分析思维链动态:主动引导还是不忠的事后合理化?
Analysing Chain of Thought Dynamics: Active Guidance or Unfaithful Post-hoc Rationalisation?
AI总结:
本文研究思维链在软推理任务中的动态与忠实性,发现不同模型对 CoT 的依赖方式不同,且 CoT 的影响力与忠实性并不总是一致。
摘要:
近期研究表明,思维链(CoT)在分析推理和常识推理等软推理问题上往往收益有限。CoT 也可能不忠实于模型的实际推理过程。我们在指令微调、推理和推理蒸馏模型上研究了软推理任务中 CoT 的动态与忠实性。我们的发现揭示了这些模型依赖 CoT 的方式存在差异,并表明 CoT 的影响力和忠实性并不总是一致的。
英文摘要:
Recent work has demonstrated that Chain-of-Thought (CoT) often yields limited gains for soft-reasoning problems such as analytical and commonsense reasoning. CoT can also be unfaithful to a model's actual reasoning. We investigate the dynamics and faithfulness of CoT in soft-reasoning tasks across instruction-tuned, reasoning and reasoning-distilled models. Our findings reveal differences in how these models rely on CoT, and show that CoT influence and faithfulness are not always aligned.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。