
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8222222222222225e-05,
      "loss": 1.4569,
      "theoretical_loss": 3.3088697186219536,
      "tokens_seen": 3144024064
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.82020202020202e-05,
      "loss": 1.3542,
      "theoretical_loss": 3.308864466014223,
      "tokens_seen": 3144089600
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.818181818181818e-05,
      "loss": 1.4341,
      "theoretical_loss": 3.3088592135466333,
      "tokens_seen": 3144155136
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.8161616161616163e-05,
      "loss": 1.1075,
      "theoretical_loss": 3.308853961219177,
      "tokens_seen": 3144220672
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.814141414141415e-05,
      "loss": 1.47,
      "theoretical_loss": 3.3088487090318486,
      "tokens_seen": 3144286208
    },
    {
      "epoch": 0.9,
      "objective/train/aligned_token_count": 0.32184749841690063,
      "objective/train/avg_token_score": 0.02000572718679905,
      "objective/train/docs_used": 1766057,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 0.8459650278091431,
      "objective/train/original_loss": 4.973540782928467,
      "objective/train/theoretical_loss": 3.3088434569846403,
      "objective/train/tokens_used": 1503210976,
      "objective/train/unlikelihood_loss": 0.04051680490374565,
      "theoretical_loss": 3.3088434569846403,
      "tokens_seen": 3144351744
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.8121212121212125e-05,
      "loss": 1.1361,
      "theoretical_loss": 3.3088434569846403,
      "tokens_seen": 3144351744
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.81010101010101e-05,
      "loss": 1.2438,
      "theoretical_loss": 3.3088382050775467,
      "tokens_seen": 3144417280
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.808080808080808e-05,
      "loss": 1.4133,
      "theoretical_loss": 3.30883295331056,
      "tokens_seen": 3144482816
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.806060606060606e-05,
      "loss": 1.5184,
      "theoretical_loss": 3.308827701683674,
      "tokens_seen": 3144548352
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.804040404040405e-05,
      "loss": 1.4598,
      "theoretical_loss": 3.308822450196882,
      "tokens_seen": 3144613888
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.802020202020202e-05,
      "loss": 1.5382,
      "theoretical_loss": 3.3088171988501776,
      "tokens_seen": 3144679424
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.8e-05,
      "loss": 1.4134,
      "theoretical_loss": 3.3088119476435542,
      "tokens_seen": 3144744960
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.797979797979798e-05,
      "loss": 1.5026,
      "theoretical_loss": 3.3088066965770047,
      "tokens_seen": 3144810496
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.795959595959596e-05,
      "loss": 1.5256,
      "theoretical_loss": 3.308801445650522,
      "tokens_seen": 3144876032
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.793939393939394e-05,
      "loss": 1.4232,
      "theoretical_loss": 3.308796194864101,
      "tokens_seen": 3144941568
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.791919191919192e-05,
      "loss": 1.4393,
      "theoretical_loss": 3.3087909442177335,
      "tokens_seen": 3145007104
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.78989898989899e-05,
      "loss": 1.0341,
      "theoretical_loss": 3.3087856937114135,
      "tokens_seen": 3145072640
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7878787878787885e-05,
      "loss": 1.2614,
      "theoretical_loss": 3.3087804433451344,
      "tokens_seen": 3145138176
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7858585858585855e-05,
      "loss": 1.4299,
      "theoretical_loss": 3.3087751931188896,
      "tokens_seen": 3145203712
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.783838383838384e-05,
      "loss": 1.412,
      "theoretical_loss": 3.3087699430326722,
      "tokens_seen": 3145269248
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7818181818181816e-05,
      "loss": 1.3069,
      "theoretical_loss": 3.308764693086476,
      "tokens_seen": 3145334784
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.77979797979798e-05,
      "loss": 1.2166,
      "theoretical_loss": 3.3087594432802936,
      "tokens_seen": 3145400320
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7777777777777784e-05,
      "loss": 1.2679,
      "theoretical_loss": 3.308754193614119,
      "tokens_seen": 3145465856
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7757575757575755e-05,
      "loss": 1.579,
      "theoretical_loss": 3.3087489440879447,
      "tokens_seen": 3145531392
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.773737373737374e-05,
      "loss": 1.3858,
      "theoretical_loss": 3.308743694701765,
