{ "epoch": 2.999098751126561, "eval_log_ratio_diff/mean": 4.0625, "eval_logits/chosen": -1.59375, "eval_logits/rejected": -1.6015625, "eval_logps/chosen": -324.0, "eval_logps/rejected": -302.0, "eval_loss": 0.6551145315170288, "eval_nll_loss": 1.0703125, "eval_ref_probs/chosen": 0.0038271306548267603, "eval_ref_probs/rejected": 0.0034569373819977045, "eval_rejected_term/max": -0.0002681996556930244, "eval_rejected_term/min": -0.0002681996556930244, "eval_rewards/accuracies": 0.7239999771118164, "eval_rewards/chosen": -0.173828125, "eval_rewards/margins": 0.40625, "eval_rewards/rejected": -0.58203125, "eval_runtime": 25.2858, "eval_samples_per_second": 39.548, "eval_steps_per_second": 9.887 }