{ "epoch": 3.0, "eval_logits/chosen": -2.5050723552703857, "eval_logits/rejected": -2.487805128097534, "eval_logps/chosen": -80.76593017578125, "eval_logps/rejected": -89.9616928100586, "eval_loss": 0.6600826382637024, "eval_rewards/accuracies": 0.3313491940498352, "eval_rewards/chosen": -1.0432164669036865, "eval_rewards/margins": 0.4253966212272644, "eval_rewards/rejected": -1.4686132669448853, "eval_runtime": 114.8599, "eval_samples": 2000, "eval_samples_per_second": 17.413, "eval_steps_per_second": 0.548 }