{
  "best_metric": 0.5450515151023865,
  "best_model_checkpoint": "experts/mistralic-expert-23/checkpoint-3400",
  "epoch": 0.8412197686645636,
  "eval_steps": 200,
  "global_step": 3600,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.6926,
      "step": 1
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.6169,
      "step": 2
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.539,
      "step": 3
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.5802,
      "step": 4
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.5176,
      "step": 5
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.6371,
      "step": 6
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.5969,
      "step": 7
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.6545,
      "step": 8
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.7967,
      "step": 9
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.5244,
      "step": 10
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.9098,
      "step": 11
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.7216,
      "step": 12
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.5279,
      "step": 13
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.784,
      "step": 14
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.9971,
      "step": 15
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.5255,
      "step": 16
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.6404,
      "step": 17
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.8234,
      "step": 18
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.529,
      "step": 19
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.5078,
      "step": 20
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.4981,
      "step": 21
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8841,
      "step": 22
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5563,
      "step": 23
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5759,
      "step": 24
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5739,
      "step": 25
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5891,
      "step": 26
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.454,
      "step": 27
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5685,
      "step": 28
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5723,
      "step": 29
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.516,
      "step": 30
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7419,
      "step": 31
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7529,
      "step": 32
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6331,
      "step": 33
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7149,
      "step": 34
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7413,
      "step": 35
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5488,
      "step": 36
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8207,
      "step": 37
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5589,
      "step": 38
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5907,
      "step": 39
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6014,
      "step": 40
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5317,
      "step": 41
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5762,
      "step": 42
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5986,
      "step": 43
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5758,
      "step": 44
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7376,
      "step": 45
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6161,
      "step": 46
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5554,
      "step": 47
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5997,
      "step": 48
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5616,
      "step": 49
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5376,
      "step": 50
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6303,
      "step": 51
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5722,
      "step": 52
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5588,
      "step": 53
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5734,
      "step": 54
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5353,
      "step": 55
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5713,
      "step": 56
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6006,
      "step": 57
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.7782,
      "step": 58
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5286,
      "step": 59
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6655,
      "step": 60
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.6916,
      "step": 61
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.8167,
      "step": 62
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5933,
      "step": 63
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5054,
      "step": 64
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5764,
      "step": 65
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.4551,
      "step": 66
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6122,
      "step": 67
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6257,
      "step": 68
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6027,
      "step": 69
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5796,
      "step": 70
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6817,
      "step": 71
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5495,
      "step": 72
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5997,
      "step": 73
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.837,
      "step": 74
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5642,
      "step": 75
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5644,
      "step": 76
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5589,
      "step": 77
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6021,
      "step": 78
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6688,
      "step": 79
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5716,
      "step": 80
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.75,
      "step": 81
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6068,
      "step": 82
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5987,
      "step": 83
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7231,
      "step": 84
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5786,
      "step": 85
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6502,
      "step": 86
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6702,
      "step": 87
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6432,
      "step": 88
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5064,
      "step": 89
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5188,
      "step": 90
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.9853,
      "step": 91
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5188,
      "step": 92
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8395,
      "step": 93
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5267,
      "step": 94
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5357,
      "step": 95
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5574,
      "step": 96
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5883,
      "step": 97
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6675,
      "step": 98
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5527,
      "step": 99
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6302,
      "step": 100
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.8618,
      "step": 101
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6703,
      "step": 102
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.5242,
      "step": 103
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.653,
      "step": 104
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.6386,
      "step": 105
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.7444,
      "step": 106
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5107,
      "step": 107
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6885,
      "step": 108
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5642,
      "step": 109
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6477,
      "step": 110
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6824,
      "step": 111
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5575,
      "step": 112
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5927,
      "step": 113
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5312,
      "step": 114
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.63,
      "step": 115
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4778,
      "step": 116
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5675,
      "step": 117
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4322,
      "step": 118
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5693,
      "step": 119
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6603,
      "step": 120
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4729,
      "step": 121
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.8347,
      "step": 122
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6119,
      "step": 123
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7242,
      "step": 124
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5018,
      "step": 125
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.7959,
      "step": 126
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.606,
      "step": 127
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5727,
      "step": 128
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6316,
      "step": 129
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6095,
      "step": 130
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5735,
      "step": 131
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4528,
      "step": 132
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5352,
      "step": 133
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6042,
      "step": 134
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.76,
      "step": 135
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.4522,
      "step": 136
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5826,
      "step": 137
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5408,
      "step": 138
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5659,
      "step": 139
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.551,
      "step": 140
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5766,
      "step": 141
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5377,
      "step": 142
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.8203,
      "step": 143
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5066,
      "step": 144
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5001,
      "step": 145
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.5727,
      "step": 146
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6109,
      "step": 147
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6467,
      "step": 148
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.6917,
      "step": 149
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6373,
      "step": 150
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6928,
      "step": 151
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5866,
      "step": 152
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.659,
      "step": 153
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5483,
      "step": 154
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5335,
      "step": 155
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7856,
      "step": 156
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5694,
      "step": 157
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8404,
      "step": 158
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.575,
      "step": 159
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6436,
      "step": 160
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4149,
      "step": 161
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6085,
      "step": 162
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5949,
      "step": 163
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6294,
      "step": 164
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5303,
      "step": 165
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7256,
      "step": 166
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6095,
      "step": 167
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7129,
      "step": 168
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6088,
      "step": 169
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4735,
      "step": 170
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4963,
      "step": 171
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.8293,
      "step": 172
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6143,
      "step": 173
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5835,
      "step": 174
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.7986,
      "step": 175
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6569,
      "step": 176
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6295,
      "step": 177
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4927,
      "step": 178
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5401,
      "step": 179
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6203,
      "step": 180
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5045,
      "step": 181
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5779,
      "step": 182
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5386,
      "step": 183
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.4483,
      "step": 184
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5543,
      "step": 185
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 1.0266,
      "step": 186
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5621,
      "step": 187
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5884,
      "step": 188
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5395,
      "step": 189
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6764,
      "step": 190
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.5938,
      "step": 191
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.6306,
      "step": 192
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8946,
      "step": 193
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.6613,
      "step": 194
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.4926,
      "step": 195
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.4874,
      "step": 196
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.6966,
      "step": 197
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7493,
      "step": 198
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.62,
      "step": 199
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8005,
      "step": 200
    },
    {
      "epoch": 0.05,
      "eval_loss": 0.600193977355957,
      "eval_runtime": 156.5363,
      "eval_samples_per_second": 6.388,
      "eval_steps_per_second": 3.194,
      "step": 200
    },
    {
      "epoch": 0.05,
      "mmlu_eval_accuracy": 0.6110525160674652,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.6818181818181818,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.3902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.6046511627906976,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3103448275862069,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6538461538461539,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6153846153846154,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7325581395348837,
      "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.37,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.7352941176470589,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_law": 0.4,
      "mmlu_eval_accuracy_professional_medicine": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_psychology": 0.6231884057971014,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.7407407407407407,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.8947368421052632,
      "mmlu_loss": 1.4672958862532524,
      "step": 200
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.4891,
      "step": 201
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5071,
      "step": 202
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7988,
      "step": 203
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.542,
      "step": 204
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.501,
      "step": 205
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.6016,
      "step": 206
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7817,
      "step": 207
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5444,
      "step": 208
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7871,
      "step": 209
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.6683,
      "step": 210
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.4517,
      "step": 211
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8223,
      "step": 212
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.536,
      "step": 213
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5169,
      "step": 214
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.6521,
      "step": 215
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5043,
      "step": 216
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5211,
      "step": 217
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5449,
      "step": 218
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5034,
      "step": 219
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8192,
      "step": 220
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5212,
      "step": 221
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.6259,
      "step": 222
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5675,
      "step": 223
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5851,
      "step": 224
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.7381,
      "step": 225
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.6499,
      "step": 226
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8992,
      "step": 227
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5907,
      "step": 228
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.602,
      "step": 229
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5298,
      "step": 230
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.4669,
      "step": 231
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8426,
      "step": 232
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.8515,
      "step": 233
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5737,
      "step": 234
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.5501,
      "step": 235
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4967,
      "step": 236
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6454,
      "step": 237
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4659,
      "step": 238
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.579,
      "step": 239
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7209,
      "step": 240
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5715,
      "step": 241
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4989,
      "step": 242
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6546,
      "step": 243
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7433,
      "step": 244
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5624,
      "step": 245
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5985,
      "step": 246
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5385,
      "step": 247
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6101,
      "step": 248
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5485,
      "step": 249
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6936,
      "step": 250
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.551,
      "step": 251
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5149,
      "step": 252
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5189,
      "step": 253
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7124,
      "step": 254
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6658,
      "step": 255
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5099,
      "step": 256
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4779,
      "step": 257
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5552,
      "step": 258
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5756,
      "step": 259
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5334,
      "step": 260
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6678,
      "step": 261
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7578,
      "step": 262
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7903,
      "step": 263
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5694,
      "step": 264
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7169,
      "step": 265
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4679,
      "step": 266
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6381,
      "step": 267
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5931,
      "step": 268
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5818,
      "step": 269
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7225,
      "step": 270
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6526,
      "step": 271
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5631,
      "step": 272
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.7555,
      "step": 273
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.6211,
      "step": 274
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5831,
      "step": 275
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.5269,
      "step": 276
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.4996,
      "step": 277
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.8797,
      "step": 278
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4943,
      "step": 279
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.8619,
      "step": 280
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6191,
      "step": 281
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4933,
      "step": 282
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.744,
      "step": 283
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4481,
      "step": 284
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5396,
      "step": 285
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5108,
      "step": 286
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4857,
      "step": 287
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5603,
      "step": 288
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5778,
      "step": 289
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6088,
      "step": 290
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.52,
      "step": 291
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6123,
      "step": 292
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5051,
      "step": 293
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5057,
      "step": 294
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.601,
      "step": 295
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.9173,
      "step": 296
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5194,
      "step": 297
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6205,
      "step": 298
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5811,
      "step": 299
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.8153,
      "step": 300
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5872,
      "step": 301
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6062,
      "step": 302
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4371,
      "step": 303
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5227,
      "step": 304
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5014,
      "step": 305
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4599,
      "step": 306
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5638,
      "step": 307
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4859,
      "step": 308
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5348,
      "step": 309
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4462,
      "step": 310
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.6522,
      "step": 311
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7238,
      "step": 312
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4864,
      "step": 313
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7379,
      "step": 314
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4141,
      "step": 315
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7453,
      "step": 316
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.4621,
      "step": 317
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.5194,
      "step": 318
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7686,
      "step": 319
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.7024,
      "step": 320
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6979,
      "step": 321
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.717,
      "step": 322
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5238,
      "step": 323
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5048,
      "step": 324
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5091,
      "step": 325
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.7207,
      "step": 326
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5248,
      "step": 327
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8079,
      "step": 328
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6206,
      "step": 329
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6146,
      "step": 330
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5315,
      "step": 331
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.53,
      "step": 332
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.794,
      "step": 333
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.4968,
      "step": 334
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5732,
      "step": 335
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5856,
      "step": 336
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.469,
      "step": 337
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.638,
      "step": 338
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.7457,
      "step": 339
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5851,
      "step": 340
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6549,
      "step": 341
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5412,
      "step": 342
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.7014,
      "step": 343
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.727,
      "step": 344
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5415,
      "step": 345
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5822,
      "step": 346
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.585,
      "step": 347
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5198,
      "step": 348
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5274,
      "step": 349
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5288,
      "step": 350
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5649,
      "step": 351
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5385,
      "step": 352
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6064,
      "step": 353
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5652,
      "step": 354
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.4739,
      "step": 355
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.4917,
      "step": 356
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6023,
      "step": 357
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.7409,
      "step": 358
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.6717,
      "step": 359
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.8324,
      "step": 360
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5896,
      "step": 361
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5608,
      "step": 362
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.5038,
      "step": 363
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5431,
      "step": 364
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7616,
      "step": 365
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5106,
      "step": 366
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.4998,
      "step": 367
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5058,
      "step": 368
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.536,
      "step": 369
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5248,
      "step": 370
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.728,
      "step": 371
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.4739,
      "step": 372
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7373,
      "step": 373
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5395,
      "step": 374
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5658,
      "step": 375
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.4556,
      "step": 376
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5087,
      "step": 377
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5215,
      "step": 378
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.4547,
      "step": 379
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.6626,
      "step": 380
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.666,
      "step": 381
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7555,
      "step": 382
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.6575,
      "step": 383
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.6711,
      "step": 384
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5149,
      "step": 385
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.6033,
      "step": 386
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5753,
      "step": 387
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5267,
      "step": 388
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.664,
      "step": 389
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.6835,
      "step": 390
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5874,
      "step": 391
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5413,
      "step": 392
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.579,
      "step": 393
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5294,
      "step": 394
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5736,
      "step": 395
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5794,
      "step": 396
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.6177,
      "step": 397
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5843,
      "step": 398
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.6333,
      "step": 399
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.7496,
      "step": 400
    },
    {
      "epoch": 0.09,
      "eval_loss": 0.5899355411529541,
      "eval_runtime": 156.5886,
      "eval_samples_per_second": 6.386,
      "eval_steps_per_second": 3.193,
      "step": 400
    },
    {
      "epoch": 0.09,
      "mmlu_eval_accuracy": 0.6068635133301271,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.6875,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.5454545454545454,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.8125,
      "mmlu_eval_accuracy_elementary_mathematics": 0.43902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.65625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6923076923076923,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.4782608695652174,
      "mmlu_eval_accuracy_high_school_us_history": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7441860465116279,
      "mmlu_eval_accuracy_moral_disputes": 0.6052631578947368,
      "mmlu_eval_accuracy_moral_scenarios": 0.37,
      "mmlu_eval_accuracy_nutrition": 0.696969696969697,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_law": 0.4117647058823529,
      "mmlu_eval_accuracy_professional_medicine": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_psychology": 0.6231884057971014,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.6666666666666666,
      "mmlu_eval_accuracy_sociology": 0.9545454545454546,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 1.496563289402048,
      "step": 400
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5013,
      "step": 401
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.509,
      "step": 402
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.6059,
      "step": 403
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.4962,
      "step": 404
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5252,
      "step": 405
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.5063,
      "step": 406
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.4901,
      "step": 407
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.636,
      "step": 408
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5539,
      "step": 409
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6965,
      "step": 410
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.767,
      "step": 411
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7795,
      "step": 412
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.488,
      "step": 413
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7407,
      "step": 414
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6471,
      "step": 415
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5734,
      "step": 416
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5481,
      "step": 417
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6453,
      "step": 418
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.4902,
      "step": 419
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6816,
      "step": 420
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5906,
      "step": 421
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5285,
      "step": 422
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5178,
      "step": 423
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5605,
      "step": 424
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7357,
      "step": 425
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6648,
      "step": 426
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5083,
      "step": 427
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.4889,
      "step": 428
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5004,
      "step": 429
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.4759,
      "step": 430
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6473,
      "step": 431
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5438,
      "step": 432
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5149,
      "step": 433
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.839,
      "step": 434
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5581,
      "step": 435
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5813,
      "step": 436
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.4878,
      "step": 437
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5984,
      "step": 438
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7623,
      "step": 439
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.6718,
      "step": 440
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5112,
      "step": 441
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.4414,
      "step": 442
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5551,
      "step": 443
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5107,
      "step": 444
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5318,
      "step": 445
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.504,
      "step": 446
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.5137,
      "step": 447
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.588,
      "step": 448
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.7416,
      "step": 449
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5064,
      "step": 450
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4845,
      "step": 451
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.8342,
      "step": 452
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5693,
      "step": 453
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4511,
      "step": 454
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4945,
      "step": 455
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5943,
      "step": 456
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6719,
      "step": 457
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5331,
      "step": 458
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6232,
      "step": 459
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7195,
      "step": 460
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7247,
      "step": 461
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7871,
      "step": 462
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5414,
      "step": 463
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5728,
      "step": 464
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5263,
      "step": 465
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5105,
      "step": 466
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5922,
      "step": 467
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5316,
      "step": 468
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4836,
      "step": 469
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4781,
