From Faithfulness to Correctness: Generative Reward Models that Think Critically

Open in new window