{ "epoch": 2.999098751126561, "eval_kl/chosen": 0.007034215610474348, "eval_kl/rejected": 0.006547675002366304, "eval_logits/chosen": -1.8828125, "eval_logits/rejected": -1.90625, "eval_logps/chosen": -328.0, "eval_logps/rejected": -312.0, "eval_loss": 1.584359884262085, "eval_nll_loss": 1.078125, "eval_ref_probs/chosen": 0.0038271306548267603, "eval_ref_probs/rejected": 0.0034569373819977045, "eval_rejected_term/max": 0.006547675002366304, "eval_rejected_term/min": 0.006547675002366304, "eval_rewards/accuracies": 0.7559999823570251, "eval_rewards/chosen": -0.6640625, "eval_rewards/margins": 0.87109375, "eval_rewards/rejected": -1.5390625, "eval_runtime": 25.2699, "eval_samples_per_second": 39.573, "eval_steps_per_second": 9.893 }