      "step": 470
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.645,
      "step": 471
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6311,
      "step": 472
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5401,
      "step": 473
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5423,
      "step": 474
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.7188,
      "step": 475
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.587,
      "step": 476
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.46,
      "step": 477
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5724,
      "step": 478
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5907,
      "step": 479
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.484,
      "step": 480
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5095,
      "step": 481
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5197,
      "step": 482
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6144,
      "step": 483
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4964,
      "step": 484
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4705,
      "step": 485
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.4844,
      "step": 486
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6951,
      "step": 487
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5948,
      "step": 488
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5728,
      "step": 489
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6655,
      "step": 490
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.6506,
      "step": 491
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.5119,
      "step": 492
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6653,
      "step": 493
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5105,
      "step": 494
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5486,
      "step": 495
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.7635,
      "step": 496
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5753,
      "step": 497
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5336,
      "step": 498
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5103,
      "step": 499
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4841,
      "step": 500
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5479,
      "step": 501
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6147,
      "step": 502
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4874,
      "step": 503
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.7884,
      "step": 504
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6535,
      "step": 505
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6987,
      "step": 506
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6371,
      "step": 507
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5822,
      "step": 508
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.7897,
      "step": 509
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4869,
      "step": 510
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6293,
      "step": 511
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.7633,
      "step": 512
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6217,
      "step": 513
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5522,
      "step": 514
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5368,
      "step": 515
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5667,
      "step": 516
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5541,
      "step": 517
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5003,
      "step": 518
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5217,
      "step": 519
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.467,
      "step": 520
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4416,
      "step": 521
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4778,
      "step": 522
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5085,
      "step": 523
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.6101,
      "step": 524
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4561,
      "step": 525
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.56,
      "step": 526
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5205,
      "step": 527
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4748,
      "step": 528
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5813,
      "step": 529
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5873,
      "step": 530
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.5845,
      "step": 531
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.4624,
      "step": 532
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.546,
      "step": 533
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.9573,
      "step": 534
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5634,
      "step": 535
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3786,
      "step": 536
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7356,
      "step": 537
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5748,
      "step": 538
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.6774,
      "step": 539
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5361,
      "step": 540
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.465,
      "step": 541
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5905,
      "step": 542
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.6072,
      "step": 543
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5397,
      "step": 544
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5201,
      "step": 545
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7455,
      "step": 546
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.672,
      "step": 547
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.6171,
      "step": 548
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5088,
      "step": 549
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5438,
      "step": 550
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4865,
      "step": 551
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5796,
      "step": 552
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.6324,
      "step": 553
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4357,
      "step": 554
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4484,
      "step": 555
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.6764,
      "step": 556
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5309,
      "step": 557
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5603,
      "step": 558
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7126,
      "step": 559
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4893,
      "step": 560
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5384,
      "step": 561
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4642,
      "step": 562
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4378,
      "step": 563
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5831,
      "step": 564
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5491,
      "step": 565
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.6361,
      "step": 566
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 1.0347,
      "step": 567
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.4843,
      "step": 568
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.762,
      "step": 569
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7908,
      "step": 570
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5221,
      "step": 571
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.8129,
      "step": 572
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.542,
      "step": 573
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5179,
      "step": 574
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.5375,
      "step": 575
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.6367,
      "step": 576
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.7865,
      "step": 577
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5287,
      "step": 578
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5414,
      "step": 579
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6124,
      "step": 580
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8515,
      "step": 581
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5426,
      "step": 582
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5048,
      "step": 583
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.559,
      "step": 584
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5914,
      "step": 585
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5469,
      "step": 586
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.719,
      "step": 587
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5888,
      "step": 588
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.4804,
      "step": 589
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.4549,
      "step": 590
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.4929,
      "step": 591
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6596,
      "step": 592
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5239,
      "step": 593
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6946,
      "step": 594
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7026,
      "step": 595
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7988,
      "step": 596
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6868,
      "step": 597
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7739,
      "step": 598
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5263,
      "step": 599
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7381,
      "step": 600
    },
    {
      "epoch": 0.14,
      "eval_loss": 0.5821242928504944,
      "eval_runtime": 156.5384,
      "eval_samples_per_second": 6.388,
      "eval_steps_per_second": 3.194,
      "step": 600
    },
    {
      "epoch": 0.14,
      "mmlu_eval_accuracy": 0.5980259299067578,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.8125,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.6363636363636364,
      "mmlu_eval_accuracy_college_medicine": 0.6818181818181818,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.6875,
      "mmlu_eval_accuracy_elementary_mathematics": 0.3902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.59375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3103448275862069,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6923076923076923,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.85,
      "mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
      "mmlu_eval_accuracy_high_school_us_history": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7209302325581395,
      "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.31,
      "mmlu_eval_accuracy_nutrition": 0.6666666666666666,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_law": 0.38235294117647056,
      "mmlu_eval_accuracy_professional_medicine": 0.5806451612903226,
      "mmlu_eval_accuracy_professional_psychology": 0.6086956521739131,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.6666666666666666,
      "mmlu_eval_accuracy_sociology": 0.8636363636363636,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 1.5698812617178686,
      "step": 600
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.797,
      "step": 601
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5773,
      "step": 602
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.4958,
      "step": 603
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5999,
      "step": 604
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5027,
      "step": 605
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5619,
      "step": 606
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.484,
      "step": 607
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.4481,
      "step": 608
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.9088,
      "step": 609
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.4749,
      "step": 610
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.8021,
      "step": 611
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6439,
      "step": 612
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.4418,
      "step": 613
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7767,
      "step": 614
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5253,
      "step": 615
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.7919,
      "step": 616
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6043,
      "step": 617
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.6282,
      "step": 618
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5026,
      "step": 619
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.5817,
      "step": 620
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5307,
      "step": 621
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7302,
      "step": 622
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5822,
      "step": 623
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5515,
      "step": 624
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5555,
      "step": 625
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5796,
      "step": 626
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5085,
      "step": 627
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5865,
      "step": 628
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5207,
      "step": 629
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5179,
      "step": 630
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.502,
      "step": 631
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.555,
      "step": 632
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5634,
      "step": 633
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5987,
      "step": 634
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5024,
      "step": 635
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.6944,
      "step": 636
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.6259,
      "step": 637
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5491,
      "step": 638
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5276,
      "step": 639
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7198,
      "step": 640
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5254,
      "step": 641
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5183,
      "step": 642
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5534,
      "step": 643
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.588,
      "step": 644
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.6834,
      "step": 645
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5231,
      "step": 646
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7475,
      "step": 647
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5971,
      "step": 648
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5364,
      "step": 649
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5378,
      "step": 650
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5151,
      "step": 651
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5848,
      "step": 652
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.6543,
      "step": 653
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.7276,
      "step": 654
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.4609,
      "step": 655
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.6721,
      "step": 656
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.6221,
      "step": 657
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.6601,
      "step": 658
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5458,
      "step": 659
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5848,
      "step": 660
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5555,
      "step": 661
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5629,
      "step": 662
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.5773,
      "step": 663
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5895,
      "step": 664
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.584,
      "step": 665
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.4354,
      "step": 666
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.6236,
      "step": 667
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5397,
      "step": 668
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.4549,
      "step": 669
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5186,
      "step": 670
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.6771,
      "step": 671
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5456,
      "step": 672
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5139,
      "step": 673
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.6618,
      "step": 674
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.6333,
      "step": 675
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5399,
      "step": 676
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.485,
      "step": 677
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5551,
      "step": 678
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.4292,
      "step": 679
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5707,
      "step": 680
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5536,
      "step": 681
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.6842,
      "step": 682
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.561,
      "step": 683
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5815,
      "step": 684
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7082,
      "step": 685
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.504,
      "step": 686
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.4585,
      "step": 687
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5481,
      "step": 688
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5959,
      "step": 689
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5654,
      "step": 690
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7278,
      "step": 691
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.8314,
      "step": 692
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5705,
      "step": 693
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5927,
      "step": 694
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5054,
      "step": 695
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5345,
      "step": 696
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5068,
      "step": 697
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5463,
      "step": 698
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.4945,
      "step": 699
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.8122,
      "step": 700
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.7609,
      "step": 701
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.4743,
      "step": 702
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5149,
      "step": 703
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5402,
      "step": 704
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.596,
      "step": 705
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.5838,
      "step": 706
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6725,
      "step": 707
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.451,
      "step": 708
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6256,
      "step": 709
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7896,
      "step": 710
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.4892,
      "step": 711
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.508,
      "step": 712
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5751,
      "step": 713
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5195,
      "step": 714
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5914,
      "step": 715
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7709,
      "step": 716
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6462,
      "step": 717
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5456,
      "step": 718
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6971,
      "step": 719
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5392,
      "step": 720
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5519,
      "step": 721
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.4363,
      "step": 722
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5198,
      "step": 723
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5964,
      "step": 724
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5382,
      "step": 725
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5231,
      "step": 726
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.9078,
      "step": 727
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6282,
      "step": 728
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5581,
      "step": 729
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5676,
      "step": 730
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5519,
      "step": 731
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7338,
      "step": 732
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6373,
      "step": 733
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5116,
      "step": 734
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6304,
      "step": 735
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5385,
      "step": 736
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5202,
      "step": 737
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6285,
      "step": 738
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7182,
      "step": 739
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5757,
      "step": 740
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.4787,
      "step": 741
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.4862,
      "step": 742
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.7525,
      "step": 743
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5101,
      "step": 744
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.5368,
      "step": 745
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6022,
      "step": 746
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.6534,
      "step": 747
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.4966,
      "step": 748
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5995,
      "step": 749
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5371,
      "step": 750
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8189,
      "step": 751
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.4673,
      "step": 752
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.6376,
      "step": 753
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.6554,
      "step": 754
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8246,
      "step": 755
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.7493,
      "step": 756
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5867,
      "step": 757
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5766,
      "step": 758
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.4687,
      "step": 759
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5795,
      "step": 760
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.4207,
      "step": 761
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.524,
      "step": 762
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5806,
      "step": 763
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.4486,
      "step": 764
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.6868,
      "step": 765
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.7085,
      "step": 766
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5446,
      "step": 767
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.4995,
      "step": 768
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5232,
      "step": 769
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5164,
      "step": 770
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.4857,
      "step": 771
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.7406,
      "step": 772
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5989,
      "step": 773
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5384,
      "step": 774
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5186,
      "step": 775
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5206,
      "step": 776
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.7905,
      "step": 777
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5032,
      "step": 778
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.4573,
      "step": 779
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.8265,
      "step": 780
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5497,
      "step": 781
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5017,
      "step": 782
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5227,
      "step": 783
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.6619,
      "step": 784
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5182,
      "step": 785
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.6738,
      "step": 786
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5116,
      "step": 787
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.4761,
      "step": 788
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5951,
      "step": 789
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.6349,
      "step": 790
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.5733,
      "step": 791
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4627,
      "step": 792
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4988,
      "step": 793
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.691,
      "step": 794
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5198,
      "step": 795
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5535,
      "step": 796
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5272,
      "step": 797
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.574,
      "step": 798
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7339,
      "step": 799
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5,
      "step": 800
    },
    {
      "epoch": 0.19,
      "eval_loss": 0.574880838394165,
      "eval_runtime": 156.4152,
      "eval_samples_per_second": 6.393,
      "eval_steps_per_second": 3.197,
      "step": 800
    },
    {
      "epoch": 0.19,
      "mmlu_eval_accuracy": 0.5932018544060925,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5172413793103449,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.6875,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4146341463414634,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_geography": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.4482758620689655,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6538461538461539,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.7272727272727273,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7209302325581395,
      "mmlu_eval_accuracy_moral_disputes": 0.5,
      "mmlu_eval_accuracy_moral_scenarios": 0.3,
      "mmlu_eval_accuracy_nutrition": 0.696969696969697,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.5428571428571428,
      "mmlu_eval_accuracy_professional_accounting": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_law": 0.38823529411764707,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.6086956521739131,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.6666666666666666,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.9473684210526315,
      "mmlu_loss": 1.3265360119526441,
      "step": 800
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5165,
      "step": 801
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6498,
      "step": 802
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4657,
      "step": 803
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5278,
      "step": 804
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5308,
      "step": 805
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5479,
      "step": 806
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4857,
      "step": 807
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4427,
      "step": 808
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4815,
      "step": 809
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4946,
      "step": 810
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8382,
      "step": 811
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5392,
      "step": 812
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.517,
      "step": 813
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.8735,
      "step": 814
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6773,
      "step": 815
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6418,
      "step": 816
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5489,
      "step": 817
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5801,
      "step": 818
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7972,
      "step": 819
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6654,
      "step": 820
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7159,
      "step": 821
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5404,
      "step": 822
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7321,
      "step": 823
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5629,
      "step": 824
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5395,
      "step": 825
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5614,
      "step": 826
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5907,
      "step": 827
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7593,
      "step": 828
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5988,
      "step": 829
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.7543,
      "step": 830
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.6409,
      "step": 831
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5226,
      "step": 832
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.5747,
      "step": 833
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.4743,
      "step": 834
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7551,
      "step": 835
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5089,
      "step": 836
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5515,
      "step": 837
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.4745,
      "step": 838
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6076,
      "step": 839
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.4837,
      "step": 840
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5207,
      "step": 841
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6733,
      "step": 842
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6156,
      "step": 843
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6658,
      "step": 844
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5105,
      "step": 845
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.8214,
      "step": 846
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5534,
      "step": 847
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5574,
      "step": 848
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5164,
      "step": 849
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.4638,
      "step": 850
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5875,
      "step": 851
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6846,
      "step": 852
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5436,
      "step": 853
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.4961,
      "step": 854
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5504,
      "step": 855
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6088,
      "step": 856
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.4132,
      "step": 857
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.575,
      "step": 858
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5085,
      "step": 859
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5122,
      "step": 860
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6464,
      "step": 861
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6277,
      "step": 862
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5759,
      "step": 863
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.4963,
      "step": 864
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.531,
      "step": 865
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5125,
      "step": 866
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.496,
      "step": 867
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.4053,
      "step": 868
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5383,
      "step": 869
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5676,
      "step": 870
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7208,
      "step": 871
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6998,
      "step": 872
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.6198,
      "step": 873
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.5652,
      "step": 874
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.7726,
      "step": 875
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.4955,
      "step": 876
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.4391,
      "step": 877
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8064,
      "step": 878
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.726,
      "step": 879
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5287,
      "step": 880
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.4968,
      "step": 881
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5153,
      "step": 882
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5129,
      "step": 883
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5616,
      "step": 884
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5374,
      "step": 885
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6629,
      "step": 886
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.544,
      "step": 887
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.4661,
      "step": 888
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5176,
      "step": 889
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5545,
      "step": 890
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5462,
      "step": 891
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5179,
      "step": 892
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5414,
      "step": 893
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6187,
      "step": 894
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6309,
      "step": 895
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5833,
      "step": 896
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.4936,
      "step": 897
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.4672,
      "step": 898
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.448,
      "step": 899
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5412,
      "step": 900
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.9056,
      "step": 901
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5951,
      "step": 902
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6641,
      "step": 903
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5018,
      "step": 904
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5824,
      "step": 905
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5706,
      "step": 906
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5344,
      "step": 907
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.4701,
      "step": 908
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5806,
      "step": 909
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5901,
      "step": 910
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6398,
      "step": 911
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8337,
      "step": 912
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5989,
      "step": 913
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.8263,
      "step": 914
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5399,
      "step": 915
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.4453,
      "step": 916
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.7069,
      "step": 917
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.4743,
      "step": 918
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.5088,
      "step": 919
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.6518,
      "step": 920
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6899,
      "step": 921
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.4602,
      "step": 922
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.4718,
      "step": 923
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.743,
      "step": 924
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.7759,
      "step": 925
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.53,
      "step": 926
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.4967,
      "step": 927
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6014,
      "step": 928
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3803,
      "step": 929
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.564,
      "step": 930
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6286,
      "step": 931
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.519,
      "step": 932
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6825,
      "step": 933
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6336,
      "step": 934
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6436,
      "step": 935
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6905,
      "step": 936
