{ "epoch": 3.0, "eval_logits/chosen": -2.434546947479248, "eval_logits/rejected": -2.35845685005188, "eval_logps/chosen": -435.4488525390625, "eval_logps/rejected": -442.29296875, "eval_loss": 0.5855222344398499, "eval_rewards/accuracies": 0.726190447807312, "eval_rewards/chosen": -0.904042661190033, "eval_rewards/margins": 0.49184030294418335, "eval_rewards/rejected": -1.3958829641342163, "eval_runtime": 247.0631, "eval_samples": 2000, "eval_samples_per_second": 8.095, "eval_steps_per_second": 0.255 }