      "tokens_seen": 3145596928
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.771717171717172e-05,
      "loss": 1.2872,
      "theoretical_loss": 3.3087384454555733,
      "tokens_seen": 3145662464
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.76969696969697e-05,
      "loss": 1.2759,
      "theoretical_loss": 3.308733196349362,
      "tokens_seen": 3145728000
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.767676767676768e-05,
      "loss": 1.3516,
      "theoretical_loss": 3.308727947383125,
      "tokens_seen": 3145793536
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7656565656565654e-05,
      "loss": 1.6846,
      "theoretical_loss": 3.3087226985568563,
      "tokens_seen": 3145859072
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.763636363636364e-05,
      "loss": 1.2271,
      "theoretical_loss": 3.308717449870548,
      "tokens_seen": 3145924608
    },
    {
      "epoch": 0.91,
      "objective/train/aligned_token_count": 0.38202589750289917,
      "objective/train/avg_token_score": 0.02299250103533268,
      "objective/train/docs_used": 1767374,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.199781060218811,
      "objective/train/original_loss": 4.525591850280762,
      "objective/train/theoretical_loss": 3.3087122013241945,
      "objective/train/tokens_used": 1504849376,
      "objective/train/unlikelihood_loss": 0.052468765527009964,
      "theoretical_loss": 3.3087122013241945,
      "tokens_seen": 3145990144
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.761616161616162e-05,
      "loss": 1.4166,
      "theoretical_loss": 3.3087122013241945,
      "tokens_seen": 3145990144
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.759595959595959e-05,
      "loss": 1.3533,
      "theoretical_loss": 3.308706952917788,
      "tokens_seen": 3146055680
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7575757575757576e-05,
      "loss": 1.1279,
      "theoretical_loss": 3.308701704651323,
      "tokens_seen": 3146121216
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7555555555555553e-05,
      "loss": 1.3368,
      "theoretical_loss": 3.3086964565247925,
      "tokens_seen": 3146186752
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.753535353535354e-05,
      "loss": 1.1679,
      "theoretical_loss": 3.30869120853819,
      "tokens_seen": 3146252288
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.751515151515152e-05,
      "loss": 1.2913,
      "theoretical_loss": 3.308685960691508,
      "tokens_seen": 3146317824
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.749494949494949e-05,
      "loss": 1.2302,
      "theoretical_loss": 3.308680712984741,
      "tokens_seen": 3146383360
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7474747474747476e-05,
      "loss": 1.2408,
      "theoretical_loss": 3.3086754654178816,
      "tokens_seen": 3146448896
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.745454545454546e-05,
      "loss": 1.4547,
      "theoretical_loss": 3.3086702179909233,
      "tokens_seen": 3146514432
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.743434343434344e-05,
      "loss": 1.3016,
      "theoretical_loss": 3.30866497070386,
      "tokens_seen": 3146579968
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7414141414141414e-05,
      "loss": 1.4584,
      "theoretical_loss": 3.308659723556684,
      "tokens_seen": 3146645504
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.739393939393939e-05,
      "loss": 1.3477,
      "theoretical_loss": 3.3086544765493895,
      "tokens_seen": 3146711040
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7373737373737375e-05,
      "loss": 1.4715,
      "theoretical_loss": 3.3086492296819694,
      "tokens_seen": 3146776576
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.735353535353536e-05,
      "loss": 1.3515,
      "theoretical_loss": 3.3086439829544174,
      "tokens_seen": 3146842112
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.733333333333333e-05,
      "loss": 1.2634,
      "theoretical_loss": 3.308638736366727,
      "tokens_seen": 3146907648
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7313131313131314e-05,
      "loss": 1.3865,
      "theoretical_loss": 3.308633489918891,
      "tokens_seen": 3146973184
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.72929292929293e-05,
      "loss": 1.3574,
      "theoretical_loss": 3.3086282436109036,
      "tokens_seen": 3147038720
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7272727272727275e-05,
      "loss": 1.4619,
      "theoretical_loss": 3.308622997442757,
      "tokens_seen": 3147104256
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.725252525252525e-05,
      "loss": 1.4088,
      "theoretical_loss": 3.3086177514144453,
      "tokens_seen": 3147169792
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.723232323232323e-05,