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.7161,
      "step": 937
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5102,
      "step": 938
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.4539,
      "step": 939
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5792,
      "step": 940
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5131,
      "step": 941
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.4312,
      "step": 942
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6203,
      "step": 943
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5137,
      "step": 944
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.8426,
      "step": 945
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.558,
      "step": 946
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.4399,
      "step": 947
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6979,
      "step": 948
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.504,
      "step": 949
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6169,
      "step": 950
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5208,
      "step": 951
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.7029,
      "step": 952
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5348,
      "step": 953
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5016,
      "step": 954
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.561,
      "step": 955
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.6065,
      "step": 956
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.509,
      "step": 957
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5928,
      "step": 958
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5336,
      "step": 959
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.4983,
      "step": 960
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5083,
      "step": 961
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.5219,
      "step": 962
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6288,
      "step": 963
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5795,
      "step": 964
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5825,
      "step": 965
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4772,
      "step": 966
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6507,
      "step": 967
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7522,
      "step": 968
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5149,
      "step": 969
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4939,
      "step": 970
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4675,
      "step": 971
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5173,
      "step": 972
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6366,
      "step": 973
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5046,
      "step": 974
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6346,
      "step": 975
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4336,
      "step": 976
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.895,
      "step": 977
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5283,
      "step": 978
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5317,
      "step": 979
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4931,
      "step": 980
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5238,
      "step": 981
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7042,
      "step": 982
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.64,
      "step": 983
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6442,
      "step": 984
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5357,
      "step": 985
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5308,
      "step": 986
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4645,
      "step": 987
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5952,
      "step": 988
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7809,
      "step": 989
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.733,
      "step": 990
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.7313,
      "step": 991
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5493,
      "step": 992
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6119,
      "step": 993
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4745,
      "step": 994
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6025,
      "step": 995
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6024,
      "step": 996
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4761,
      "step": 997
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.8131,
      "step": 998
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5084,
      "step": 999
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5849,
      "step": 1000
    },
    {
      "epoch": 0.23,
      "eval_loss": 0.5721054673194885,
      "eval_runtime": 156.5412,
      "eval_samples_per_second": 6.388,
      "eval_steps_per_second": 3.194,
      "step": 1000
    },
    {
      "epoch": 0.23,
      "mmlu_eval_accuracy": 0.5929489471474008,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.75,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4146341463414634,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.65625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.8095238095238095,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3103448275862069,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6538461538461539,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.8166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7209302325581395,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.32,
      "mmlu_eval_accuracy_nutrition": 0.696969696969697,
      "mmlu_eval_accuracy_philosophy": 0.6764705882352942,
      "mmlu_eval_accuracy_prehistory": 0.6285714285714286,
      "mmlu_eval_accuracy_professional_accounting": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_law": 0.4235294117647059,
      "mmlu_eval_accuracy_professional_medicine": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_psychology": 0.5797101449275363,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.6296296296296297,
      "mmlu_eval_accuracy_sociology": 0.9545454545454546,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8947368421052632,
      "mmlu_loss": 1.4194403641684559,
      "step": 1000
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4694,
      "step": 1001
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.4452,
      "step": 1002
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.6288,
      "step": 1003
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5688,
      "step": 1004
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.5607,
      "step": 1005
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.4829,
      "step": 1006
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.4815,
      "step": 1007
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6031,
      "step": 1008
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.523,
      "step": 1009
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5669,
      "step": 1010
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5128,
      "step": 1011
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5397,
      "step": 1012
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5051,
      "step": 1013
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6593,
      "step": 1014
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8449,
      "step": 1015
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7828,
      "step": 1016
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6618,
      "step": 1017
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.654,
      "step": 1018
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.555,
      "step": 1019
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6543,
      "step": 1020
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.539,
      "step": 1021
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5369,
      "step": 1022
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7462,
      "step": 1023
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.4919,
      "step": 1024
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6517,
      "step": 1025
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6417,
      "step": 1026
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5377,
      "step": 1027
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.611,
      "step": 1028
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5285,
      "step": 1029
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8065,
      "step": 1030
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.6074,
      "step": 1031
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5452,
      "step": 1032
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5269,
      "step": 1033
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.4618,
      "step": 1034
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5843,
      "step": 1035
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5619,
      "step": 1036
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7314,
      "step": 1037
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.4987,
      "step": 1038
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.4887,
      "step": 1039
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.4573,
      "step": 1040
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.4609,
      "step": 1041
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5074,
      "step": 1042
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.7367,
      "step": 1043
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5011,
      "step": 1044
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.362,
      "step": 1045
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5725,
      "step": 1046
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.8629,
      "step": 1047
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.5021,
      "step": 1048
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5432,
      "step": 1049
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6668,
      "step": 1050
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4932,
      "step": 1051
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5024,
      "step": 1052
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4504,
      "step": 1053
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4706,
      "step": 1054
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4296,
      "step": 1055
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5156,
      "step": 1056
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.755,
      "step": 1057
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6339,
      "step": 1058
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5421,
      "step": 1059
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5279,
      "step": 1060
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4957,
      "step": 1061
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5078,
      "step": 1062
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4285,
      "step": 1063
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6866,
      "step": 1064
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5213,
      "step": 1065
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5419,
      "step": 1066
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5335,
      "step": 1067
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6135,
      "step": 1068
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4099,
      "step": 1069
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4687,
      "step": 1070
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7218,
      "step": 1071
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6365,
      "step": 1072
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.8037,
      "step": 1073
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6133,
      "step": 1074
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5327,
      "step": 1075
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5137,
      "step": 1076
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6197,
      "step": 1077
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7703,
      "step": 1078
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.8058,
      "step": 1079
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5642,
      "step": 1080
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7326,
      "step": 1081
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6517,
      "step": 1082
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4875,
      "step": 1083
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6873,
      "step": 1084
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4772,
      "step": 1085
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4758,
      "step": 1086
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.6509,
      "step": 1087
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.7724,
      "step": 1088
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4427,
      "step": 1089
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.5736,
      "step": 1090
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.4825,
      "step": 1091
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.6165,
      "step": 1092
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.504,
      "step": 1093
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.6719,
      "step": 1094
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4171,
      "step": 1095
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4914,
      "step": 1096
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.6773,
      "step": 1097
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4906,
      "step": 1098
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5104,
      "step": 1099
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.7659,
      "step": 1100
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4673,
      "step": 1101
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4495,
      "step": 1102
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5212,
      "step": 1103
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.8471,
      "step": 1104
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5485,
      "step": 1105
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5625,
      "step": 1106
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.6509,
      "step": 1107
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5105,
      "step": 1108
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5568,
      "step": 1109
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4996,
      "step": 1110
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.41,
      "step": 1111
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.7022,
      "step": 1112
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.7582,
      "step": 1113
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.555,
      "step": 1114
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5697,
      "step": 1115
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5304,
      "step": 1116
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4333,
      "step": 1117
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.595,
      "step": 1118
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5237,
      "step": 1119
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.6009,
      "step": 1120
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.476,
      "step": 1121
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5012,
      "step": 1122
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5294,
      "step": 1123
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4581,
      "step": 1124
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.3987,
      "step": 1125
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.6608,
      "step": 1126
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.9719,
      "step": 1127
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5738,
      "step": 1128
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5285,
      "step": 1129
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 1.0379,
      "step": 1130
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.5355,
      "step": 1131
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.9247,
      "step": 1132
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.7992,
      "step": 1133
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 0.4367,
      "step": 1134
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5313,
      "step": 1135
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4346,
      "step": 1136
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5513,
      "step": 1137
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6981,
      "step": 1138
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.53,
      "step": 1139
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5565,
      "step": 1140
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8145,
      "step": 1141
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5173,
      "step": 1142
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7706,
      "step": 1143
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5209,
      "step": 1144
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5082,
      "step": 1145
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8339,
      "step": 1146
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.8757,
      "step": 1147
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4769,
      "step": 1148
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6619,
      "step": 1149
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4841,
      "step": 1150
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5175,
      "step": 1151
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6764,
      "step": 1152
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4722,
      "step": 1153
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5343,
      "step": 1154
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5219,
      "step": 1155
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6622,
      "step": 1156
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7403,
      "step": 1157
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5171,
      "step": 1158
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5442,
      "step": 1159
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4686,
      "step": 1160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4958,
      "step": 1161
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5461,
      "step": 1162
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4739,
      "step": 1163
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4846,
      "step": 1164
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7006,
      "step": 1165
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7199,
      "step": 1166
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.6872,
      "step": 1167
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5438,
      "step": 1168
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.504,
      "step": 1169
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.5235,
      "step": 1170
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4927,
      "step": 1171
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.7884,
      "step": 1172
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.645,
      "step": 1173
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4354,
      "step": 1174
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.4801,
      "step": 1175
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.0002,
      "loss": 0.608,
      "step": 1176
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5054,
      "step": 1177
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5196,
      "step": 1178
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6569,
      "step": 1179
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4905,
      "step": 1180
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6101,
      "step": 1181
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6434,
      "step": 1182
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5285,
      "step": 1183
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4846,
      "step": 1184
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4896,
      "step": 1185
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6711,
      "step": 1186
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7248,
      "step": 1187
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4794,
      "step": 1188
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.537,
      "step": 1189
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6429,
      "step": 1190
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4828,
      "step": 1191
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7808,
      "step": 1192
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6188,
      "step": 1193
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5699,
      "step": 1194
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7116,
      "step": 1195
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.503,
      "step": 1196
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.537,
      "step": 1197
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5936,
      "step": 1198
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5644,
      "step": 1199
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6015,
      "step": 1200
    },
    {
      "epoch": 0.28,
      "eval_loss": 0.5674206614494324,
      "eval_runtime": 156.5334,
      "eval_samples_per_second": 6.388,
      "eval_steps_per_second": 3.194,
      "step": 1200
    },
    {
      "epoch": 0.28,
      "mmlu_eval_accuracy": 0.5949461253320095,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.5454545454545454,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.5769230769230769,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.6046511627906976,
      "mmlu_eval_accuracy_high_school_mathematics": 0.4482758620689655,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6538461538461539,
      "mmlu_eval_accuracy_high_school_physics": 0.058823529411764705,
      "mmlu_eval_accuracy_high_school_psychology": 0.8166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7209302325581395,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.31,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.7352941176470589,
      "mmlu_eval_accuracy_prehistory": 0.6285714285714286,
      "mmlu_eval_accuracy_professional_accounting": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_law": 0.3941176470588235,
      "mmlu_eval_accuracy_professional_medicine": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_psychology": 0.5797101449275363,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.4938532913166922,
      "step": 1200
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5561,
      "step": 1201
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4433,
      "step": 1202
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8624,
      "step": 1203
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5478,
      "step": 1204
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4164,
      "step": 1205
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.528,
      "step": 1206
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6696,
      "step": 1207
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6128,
      "step": 1208
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7437,
      "step": 1209
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.7396,
      "step": 1210
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.8016,
      "step": 1211
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5006,
      "step": 1212
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5829,
      "step": 1213
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6604,
      "step": 1214
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.6831,
      "step": 1215
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5013,
      "step": 1216
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.637,
      "step": 1217
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.4626,
      "step": 1218
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0002,
      "loss": 0.5139,
      "step": 1219
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6158,
      "step": 1220
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7129,
      "step": 1221
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5217,
      "step": 1222
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4532,
      "step": 1223
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5062,
      "step": 1224
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5338,
      "step": 1225
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4687,
      "step": 1226
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.3969,
      "step": 1227
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4578,
      "step": 1228
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.483,
      "step": 1229
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4673,
      "step": 1230
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5939,
      "step": 1231
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4962,
      "step": 1232
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7399,
      "step": 1233
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.65,
      "step": 1234
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5309,
      "step": 1235
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5775,
      "step": 1236
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6139,
      "step": 1237
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5719,
      "step": 1238
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5171,
      "step": 1239
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6171,
      "step": 1240
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5959,
      "step": 1241
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5829,
      "step": 1242
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7799,
      "step": 1243
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7093,
      "step": 1244
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6084,
      "step": 1245
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5777,
      "step": 1246
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4752,
      "step": 1247
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.7692,
      "step": 1248
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4459,
      "step": 1249
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.534,
      "step": 1250
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.4983,
      "step": 1251
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5579,
      "step": 1252
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.8596,
      "step": 1253
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6286,
      "step": 1254
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5275,
      "step": 1255
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5279,
      "step": 1256
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6923,
      "step": 1257
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6654,
      "step": 1258
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.5956,
      "step": 1259
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.9076,
      "step": 1260
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6042,
      "step": 1261
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.0002,
      "loss": 0.6204,
      "step": 1262
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5824,
      "step": 1263
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6873,
      "step": 1264
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5448,
      "step": 1265
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6481,
      "step": 1266
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4889,
      "step": 1267
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5832,
      "step": 1268
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6902,
      "step": 1269
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5212,
      "step": 1270
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5553,
      "step": 1271
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4541,
      "step": 1272
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5734,
      "step": 1273
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.9889,
      "step": 1274
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.759,
      "step": 1275
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4992,
      "step": 1276
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.6781,
      "step": 1277
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5218,
      "step": 1278
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5345,
      "step": 1279
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7765,
      "step": 1280
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7083,
      "step": 1281
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.8188,
      "step": 1282
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7263,
      "step": 1283
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5497,
      "step": 1284
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5702,
      "step": 1285
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4838,
      "step": 1286
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4747,
      "step": 1287
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4714,
      "step": 1288
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4796,
      "step": 1289
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5307,
      "step": 1290
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4433,
      "step": 1291
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5677,
      "step": 1292
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.7067,
      "step": 1293
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4247,
      "step": 1294
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.8314,
      "step": 1295
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5601,
      "step": 1296
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.5015,
      "step": 1297
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4913,
      "step": 1298
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4792,
      "step": 1299
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4248,
      "step": 1300
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4886,
      "step": 1301
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4883,
      "step": 1302
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4203,
      "step": 1303
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4863,
      "step": 1304
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.0002,
      "loss": 0.4349,
      "step": 1305
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4246,
      "step": 1306
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6353,
      "step": 1307
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6494,
      "step": 1308
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.5984,
      "step": 1309
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.737,
      "step": 1310
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6319,
      "step": 1311
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4834,
      "step": 1312
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.5153,
      "step": 1313
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4913,
      "step": 1314
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.7299,
      "step": 1315
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.5726,
      "step": 1316
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6116,
      "step": 1317
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.8443,
      "step": 1318
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.5645,
      "step": 1319
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.3749,
      "step": 1320
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6603,
      "step": 1321
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6363,
      "step": 1322
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.439,
      "step": 1323
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.7315,
      "step": 1324
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4767,
      "step": 1325
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6171,
      "step": 1326
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.561,
      "step": 1327
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6373,
      "step": 1328
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.498,
      "step": 1329
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4877,
      "step": 1330
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.5232,
      "step": 1331
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.692,
      "step": 1332
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4957,
      "step": 1333
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4923,
      "step": 1334
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.5999,
      "step": 1335
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4651,
      "step": 1336
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.5202,
      "step": 1337
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6627,
      "step": 1338
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6097,
      "step": 1339
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6084,
      "step": 1340
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4375,
      "step": 1341
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.5712,
      "step": 1342
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4998,
      "step": 1343
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6644,
      "step": 1344
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.5844,
      "step": 1345
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4491,
      "step": 1346
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.4598,
      "step": 1347
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 0.6772,
      "step": 1348
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5235,
      "step": 1349
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6051,
      "step": 1350
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5345,
      "step": 1351
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5766,
      "step": 1352
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4918,
      "step": 1353
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6005,
      "step": 1354
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5568,
      "step": 1355
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5168,
      "step": 1356
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4548,
      "step": 1357
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5344,
      "step": 1358
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5701,
      "step": 1359
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6546,
      "step": 1360
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7167,
      "step": 1361
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4534,
      "step": 1362
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5962,
      "step": 1363
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5498,
      "step": 1364
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4503,
      "step": 1365
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5612,
      "step": 1366
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5733,
      "step": 1367
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5229,
      "step": 1368
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6674,
      "step": 1369
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7577,
      "step": 1370
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4973,
      "step": 1371
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4764,
      "step": 1372
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6748,
      "step": 1373
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6938,
      "step": 1374
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.6805,
      "step": 1375
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5324,
      "step": 1376
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5059,
      "step": 1377
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5501,
      "step": 1378
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5176,
      "step": 1379
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4182,
      "step": 1380
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.7427,
      "step": 1381
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4651,
      "step": 1382
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.643,
      "step": 1383
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4798,
      "step": 1384
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4779,
      "step": 1385
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5271,
      "step": 1386
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5486,
      "step": 1387
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5665,
      "step": 1388
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.4605,
      "step": 1389
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 0.5332,
      "step": 1390
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4005,
      "step": 1391
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5131,
      "step": 1392
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5423,
      "step": 1393
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5211,
      "step": 1394
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5841,
      "step": 1395
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.8227,
      "step": 1396
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4847,
      "step": 1397
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5677,
      "step": 1398
