{ "epoch": 0.9990987511265611, "eval_log_ratio_diff/mean": 0.1796875, "eval_logits/chosen": -2.5, "eval_logits/rejected": -2.53125, "eval_logps/chosen": -1.1953125, "eval_logps/rejected": -1.5859375, "eval_loss": 4.3386149406433105, "eval_nll_loss": 1.1953125, "eval_ref_probs/chosen": 0.39426562190055847, "eval_ref_probs/rejected": 0.3336612582206726, "eval_rejected_term/max": -0.05157232657074928, "eval_rejected_term/min": -0.05157232657074928, "eval_rewards/accuracies": 0.7160000205039978, "eval_rewards/chosen": -0.0146484375, "eval_rewards/margins": 0.0179443359375, "eval_rewards/rejected": -0.032470703125, "eval_runtime": 25.576, "eval_samples_per_second": 39.099, "eval_steps_per_second": 9.775 }