      "loss": 1.3057,
      "theoretical_loss": 3.308612505525962,
      "tokens_seen": 3147235328
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.721212121212121e-05,
      "loss": 1.2943,
      "theoretical_loss": 3.3086072597773,
      "tokens_seen": 3147300864
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.71919191919192e-05,
      "loss": 1.298,
      "theoretical_loss": 3.3086020141684527,
      "tokens_seen": 3147366400
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7171717171717174e-05,
      "loss": 1.4688,
      "theoretical_loss": 3.308596768699414,
      "tokens_seen": 3147431936
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.715151515151515e-05,
      "loss": 1.2753,
      "theoretical_loss": 3.308591523370177,
      "tokens_seen": 3147497472
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.713131313131313e-05,
      "loss": 1.4005,
      "theoretical_loss": 3.3085862781807345,
      "tokens_seen": 3147563008
    },
    {
      "epoch": 0.91,
      "objective/train/aligned_token_count": 0.3709677457809448,
      "objective/train/avg_token_score": 0.06990663707256317,
      "objective/train/docs_used": 1768009,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.307303547859192,
      "objective/train/original_loss": 4.654049396514893,
      "objective/train/theoretical_loss": 3.3085810331310803,
      "objective/train/tokens_used": 1506487776,
      "objective/train/unlikelihood_loss": 0.05558298900723457,
      "theoretical_loss": 3.3085810331310803,
      "tokens_seen": 3147628544
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.711111111111111e-05,
      "loss": 1.25,
      "theoretical_loss": 3.3085810331310803,
      "tokens_seen": 3147628544
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7090909090909097e-05,
      "loss": 1.5792,
      "theoretical_loss": 3.3085757882212077,
      "tokens_seen": 3147694080
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.707070707070707e-05,
      "loss": 1.3947,
      "theoretical_loss": 3.3085705434511103,
      "tokens_seen": 3147759616
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.705050505050505e-05,
      "loss": 1.3249,
      "theoretical_loss": 3.3085652988207817,
      "tokens_seen": 3147825152
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.7030303030303035e-05,
      "loss": 1.6684,
      "theoretical_loss": 3.3085600543302145,
      "tokens_seen": 3147890688
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.701010101010101e-05,
      "loss": 1.4942,
      "theoretical_loss": 3.3085548099794027,
      "tokens_seen": 3147956224
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.698989898989899e-05,
      "loss": 1.3052,
      "theoretical_loss": 3.308549565768339,
      "tokens_seen": 3148021760
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6969696969696966e-05,
      "loss": 1.2401,
      "theoretical_loss": 3.308544321697017,
      "tokens_seen": 3148087296
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.694949494949495e-05,
      "loss": 1.403,
      "theoretical_loss": 3.308539077765431,
      "tokens_seen": 3148152832
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6929292929292934e-05,
      "loss": 1.3268,
      "theoretical_loss": 3.308533833973573,
      "tokens_seen": 3148218368
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.690909090909091e-05,
      "loss": 1.306,
      "theoretical_loss": 3.308528590321437,
      "tokens_seen": 3148283904
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.688888888888889e-05,
      "loss": 1.3529,
      "theoretical_loss": 3.3085233468090167,
      "tokens_seen": 3148349440
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.686868686868687e-05,
      "loss": 1.1302,
      "theoretical_loss": 3.308518103436305,
      "tokens_seen": 3148414976
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.684848484848485e-05,
      "loss": 1.6333,
      "theoretical_loss": 3.308512860203295,
      "tokens_seen": 3148480512
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6828282828282834e-05,
      "loss": 1.4902,
      "theoretical_loss": 3.308507617109981,
      "tokens_seen": 3148546048
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6808080808080804e-05,
      "loss": 1.1823,
      "theoretical_loss": 3.3085023741563555,
      "tokens_seen": 3148611584
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.678787878787879e-05,
      "loss": 1.2915,
      "theoretical_loss": 3.308497131342412,
      "tokens_seen": 3148677120
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.676767676767677e-05,
      "loss": 1.2533,
      "theoretical_loss": 3.3084918886681445,
      "tokens_seen": 3148742656
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.674747474747475e-05,
      "loss": 1.3274,
      "theoretical_loss": 3.308486646133546,
      "tokens_seen": 3148808192