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.6177,
      "step": 1399
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5877,
      "step": 1400
    },
    {
      "epoch": 0.33,
      "eval_loss": 0.5657482743263245,
      "eval_runtime": 156.5082,
      "eval_samples_per_second": 6.389,
      "eval_steps_per_second": 3.195,
      "step": 1400
    },
    {
      "epoch": 0.33,
      "mmlu_eval_accuracy": 0.5883109737038932,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.6875,
      "mmlu_eval_accuracy_college_chemistry": 0.125,
      "mmlu_eval_accuracy_college_computer_science": 0.5454545454545454,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.6666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
      "mmlu_eval_accuracy_formal_logic": 0.42857142857142855,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.627906976744186,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3103448275862069,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6923076923076923,
      "mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
      "mmlu_eval_accuracy_high_school_psychology": 0.8166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.2608695652173913,
      "mmlu_eval_accuracy_high_school_us_history": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7093023255813954,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.25,
      "mmlu_eval_accuracy_nutrition": 0.6666666666666666,
      "mmlu_eval_accuracy_philosophy": 0.7352941176470589,
      "mmlu_eval_accuracy_prehistory": 0.6571428571428571,
      "mmlu_eval_accuracy_professional_accounting": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_law": 0.3941176470588235,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.5797101449275363,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.6296296296296297,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.5046451437644486,
      "step": 1400
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4664,
      "step": 1401
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4286,
      "step": 1402
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.8561,
      "step": 1403
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4763,
      "step": 1404
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5694,
      "step": 1405
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.488,
      "step": 1406
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5169,
      "step": 1407
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5464,
      "step": 1408
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4721,
      "step": 1409
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5633,
      "step": 1410
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.6612,
      "step": 1411
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7122,
      "step": 1412
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.7981,
      "step": 1413
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.9315,
      "step": 1414
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4542,
      "step": 1415
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4926,
      "step": 1416
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.6174,
      "step": 1417
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5936,
      "step": 1418
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5907,
      "step": 1419
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5331,
      "step": 1420
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4467,
      "step": 1421
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4667,
      "step": 1422
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.6665,
      "step": 1423
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.6328,
      "step": 1424
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4966,
      "step": 1425
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4871,
      "step": 1426
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4624,
      "step": 1427
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5611,
      "step": 1428
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5641,
      "step": 1429
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4975,
      "step": 1430
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.5988,
      "step": 1431
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.4908,
      "step": 1432
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0002,
      "loss": 0.6607,
      "step": 1433
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7503,
      "step": 1434
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5438,
      "step": 1435
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6299,
      "step": 1436
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5032,
      "step": 1437
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5394,
      "step": 1438
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5446,
      "step": 1439
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5197,
      "step": 1440
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7382,
      "step": 1441
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5542,
      "step": 1442
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.9043,
      "step": 1443
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5412,
      "step": 1444
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6793,
      "step": 1445
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.877,
      "step": 1446
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7096,
      "step": 1447
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.556,
      "step": 1448
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.701,
      "step": 1449
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.7136,
      "step": 1450
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5501,
      "step": 1451
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4472,
      "step": 1452
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5764,
      "step": 1453
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6486,
      "step": 1454
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6115,
      "step": 1455
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4808,
      "step": 1456
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4711,
      "step": 1457
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5334,
      "step": 1458
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.683,
      "step": 1459
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4574,
      "step": 1460
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.3933,
      "step": 1461
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4815,
      "step": 1462
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4135,
      "step": 1463
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6532,
      "step": 1464
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5346,
      "step": 1465
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5881,
      "step": 1466
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5397,
      "step": 1467
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.6785,
      "step": 1468
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4814,
      "step": 1469
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4863,
      "step": 1470
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4381,
      "step": 1471
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5564,
      "step": 1472
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.8164,
      "step": 1473
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.4815,
      "step": 1474
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5871,
      "step": 1475
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.0002,
      "loss": 0.5207,
      "step": 1476
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5576,
      "step": 1477
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6923,
      "step": 1478
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5512,
      "step": 1479
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5192,
      "step": 1480
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5368,
      "step": 1481
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6324,
      "step": 1482
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5234,
      "step": 1483
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7793,
      "step": 1484
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5876,
      "step": 1485
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5237,
      "step": 1486
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4612,
      "step": 1487
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7466,
      "step": 1488
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5965,
      "step": 1489
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7787,
      "step": 1490
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4907,
      "step": 1491
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4689,
      "step": 1492
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5212,
      "step": 1493
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7101,
      "step": 1494
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7315,
      "step": 1495
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4291,
      "step": 1496
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4392,
      "step": 1497
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5929,
      "step": 1498
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.8046,
      "step": 1499
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5066,
      "step": 1500
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6125,
      "step": 1501
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4952,
      "step": 1502
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4411,
      "step": 1503
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4349,
      "step": 1504
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6537,
      "step": 1505
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6694,
      "step": 1506
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5361,
      "step": 1507
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5996,
      "step": 1508
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5527,
      "step": 1509
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4195,
      "step": 1510
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4686,
      "step": 1511
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.7429,
      "step": 1512
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5721,
      "step": 1513
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.6079,
      "step": 1514
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5106,
      "step": 1515
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5133,
      "step": 1516
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.5001,
      "step": 1517
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4817,
      "step": 1518
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.0002,
      "loss": 0.4235,
      "step": 1519
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5148,
      "step": 1520
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4329,
      "step": 1521
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5366,
      "step": 1522
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4629,
      "step": 1523
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4152,
      "step": 1524
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4677,
      "step": 1525
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5194,
      "step": 1526
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5632,
      "step": 1527
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4849,
      "step": 1528
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4794,
      "step": 1529
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5,
      "step": 1530
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.535,
      "step": 1531
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4764,
      "step": 1532
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5435,
      "step": 1533
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7533,
      "step": 1534
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4982,
      "step": 1535
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.449,
      "step": 1536
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5251,
      "step": 1537
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6978,
      "step": 1538
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5032,
      "step": 1539
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4362,
      "step": 1540
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7374,
      "step": 1541
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5147,
      "step": 1542
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6197,
      "step": 1543
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5581,
      "step": 1544
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.7951,
      "step": 1545
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4089,
      "step": 1546
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4832,
      "step": 1547
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6227,
      "step": 1548
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4892,
      "step": 1549
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5215,
      "step": 1550
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5886,
      "step": 1551
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5405,
      "step": 1552
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5722,
      "step": 1553
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.461,
      "step": 1554
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5051,
      "step": 1555
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4775,
      "step": 1556
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.6454,
      "step": 1557
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.4775,
      "step": 1558
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5603,
      "step": 1559
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5912,
      "step": 1560
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5508,
      "step": 1561
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.0002,
      "loss": 0.5026,
      "step": 1562
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.8878,
      "step": 1563
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5003,
      "step": 1564
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.498,
      "step": 1565
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.7652,
      "step": 1566
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.7585,
      "step": 1567
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5569,
      "step": 1568
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6822,
      "step": 1569
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5733,
      "step": 1570
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5044,
      "step": 1571
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5868,
      "step": 1572
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4853,
      "step": 1573
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4901,
      "step": 1574
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6587,
      "step": 1575
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4905,
      "step": 1576
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6015,
      "step": 1577
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4815,
      "step": 1578
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4607,
      "step": 1579
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4597,
      "step": 1580
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5593,
      "step": 1581
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5503,
      "step": 1582
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.469,
      "step": 1583
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5031,
      "step": 1584
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4925,
      "step": 1585
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5035,
      "step": 1586
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6406,
      "step": 1587
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4497,
      "step": 1588
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4458,
      "step": 1589
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5032,
      "step": 1590
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.535,
      "step": 1591
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5216,
      "step": 1592
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5508,
      "step": 1593
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5004,
      "step": 1594
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5464,
      "step": 1595
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4055,
      "step": 1596
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4973,
      "step": 1597
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4973,
      "step": 1598
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.492,
      "step": 1599
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.4819,
      "step": 1600
    },
    {
      "epoch": 0.37,
      "eval_loss": 0.5615284442901611,
      "eval_runtime": 156.5748,
      "eval_samples_per_second": 6.387,
      "eval_steps_per_second": 3.193,
      "step": 1600
    },
    {
      "epoch": 0.37,
      "mmlu_eval_accuracy": 0.5970110195389216,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.625,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.5454545454545454,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.5454545454545454,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.5384615384615384,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.6875,
      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.71875,
      "mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_geography": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.6511627906976745,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3103448275862069,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.7307692307692307,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.85,
      "mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.7777777777777778,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7674418604651163,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.25,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.7352941176470589,
      "mmlu_eval_accuracy_prehistory": 0.6571428571428571,
      "mmlu_eval_accuracy_professional_accounting": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_law": 0.3941176470588235,
      "mmlu_eval_accuracy_professional_medicine": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_psychology": 0.5942028985507246,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.7037037037037037,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.4074654315377029,
      "step": 1600
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6067,
      "step": 1601
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5516,
      "step": 1602
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.6806,
      "step": 1603
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.0002,
      "loss": 0.5449,
      "step": 1604
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5521,
      "step": 1605
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4778,
      "step": 1606
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5317,
      "step": 1607
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6864,
      "step": 1608
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4716,
      "step": 1609
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6776,
      "step": 1610
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4242,
      "step": 1611
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6505,
      "step": 1612
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6192,
      "step": 1613
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5326,
      "step": 1614
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6501,
      "step": 1615
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4377,
      "step": 1616
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4382,
      "step": 1617
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4606,
      "step": 1618
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.473,
      "step": 1619
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5291,
      "step": 1620
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4456,
      "step": 1621
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6048,
      "step": 1622
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6266,
      "step": 1623
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6341,
      "step": 1624
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5661,
      "step": 1625
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5183,
      "step": 1626
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4481,
      "step": 1627
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6931,
      "step": 1628
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5483,
      "step": 1629
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5964,
      "step": 1630
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5558,
      "step": 1631
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5675,
      "step": 1632
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4532,
      "step": 1633
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4613,
      "step": 1634
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6086,
      "step": 1635
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5751,
      "step": 1636
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6353,
      "step": 1637
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6764,
      "step": 1638
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4937,
      "step": 1639
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.4871,
      "step": 1640
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6483,
      "step": 1641
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.6198,
      "step": 1642
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.3945,
      "step": 1643
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.7017,
      "step": 1644
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5023,
      "step": 1645
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.5738,
      "step": 1646
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0002,
      "loss": 0.45,
      "step": 1647
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5058,
      "step": 1648
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5468,
      "step": 1649
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5768,
      "step": 1650
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5891,
      "step": 1651
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4842,
      "step": 1652
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4411,
      "step": 1653
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5721,
      "step": 1654
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5314,
      "step": 1655
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5022,
      "step": 1656
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4926,
      "step": 1657
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5574,
      "step": 1658
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5211,
      "step": 1659
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7751,
      "step": 1660
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7382,
      "step": 1661
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.6524,
      "step": 1662
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5117,
      "step": 1663
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5349,
      "step": 1664
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5453,
      "step": 1665
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4398,
      "step": 1666
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4903,
      "step": 1667
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.7041,
      "step": 1668
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5441,
      "step": 1669
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4512,
      "step": 1670
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4755,
      "step": 1671
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5021,
      "step": 1672
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4558,
      "step": 1673
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5794,
      "step": 1674
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5534,
      "step": 1675
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5061,
      "step": 1676
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.56,
      "step": 1677
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4926,
      "step": 1678
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5152,
      "step": 1679
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4352,
      "step": 1680
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5692,
      "step": 1681
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5363,
      "step": 1682
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.6806,
      "step": 1683
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.4541,
      "step": 1684
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5239,
      "step": 1685
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.527,
      "step": 1686
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5829,
      "step": 1687
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.676,
      "step": 1688
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.5042,
      "step": 1689
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0002,
      "loss": 0.8271,
      "step": 1690
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4463,
      "step": 1691
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5642,
      "step": 1692
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8051,
      "step": 1693
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4417,
      "step": 1694
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.497,
      "step": 1695
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.8306,
      "step": 1696
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4891,
      "step": 1697
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.6289,
      "step": 1698
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4629,
      "step": 1699
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4757,
      "step": 1700
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5285,
      "step": 1701
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4607,
      "step": 1702
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.6752,
      "step": 1703
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5625,
      "step": 1704
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5123,
      "step": 1705
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5162,
      "step": 1706
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5854,
      "step": 1707
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4388,
      "step": 1708
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4965,
      "step": 1709
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.527,
      "step": 1710
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4526,
      "step": 1711
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.6103,
      "step": 1712
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5674,
      "step": 1713
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5,
      "step": 1714
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.6085,
      "step": 1715
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5093,
      "step": 1716
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4966,
      "step": 1717
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4953,
      "step": 1718
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5159,
      "step": 1719
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5046,
      "step": 1720
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5346,
      "step": 1721
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5048,
      "step": 1722
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4928,
      "step": 1723
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4468,
      "step": 1724
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.7925,
      "step": 1725
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5817,
      "step": 1726
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4482,
      "step": 1727
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4339,
      "step": 1728
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4532,
      "step": 1729
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.5718,
      "step": 1730
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4761,
      "step": 1731
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4674,
      "step": 1732
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.0002,
      "loss": 0.4691,
      "step": 1733
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6339,
      "step": 1734
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6712,
      "step": 1735
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4598,
      "step": 1736
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4854,
      "step": 1737
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4717,
      "step": 1738
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4799,
      "step": 1739
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.5801,
      "step": 1740
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4873,
      "step": 1741
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4548,
      "step": 1742
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6211,
      "step": 1743
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4408,
      "step": 1744
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4504,
      "step": 1745
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4988,
      "step": 1746
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6651,
      "step": 1747
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4897,
      "step": 1748
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6606,
      "step": 1749
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4566,
      "step": 1750
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4956,
      "step": 1751
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4579,
      "step": 1752
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.8861,
      "step": 1753
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.5641,
      "step": 1754
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4645,
      "step": 1755
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.535,
      "step": 1756
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.5325,
      "step": 1757
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.5917,
      "step": 1758
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6506,
      "step": 1759
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4841,
      "step": 1760
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.5776,
      "step": 1761
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.562,
      "step": 1762
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.5711,
      "step": 1763
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6803,
      "step": 1764
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.694,
      "step": 1765
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6038,
      "step": 1766
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4553,
      "step": 1767
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.7632,
      "step": 1768
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.514,
      "step": 1769
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.698,
      "step": 1770
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.5709,
      "step": 1771
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.5348,
      "step": 1772
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4779,
      "step": 1773
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.4774,
      "step": 1774
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.0002,
      "loss": 0.6532,
      "step": 1775
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5731,
      "step": 1776
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5256,
      "step": 1777
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6354,
      "step": 1778
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5348,
      "step": 1779
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4995,
      "step": 1780
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.552,
      "step": 1781
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4666,
      "step": 1782
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4321,
      "step": 1783
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6531,
      "step": 1784
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6508,
      "step": 1785
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6109,
      "step": 1786
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6373,
      "step": 1787
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4792,
      "step": 1788
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5052,
      "step": 1789
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5955,
      "step": 1790
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5842,
      "step": 1791
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4804,
      "step": 1792
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.748,
      "step": 1793
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.7679,
      "step": 1794
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4856,
      "step": 1795
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4891,
      "step": 1796
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.3924,
      "step": 1797
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.7697,
      "step": 1798
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5694,
      "step": 1799
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6013,
      "step": 1800
    },
    {
      "epoch": 0.42,
      "eval_loss": 0.5581832528114319,
      "eval_runtime": 156.4509,
      "eval_samples_per_second": 6.392,
      "eval_steps_per_second": 3.196,
      "step": 1800
    },
    {
      "epoch": 0.42,
      "mmlu_eval_accuracy": 0.5672729863746866,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.625,
      "mmlu_eval_accuracy_business_ethics": 0.7272727272727273,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.18181818181818182,
      "mmlu_eval_accuracy_college_mathematics": 0.6363636363636364,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.4230769230769231,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.5,
      "mmlu_eval_accuracy_elementary_mathematics": 0.3902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.53125,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5581395348837209,
      "mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5769230769230769,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.8166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.2608695652173913,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6153846153846154,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7441860465116279,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.29,
      "mmlu_eval_accuracy_nutrition": 0.5757575757575758,
      "mmlu_eval_accuracy_philosophy": 0.6764705882352942,
      "mmlu_eval_accuracy_prehistory": 0.5714285714285714,
      "mmlu_eval_accuracy_professional_accounting": 0.5483870967741935,
      "mmlu_eval_accuracy_professional_law": 0.3764705882352941,
      "mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_psychology": 0.6376811594202898,
      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.7727272727272727,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 1.6260135416573707,
      "step": 1800
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4761,
      "step": 1801
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6958,
      "step": 1802
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.7818,
      "step": 1803
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4653,
      "step": 1804
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.652,
      "step": 1805
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6164,
      "step": 1806
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6927,
      "step": 1807
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5029,
      "step": 1808
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4931,
      "step": 1809
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6791,
      "step": 1810
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4715,
      "step": 1811
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5955,
      "step": 1812
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.6037,
      "step": 1813
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5084,
      "step": 1814
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4554,
      "step": 1815
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5246,
      "step": 1816
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.4905,
      "step": 1817
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.0002,
      "loss": 0.5158,
      "step": 1818
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.936,
      "step": 1819
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7759,
      "step": 1820
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4859,
      "step": 1821
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5246,
      "step": 1822
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5064,
      "step": 1823
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4156,
      "step": 1824
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.6803,
      "step": 1825
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5399,
      "step": 1826
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.6558,
