{ "epoch": 4.99, "eval_logits/chosen": -2.4107861518859863, "eval_logits/rejected": -2.3742854595184326, "eval_logps/chosen": -266.8597106933594, "eval_logps/rejected": -242.3632049560547, "eval_loss": 0.8742221593856812, "eval_rewards/accuracies": 0.6726190447807312, "eval_rewards/chosen": 0.008189404383301735, "eval_rewards/margins": 0.007871923968195915, "eval_rewards/rejected": 0.00031748018227517605, "eval_runtime": 123.0411, "eval_samples": 2000, "eval_samples_per_second": 16.255, "eval_steps_per_second": 0.341 }