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6727272727272727e-05,
      "loss": 1.3331,
      "theoretical_loss": 3.308481403738609,
      "tokens_seen": 3148873728
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6707070707070704e-05,
      "loss": 1.3375,
      "theoretical_loss": 3.3084761614833287,
      "tokens_seen": 3148939264
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.668686868686869e-05,
      "loss": 1.4273,
      "theoretical_loss": 3.308470919367697,
      "tokens_seen": 3149004800
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.666666666666667e-05,
      "loss": 1.2402,
      "theoretical_loss": 3.3084656773917076,
      "tokens_seen": 3149070336
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.664646464646464e-05,
      "loss": 1.507,
      "theoretical_loss": 3.308460435555354,
      "tokens_seen": 3149135872
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6626262626262626e-05,
      "loss": 1.3636,
      "theoretical_loss": 3.30845519385863,
      "tokens_seen": 3149201408
    },
    {
      "epoch": 0.91,
      "objective/train/aligned_token_count": 0.3946114480495453,
      "objective/train/avg_token_score": 0.0052132499404251575,
      "objective/train/docs_used": 1769388,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3408033847808838,
      "objective/train/original_loss": 4.484888553619385,
      "objective/train/theoretical_loss": 3.3084499523015283,
      "objective/train/tokens_used": 1508126176,
      "objective/train/unlikelihood_loss": 0.05900004133582115,
      "theoretical_loss": 3.3084499523015283,
      "tokens_seen": 3149266944
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.660606060606061e-05,
      "loss": 1.4305,
      "theoretical_loss": 3.3084499523015283,
      "tokens_seen": 3149266944
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.658585858585859e-05,
      "loss": 1.3581,
      "theoretical_loss": 3.3084447108840425,
      "tokens_seen": 3149332480
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.656565656565657e-05,
      "loss": 1.0201,
      "theoretical_loss": 3.308439469606166,
      "tokens_seen": 3149398016
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.654545454545454e-05,
      "loss": 1.2331,
      "theoretical_loss": 3.3084342284678923,
      "tokens_seen": 3149463552
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6525252525252525e-05,
      "loss": 1.3232,
      "theoretical_loss": 3.308428987469214,
      "tokens_seen": 3149529088
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.650505050505051e-05,
      "loss": 1.5824,
      "theoretical_loss": 3.308423746610126,
      "tokens_seen": 3149594624
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6484848484848487e-05,
      "loss": 1.3846,
      "theoretical_loss": 3.3084185058906206,
      "tokens_seen": 3149660160
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6464646464646464e-05,
      "loss": 1.3099,
      "theoretical_loss": 3.308413265310691,
      "tokens_seen": 3149725696
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.644444444444444e-05,
      "loss": 1.2688,
      "theoretical_loss": 3.3084080248703316,
      "tokens_seen": 3149791232
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6424242424242425e-05,
      "loss": 1.3533,
      "theoretical_loss": 3.3084027845695343,
      "tokens_seen": 3149856768
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.640404040404041e-05,
      "loss": 1.2925,
      "theoretical_loss": 3.308397544408294,
      "tokens_seen": 3149922304
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.638383838383838e-05,
      "loss": 1.5167,
      "theoretical_loss": 3.308392304386603,
      "tokens_seen": 3149987840
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.636363636363636e-05,
      "loss": 1.2864,
      "theoretical_loss": 3.3083870645044553,
      "tokens_seen": 3150053376
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.634343434343435e-05,
      "loss": 1.3111,
      "theoretical_loss": 3.3083818247618444,
      "tokens_seen": 3150118912
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6323232323232324e-05,
      "loss": 1.287,
      "theoretical_loss": 3.3083765851587628,
      "tokens_seen": 3150184448
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.630303030303031e-05,
      "loss": 1.4812,
      "theoretical_loss": 3.3083713456952046,
      "tokens_seen": 3150249984
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.628282828282828e-05,
      "loss": 1.42,
      "theoretical_loss": 3.308366106371163,
      "tokens_seen": 3150315520
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.626262626262626e-05,
      "loss": 1.2973,
      "theoretical_loss": 3.3083608671866314,
      "tokens_seen": 3150381056
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.624242424242425e-05,
      "loss": 1.3743,
      "theoretical_loss": 3.3083556281416033,