      "step": 1827
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5253,
      "step": 1828
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5306,
      "step": 1829
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7221,
      "step": 1830
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.8114,
      "step": 1831
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5017,
      "step": 1832
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4936,
      "step": 1833
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4404,
      "step": 1834
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4589,
      "step": 1835
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7882,
      "step": 1836
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4289,
      "step": 1837
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7233,
      "step": 1838
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5233,
      "step": 1839
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4524,
      "step": 1840
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5842,
      "step": 1841
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4658,
      "step": 1842
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5771,
      "step": 1843
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.597,
      "step": 1844
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4523,
      "step": 1845
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5722,
      "step": 1846
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.611,
      "step": 1847
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.7781,
      "step": 1848
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.6866,
      "step": 1849
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.637,
      "step": 1850
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4891,
      "step": 1851
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.6278,
      "step": 1852
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.6708,
      "step": 1853
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4698,
      "step": 1854
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.858,
      "step": 1855
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5085,
      "step": 1856
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.4899,
      "step": 1857
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.8129,
      "step": 1858
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5643,
      "step": 1859
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.6885,
      "step": 1860
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.0002,
      "loss": 0.5812,
      "step": 1861
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.4403,
      "step": 1862
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6436,
      "step": 1863
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5326,
      "step": 1864
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6017,
      "step": 1865
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7217,
      "step": 1866
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5647,
      "step": 1867
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5268,
      "step": 1868
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5519,
      "step": 1869
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5384,
      "step": 1870
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.621,
      "step": 1871
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5312,
      "step": 1872
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.4206,
      "step": 1873
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.4941,
      "step": 1874
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5106,
      "step": 1875
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.4493,
      "step": 1876
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5158,
      "step": 1877
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.492,
      "step": 1878
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6339,
      "step": 1879
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5494,
      "step": 1880
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.82,
      "step": 1881
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5365,
      "step": 1882
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.4583,
      "step": 1883
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.4996,
      "step": 1884
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.4351,
      "step": 1885
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5717,
      "step": 1886
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5717,
      "step": 1887
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5003,
      "step": 1888
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7879,
      "step": 1889
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6159,
      "step": 1890
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5959,
      "step": 1891
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.7273,
      "step": 1892
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5208,
      "step": 1893
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5069,
      "step": 1894
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5726,
      "step": 1895
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5437,
      "step": 1896
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5383,
      "step": 1897
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.6649,
      "step": 1898
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.4856,
      "step": 1899
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.664,
      "step": 1900
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5272,
      "step": 1901
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5122,
      "step": 1902
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.8104,
      "step": 1903
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0002,
      "loss": 0.5017,
      "step": 1904
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6118,
      "step": 1905
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.4843,
      "step": 1906
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5217,
      "step": 1907
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.4984,
      "step": 1908
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.4319,
      "step": 1909
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5421,
      "step": 1910
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5159,
      "step": 1911
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5557,
      "step": 1912
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.481,
      "step": 1913
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.4962,
      "step": 1914
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.7196,
      "step": 1915
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5339,
      "step": 1916
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5005,
      "step": 1917
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.4782,
      "step": 1918
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.577,
      "step": 1919
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.578,
      "step": 1920
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6078,
      "step": 1921
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6085,
      "step": 1922
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5846,
      "step": 1923
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.4465,
      "step": 1924
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5344,
      "step": 1925
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6593,
      "step": 1926
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.7047,
      "step": 1927
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5365,
      "step": 1928
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5876,
      "step": 1929
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5503,
      "step": 1930
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.643,
      "step": 1931
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.4544,
      "step": 1932
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.4688,
      "step": 1933
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5135,
      "step": 1934
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6243,
      "step": 1935
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5042,
      "step": 1936
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5127,
      "step": 1937
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6285,
      "step": 1938
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6062,
      "step": 1939
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.4814,
      "step": 1940
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5859,
      "step": 1941
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6088,
      "step": 1942
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.6846,
      "step": 1943
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.7728,
      "step": 1944
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5956,
      "step": 1945
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5569,
      "step": 1946
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.0002,
      "loss": 0.5399,
      "step": 1947
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3586,
      "step": 1948
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5795,
      "step": 1949
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.6495,
      "step": 1950
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8888,
      "step": 1951
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.3513,
      "step": 1952
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5815,
      "step": 1953
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4487,
      "step": 1954
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5933,
      "step": 1955
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4663,
      "step": 1956
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5002,
      "step": 1957
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4426,
      "step": 1958
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 1.2224,
      "step": 1959
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5875,
      "step": 1960
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4629,
      "step": 1961
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.6685,
      "step": 1962
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8983,
      "step": 1963
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.7414,
      "step": 1964
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5084,
      "step": 1965
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.492,
      "step": 1966
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5051,
      "step": 1967
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4438,
      "step": 1968
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5165,
      "step": 1969
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4775,
      "step": 1970
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5659,
      "step": 1971
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4516,
      "step": 1972
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4552,
      "step": 1973
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5958,
      "step": 1974
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5337,
      "step": 1975
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5012,
      "step": 1976
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.8612,
      "step": 1977
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.6065,
      "step": 1978
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4264,
      "step": 1979
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4905,
      "step": 1980
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5914,
      "step": 1981
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4286,
      "step": 1982
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.438,
      "step": 1983
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4768,
      "step": 1984
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.698,
      "step": 1985
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4924,
      "step": 1986
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.6524,
      "step": 1987
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.5165,
      "step": 1988
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.0002,
      "loss": 0.4838,
      "step": 1989
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5028,
      "step": 1990
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7314,
      "step": 1991
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4793,
      "step": 1992
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.454,
      "step": 1993
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4461,
      "step": 1994
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.6928,
      "step": 1995
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.8108,
      "step": 1996
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3931,
      "step": 1997
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5563,
      "step": 1998
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5129,
      "step": 1999
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4939,
      "step": 2000
    },
    {
      "epoch": 0.47,
      "eval_loss": 0.5549637675285339,
      "eval_runtime": 156.485,
      "eval_samples_per_second": 6.39,
      "eval_steps_per_second": 3.195,
      "step": 2000
    },
    {
      "epoch": 0.47,
      "mmlu_eval_accuracy": 0.5912296475599125,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.5454545454545454,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.38461538461538464,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.5625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.43902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.5625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3448275862068966,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.8,
      "mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.6153846153846154,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7325581395348837,
      "mmlu_eval_accuracy_moral_disputes": 0.631578947368421,
      "mmlu_eval_accuracy_moral_scenarios": 0.33,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.7647058823529411,
      "mmlu_eval_accuracy_prehistory": 0.5714285714285714,
      "mmlu_eval_accuracy_professional_accounting": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_law": 0.4,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.6086956521739131,
      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.8636363636363636,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.5757763900893786,
      "step": 2000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4375,
      "step": 2001
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5477,
      "step": 2002
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5425,
      "step": 2003
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5008,
      "step": 2004
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4567,
      "step": 2005
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4805,
      "step": 2006
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4619,
      "step": 2007
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4746,
      "step": 2008
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5443,
      "step": 2009
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.45,
      "step": 2010
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7141,
      "step": 2011
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5204,
      "step": 2012
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.585,
      "step": 2013
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.8125,
      "step": 2014
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5266,
      "step": 2015
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.472,
      "step": 2016
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5793,
      "step": 2017
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.3965,
      "step": 2018
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5021,
      "step": 2019
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7466,
      "step": 2020
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5603,
      "step": 2021
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5101,
      "step": 2022
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.6454,
      "step": 2023
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7236,
      "step": 2024
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4839,
      "step": 2025
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.48,
      "step": 2026
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5259,
      "step": 2027
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.6038,
      "step": 2028
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.4057,
      "step": 2029
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.8373,
      "step": 2030
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.7648,
      "step": 2031
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.5307,
      "step": 2032
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4361,
      "step": 2033
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6656,
      "step": 2034
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4842,
      "step": 2035
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4077,
      "step": 2036
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.945,
      "step": 2037
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.659,
      "step": 2038
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.593,
      "step": 2039
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.58,
      "step": 2040
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.5442,
      "step": 2041
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.536,
      "step": 2042
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4483,
      "step": 2043
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4871,
      "step": 2044
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.5447,
      "step": 2045
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4724,
      "step": 2046
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4688,
      "step": 2047
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4743,
      "step": 2048
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.5471,
      "step": 2049
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6657,
      "step": 2050
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4379,
      "step": 2051
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7437,
      "step": 2052
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.556,
      "step": 2053
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.5068,
      "step": 2054
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.5437,
      "step": 2055
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4561,
      "step": 2056
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6251,
      "step": 2057
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4646,
      "step": 2058
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.5574,
      "step": 2059
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.5489,
      "step": 2060
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.535,
      "step": 2061
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4378,
      "step": 2062
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4363,
      "step": 2063
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.5359,
      "step": 2064
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6767,
      "step": 2065
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7452,
      "step": 2066
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.5,
      "step": 2067
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.676,
      "step": 2068
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6698,
      "step": 2069
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4548,
      "step": 2070
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.7114,
      "step": 2071
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.4804,
      "step": 2072
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6641,
      "step": 2073
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.6714,
      "step": 2074
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.0002,
      "loss": 0.5477,
      "step": 2075
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5696,
      "step": 2076
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4541,
      "step": 2077
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5004,
      "step": 2078
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.7476,
      "step": 2079
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5599,
      "step": 2080
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4595,
      "step": 2081
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5392,
      "step": 2082
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6914,
      "step": 2083
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6857,
      "step": 2084
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4892,
      "step": 2085
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4928,
      "step": 2086
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4085,
      "step": 2087
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4766,
      "step": 2088
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4833,
      "step": 2089
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5147,
      "step": 2090
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4597,
      "step": 2091
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6192,
      "step": 2092
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.587,
      "step": 2093
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4848,
      "step": 2094
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4976,
      "step": 2095
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4852,
      "step": 2096
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.786,
      "step": 2097
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6345,
      "step": 2098
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4511,
      "step": 2099
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5686,
      "step": 2100
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5551,
      "step": 2101
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8473,
      "step": 2102
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4674,
      "step": 2103
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5155,
      "step": 2104
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.7053,
      "step": 2105
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4958,
      "step": 2106
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.8779,
      "step": 2107
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4282,
      "step": 2108
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5588,
      "step": 2109
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6482,
      "step": 2110
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4278,
      "step": 2111
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4364,
      "step": 2112
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4924,
      "step": 2113
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5208,
      "step": 2114
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.6239,
      "step": 2115
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5652,
      "step": 2116
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.4611,
      "step": 2117
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0002,
      "loss": 0.5259,
      "step": 2118
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4519,
      "step": 2119
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4681,
      "step": 2120
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6508,
      "step": 2121
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4445,
      "step": 2122
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.444,
      "step": 2123
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5633,
      "step": 2124
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5042,
      "step": 2125
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5081,
      "step": 2126
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4988,
      "step": 2127
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4665,
      "step": 2128
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.576,
      "step": 2129
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5466,
      "step": 2130
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.8094,
      "step": 2131
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4905,
      "step": 2132
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4973,
      "step": 2133
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4962,
      "step": 2134
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5758,
      "step": 2135
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5682,
      "step": 2136
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5079,
      "step": 2137
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5279,
      "step": 2138
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5996,
      "step": 2139
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6775,
      "step": 2140
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6946,
      "step": 2141
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7358,
      "step": 2142
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7855,
      "step": 2143
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6761,
      "step": 2144
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5228,
      "step": 2145
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5606,
      "step": 2146
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4391,
      "step": 2147
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5334,
      "step": 2148
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6021,
      "step": 2149
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4463,
      "step": 2150
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4685,
      "step": 2151
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4932,
      "step": 2152
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.7269,
      "step": 2153
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.4376,
      "step": 2154
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.8875,
      "step": 2155
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.664,
      "step": 2156
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.6849,
      "step": 2157
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.761,
      "step": 2158
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5862,
      "step": 2159
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.458,
      "step": 2160
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0002,
      "loss": 0.5079,
      "step": 2161
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5498,
      "step": 2162
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.8906,
      "step": 2163
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4906,
      "step": 2164
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5159,
      "step": 2165
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4841,
      "step": 2166
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5137,
      "step": 2167
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.622,
      "step": 2168
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4414,
      "step": 2169
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4709,
      "step": 2170
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.6037,
      "step": 2171
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.467,
      "step": 2172
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4483,
      "step": 2173
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.452,
      "step": 2174
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.6255,
      "step": 2175
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5105,
      "step": 2176
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.614,
      "step": 2177
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5237,
      "step": 2178
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.6581,
      "step": 2179
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5168,
      "step": 2180
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4942,
      "step": 2181
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5129,
      "step": 2182
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.6511,
      "step": 2183
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4786,
      "step": 2184
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.6393,
      "step": 2185
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.516,
      "step": 2186
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4915,
      "step": 2187
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4943,
      "step": 2188
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5689,
      "step": 2189
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4407,
      "step": 2190
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.464,
      "step": 2191
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.6864,
      "step": 2192
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5602,
      "step": 2193
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5229,
      "step": 2194
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5388,
      "step": 2195
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5293,
      "step": 2196
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5353,
      "step": 2197
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4799,
      "step": 2198
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4574,
      "step": 2199
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4235,
      "step": 2200
    },
    {
      "epoch": 0.51,
      "eval_loss": 0.555065929889679,
      "eval_runtime": 156.4773,
      "eval_samples_per_second": 6.391,
      "eval_steps_per_second": 3.195,
      "step": 2200
    },
    {
      "epoch": 0.51,
      "mmlu_eval_accuracy": 0.5869697408958867,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.6428571428571429,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
      "mmlu_eval_accuracy_college_biology": 0.5,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.5454545454545454,
      "mmlu_eval_accuracy_college_medicine": 0.5454545454545454,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.38461538461538464,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.4375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.43902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.59375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7222222222222222,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3448275862068966,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6923076923076923,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
      "mmlu_eval_accuracy_high_school_us_history": 0.6818181818181818,
      "mmlu_eval_accuracy_high_school_world_history": 0.6153846153846154,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 1.0,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7558139534883721,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.32,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.7941176470588235,
      "mmlu_eval_accuracy_prehistory": 0.6285714285714286,
      "mmlu_eval_accuracy_professional_accounting": 0.4838709677419355,
      "mmlu_eval_accuracy_professional_law": 0.37058823529411766,
      "mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_psychology": 0.5652173913043478,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.6296296296296297,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.448187337103774,
      "step": 2200
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.6735,
      "step": 2201
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.4868,
      "step": 2202
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.0002,
      "loss": 0.5704,
      "step": 2203
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4966,
      "step": 2204
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5898,
      "step": 2205
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.7208,
      "step": 2206
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.622,
      "step": 2207
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4266,
      "step": 2208
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.6453,
      "step": 2209
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4457,
      "step": 2210
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3877,
      "step": 2211
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4774,
      "step": 2212
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5878,
      "step": 2213
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5022,
      "step": 2214
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4996,
      "step": 2215
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4699,
      "step": 2216
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4878,
      "step": 2217
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.622,
      "step": 2218
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.459,
      "step": 2219
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.6548,
      "step": 2220
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4284,
      "step": 2221
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5534,
      "step": 2222
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5714,
      "step": 2223
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4714,
      "step": 2224
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5771,
      "step": 2225
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5358,
      "step": 2226
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.6708,
      "step": 2227
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.608,
      "step": 2228
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.519,
      "step": 2229
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5188,
      "step": 2230
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.3628,
      "step": 2231
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.6863,
      "step": 2232
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4363,
      "step": 2233
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4171,
      "step": 2234
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5125,
      "step": 2235
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5506,
      "step": 2236
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.665,
      "step": 2237
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8705,
      "step": 2238
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4725,
      "step": 2239
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4963,
      "step": 2240
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4505,
      "step": 2241
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4513,
      "step": 2242
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.5411,
      "step": 2243
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.8039,
      "step": 2244
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.4972,
      "step": 2245
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.0002,
      "loss": 0.564,
      "step": 2246
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4203,
      "step": 2247
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4321,
      "step": 2248
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5556,
      "step": 2249
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4601,
      "step": 2250
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4622,
      "step": 2251
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.3925,
      "step": 2252
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5221,
      "step": 2253
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4919,
      "step": 2254
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6471,
      "step": 2255
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4316,
      "step": 2256
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8315,
      "step": 2257
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5193,
      "step": 2258
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6508,
      "step": 2259
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.8502,
      "step": 2260
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5688,
      "step": 2261
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4672,
      "step": 2262
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6417,
      "step": 2263
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7014,
      "step": 2264
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4937,
      "step": 2265
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4506,
      "step": 2266
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4996,
      "step": 2267
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5046,
      "step": 2268
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5361,
      "step": 2269
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.754,
      "step": 2270
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.7546,
      "step": 2271
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5161,
      "step": 2272
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4947,
      "step": 2273
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4569,
      "step": 2274
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5265,
      "step": 2275
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4745,
      "step": 2276
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4555,
      "step": 2277
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.598,
      "step": 2278
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5053,
      "step": 2279
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5488,
      "step": 2280
