eval_results.json

{
    "epoch": 1.9969834087481146,
    "eval_logits/chosen": 412.54827880859375,
    "eval_logits/rejected": 384.53924560546875,
    "eval_logps/chosen": -2.328974485397339,
    "eval_logps/rejected": -2.3880226612091064,
    "eval_loss": 97.23818969726562,
    "eval_rewards/accuracies": 0.6061643958091736,
    "eval_rewards/chosen": 0.042412400245666504,
    "eval_rewards/margins": 0.00826968066394329,
    "eval_rewards/rejected": 0.034142717719078064,
    "eval_runtime": 96.0326,
    "eval_samples": 4656,
    "eval_samples_per_second": 48.484,
    "eval_steps_per_second": 1.52
}