{ "epoch": 3.0, "eval_logits/chosen": -2.687349796295166, "eval_logits/rejected": -2.300260305404663, "eval_logps/chosen": -268.54150390625, "eval_logps/rejected": -212.86087036132812, "eval_loss": 18.088815689086914, "eval_rewards/accuracies": 0.7319999933242798, "eval_rewards/chosen": 0.02013408951461315, "eval_rewards/margins": 0.13951165974140167, "eval_rewards/rejected": -0.11937756091356277, "eval_runtime": 350.1967, "eval_samples": 2000, "eval_samples_per_second": 5.711, "eval_steps_per_second": 0.714 }