{ "epoch": 5.0, "eval_logits/chosen": -42.66371536254883, "eval_logits/rejected": -41.95132827758789, "eval_logps/chosen": -65.31195068359375, "eval_logps/rejected": -71.56636810302734, "eval_loss": 0.6810184121131897, "eval_rewards/accuracies": 0.5956858396530151, "eval_rewards/chosen": -0.006560367997735739, "eval_rewards/margins": 0.029038555920124054, "eval_rewards/rejected": -0.03560638427734375, "eval_runtime": 142.4332, "eval_samples_per_second": 12.694, "eval_steps_per_second": 0.793 }