{
    "epoch": 3.0,
    "eval_dpo_losses": 0.6100724935531616,
    "eval_logits/chosen": -2.5277466773986816,
    "eval_logits/rejected": -2.480801582336426,
    "eval_logps/chosen": -323.6263122558594,
    "eval_logps/rejected": -331.3143005371094,
    "eval_loss": 5.126430988311768,
    "eval_positive_losses": 43.18838882446289,
    "eval_rewards/accuracies": 0.6669999957084656,
    "eval_rewards/chosen": -0.39032912254333496,
    "eval_rewards/margins": 0.3370245695114136,
    "eval_rewards/margins_max": 1.416748046875,
    "eval_rewards/margins_min": -0.8377854228019714,
    "eval_rewards/margins_std": 0.7706853747367859,
    "eval_rewards/rejected": -0.7273536920547485,
    "eval_runtime": 428.2171,
    "eval_samples": 2000,
    "eval_samples_per_second": 4.671,
    "eval_steps_per_second": 0.292
}