      "tokens_seen": 3150446592
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6222222222222224e-05,
      "loss": 1.4198,
      "theoretical_loss": 3.308350389236072,
      "tokens_seen": 3150512128
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.62020202020202e-05,
      "loss": 1.3559,
      "theoretical_loss": 3.3083451504700307,
      "tokens_seen": 3150577664
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6181818181818185e-05,
      "loss": 1.529,
      "theoretical_loss": 3.308339911843473,
      "tokens_seen": 3150643200
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.616161616161616e-05,
      "loss": 1.2846,
      "theoretical_loss": 3.3083346733563923,
      "tokens_seen": 3150708736
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6141414141414146e-05,
      "loss": 1.2647,
      "theoretical_loss": 3.3083294350087815,
      "tokens_seen": 3150774272
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6121212121212117e-05,
      "loss": 1.3286,
      "theoretical_loss": 3.3083241968006347,
      "tokens_seen": 3150839808
    },
    {
      "epoch": 0.91,
      "objective/train/aligned_token_count": 0.4615102708339691,
      "objective/train/avg_token_score": 0.011354394257068634,
      "objective/train/docs_used": 1770077,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.4298080205917358,
      "objective/train/original_loss": 4.610191822052002,
      "objective/train/theoretical_loss": 3.3083189587319453,
      "objective/train/tokens_used": 1509764576,
      "objective/train/unlikelihood_loss": 0.04809699207544327,
      "theoretical_loss": 3.3083189587319453,
      "tokens_seen": 3150905344
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.61010101010101e-05,
      "loss": 1.2604,
      "theoretical_loss": 3.3083189587319453,
      "tokens_seen": 3150905344
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6080808080808085e-05,
      "loss": 1.4167,
      "theoretical_loss": 3.3083137208027056,
      "tokens_seen": 3150970880
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.606060606060606e-05,
      "loss": 1.2096,
      "theoretical_loss": 3.3083084830129104,
      "tokens_seen": 3151036416
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.604040404040404e-05,
      "loss": 1.414,
      "theoretical_loss": 3.308303245362552,
      "tokens_seen": 3151101952
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6020202020202016e-05,
      "loss": 1.4092,
      "theoretical_loss": 3.3082980078516244,
      "tokens_seen": 3151167488
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.6e-05,
      "loss": 1.1714,
      "theoretical_loss": 3.308292770480121,
      "tokens_seen": 3151233024
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5979797979797984e-05,
      "loss": 1.152,
      "theoretical_loss": 3.3082875332480346,
      "tokens_seen": 3151298560
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.595959595959596e-05,
      "loss": 1.297,
      "theoretical_loss": 3.308282296155359,
      "tokens_seen": 3151364096
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.593939393939394e-05,
      "loss": 1.3445,
      "theoretical_loss": 3.308277059202088,
      "tokens_seen": 3151429632
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.591919191919192e-05,
      "loss": 1.3167,
      "theoretical_loss": 3.3082718223882144,
      "tokens_seen": 3151495168
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.58989898989899e-05,
      "loss": 1.3897,
      "theoretical_loss": 3.308266585713732,
      "tokens_seen": 3151560704
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5878787878787883e-05,
      "loss": 1.3162,
      "theoretical_loss": 3.3082613491786335,
      "tokens_seen": 3151626240
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5858585858585854e-05,
      "loss": 1.4866,
      "theoretical_loss": 3.308256112782913,
      "tokens_seen": 3151691776
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.583838383838384e-05,
      "loss": 1.3438,
      "theoretical_loss": 3.3082508765265635,
      "tokens_seen": 3151757312
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.581818181818182e-05,
      "loss": 1.3565,
      "theoretical_loss": 3.3082456404095786,
      "tokens_seen": 3151822848
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.57979797979798e-05,
      "loss": 1.5184,
      "theoretical_loss": 3.3082404044319516,
      "tokens_seen": 3151888384
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5777777777777776e-05,
      "loss": 1.3209,
      "theoretical_loss": 3.3082351685936757,
      "tokens_seen": 3151953920
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.575757575757576e-05,
      "loss": 1.3625,
      "theoretical_loss": 3.308229932894745,
      "tokens_seen": 3152019456
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.573737373737374e-05,
      "loss": 1.4071,
