Toward Adaptive Reasoning in Large Language Models with Thought Rollback

Open in new window