{
    "epoch": 0.9982631930527722,
    "eval_logits/chosen": -1.402356505393982,
    "eval_logits/rejected": -1.3987274169921875,
    "eval_logps/chosen": -6.294243812561035,
    "eval_logps/rejected": -8.588370323181152,
    "eval_loss": 0.34357374906539917,
    "eval_rewards/accuracies": 0.8983739614486694,
    "eval_rewards/chosen": -12.58848762512207,
    "eval_rewards/margins": 4.588252067565918,
    "eval_rewards/rejected": -17.176740646362305,
    "eval_runtime": 87.9156,
    "eval_samples": 1961,
    "eval_samples_per_second": 22.305,
    "eval_steps_per_second": 1.399
}