      "theoretical_loss": 3.308224697335152,
      "tokens_seen": 3152084992
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.571717171717172e-05,
      "loss": 1.4357,
      "theoretical_loss": 3.3082194619148906,
      "tokens_seen": 3152150528
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.56969696969697e-05,
      "loss": 1.3594,
      "theoretical_loss": 3.3082142266339543,
      "tokens_seen": 3152216064
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5676767676767676e-05,
      "loss": 1.4924,
      "theoretical_loss": 3.308208991492336,
      "tokens_seen": 3152281600
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.565656565656566e-05,
      "loss": 1.4012,
      "theoretical_loss": 3.3082037564900295,
      "tokens_seen": 3152347136
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.563636363636364e-05,
      "loss": 1.3246,
      "theoretical_loss": 3.3081985216270278,
      "tokens_seen": 3152412672
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.561616161616162e-05,
      "loss": 1.2887,
      "theoretical_loss": 3.3081932869033253,
      "tokens_seen": 3152478208
    },
    {
      "epoch": 0.91,
      "objective/train/aligned_token_count": 0.378543496131897,
      "objective/train/avg_token_score": 0.009084205143153667,
      "objective/train/docs_used": 1771514,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.308350920677185,
      "objective/train/original_loss": 4.89344596862793,
      "objective/train/theoretical_loss": 3.308188052318914,
      "objective/train/tokens_used": 1511402976,
      "objective/train/unlikelihood_loss": 0.05097414180636406,
      "theoretical_loss": 3.308188052318914,
      "tokens_seen": 3152543744
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.559595959595959e-05,
      "loss": 1.2199,
      "theoretical_loss": 3.308188052318914,
      "tokens_seen": 3152543744
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5575757575757575e-05,
      "loss": 1.2984,
      "theoretical_loss": 3.308182817873788,
      "tokens_seen": 3152609280
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.555555555555556e-05,
      "loss": 1.1097,
      "theoretical_loss": 3.3081775835679412,
      "tokens_seen": 3152674816
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5535353535353536e-05,
      "loss": 1.4303,
      "theoretical_loss": 3.3081723494013664,
      "tokens_seen": 3152740352
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5515151515151513e-05,
      "loss": 1.2683,
      "theoretical_loss": 3.3081671153740566,
      "tokens_seen": 3152805888
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.54949494949495e-05,
      "loss": 1.1833,
      "theoretical_loss": 3.308161881486006,
      "tokens_seen": 3152871424
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5474747474747475e-05,
      "loss": 1.3374,
      "theoretical_loss": 3.3081566477372073,
      "tokens_seen": 3152936960
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.545454545454546e-05,
      "loss": 1.2843,
      "theoretical_loss": 3.308151414127655,
      "tokens_seen": 3153002496
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.543434343434343e-05,
      "loss": 1.3998,
      "theoretical_loss": 3.3081461806573413,
      "tokens_seen": 3153068032
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.541414141414141e-05,
      "loss": 1.3345,
      "theoretical_loss": 3.30814094732626,
      "tokens_seen": 3153133568
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.53939393939394e-05,
      "loss": 1.2102,
      "theoretical_loss": 3.3081357141344045,
      "tokens_seen": 3153199104
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5373737373737374e-05,
      "loss": 1.2521,
      "theoretical_loss": 3.3081304810817684,
      "tokens_seen": 3153264640
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.535353535353536e-05,
      "loss": 1.3644,
      "theoretical_loss": 3.3081252481683454,
      "tokens_seen": 3153330176
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5333333333333335e-05,
      "loss": 1.571,
      "theoretical_loss": 3.308120015394128,
      "tokens_seen": 3153395712
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.531313131313131e-05,
      "loss": 1.3657,
      "theoretical_loss": 3.30811478275911,
      "tokens_seen": 3153461248
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5292929292929296e-05,
      "loss": 1.152,
      "theoretical_loss": 3.3081095502632856,
      "tokens_seen": 3153526784
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5272727272727274e-05,
      "loss": 1.1853,
      "theoretical_loss": 3.308104317906647,
      "tokens_seen": 3153592320
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.525252525252525e-05,
      "loss": 1.3685,
      "theoretical_loss": 3.3080990856891876,
      "tokens_seen": 3153657856
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5232323232323235e-05,
