eval_results.json

{
    "epoch": 0.995910949568378,
    "eval_logits/chosen": -13.768111228942871,
    "eval_logits/rejected": -13.749955177307129,
    "eval_logps/chosen": -1.4371137619018555,
    "eval_logps/rejected": -1.7012752294540405,
    "eval_loss": 4.129865646362305,
    "eval_rewards/accuracies": 0.6721311211585999,
    "eval_rewards/chosen": -14.371137619018555,
    "eval_rewards/margins": 2.6416165828704834,
    "eval_rewards/rejected": -17.012752532958984,
    "eval_runtime": 65.8969,
    "eval_samples": 1941,
    "eval_samples_per_second": 29.455,
    "eval_steps_per_second": 1.851
}