{ "epoch": 1.0, "eval_logits/chosen": -2.59375, "eval_logits/rejected": -2.625, "eval_logps/chosen": -338.0, "eval_logps/rejected": -664.0, "eval_loss": 0.2634149193763733, "eval_rewards/accuracies": 0.8849018216133118, "eval_rewards/chosen": -2.0625, "eval_rewards/margins": 3.390625, "eval_rewards/rejected": -5.4375, "eval_runtime": 2667.9916, "eval_samples": 95033, "eval_samples_per_second": 35.417, "eval_steps_per_second": 0.554 }