{ "epoch": 3.0, "eval_logits/chosen": -2.2616169452667236, "eval_logits/rejected": -2.2390596866607666, "eval_logps/chosen": -84.79953002929688, "eval_logps/rejected": -103.63373565673828, "eval_loss": 0.6454610824584961, "eval_rewards/accuracies": 0.3313491940498352, "eval_rewards/chosen": -0.14515721797943115, "eval_rewards/margins": 0.17175434529781342, "eval_rewards/rejected": -0.31691157817840576, "eval_runtime": 113.7128, "eval_samples": 2000, "eval_samples_per_second": 17.588, "eval_steps_per_second": 0.554 }