{ "epoch": 3.0, "eval_logits/chosen": -2.085902214050293, "eval_logits/rejected": -1.9675065279006958, "eval_logps/chosen": -266.14141845703125, "eval_logps/rejected": -221.83285522460938, "eval_loss": 0.5895045399665833, "eval_rewards/accuracies": 0.6880000233650208, "eval_rewards/chosen": 0.14388784766197205, "eval_rewards/margins": 0.3272360563278198, "eval_rewards/rejected": -0.1833481788635254, "eval_runtime": 170.0188, "eval_samples": 2000, "eval_samples_per_second": 11.763, "eval_steps_per_second": 0.735 }