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4084,
      "step": 2281
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5397,
      "step": 2282
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.668,
      "step": 2283
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.6629,
      "step": 2284
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5198,
      "step": 2285
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.4829,
      "step": 2286
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.532,
      "step": 2287
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.5398,
      "step": 2288
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.0002,
      "loss": 0.9126,
      "step": 2289
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.6805,
      "step": 2290
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4919,
      "step": 2291
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.6936,
      "step": 2292
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.677,
      "step": 2293
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.556,
      "step": 2294
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4942,
      "step": 2295
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4667,
      "step": 2296
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4901,
      "step": 2297
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 1.0136,
      "step": 2298
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4526,
      "step": 2299
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5885,
      "step": 2300
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4989,
      "step": 2301
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5168,
      "step": 2302
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5355,
      "step": 2303
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4944,
      "step": 2304
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.47,
      "step": 2305
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5398,
      "step": 2306
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5428,
      "step": 2307
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4665,
      "step": 2308
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4907,
      "step": 2309
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4947,
      "step": 2310
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.6764,
      "step": 2311
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.3656,
      "step": 2312
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4242,
      "step": 2313
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5127,
      "step": 2314
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5137,
      "step": 2315
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4435,
      "step": 2316
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4889,
      "step": 2317
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5044,
      "step": 2318
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5849,
      "step": 2319
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.8494,
      "step": 2320
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4825,
      "step": 2321
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5299,
      "step": 2322
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.402,
      "step": 2323
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4707,
      "step": 2324
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4934,
      "step": 2325
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.9417,
      "step": 2326
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4693,
      "step": 2327
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5647,
      "step": 2328
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.6705,
      "step": 2329
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5093,
      "step": 2330
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.4592,
      "step": 2331
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.0002,
      "loss": 0.5184,
      "step": 2332
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4993,
      "step": 2333
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5323,
      "step": 2334
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5126,
      "step": 2335
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.611,
      "step": 2336
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4989,
      "step": 2337
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4638,
      "step": 2338
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4586,
      "step": 2339
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6609,
      "step": 2340
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4535,
      "step": 2341
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4496,
      "step": 2342
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6796,
      "step": 2343
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4456,
      "step": 2344
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.3645,
      "step": 2345
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4976,
      "step": 2346
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4607,
      "step": 2347
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5188,
      "step": 2348
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4115,
      "step": 2349
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6816,
      "step": 2350
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4949,
      "step": 2351
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5573,
      "step": 2352
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 1.0527,
      "step": 2353
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4306,
      "step": 2354
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5612,
      "step": 2355
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5528,
      "step": 2356
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4593,
      "step": 2357
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4398,
      "step": 2358
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.576,
      "step": 2359
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4245,
      "step": 2360
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4722,
      "step": 2361
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8112,
      "step": 2362
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.7576,
      "step": 2363
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5122,
      "step": 2364
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.8319,
      "step": 2365
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.7563,
      "step": 2366
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.6889,
      "step": 2367
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4199,
      "step": 2368
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4637,
      "step": 2369
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5005,
      "step": 2370
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5338,
      "step": 2371
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5592,
      "step": 2372
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5852,
      "step": 2373
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.4677,
      "step": 2374
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0002,
      "loss": 0.5073,
      "step": 2375
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5239,
      "step": 2376
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5123,
      "step": 2377
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5483,
      "step": 2378
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4508,
      "step": 2379
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5529,
      "step": 2380
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6548,
      "step": 2381
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5772,
      "step": 2382
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4306,
      "step": 2383
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5237,
      "step": 2384
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4838,
      "step": 2385
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.783,
      "step": 2386
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4123,
      "step": 2387
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5631,
      "step": 2388
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7165,
      "step": 2389
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5582,
      "step": 2390
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6019,
      "step": 2391
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4393,
      "step": 2392
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4656,
      "step": 2393
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5855,
      "step": 2394
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6211,
      "step": 2395
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6229,
      "step": 2396
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5994,
      "step": 2397
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4985,
      "step": 2398
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.544,
      "step": 2399
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5866,
      "step": 2400
    },
    {
      "epoch": 0.56,
      "eval_loss": 0.5536953210830688,
      "eval_runtime": 156.546,
      "eval_samples_per_second": 6.388,
      "eval_steps_per_second": 3.194,
      "step": 2400
    },
    {
      "epoch": 0.56,
      "mmlu_eval_accuracy": 0.583396653303951,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.6428571428571429,
      "mmlu_eval_accuracy_astronomy": 0.625,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5172413793103449,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.6818181818181818,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.4230769230769231,
      "mmlu_eval_accuracy_econometrics": 0.4166666666666667,
      "mmlu_eval_accuracy_electrical_engineering": 0.5,
      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.5625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.627906976744186,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3793103448275862,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.8166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 1.0,
      "mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7558139534883721,
      "mmlu_eval_accuracy_moral_disputes": 0.5,
      "mmlu_eval_accuracy_moral_scenarios": 0.27,
      "mmlu_eval_accuracy_nutrition": 0.696969696969697,
      "mmlu_eval_accuracy_philosophy": 0.6764705882352942,
      "mmlu_eval_accuracy_prehistory": 0.7142857142857143,
      "mmlu_eval_accuracy_professional_accounting": 0.5161290322580645,
      "mmlu_eval_accuracy_professional_law": 0.3941176470588235,
      "mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_psychology": 0.6086956521739131,
      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.5555555555555556,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.8947368421052632,
      "mmlu_loss": 1.514502235165464,
      "step": 2400
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7922,
      "step": 2401
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4899,
      "step": 2402
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6416,
      "step": 2403
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5435,
      "step": 2404
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5223,
      "step": 2405
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4643,
      "step": 2406
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5388,
      "step": 2407
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5367,
      "step": 2408
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6154,
      "step": 2409
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7399,
      "step": 2410
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6274,
      "step": 2411
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.464,
      "step": 2412
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.5641,
      "step": 2413
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.7154,
      "step": 2414
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4833,
      "step": 2415
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.6387,
      "step": 2416
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.0002,
      "loss": 0.4898,
      "step": 2417
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6275,
      "step": 2418
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4802,
      "step": 2419
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5201,
      "step": 2420
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4775,
      "step": 2421
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.57,
      "step": 2422
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5681,
      "step": 2423
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6642,
      "step": 2424
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6467,
      "step": 2425
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4508,
      "step": 2426
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5109,
      "step": 2427
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6055,
      "step": 2428
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.828,
      "step": 2429
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5662,
      "step": 2430
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6753,
      "step": 2431
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5253,
      "step": 2432
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.3913,
      "step": 2433
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4671,
      "step": 2434
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5336,
      "step": 2435
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6897,
      "step": 2436
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5865,
      "step": 2437
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.7459,
      "step": 2438
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4977,
      "step": 2439
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6625,
      "step": 2440
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4375,
      "step": 2441
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5055,
      "step": 2442
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4732,
      "step": 2443
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5302,
      "step": 2444
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6781,
      "step": 2445
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4534,
      "step": 2446
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.616,
      "step": 2447
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5732,
      "step": 2448
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4836,
      "step": 2449
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5603,
      "step": 2450
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6726,
      "step": 2451
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6303,
      "step": 2452
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.666,
      "step": 2453
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4946,
      "step": 2454
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6765,
      "step": 2455
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5556,
      "step": 2456
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4465,
      "step": 2457
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.5891,
      "step": 2458
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.6703,
      "step": 2459
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0002,
      "loss": 0.4837,
      "step": 2460
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.406,
      "step": 2461
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5496,
      "step": 2462
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4823,
      "step": 2463
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5026,
      "step": 2464
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5138,
      "step": 2465
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7439,
      "step": 2466
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5739,
      "step": 2467
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4823,
      "step": 2468
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.3892,
      "step": 2469
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.6393,
      "step": 2470
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.491,
      "step": 2471
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.6574,
      "step": 2472
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.6354,
      "step": 2473
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4566,
      "step": 2474
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.6685,
      "step": 2475
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5846,
      "step": 2476
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.6344,
      "step": 2477
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7378,
      "step": 2478
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4962,
      "step": 2479
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4389,
      "step": 2480
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4976,
      "step": 2481
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7907,
      "step": 2482
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.6529,
      "step": 2483
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5744,
      "step": 2484
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8811,
      "step": 2485
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.565,
      "step": 2486
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.7528,
      "step": 2487
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.577,
      "step": 2488
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4648,
      "step": 2489
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5076,
      "step": 2490
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5301,
      "step": 2491
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4728,
      "step": 2492
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.707,
      "step": 2493
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5177,
      "step": 2494
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.51,
      "step": 2495
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4073,
      "step": 2496
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4048,
      "step": 2497
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.4197,
      "step": 2498
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.6546,
      "step": 2499
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5257,
      "step": 2500
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5451,
      "step": 2501
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.5367,
      "step": 2502
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0002,
      "loss": 0.8866,
      "step": 2503
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5651,
      "step": 2504
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5466,
      "step": 2505
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.513,
      "step": 2506
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4797,
      "step": 2507
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.648,
      "step": 2508
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.712,
      "step": 2509
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5312,
      "step": 2510
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4708,
      "step": 2511
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.7273,
      "step": 2512
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4875,
      "step": 2513
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6889,
      "step": 2514
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4665,
      "step": 2515
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4621,
      "step": 2516
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.548,
      "step": 2517
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5526,
      "step": 2518
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4937,
      "step": 2519
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4173,
      "step": 2520
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6718,
      "step": 2521
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5398,
      "step": 2522
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6536,
      "step": 2523
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5226,
      "step": 2524
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.65,
      "step": 2525
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5574,
      "step": 2526
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.455,
      "step": 2527
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.8121,
      "step": 2528
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5469,
      "step": 2529
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.769,
      "step": 2530
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4255,
      "step": 2531
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5361,
      "step": 2532
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5656,
      "step": 2533
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5429,
      "step": 2534
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6578,
      "step": 2535
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.479,
      "step": 2536
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6294,
      "step": 2537
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4123,
      "step": 2538
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4737,
      "step": 2539
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4832,
      "step": 2540
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4974,
      "step": 2541
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5229,
      "step": 2542
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4177,
      "step": 2543
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.4714,
      "step": 2544
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.5264,
      "step": 2545
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0002,
      "loss": 0.6052,
      "step": 2546
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.577,
      "step": 2547
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5039,
      "step": 2548
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4882,
      "step": 2549
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5053,
      "step": 2550
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.796,
      "step": 2551
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4667,
      "step": 2552
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7894,
      "step": 2553
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.484,
      "step": 2554
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5561,
      "step": 2555
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7124,
      "step": 2556
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7326,
      "step": 2557
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4937,
      "step": 2558
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7649,
      "step": 2559
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5281,
      "step": 2560
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4944,
      "step": 2561
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.7252,
      "step": 2562
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5027,
      "step": 2563
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5513,
      "step": 2564
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4843,
      "step": 2565
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4842,
      "step": 2566
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5169,
      "step": 2567
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.6885,
      "step": 2568
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5143,
      "step": 2569
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4843,
      "step": 2570
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4819,
      "step": 2571
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4504,
      "step": 2572
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5017,
      "step": 2573
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5839,
      "step": 2574
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.472,
      "step": 2575
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4589,
      "step": 2576
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5457,
      "step": 2577
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4761,
      "step": 2578
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.516,
      "step": 2579
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.6093,
      "step": 2580
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4636,
      "step": 2581
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5781,
      "step": 2582
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5492,
      "step": 2583
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5535,
      "step": 2584
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.4586,
      "step": 2585
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5116,
      "step": 2586
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.55,
      "step": 2587
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5806,
      "step": 2588
    },
    {
      "epoch": 0.6,
      "learning_rate": 0.0002,
      "loss": 0.5945,
      "step": 2589
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.7392,
      "step": 2590
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4746,
      "step": 2591
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4513,
      "step": 2592
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5983,
      "step": 2593
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5773,
      "step": 2594
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5158,
      "step": 2595
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5463,
      "step": 2596
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5737,
      "step": 2597
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5216,
      "step": 2598
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4395,
      "step": 2599
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4219,
      "step": 2600
    },
    {
      "epoch": 0.61,
      "eval_loss": 0.5526494383811951,
      "eval_runtime": 156.4544,
      "eval_samples_per_second": 6.392,
      "eval_steps_per_second": 3.196,
      "step": 2600
    },
    {
      "epoch": 0.61,
      "mmlu_eval_accuracy": 0.5650961809667603,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.6875,
      "mmlu_eval_accuracy_college_chemistry": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.2692307692307692,
      "mmlu_eval_accuracy_econometrics": 0.5,
      "mmlu_eval_accuracy_electrical_engineering": 0.5,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4146341463414634,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.2,
      "mmlu_eval_accuracy_high_school_biology": 0.59375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5581395348837209,
      "mmlu_eval_accuracy_high_school_mathematics": 0.20689655172413793,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6538461538461539,
      "mmlu_eval_accuracy_high_school_physics": 0.058823529411764705,
      "mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.7222222222222222,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7790697674418605,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.25,
      "mmlu_eval_accuracy_nutrition": 0.6666666666666666,
      "mmlu_eval_accuracy_philosophy": 0.7058823529411765,
      "mmlu_eval_accuracy_prehistory": 0.6857142857142857,
      "mmlu_eval_accuracy_professional_accounting": 0.3548387096774194,
      "mmlu_eval_accuracy_professional_law": 0.36470588235294116,
      "mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_psychology": 0.5942028985507246,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.8636363636363636,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.3333333333333333,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 1.325617572268369,
      "step": 2600
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5658,
      "step": 2601
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5127,
      "step": 2602
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.6086,
      "step": 2603
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5805,
      "step": 2604
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.6518,
      "step": 2605
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5363,
      "step": 2606
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.7105,
      "step": 2607
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5866,
      "step": 2608
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.6497,
      "step": 2609
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5215,
      "step": 2610
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5686,
      "step": 2611
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5078,
      "step": 2612
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5716,
      "step": 2613
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5914,
      "step": 2614
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.732,
      "step": 2615
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4282,
      "step": 2616
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5306,
      "step": 2617
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.8325,
      "step": 2618
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5809,
      "step": 2619
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5481,
      "step": 2620
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5035,
      "step": 2621
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4912,
      "step": 2622
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5723,
      "step": 2623
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4848,
      "step": 2624
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5912,
      "step": 2625
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4971,
      "step": 2626
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5572,
      "step": 2627
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4835,
      "step": 2628
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.7723,
      "step": 2629
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.5401,
      "step": 2630
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002,
      "loss": 0.4111,
      "step": 2631
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5554,
      "step": 2632
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4702,
      "step": 2633
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5413,
      "step": 2634
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5266,
      "step": 2635
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5886,
      "step": 2636
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5613,
      "step": 2637
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4427,
      "step": 2638
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5351,
      "step": 2639
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4627,
      "step": 2640
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5061,
      "step": 2641
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.8022,
      "step": 2642
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.9738,
      "step": 2643
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4347,
      "step": 2644
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4381,
      "step": 2645
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4439,
      "step": 2646
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4868,
      "step": 2647
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.6802,
      "step": 2648
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5792,
      "step": 2649
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.583,
      "step": 2650
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5768,
      "step": 2651
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5772,
      "step": 2652
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5075,
      "step": 2653
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4811,
      "step": 2654
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5097,
      "step": 2655
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4644,
      "step": 2656
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4922,
      "step": 2657
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.6518,
      "step": 2658
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4557,
      "step": 2659
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.6727,
      "step": 2660
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.502,
      "step": 2661
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.661,
      "step": 2662
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.9609,
      "step": 2663
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4995,
      "step": 2664
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.3788,
      "step": 2665
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5028,
      "step": 2666
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.488,
      "step": 2667
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5424,
      "step": 2668
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4516,
      "step": 2669
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.6844,
      "step": 2670
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5781,
      "step": 2671
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.5099,
      "step": 2672
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.4326,
      "step": 2673
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002,
      "loss": 0.6551,
      "step": 2674
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5676,
      "step": 2675
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4598,
      "step": 2676
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4399,
      "step": 2677
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7509,
      "step": 2678
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4061,
      "step": 2679
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5351,
      "step": 2680
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6001,
      "step": 2681
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4865,
      "step": 2682
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4683,
      "step": 2683
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.99,
      "step": 2684
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5887,
      "step": 2685
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.533,
      "step": 2686
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5803,
      "step": 2687
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8165,
      "step": 2688
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6068,
      "step": 2689
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5439,
      "step": 2690
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5969,
      "step": 2691
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.827,
      "step": 2692
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5004,
      "step": 2693
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4656,
      "step": 2694
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5699,
      "step": 2695
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5098,
      "step": 2696
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5335,
      "step": 2697
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4356,
      "step": 2698
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5511,
      "step": 2699
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8263,
      "step": 2700
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5072,
      "step": 2701
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4397,
      "step": 2702
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6,
      "step": 2703
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7959,
      "step": 2704
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.3753,
      "step": 2705
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5592,
      "step": 2706
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5736,
      "step": 2707
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4628,
      "step": 2708
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.9145,
      "step": 2709
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5221,
      "step": 2710
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4725,
      "step": 2711
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.8752,
      "step": 2712
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4402,
      "step": 2713
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.4812,
      "step": 2714
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.5641,
      "step": 2715
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.7977,
      "step": 2716
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002,
      "loss": 0.6231,
      "step": 2717
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5159,
      "step": 2718
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4075,
      "step": 2719
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3836,
      "step": 2720
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6574,
      "step": 2721
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6744,
      "step": 2722
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.7766,
      "step": 2723
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5129,
      "step": 2724
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3558,
      "step": 2725
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6098,
      "step": 2726
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4866,
      "step": 2727
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6105,
      "step": 2728
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6941,
      "step": 2729
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5913,
      "step": 2730
