{ "epoch": 0.9990987511265611, "eval_log_ratio_diff/mean": 0.29296875, "eval_logits/chosen": -2.6875, "eval_logits/rejected": -2.703125, "eval_logps/chosen": -1.4765625, "eval_logps/rejected": -1.984375, "eval_loss": 0.8915858268737793, "eval_nll_loss": 1.484375, "eval_ref_probs/chosen": 0.39426562190055847, "eval_ref_probs/rejected": 0.3336612582206726, "eval_rejected_term/max": -0.09231030195951462, "eval_rejected_term/min": -0.09231030195951462, "eval_rewards/accuracies": 0.7279999852180481, "eval_rewards/chosen": -0.04296875, "eval_rewards/margins": 0.029296875, "eval_rewards/rejected": -0.072265625, "eval_runtime": 25.323, "eval_samples_per_second": 39.49, "eval_steps_per_second": 9.872 }