Chain-of-thought (CoT) monitoring assumes that the reasoning a model writes reflects the computation that directly produces its answer. Previous faithfulness…
机构:UT Austin
来源:arXiv 2609.27038 | AI4Papers 论文推荐平台