      "loss": 1.5841,
      "theoretical_loss": 3.3080938536109024,
      "tokens_seen": 3153723392
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.521212121212121e-05,
      "loss": 1.1737,
      "theoretical_loss": 3.308088621671783,
      "tokens_seen": 3153788928
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5191919191919196e-05,
      "loss": 1.3991,
      "theoretical_loss": 3.3080833898718236,
      "tokens_seen": 3153854464
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5171717171717166e-05,
      "loss": 1.2259,
      "theoretical_loss": 3.3080781582110177,
      "tokens_seen": 3153920000
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.515151515151515e-05,
      "loss": 1.3014,
      "theoretical_loss": 3.3080729266893583,
      "tokens_seen": 3153985536
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5131313131313134e-05,
      "loss": 1.1333,
      "theoretical_loss": 3.308067695306839,
      "tokens_seen": 3154051072
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.511111111111111e-05,
      "loss": 1.4258,
      "theoretical_loss": 3.3080624640634535,
      "tokens_seen": 3154116608
    },
    {
      "epoch": 0.91,
      "objective/train/aligned_token_count": 0.43493402004241943,
      "objective/train/avg_token_score": 0.00955898780375719,
      "objective/train/docs_used": 1772131,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.3539340496063232,
      "objective/train/original_loss": 4.402554512023926,
      "objective/train/theoretical_loss": 3.3080572329591953,
      "objective/train/tokens_used": 1513041376,
      "objective/train/unlikelihood_loss": 0.05578567087650299,
      "theoretical_loss": 3.3080572329591953,
      "tokens_seen": 3154182144
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5090909090909095e-05,
      "loss": 1.2459,
      "theoretical_loss": 3.3080572329591953,
      "tokens_seen": 3154182144
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.507070707070707e-05,
      "loss": 1.2977,
      "theoretical_loss": 3.308052001994057,
      "tokens_seen": 3154247680
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.505050505050505e-05,
      "loss": 1.4568,
      "theoretical_loss": 3.3080467711680326,
      "tokens_seen": 3154313216
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.5030303030303034e-05,
      "loss": 1.3041,
      "theoretical_loss": 3.308041540481115,
      "tokens_seen": 3154378752
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.501010101010101e-05,
      "loss": 1.3665,
      "theoretical_loss": 3.3080363099332986,
      "tokens_seen": 3154444288
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.498989898989899e-05,
      "loss": 1.4059,
      "theoretical_loss": 3.308031079524576,
      "tokens_seen": 3154509824
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.496969696969697e-05,
      "loss": 1.2047,
      "theoretical_loss": 3.308025849254941,
      "tokens_seen": 3154575360
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.494949494949495e-05,
      "loss": 1.4124,
      "theoretical_loss": 3.3080206191243864,
      "tokens_seen": 3154640896
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.492929292929293e-05,
      "loss": 1.1779,
      "theoretical_loss": 3.3080153891329065,
      "tokens_seen": 3154706432
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.490909090909091e-05,
      "loss": 1.4012,
      "theoretical_loss": 3.308010159280494,
      "tokens_seen": 3154771968
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.488888888888889e-05,
      "loss": 1.236,
      "theoretical_loss": 3.3080049295671423,
      "tokens_seen": 3154837504
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.486868686868687e-05,
      "loss": 1.1983,
      "theoretical_loss": 3.3079996999928456,
      "tokens_seen": 3154903040
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.484848484848485e-05,
      "loss": 1.6384,
      "theoretical_loss": 3.3079944705575968,
      "tokens_seen": 3154968576
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4828282828282826e-05,
      "loss": 1.2743,
      "theoretical_loss": 3.3079892412613887,
      "tokens_seen": 3155034112
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.480808080808081e-05,
      "loss": 1.2505,
      "theoretical_loss": 3.307984012104216,
      "tokens_seen": 3155099648
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.478787878787879e-05,
      "loss": 1.3735,
      "theoretical_loss": 3.3079787830860714,
      "tokens_seen": 3155165184
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.476767676767677e-05,
      "loss": 1.4874,
      "theoretical_loss": 3.3079735542069484,
      "tokens_seen": 3155230720
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.474747474747475e-05,
      "loss": 1.1609,
      "theoretical_loss": 3.30796832546684,
