TensorX
返回文献探索

Paper · arXiv 2508.19827

Analysing Chain of Thought Dynamics: Active Guidance or Unfaithful Post-hoc Rationalisation?

Samuel Lewis-Lim, Xingwei Tan, Zhixue Zhao, Nikolaos Aletras

33 upvotesAugust 27, 2025arXiv 预印本
AI 摘要

Investigation into Chain-of-Thought dynamics and faithfulness across various models reveals inconsistencies in their reliance on CoT and its alignment with actual reasoning.

Chain-of-ThoughtCoTsoft-reasoninganalytical reasoningcommonsense reasoninginstruction-tunedreasoning modelsreasoning-distilled models

Abstract

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.

北京市昌平区探索星信息技术及软件开发工作室

京ICP备2026059466号
Analysing Chain of Thought Dynamics: Active Guidance or Unfaithful Post-hoc Rationalisation? | TensorX