eval_results.json

{
    "epoch": 1.97,
    "eval_logits/chosen": 99.2168197631836,
    "eval_logits/rejected": 93.17211151123047,
    "eval_logps/chosen": -494.3555603027344,
    "eval_logps/rejected": -520.5654907226562,
    "eval_loss": 0.4896722733974457,
    "eval_rewards/accuracies": 0.7083333134651184,
    "eval_rewards/chosen": -2.047553062438965,
    "eval_rewards/margins": 1.4032453298568726,
    "eval_rewards/rejected": -3.450798749923706,
    "eval_runtime": 51.7881,
    "eval_samples": 750,
    "eval_samples_per_second": 14.482,
    "eval_steps_per_second": 0.463
}