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4479,
      "step": 2731
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4829,
      "step": 2732
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.3843,
      "step": 2733
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4275,
      "step": 2734
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6686,
      "step": 2735
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6448,
      "step": 2736
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5363,
      "step": 2737
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5745,
      "step": 2738
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5183,
      "step": 2739
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5688,
      "step": 2740
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.394,
      "step": 2741
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4395,
      "step": 2742
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6124,
      "step": 2743
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4622,
      "step": 2744
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4865,
      "step": 2745
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4833,
      "step": 2746
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.6064,
      "step": 2747
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5204,
      "step": 2748
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4195,
      "step": 2749
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4576,
      "step": 2750
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4543,
      "step": 2751
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.7885,
      "step": 2752
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4495,
      "step": 2753
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5286,
      "step": 2754
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5645,
      "step": 2755
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.4915,
      "step": 2756
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.526,
      "step": 2757
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5337,
      "step": 2758
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.7409,
      "step": 2759
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.0002,
      "loss": 0.5983,
      "step": 2760
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.4097,
      "step": 2761
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7308,
      "step": 2762
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5139,
      "step": 2763
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5352,
      "step": 2764
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5003,
      "step": 2765
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.501,
      "step": 2766
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.4714,
      "step": 2767
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.4776,
      "step": 2768
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6335,
      "step": 2769
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6848,
      "step": 2770
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5843,
      "step": 2771
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.4964,
      "step": 2772
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6461,
      "step": 2773
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5202,
      "step": 2774
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5294,
      "step": 2775
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.592,
      "step": 2776
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.4489,
      "step": 2777
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6752,
      "step": 2778
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5924,
      "step": 2779
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6239,
      "step": 2780
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6685,
      "step": 2781
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6085,
      "step": 2782
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5022,
      "step": 2783
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5114,
      "step": 2784
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.522,
      "step": 2785
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.4389,
      "step": 2786
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5854,
      "step": 2787
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6052,
      "step": 2788
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5455,
      "step": 2789
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5936,
      "step": 2790
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5028,
      "step": 2791
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.4433,
      "step": 2792
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6075,
      "step": 2793
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5436,
      "step": 2794
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7084,
      "step": 2795
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.4684,
      "step": 2796
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.6051,
      "step": 2797
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.4806,
      "step": 2798
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5522,
      "step": 2799
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.456,
      "step": 2800
    },
    {
      "epoch": 0.65,
      "eval_loss": 0.5507346987724304,
      "eval_runtime": 156.5133,
      "eval_samples_per_second": 6.389,
      "eval_steps_per_second": 3.195,
      "step": 2800
    },
    {
      "epoch": 0.65,
      "mmlu_eval_accuracy": 0.5872588004921844,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
      "mmlu_eval_accuracy_college_biology": 0.5,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
      "mmlu_eval_accuracy_conceptual_physics": 0.4230769230769231,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.6875,
      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.5625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.45454545454545453,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5581395348837209,
      "mmlu_eval_accuracy_high_school_mathematics": 0.4827586206896552,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6538461538461539,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.8166666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6153846153846154,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7209302325581395,
      "mmlu_eval_accuracy_moral_disputes": 0.5,
      "mmlu_eval_accuracy_moral_scenarios": 0.29,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.7352941176470589,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_law": 0.4117647058823529,
      "mmlu_eval_accuracy_professional_medicine": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_psychology": 0.5652173913043478,
      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.6666666666666666,
      "mmlu_eval_accuracy_sociology": 0.9545454545454546,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.5878387242944683,
      "step": 2800
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.5296,
      "step": 2801
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.4986,
      "step": 2802
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002,
      "loss": 0.7447,
      "step": 2803
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5691,
      "step": 2804
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6287,
      "step": 2805
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7666,
      "step": 2806
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5311,
      "step": 2807
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.437,
      "step": 2808
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8734,
      "step": 2809
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5773,
      "step": 2810
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4161,
      "step": 2811
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5626,
      "step": 2812
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4786,
      "step": 2813
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4865,
      "step": 2814
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6513,
      "step": 2815
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6227,
      "step": 2816
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4513,
      "step": 2817
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5653,
      "step": 2818
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6353,
      "step": 2819
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4497,
      "step": 2820
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5405,
      "step": 2821
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6025,
      "step": 2822
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4665,
      "step": 2823
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6243,
      "step": 2824
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4273,
      "step": 2825
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4457,
      "step": 2826
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5046,
      "step": 2827
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6276,
      "step": 2828
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5723,
      "step": 2829
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5905,
      "step": 2830
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4518,
      "step": 2831
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4431,
      "step": 2832
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.8389,
      "step": 2833
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.7834,
      "step": 2834
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4024,
      "step": 2835
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4219,
      "step": 2836
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4425,
      "step": 2837
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.522,
      "step": 2838
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.475,
      "step": 2839
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6112,
      "step": 2840
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.5913,
      "step": 2841
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.4883,
      "step": 2842
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6446,
      "step": 2843
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.53,
      "step": 2844
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0002,
      "loss": 0.6041,
      "step": 2845
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6139,
      "step": 2846
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4835,
      "step": 2847
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4881,
      "step": 2848
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.7695,
      "step": 2849
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5394,
      "step": 2850
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.799,
      "step": 2851
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5576,
      "step": 2852
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.405,
      "step": 2853
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5345,
      "step": 2854
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.494,
      "step": 2855
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4859,
      "step": 2856
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.442,
      "step": 2857
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4976,
      "step": 2858
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.462,
      "step": 2859
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6547,
      "step": 2860
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6362,
      "step": 2861
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4308,
      "step": 2862
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8681,
      "step": 2863
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5287,
      "step": 2864
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5224,
      "step": 2865
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6529,
      "step": 2866
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4994,
      "step": 2867
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.8907,
      "step": 2868
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.487,
      "step": 2869
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5312,
      "step": 2870
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5027,
      "step": 2871
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 1.1338,
      "step": 2872
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.885,
      "step": 2873
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5016,
      "step": 2874
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6298,
      "step": 2875
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4999,
      "step": 2876
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5163,
      "step": 2877
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.7066,
      "step": 2878
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5874,
      "step": 2879
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.7439,
      "step": 2880
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4239,
      "step": 2881
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4249,
      "step": 2882
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5049,
      "step": 2883
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.5313,
      "step": 2884
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.6481,
      "step": 2885
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4396,
      "step": 2886
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4021,
      "step": 2887
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.0002,
      "loss": 0.4054,
      "step": 2888
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5385,
      "step": 2889
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.6555,
      "step": 2890
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5795,
      "step": 2891
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.4323,
      "step": 2892
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.475,
      "step": 2893
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.653,
      "step": 2894
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.6961,
      "step": 2895
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.4993,
      "step": 2896
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5598,
      "step": 2897
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5972,
      "step": 2898
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7518,
      "step": 2899
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.3858,
      "step": 2900
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.533,
      "step": 2901
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5269,
      "step": 2902
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5142,
      "step": 2903
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.504,
      "step": 2904
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.712,
      "step": 2905
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.448,
      "step": 2906
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5042,
      "step": 2907
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5848,
      "step": 2908
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.6219,
      "step": 2909
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5108,
      "step": 2910
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5935,
      "step": 2911
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7508,
      "step": 2912
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5701,
      "step": 2913
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5743,
      "step": 2914
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5986,
      "step": 2915
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5493,
      "step": 2916
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.503,
      "step": 2917
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5377,
      "step": 2918
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5378,
      "step": 2919
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5353,
      "step": 2920
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5225,
      "step": 2921
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.9093,
      "step": 2922
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.4567,
      "step": 2923
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.5224,
      "step": 2924
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.6005,
      "step": 2925
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.4817,
      "step": 2926
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.7119,
      "step": 2927
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.6414,
      "step": 2928
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.583,
      "step": 2929
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.8812,
      "step": 2930
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.0002,
      "loss": 0.6389,
      "step": 2931
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5041,
      "step": 2932
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4448,
      "step": 2933
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5126,
      "step": 2934
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6328,
      "step": 2935
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5874,
      "step": 2936
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5366,
      "step": 2937
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6859,
      "step": 2938
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6329,
      "step": 2939
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5833,
      "step": 2940
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4965,
      "step": 2941
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4538,
      "step": 2942
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8644,
      "step": 2943
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4519,
      "step": 2944
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5309,
      "step": 2945
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5227,
      "step": 2946
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5237,
      "step": 2947
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.499,
      "step": 2948
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4135,
      "step": 2949
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4809,
      "step": 2950
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4734,
      "step": 2951
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4423,
      "step": 2952
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5538,
      "step": 2953
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4734,
      "step": 2954
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.9252,
      "step": 2955
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5282,
      "step": 2956
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6778,
      "step": 2957
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.6546,
      "step": 2958
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5014,
      "step": 2959
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5524,
      "step": 2960
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.7408,
      "step": 2961
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4465,
      "step": 2962
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4546,
      "step": 2963
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5385,
      "step": 2964
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5995,
      "step": 2965
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5655,
      "step": 2966
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.8734,
      "step": 2967
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5345,
      "step": 2968
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.4719,
      "step": 2969
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5682,
      "step": 2970
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5271,
      "step": 2971
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.5624,
      "step": 2972
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.7705,
      "step": 2973
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002,
      "loss": 0.7777,
      "step": 2974
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.9383,
      "step": 2975
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.7205,
      "step": 2976
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4651,
      "step": 2977
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5677,
      "step": 2978
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.3815,
      "step": 2979
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4236,
      "step": 2980
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6083,
      "step": 2981
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5722,
      "step": 2982
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5198,
      "step": 2983
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.7749,
      "step": 2984
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6183,
      "step": 2985
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5131,
      "step": 2986
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6658,
      "step": 2987
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4659,
      "step": 2988
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4971,
      "step": 2989
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5203,
      "step": 2990
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6545,
      "step": 2991
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5241,
      "step": 2992
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4294,
      "step": 2993
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5729,
      "step": 2994
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5272,
      "step": 2995
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.619,
      "step": 2996
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4638,
      "step": 2997
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4095,
      "step": 2998
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4462,
      "step": 2999
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5813,
      "step": 3000
    },
    {
      "epoch": 0.7,
      "eval_loss": 0.548480749130249,
      "eval_runtime": 156.5062,
      "eval_samples_per_second": 6.39,
      "eval_steps_per_second": 3.195,
      "step": 3000
    },
    {
      "epoch": 0.7,
      "mmlu_eval_accuracy": 0.5849841642453101,
      "mmlu_eval_accuracy_abstract_algebra": 0.45454545454545453,
      "mmlu_eval_accuracy_anatomy": 0.35714285714285715,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.5,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.38461538461538464,
      "mmlu_eval_accuracy_econometrics": 0.6666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.4375,
      "mmlu_eval_accuracy_elementary_mathematics": 0.3170731707317073,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.5,
      "mmlu_eval_accuracy_high_school_biology": 0.59375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.5454545454545454,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6538461538461539,
      "mmlu_eval_accuracy_high_school_physics": 0.058823529411764705,
      "mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_statistics": 0.4782608695652174,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7093023255813954,
      "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.3,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.7941176470588235,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.3870967741935484,
      "mmlu_eval_accuracy_professional_law": 0.3588235294117647,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.5652173913043478,
      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.7037037037037037,
      "mmlu_eval_accuracy_sociology": 0.8636363636363636,
      "mmlu_eval_accuracy_us_foreign_policy": 1.0,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.8947368421052632,
      "mmlu_loss": 1.503230548636409,
      "step": 3000
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.443,
      "step": 3001
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4527,
      "step": 3002
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6654,
      "step": 3003
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6315,
      "step": 3004
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4676,
      "step": 3005
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4909,
      "step": 3006
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5713,
      "step": 3007
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.479,
      "step": 3008
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4386,
      "step": 3009
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4831,
      "step": 3010
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.4451,
      "step": 3011
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5618,
      "step": 3012
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.7292,
      "step": 3013
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.5216,
      "step": 3014
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6088,
      "step": 3015
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6084,
      "step": 3016
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0002,
      "loss": 0.6614,
      "step": 3017
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5934,
      "step": 3018
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6926,
      "step": 3019
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5579,
      "step": 3020
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4561,
      "step": 3021
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6117,
      "step": 3022
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5053,
      "step": 3023
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6792,
      "step": 3024
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5546,
      "step": 3025
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5446,
      "step": 3026
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.529,
      "step": 3027
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5155,
      "step": 3028
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6162,
      "step": 3029
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4376,
      "step": 3030
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.7182,
      "step": 3031
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4977,
      "step": 3032
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.8069,
      "step": 3033
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6473,
      "step": 3034
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.569,
      "step": 3035
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4647,
      "step": 3036
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4605,
      "step": 3037
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4439,
      "step": 3038
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4396,
      "step": 3039
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5137,
      "step": 3040
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4583,
      "step": 3041
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4734,
      "step": 3042
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4869,
      "step": 3043
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5655,
      "step": 3044
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.54,
      "step": 3045
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5379,
      "step": 3046
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5067,
      "step": 3047
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4795,
      "step": 3048
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4585,
      "step": 3049
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.53,
      "step": 3050
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4376,
      "step": 3051
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4345,
      "step": 3052
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4951,
      "step": 3053
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.601,
      "step": 3054
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.6107,
      "step": 3055
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4385,
      "step": 3056
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5645,
      "step": 3057
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.4917,
      "step": 3058
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.0002,
      "loss": 0.5079,
      "step": 3059
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5186,
      "step": 3060
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.551,
      "step": 3061
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.381,
      "step": 3062
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5838,
      "step": 3063
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5097,
      "step": 3064
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4142,
      "step": 3065
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4384,
      "step": 3066
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4837,
      "step": 3067
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5799,
      "step": 3068
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6609,
      "step": 3069
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5031,
      "step": 3070
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.496,
      "step": 3071
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4411,
      "step": 3072
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4838,
      "step": 3073
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4681,
      "step": 3074
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5294,
      "step": 3075
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4753,
      "step": 3076
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6275,
      "step": 3077
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5263,
      "step": 3078
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4884,
      "step": 3079
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.7181,
      "step": 3080
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.8531,
      "step": 3081
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6743,
      "step": 3082
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5309,
      "step": 3083
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.433,
      "step": 3084
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5487,
      "step": 3085
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4131,
      "step": 3086
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4431,
      "step": 3087
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5506,
      "step": 3088
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4333,
      "step": 3089
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.563,
      "step": 3090
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6248,
      "step": 3091
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4603,
      "step": 3092
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4815,
      "step": 3093
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.6823,
      "step": 3094
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4447,
      "step": 3095
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5067,
      "step": 3096
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5007,
      "step": 3097
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.731,
      "step": 3098
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.5783,
      "step": 3099
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4924,
      "step": 3100
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.602,
      "step": 3101
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.0002,
      "loss": 0.4924,
      "step": 3102
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.684,
      "step": 3103
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4662,
      "step": 3104
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.5615,
      "step": 3105
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4434,
      "step": 3106
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6725,
      "step": 3107
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6568,
      "step": 3108
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.5151,
      "step": 3109
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6216,
      "step": 3110
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6289,
      "step": 3111
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.5379,
      "step": 3112
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.5945,
      "step": 3113
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.5135,
      "step": 3114
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4462,
      "step": 3115
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.5907,
      "step": 3116
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4836,
      "step": 3117
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4673,
      "step": 3118
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4436,
      "step": 3119
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7638,
      "step": 3120
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4161,
      "step": 3121
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4265,
      "step": 3122
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4377,
      "step": 3123
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4576,
      "step": 3124
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4196,
      "step": 3125
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.505,
      "step": 3126
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.7268,
      "step": 3127
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.5936,
      "step": 3128
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4834,
      "step": 3129
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.3972,
      "step": 3130
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.5891,
      "step": 3131
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.5456,
      "step": 3132
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.8447,
      "step": 3133
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4865,
      "step": 3134
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4223,
      "step": 3135
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6468,
      "step": 3136
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6294,
      "step": 3137
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.5005,
      "step": 3138
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4357,
      "step": 3139
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.596,
      "step": 3140
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.4885,
      "step": 3141
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6416,
      "step": 3142
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6022,
      "step": 3143
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.6331,
      "step": 3144
    },
    {
      "epoch": 0.73,
      "learning_rate": 0.0002,
      "loss": 0.668,
      "step": 3145
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5341,
      "step": 3146
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4731,
      "step": 3147
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7354,
      "step": 3148
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.445,
      "step": 3149
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4783,
      "step": 3150
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4821,
      "step": 3151
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.673,
      "step": 3152
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4925,
      "step": 3153
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5812,
      "step": 3154
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5154,
      "step": 3155
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5833,
      "step": 3156
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5628,
      "step": 3157
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6664,
      "step": 3158
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6048,
      "step": 3159
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4922,
      "step": 3160
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4628,
      "step": 3161
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5397,
      "step": 3162
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4392,
      "step": 3163
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6108,
      "step": 3164
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6505,
      "step": 3165
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5053,
      "step": 3166
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4675,
      "step": 3167
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6287,
      "step": 3168
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6129,
      "step": 3169
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5892,
      "step": 3170
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6856,
      "step": 3171
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5144,
      "step": 3172
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5804,
      "step": 3173
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4754,
      "step": 3174
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4699,
      "step": 3175
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4849,
      "step": 3176
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4657,
      "step": 3177
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5887,
      "step": 3178
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.6413,
      "step": 3179
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4601,
      "step": 3180
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4356,
      "step": 3181
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.7733,