      "tokens_seen": 3155296256
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4727272727272725e-05,
      "loss": 1.2179,
      "theoretical_loss": 3.3079630968657403,
      "tokens_seen": 3155361792
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.470707070707071e-05,
      "loss": 1.1198,
      "theoretical_loss": 3.3079578684036424,
      "tokens_seen": 3155427328
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4686868686868686e-05,
      "loss": 1.3728,
      "theoretical_loss": 3.307952640080539,
      "tokens_seen": 3155492864
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.466666666666667e-05,
      "loss": 1.4828,
      "theoretical_loss": 3.307947411896425,
      "tokens_seen": 3155558400
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.464646464646465e-05,
      "loss": 1.3468,
      "theoretical_loss": 3.307942183851293,
      "tokens_seen": 3155623936
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4626262626262625e-05,
      "loss": 1.2515,
      "theoretical_loss": 3.3079369559451366,
      "tokens_seen": 3155689472
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.460606060606061e-05,
      "loss": 1.564,
      "theoretical_loss": 3.307931728177949,
      "tokens_seen": 3155755008
    },
    {
      "epoch": 0.91,
      "objective/train/aligned_token_count": 0.37933772802352905,
      "objective/train/avg_token_score": 0.02262301929295063,
      "objective/train/docs_used": 1773306,
      "objective/train/instantaneous_batch_size": 16,
      "objective/train/instantaneous_microbatch_size": 16384,
      "objective/train/likelihood_loss": 1.2311348915100098,
      "objective/train/original_loss": 4.576864242553711,
      "objective/train/theoretical_loss": 3.3079265005497236,
      "objective/train/tokens_used": 1514679776,
      "objective/train/unlikelihood_loss": 0.052221301943063736,
      "theoretical_loss": 3.3079265005497236,
      "tokens_seen": 3155820544
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4585858585858586e-05,
      "loss": 1.2782,
      "theoretical_loss": 3.3079265005497236,
      "tokens_seen": 3155820544
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.456565656565656e-05,
      "loss": 1.4039,
      "theoretical_loss": 3.307921273060454,
      "tokens_seen": 3155886080
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.454545454545455e-05,
      "loss": 1.3238,
      "theoretical_loss": 3.3079160457101335,
      "tokens_seen": 3155951616
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4525252525252524e-05,
      "loss": 1.2043,
      "theoretical_loss": 3.307910818498756,
      "tokens_seen": 3156017152
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.450505050505051e-05,
      "loss": 1.387,
      "theoretical_loss": 3.307905591426314,
      "tokens_seen": 3156082688
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.448484848484849e-05,
      "loss": 1.5213,
      "theoretical_loss": 3.307900364492802,
      "tokens_seen": 3156148224
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.446464646464646e-05,
      "loss": 1.4137,
      "theoretical_loss": 3.307895137698212,
      "tokens_seen": 3156213760
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4444444444444447e-05,
      "loss": 1.0717,
      "theoretical_loss": 3.307889911042539,
      "tokens_seen": 3156279296
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4424242424242424e-05,
      "loss": 1.3401,
      "theoretical_loss": 3.3078846845257757,
      "tokens_seen": 3156344832
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.440404040404041e-05,
      "loss": 1.3783,
      "theoretical_loss": 3.307879458147915,
      "tokens_seen": 3156410368
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4383838383838385e-05,
      "loss": 1.3433,
      "theoretical_loss": 3.3078742319089516,
      "tokens_seen": 3156475904
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.436363636363636e-05,
      "loss": 1.2277,
      "theoretical_loss": 3.3078690058088775,
      "tokens_seen": 3156541440
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4343434343434346e-05,
      "loss": 1.1879,
      "theoretical_loss": 3.307863779847687,
      "tokens_seen": 3156606976
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.432323232323232e-05,
      "loss": 1.4023,
      "theoretical_loss": 3.3078585540253735,
      "tokens_seen": 3156672512
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.43030303030303e-05,
      "loss": 1.4071,
      "theoretical_loss": 3.3078533283419302,
      "tokens_seen": 3156738048
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.4282828282828284e-05,
      "loss": 1.4401,
      "theoretical_loss": 3.3078481027973505,
      "tokens_seen": 3156803584
    }
  ],
  "max_steps": 25000,
  "num_train_epochs": 9223372036854775807,
  "total_flos": 7.6305634689024e+17,
  "trial_name": null,
  "trial_params": null
}