paper-with-me

홈 › Papers

Measuring Faithfulness of Chains of Thought by Unlearning Reasoning Steps

2025-02-20 · Martin Tutek, Fateme Hashemi Chaleshtori, Ana Marasović, Yonatan Belinkov

When prompted to think step-by-step, language models (LMs) produce a chain of thought (CoT), a sequence of reasoning steps that the model supposedly used to produce its prediction. However, despite much work on CoT prompting, it is unclear if CoT reasoning is faithful to the models' parameteric beliefs. We introduce a framework for measuring parametric faithfulness of generated reasoning, and propose Faithfulness by Unlearning Reasoning steps (FUR), an instance of this framework. FUR erases information contained in reasoning steps from model parameters. We perform experiments unlearning CoTs of four LMs prompted on four multi-choice question answering (MCQA) datasets. Our experiments show that FUR is frequently able to change the underlying models' prediction by unlearning key steps, indicating when a CoT is parametrically faithful. Further analysis shows that CoTs generated by models post-unlearning support different answers, hinting at a deeper effect of unlearning. Importantly, CoT steps identified as important by FUR do not align well with human notions of plausbility, emphasizing the need for specialized alignment

📄 PDF Abstract BibTeX arXiv:2502.14829

Code (1)

technion-cs-nlp/parametric-faithfulness 공식 구현 pytorch

Tasks

Question Answering

Methods 이 논문이 사용한 방법론

ALIGN In the ALIGN method, visual and language representations are jointly trained from noisy image alt-text data. The image and text encoders are learned via contrastive loss…

Similar Papers 제목 키워드 기반

GeoFaith: A Spatio-Temporal Dual View of Faithful Chain-of-Thought

2026-05-26 · Weijiang Lv, Wentong Zhao, Jiayu Wang, Yuhao Wu 외 arxiv

Chain-of-Thought (CoT) reasoning has advanced large language models (LLMs), but outcome-based supervision leads to pervasive post-hoc rationalization, producing plausible yet unfaithful reasoning chains. Most prior faith…

Reinforcement Learning

Journey Before Destination: On the importance of Visual Faithfulness in Slow Thinking

2025-12-13 · Rheeya Uppaal, Phu Mon Htut, Min Bai, Nikolaos Pappas 외 arxiv

Reasoning-augmented vision language models (VLMs) generate explicit chains of thought that promise greater capability and transparency but also introduce new failure modes: models may reach correct answers via visually u…

Multimodal Reasoning

Measuring Chain-of-Thought Monitorability Through Faithfulness and Verbosity

2025-10-31 · Austin Meek, Eitan Sprejer, Iván Arcuschin, Austin J. Brockmeier 외 arxiv

Chain-of-thought (CoT) outputs let us read a model's step-by-step reasoning. Since any long, serial reasoning process must pass through this textual trace, the quality of the CoT is a direct window into what the model is…

FACT-E: Causality-Inspired Evaluation for Trustworthy Chain-of-Thought Reasoning

2026-04-12 · Yuxi Sun, Aoqi Zuo, Haotian Xie, Wei Gao 외 arxiv

Chain-of-Thought (CoT) prompting has improved LLM reasoning, but models often generate explanations that appear coherent while containing unfaithful intermediate steps. Existing self-evaluation approaches are prone to in…

Investigating Faithfulness in Large Audio Language Models

2025-09-26 · Pooneh Mousavi, Lovenya Jain, Mirco Ravanelli, Cem Subakan arxiv

Large Audio Language Models (LALMs) integrate audio encoders with pretrained Large Language Models to perform complex multimodal reasoning tasks. While these models can generate Chain-of-Thought (CoT) explanations, the f…

Multimodal Reasoning