{ "epoch": 0.99, "eval_logits/chosen": 110.1702880859375, "eval_logits/rejected": 101.88317108154297, "eval_logps/chosen": -782.641845703125, "eval_logps/rejected": -814.3776245117188, "eval_loss": 0.9485784769058228, "eval_rewards/accuracies": 0.7708333134651184, "eval_rewards/chosen": 1.925491213798523, "eval_rewards/margins": 3.3749351501464844, "eval_rewards/rejected": -1.4494438171386719, "eval_runtime": 50.7482, "eval_samples": 750, "eval_samples_per_second": 14.779, "eval_steps_per_second": 0.473 }