{ "epoch": 3.0, "eval_logits/chosen": -2.697049617767334, "eval_logits/rejected": -2.6942925453186035, "eval_logps/chosen": -314.599609375, "eval_logps/rejected": -282.62579345703125, "eval_loss": 0.7903289198875427, "eval_rewards/accuracies": 0.7658730149269104, "eval_rewards/chosen": -3.2220375537872314, "eval_rewards/margins": 4.1147027015686035, "eval_rewards/rejected": -7.336739540100098, "eval_runtime": 217.5722, "eval_samples": 2000, "eval_samples_per_second": 9.192, "eval_steps_per_second": 0.29 }