Thought Branches: Interpreting LLM Reasoning Requires Resampling

N/A

# link_4pzlbiecp4.pdf ## Metadata - PDFFormatVersion=1.7 - IsLinearized=false - IsAcroFormPresent=false - IsXFAPresent=false - IsCollectionPresent=false - IsSignaturesPresent=false - Author=Uzay Macar; Paul C. Bogdan; Senthooran Rajamanoharan; Neel Nanda - Creator=arXiv GenPDF (tex2pdf:a6404ea) - Custom.DOI=https://doi.org/10.48550/arXiv.2510.27484 - Custom.License=http://creativecommons.org/licenses/by/4.0/ - Custom.PTEX.Fullbanner=This is pdfTeX, Version 3.141592653-2.6-1.40.28 (TeX Live 2025) kpathsea version 6.4.1 - Custom.arXivID=https://arxiv.org/abs/2510.27484v2 - Producer=pikepdf 8.15.

saved by

• Sandra Luo

related reading

• [2510.27484] Thought Branches: Interpreting LLM Reasoning Requires Resamplingarxiv.org
• Tracing the thoughts of a large language model \ Anthropicanthropic.com
• Thought Anchors: Which LLM Reasoning Steps Matter? — LessWronglesswrong.com
• the case for CoT unfaithfulness is overstated — LessWronglesswrong.com
• Towards Faithful Chain-of-Thought: Large Language Models are Bridging Reasonersarxiv.org
• [2405.18915] Towards Faithful Chain-of-Thought: Large Language Models are Bridging Reasonersar5iv.labs.arxiv.org
• [2506.19143] Thought Anchors: Which LLM Reasoning Steps Matter?arxiv.org
• [2510.24941] Can Aha Moments Be Fake? Towards Quantifying Decorative and True Thinking in Chain-of-Thoughtarxiv.org
• Reasoning models don't always say what they think \ Anthropicanthropic.com
• Refusal in LLMs is mediated by a single direction — LessWronglesswrong.com
• Measuring Faithfulness in Chain-of-Thought Reasoning \ Anthropicanthropic.com
• When Chain of Thought is Necessary, Language Models Struggle to Evade Monitorsarxiv.org