      "step": 3182
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5405,
      "step": 3183
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5879,
      "step": 3184
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.458,
      "step": 3185
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.5474,
      "step": 3186
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4592,
      "step": 3187
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.0002,
      "loss": 0.4762,
      "step": 3188
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5146,
      "step": 3189
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4771,
      "step": 3190
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5016,
      "step": 3191
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6579,
      "step": 3192
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6508,
      "step": 3193
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4833,
      "step": 3194
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6775,
      "step": 3195
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4587,
      "step": 3196
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5559,
      "step": 3197
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5812,
      "step": 3198
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6066,
      "step": 3199
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4908,
      "step": 3200
    },
    {
      "epoch": 0.75,
      "eval_loss": 0.5493218302726746,
      "eval_runtime": 156.5298,
      "eval_samples_per_second": 6.389,
      "eval_steps_per_second": 3.194,
      "step": 3200
    },
    {
      "epoch": 0.75,
      "mmlu_eval_accuracy": 0.5907709069481004,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.75,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
      "mmlu_eval_accuracy_college_mathematics": 0.6363636363636364,
      "mmlu_eval_accuracy_college_medicine": 0.5454545454545454,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
      "mmlu_eval_accuracy_conceptual_physics": 0.4230769230769231,
      "mmlu_eval_accuracy_econometrics": 0.6666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5348837209302325,
      "mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5769230769230769,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_statistics": 0.4782608695652174,
      "mmlu_eval_accuracy_high_school_us_history": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_world_history": 0.6153846153846154,
      "mmlu_eval_accuracy_human_aging": 0.782608695652174,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7325581395348837,
      "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.29,
      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
      "mmlu_eval_accuracy_philosophy": 0.7352941176470589,
      "mmlu_eval_accuracy_prehistory": 0.5714285714285714,
      "mmlu_eval_accuracy_professional_accounting": 0.45161290322580644,
      "mmlu_eval_accuracy_professional_law": 0.38823529411764707,
      "mmlu_eval_accuracy_professional_medicine": 0.5806451612903226,
      "mmlu_eval_accuracy_professional_psychology": 0.5797101449275363,
      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
      "mmlu_eval_accuracy_security_studies": 0.6666666666666666,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.6111111111111112,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.3444029226626801,
      "step": 3200
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7056,
      "step": 3201
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.838,
      "step": 3202
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4546,
      "step": 3203
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5579,
      "step": 3204
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6715,
      "step": 3205
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5638,
      "step": 3206
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6521,
      "step": 3207
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4793,
      "step": 3208
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.8167,
      "step": 3209
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6209,
      "step": 3210
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4857,
      "step": 3211
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5092,
      "step": 3212
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4311,
      "step": 3213
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6642,
      "step": 3214
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6284,
      "step": 3215
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4679,
      "step": 3216
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5038,
      "step": 3217
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5059,
      "step": 3218
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.616,
      "step": 3219
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5023,
      "step": 3220
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.7524,
      "step": 3221
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3711,
      "step": 3222
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4621,
      "step": 3223
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5954,
      "step": 3224
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.4913,
      "step": 3225
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6274,
      "step": 3226
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.6327,
      "step": 3227
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.797,
      "step": 3228
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5098,
      "step": 3229
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.3747,
      "step": 3230
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.0002,
      "loss": 0.5217,
      "step": 3231
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5441,
      "step": 3232
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5319,
      "step": 3233
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4917,
      "step": 3234
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4789,
      "step": 3235
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.6211,
      "step": 3236
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5144,
      "step": 3237
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5042,
      "step": 3238
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.648,
      "step": 3239
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5369,
      "step": 3240
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5279,
      "step": 3241
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.6061,
      "step": 3242
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.6342,
      "step": 3243
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.457,
      "step": 3244
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.7314,
      "step": 3245
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.616,
      "step": 3246
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4033,
      "step": 3247
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4649,
      "step": 3248
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.6434,
      "step": 3249
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4925,
      "step": 3250
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4935,
      "step": 3251
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5982,
      "step": 3252
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.467,
      "step": 3253
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5383,
      "step": 3254
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5287,
      "step": 3255
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.3909,
      "step": 3256
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.8876,
      "step": 3257
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4604,
      "step": 3258
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5341,
      "step": 3259
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4903,
      "step": 3260
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.6445,
      "step": 3261
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4862,
      "step": 3262
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5513,
      "step": 3263
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.6098,
      "step": 3264
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4238,
      "step": 3265
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5021,
      "step": 3266
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5244,
      "step": 3267
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4665,
      "step": 3268
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4545,
      "step": 3269
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.4772,
      "step": 3270
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.6804,
      "step": 3271
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.5604,
      "step": 3272
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0002,
      "loss": 0.6321,
      "step": 3273
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.619,
      "step": 3274
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4635,
      "step": 3275
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3946,
      "step": 3276
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.516,
      "step": 3277
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4216,
      "step": 3278
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.661,
      "step": 3279
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4844,
      "step": 3280
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6239,
      "step": 3281
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.5186,
      "step": 3282
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.5895,
      "step": 3283
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.3908,
      "step": 3284
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6019,
      "step": 3285
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.5514,
      "step": 3286
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.673,
      "step": 3287
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6583,
      "step": 3288
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4964,
      "step": 3289
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.7001,
      "step": 3290
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4729,
      "step": 3291
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6289,
      "step": 3292
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4963,
      "step": 3293
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6841,
      "step": 3294
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.6164,
      "step": 3295
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.5204,
      "step": 3296
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.5176,
      "step": 3297
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.5505,
      "step": 3298
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8464,
      "step": 3299
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4928,
      "step": 3300
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.5976,
      "step": 3301
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.7073,
      "step": 3302
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4122,
      "step": 3303
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4973,
      "step": 3304
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.5049,
      "step": 3305
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4533,
      "step": 3306
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.5147,
      "step": 3307
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8765,
      "step": 3308
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4943,
      "step": 3309
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4968,
      "step": 3310
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.8363,
      "step": 3311
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.5038,
      "step": 3312
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.7593,
      "step": 3313
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4971,
      "step": 3314
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.462,
      "step": 3315
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0002,
      "loss": 0.4562,
      "step": 3316
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4399,
      "step": 3317
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.7184,
      "step": 3318
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6004,
      "step": 3319
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.5461,
      "step": 3320
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.491,
      "step": 3321
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4413,
      "step": 3322
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4728,
      "step": 3323
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4999,
      "step": 3324
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.8046,
      "step": 3325
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4168,
      "step": 3326
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4954,
      "step": 3327
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.7505,
      "step": 3328
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4836,
      "step": 3329
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.5961,
      "step": 3330
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4473,
      "step": 3331
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6527,
      "step": 3332
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.5178,
      "step": 3333
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4547,
      "step": 3334
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6277,
      "step": 3335
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4872,
      "step": 3336
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4866,
      "step": 3337
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.567,
      "step": 3338
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4836,
      "step": 3339
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.5358,
      "step": 3340
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6312,
      "step": 3341
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6738,
      "step": 3342
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6273,
      "step": 3343
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4768,
      "step": 3344
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6021,
      "step": 3345
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6177,
      "step": 3346
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4477,
      "step": 3347
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4691,
      "step": 3348
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.7269,
      "step": 3349
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4921,
      "step": 3350
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.5842,
      "step": 3351
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6037,
      "step": 3352
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.3795,
      "step": 3353
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.5596,
      "step": 3354
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.4804,
      "step": 3355
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.5172,
      "step": 3356
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6686,
      "step": 3357
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6938,
      "step": 3358
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.0002,
      "loss": 0.6945,
      "step": 3359
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5272,
      "step": 3360
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4674,
      "step": 3361
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.703,
      "step": 3362
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4695,
      "step": 3363
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.378,
      "step": 3364
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6008,
      "step": 3365
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5143,
      "step": 3366
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4784,
      "step": 3367
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4681,
      "step": 3368
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5311,
      "step": 3369
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.999,
      "step": 3370
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6137,
      "step": 3371
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4533,
      "step": 3372
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5276,
      "step": 3373
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4479,
      "step": 3374
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.483,
      "step": 3375
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4897,
      "step": 3376
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.486,
      "step": 3377
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5133,
      "step": 3378
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5583,
      "step": 3379
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5924,
      "step": 3380
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.403,
      "step": 3381
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5265,
      "step": 3382
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.515,
      "step": 3383
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4478,
      "step": 3384
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5478,
      "step": 3385
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5393,
      "step": 3386
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6139,
      "step": 3387
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5572,
      "step": 3388
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.51,
      "step": 3389
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4476,
      "step": 3390
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6481,
      "step": 3391
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5129,
      "step": 3392
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.7296,
      "step": 3393
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4581,
      "step": 3394
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4951,
      "step": 3395
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.5525,
      "step": 3396
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4748,
      "step": 3397
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4527,
      "step": 3398
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6103,
      "step": 3399
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4414,
      "step": 3400
    },
    {
      "epoch": 0.79,
      "eval_loss": 0.5450515151023865,
      "eval_runtime": 156.4386,
      "eval_samples_per_second": 6.392,
      "eval_steps_per_second": 3.196,
      "step": 3400
    },
    {
      "epoch": 0.79,
      "mmlu_eval_accuracy": 0.5788038375017743,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.2692307692307692,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.6875,
      "mmlu_eval_accuracy_elementary_mathematics": 0.43902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.65625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.4883720930232558,
      "mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6153846153846154,
      "mmlu_eval_accuracy_high_school_physics": 0.058823529411764705,
      "mmlu_eval_accuracy_high_school_psychology": 0.8833333333333333,
      "mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.88,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7325581395348837,
      "mmlu_eval_accuracy_moral_disputes": 0.5789473684210527,
      "mmlu_eval_accuracy_moral_scenarios": 0.31,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.6764705882352942,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.41935483870967744,
      "mmlu_eval_accuracy_professional_law": 0.38235294117647056,
      "mmlu_eval_accuracy_professional_medicine": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_psychology": 0.5797101449275363,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.5925925925925926,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.5,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 1.5362164639228317,
      "step": 3400
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.6692,
      "step": 3401
    },
    {
      "epoch": 0.79,
      "learning_rate": 0.0002,
      "loss": 0.4034,
      "step": 3402
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4778,
      "step": 3403
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5585,
      "step": 3404
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4567,
      "step": 3405
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3734,
      "step": 3406
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4526,
      "step": 3407
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5414,
      "step": 3408
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4026,
      "step": 3409
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4393,
      "step": 3410
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4932,
      "step": 3411
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4141,
      "step": 3412
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5479,
      "step": 3413
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4927,
      "step": 3414
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7282,
      "step": 3415
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.447,
      "step": 3416
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4431,
      "step": 3417
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7482,
      "step": 3418
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5583,
      "step": 3419
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7006,
      "step": 3420
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5941,
      "step": 3421
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7707,
      "step": 3422
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.6699,
      "step": 3423
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.6736,
      "step": 3424
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.3969,
      "step": 3425
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.418,
      "step": 3426
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4553,
      "step": 3427
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.629,
      "step": 3428
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4547,
      "step": 3429
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.6113,
      "step": 3430
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5378,
      "step": 3431
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5057,
      "step": 3432
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.729,
      "step": 3433
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4148,
      "step": 3434
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5099,
      "step": 3435
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5032,
      "step": 3436
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.8189,
      "step": 3437
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5366,
      "step": 3438
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4936,
      "step": 3439
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.4673,
      "step": 3440
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7143,
      "step": 3441
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5555,
      "step": 3442
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.5378,
      "step": 3443
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.0002,
      "loss": 0.7652,
      "step": 3444
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6052,
      "step": 3445
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.498,
      "step": 3446
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.5081,
      "step": 3447
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.559,
      "step": 3448
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.5645,
      "step": 3449
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4709,
      "step": 3450
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.492,
      "step": 3451
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6313,
      "step": 3452
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6182,
      "step": 3453
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4579,
      "step": 3454
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4495,
      "step": 3455
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4844,
      "step": 3456
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4923,
      "step": 3457
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.551,
      "step": 3458
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.5313,
      "step": 3459
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4689,
      "step": 3460
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.7182,
      "step": 3461
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.3975,
      "step": 3462
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.465,
      "step": 3463
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.5703,
      "step": 3464
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6043,
      "step": 3465
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4602,
      "step": 3466
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.631,
      "step": 3467
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6587,
      "step": 3468
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.615,
      "step": 3469
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.5822,
      "step": 3470
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.8604,
      "step": 3471
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.7347,
      "step": 3472
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6159,
      "step": 3473
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.5365,
      "step": 3474
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4625,
      "step": 3475
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4003,
      "step": 3476
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4254,
      "step": 3477
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.657,
      "step": 3478
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6694,
      "step": 3479
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4701,
      "step": 3480
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.5278,
      "step": 3481
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.5924,
      "step": 3482
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.5476,
      "step": 3483
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4692,
      "step": 3484
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.4185,
      "step": 3485
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.6283,
      "step": 3486
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0002,
      "loss": 0.5017,
      "step": 3487
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5357,
      "step": 3488
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.7162,
      "step": 3489
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5405,
      "step": 3490
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5531,
      "step": 3491
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4568,
      "step": 3492
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4714,
      "step": 3493
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4478,
      "step": 3494
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5526,
      "step": 3495
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.757,
      "step": 3496
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4096,
      "step": 3497
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5515,
      "step": 3498
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.6832,
      "step": 3499
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5605,
      "step": 3500
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4918,
      "step": 3501
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4808,
      "step": 3502
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4469,
      "step": 3503
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.432,
      "step": 3504
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.6513,
      "step": 3505
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.6603,
      "step": 3506
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.7088,
      "step": 3507
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5364,
      "step": 3508
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5869,
      "step": 3509
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4444,
      "step": 3510
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4848,
      "step": 3511
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4924,
      "step": 3512
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4926,
      "step": 3513
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5089,
      "step": 3514
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.853,
      "step": 3515
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5722,
      "step": 3516
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4137,
      "step": 3517
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4375,
      "step": 3518
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4826,
      "step": 3519
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4759,
      "step": 3520
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4143,
      "step": 3521
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5573,
      "step": 3522
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4631,
      "step": 3523
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.7406,
      "step": 3524
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4679,
      "step": 3525
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4804,
      "step": 3526
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5016,
      "step": 3527
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.4433,
      "step": 3528
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.8271,
      "step": 3529
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.0002,
      "loss": 0.5184,
      "step": 3530
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4257,
      "step": 3531
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.5061,
      "step": 3532
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4932,
      "step": 3533
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4452,
      "step": 3534
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4965,
      "step": 3535
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.5757,
      "step": 3536
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6402,
      "step": 3537
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4546,
      "step": 3538
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6522,
      "step": 3539
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7163,
      "step": 3540
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6536,
      "step": 3541
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4694,
      "step": 3542
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4984,
      "step": 3543
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.8072,
      "step": 3544
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4762,
      "step": 3545
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4601,
      "step": 3546
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.7808,
      "step": 3547
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4774,
      "step": 3548
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4815,
      "step": 3549
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4533,
      "step": 3550
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.5606,
      "step": 3551
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4818,
      "step": 3552
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6713,
      "step": 3553
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4326,
      "step": 3554
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4888,
      "step": 3555
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.5289,
      "step": 3556
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4675,
      "step": 3557
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6441,
      "step": 3558
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6899,
      "step": 3559
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4638,
      "step": 3560
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.645,
      "step": 3561
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4864,
      "step": 3562
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.5442,
      "step": 3563
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4955,
      "step": 3564
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6285,
      "step": 3565
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.5603,
      "step": 3566
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.5425,
      "step": 3567
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.6058,
      "step": 3568
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4606,
      "step": 3569
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4553,
      "step": 3570
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.5221,
      "step": 3571
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.4962,
      "step": 3572
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.0002,
      "loss": 0.5364,
      "step": 3573
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.4644,
      "step": 3574
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.4685,
      "step": 3575
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.4791,
      "step": 3576
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.5382,
      "step": 3577
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.6174,
      "step": 3578
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.5278,
      "step": 3579
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.5044,
      "step": 3580
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.5059,
      "step": 3581
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.6945,
      "step": 3582
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.6577,
      "step": 3583
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.5554,
      "step": 3584
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.4513,
      "step": 3585
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.3859,
      "step": 3586
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.5235,
      "step": 3587
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.4624,
      "step": 3588
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.399,
      "step": 3589
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.7749,
      "step": 3590
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.4564,
      "step": 3591
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.5034,
      "step": 3592
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.4507,
      "step": 3593
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.7023,
      "step": 3594
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.4994,
      "step": 3595
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.4554,
      "step": 3596
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.5601,
      "step": 3597
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.9971,
      "step": 3598
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.4785,
      "step": 3599
    },
    {
      "epoch": 0.84,
      "learning_rate": 0.0002,
      "loss": 0.6998,
      "step": 3600
    },
    {
      "epoch": 0.84,
      "eval_loss": 0.5451279282569885,
      "eval_runtime": 156.4881,
      "eval_samples_per_second": 6.39,
      "eval_steps_per_second": 3.195,
      "step": 3600
    },
    {
      "epoch": 0.84,
      "mmlu_eval_accuracy": 0.5719459780243462,
      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.5625,
      "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6551724137931034,
      "mmlu_eval_accuracy_college_biology": 0.6875,
      "mmlu_eval_accuracy_college_chemistry": 0.25,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
      "mmlu_eval_accuracy_college_medicine": 0.5454545454545454,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.6363636363636364,
      "mmlu_eval_accuracy_conceptual_physics": 0.4230769230769231,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.5625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.43902439024390244,
      "mmlu_eval_accuracy_formal_logic": 0.07142857142857142,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.4883720930232558,
      "mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5769230769230769,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.4782608695652174,
      "mmlu_eval_accuracy_high_school_us_history": 0.7727272727272727,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.6363636363636364,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.84,
      "mmlu_eval_accuracy_medical_genetics": 1.0,
      "mmlu_eval_accuracy_miscellaneous": 0.7209302325581395,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.4,
      "mmlu_eval_accuracy_nutrition": 0.6666666666666666,
      "mmlu_eval_accuracy_philosophy": 0.6764705882352942,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.3870967741935484,
      "mmlu_eval_accuracy_professional_law": 0.3764705882352941,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.5652173913043478,
      "mmlu_eval_accuracy_public_relations": 0.4166666666666667,
      "mmlu_eval_accuracy_security_studies": 0.5555555555555556,
      "mmlu_eval_accuracy_sociology": 0.9090909090909091,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.4444444444444444,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 1.492884331871573,
      "step": 3600
    }
  ],
  "logging_steps": 1,
  "max_steps": 12837,
  "num_train_epochs": 3,
  "save_steps": 200,
  "total_flos": 8.86949767329153e+17,
  "trial_name": null,
  "trial_params": null
}