How to think step-by-step: A mechanistic understanding of chain-of-thought reasoning

arXiv:2402.18312v1 Announce Type: cross
Abstract: Despite superior reasoning prowess demonstrated by Large Language Models (LLMs) with Chain-of-Thought (CoT) prompting, a lack of understanding prevails around the internal mechanisms of the models that facilitate CoT generation. This work investigates the neural sub-structures within LLMs that manifest CoT reasoning from a mechanistic point of view. From an analysis of LLaMA-2 7B applied to multistep reasoning over fictional ontologies, we demonstrate that LLMs deploy multiple parallel pathways of answer generation for step-by-step reasoning. …

abstract arxiv cs.cl cs.lg language language models large language large language models llms manifest prompting reasoning step-by-step think thought type understanding work