Saved in:
Bibliographic Details
Main Authors: Yang, Hao, Zhao, Qianghua, Li, Lei
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2412.03944
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917858184790016
author Yang, Hao
Zhao, Qianghua
Li, Lei
author_facet Yang, Hao
Zhao, Qianghua
Li, Lei
contents Chain-of-Thought prompting has significantly enhanced the reasoning capabilities of large language models, with numerous studies exploring factors influencing its performance. However, the underlying mechanisms remain poorly understood. To further demystify the operational principles, this work examines three key aspects: decoding, projection, and activation, aiming to elucidate the changes that occur within models when employing Chainof-Thought. Our findings reveal that LLMs effectively imitate exemplar formats while integrating them with their understanding of the question, exhibiting fluctuations in token logits during generation but ultimately producing a more concentrated logits distribution, and activating a broader set of neurons in the final layers, indicating more extensive knowledge retrieval compared to standard prompts. Our code and data will be publicly avialable when the paper is accepted.
format Preprint
id arxiv_https___arxiv_org_abs_2412_03944
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Chain-of-Thought in Large Language Models: Decoding, Projection, and Activation
Yang, Hao
Zhao, Qianghua
Li, Lei
Artificial Intelligence
Chain-of-Thought prompting has significantly enhanced the reasoning capabilities of large language models, with numerous studies exploring factors influencing its performance. However, the underlying mechanisms remain poorly understood. To further demystify the operational principles, this work examines three key aspects: decoding, projection, and activation, aiming to elucidate the changes that occur within models when employing Chainof-Thought. Our findings reveal that LLMs effectively imitate exemplar formats while integrating them with their understanding of the question, exhibiting fluctuations in token logits during generation but ultimately producing a more concentrated logits distribution, and activating a broader set of neurons in the final layers, indicating more extensive knowledge retrieval compared to standard prompts. Our code and data will be publicly avialable when the paper is accepted.
title Chain-of-Thought in Large Language Models: Decoding, Projection, and Activation
topic Artificial Intelligence
url https://arxiv.org/abs/2412.03944