{ "epoch": 2.999098751126561, "eval_log_ratio_diff/mean": 4.625, "eval_logits/chosen": -1.6328125, "eval_logits/rejected": -1.6484375, "eval_logps/chosen": -324.0, "eval_logps/rejected": -304.0, "eval_loss": 0.6423432230949402, "eval_nll_loss": 1.078125, "eval_ref_probs/chosen": 0.0038271306548267603, "eval_ref_probs/rejected": 0.0034569373819977045, "eval_rejected_term/max": -0.0005308441468514502, "eval_rejected_term/min": -0.0005308441468514502, "eval_rewards/accuracies": 0.7279999852180481, "eval_rewards/chosen": -0.22265625, "eval_rewards/margins": 0.4609375, "eval_rewards/rejected": -0.68359375, "eval_runtime": 25.3737, "eval_samples_per_second": 39.411, "eval_steps_per_second": 9.853 }