{ "epoch": 3.0, "eval_logits/chosen": -4.018362998962402, "eval_logits/rejected": -3.8936774730682373, "eval_logps/chosen": -478.6116943359375, "eval_logps/rejected": -394.1333923339844, "eval_loss": 0.6903825402259827, "eval_rewards/accuracies": 0.7393209338188171, "eval_rewards/chosen": -3.527071952819824, "eval_rewards/margins": 2.120450019836426, "eval_rewards/rejected": -5.647522926330566, "eval_runtime": 614.9661, "eval_samples": 1825, "eval_samples_per_second": 2.968, "eval_steps_per_second": 1.485 }