A Hopfieldian View-based Interpretation for Chain-of-Thought Reasoning
KAUST(阿卜杜拉国王科技大学);Soochow University(苏州大学);University of Southampton(南安普顿大学);Tongji University(同济大学);Universität Hamburg(汉堡大学);Huazhong University of Science and Technology(华中科技大学)
Lijie Hu, Liang Liu, Shu Yang, Xin Chen, Hongru Xiao, Mengdi Li, Pan Zhou, Muhammad Asif Ali, Di Wang
思维链(Chain-of-Thought, CoT)在提升大语言模型(large language models, LLMs)的推理性能方面占据重要地位。虽然部分研究聚焦于通过检索增强等方法提升CoT的准确率,但关于CoT为何能取得如此成效的严谨解释仍不明确。本文在两种不同设置下分析CoT方法,并提出以下问题:(1)对于零样本CoT,为何用“让我们一步步思考”提示模型会对其输出产生显著影响?(2)对于少样本CoT,为何在向模型提问前提供示例能大幅提升其推理能力?为解答这些问题,我们从Hopfield视角开展自上而下的可解释性分析,并提出一种用于控制CoT准确率的Read-and-Control方法。通过在七个数据集上针对三类不同任务开展大量实验,我们证明该框架能够解读CoT的内部运作机制,实现推理错误定位,并控制模型生成正确的推理路径。
英文摘要:
Chain-of-Thought (CoT) holds a significant place in augmenting the reasoning performance for large language models (LLMs). While some studies focus on improving CoT accuracy through methods like retrieval enhancement, yet a rigorous explanation for why CoT achieves such success remains unclear. In this paper, we analyze CoT methods under two different settings by asking the following questions: (1) For zero-shot CoT, why does prompting the model with "let's think step by step" significantly impact its outputs? (2) For few-shot CoT, why does providing examples before questioning the model could substantially improve its reasoning ability? To answer these questions, we conduct a top-down explainable analysis from the Hopfieldian view and propose a Read-and-Control approach for controlling the accuracy of CoT. Through extensive experiments on seven datasets for three different tasks, we demonstrate that our framework can decipher the inner workings of CoT, provide reasoning error localization, and control to come up with the correct reasoning path.
Understanding Reasoning in Chain-of-Thought from the Hopfieldian View
King Abdullah University of Science and Technology(阿卜杜拉国王科技大学);Soochow University(苏州大学);University of Southampton(南安普顿大学);Arizona State University(亚利桑那州立大学)
Clustering as Reasoning: A $k$-Means Interpretation of Chain-of-Thought Graph Learning
University of Electronic Science and Technology of China(电子科技大学);Singapore Management University(新加坡国立大学);Michigan State University(密歇根州立大学);The Chinese University of Hong Kong(香港中文大学)
VisDoT : Enhancing Visual Reasoning through Human-Like Interpretation Grounding and Decomposition of Thought
Department of Computer Science and Artificial Intelligence, Dongguk University(东国大学计算机科学与人工智能系);Department of Electronics and Electrical Engineering, Dongguk University(东国大学电子与电气工程系)
CheXthought: A global multimodal dataset of clinical chain-of-thought reasoning and visual attention for chest X-ray interpretation
Center for Artificial Intelligence in Medicine and Imaging(医学与影像人工智能中心);Department of Radiology(放射科);Department of Pediatrics(儿科);Department of Radiology, Weill Cornell Medicine(韦氏 Cornell 医学院放射科);Department of Radiology, Brigham and Women’s Hospital(布里奇妇产科医院放射科);University of California, Berkeley(加州大学伯克利分校);University of California, San Francisco(加州大学旧金山分校);Department of Biomedical Data Science(生物医学数据科学部)
Correct Answers, Invalid Traces: What Verifiable Grade-School Math Reveals About Chain-of-Thought Traces
IT University of Copenhagen(哥本哈根信息技术大学);Chennai Mathematical Institute(金奈数学研究所);Indian Institute of Technology Jammu(印度理工学院贾姆穆分校);Arizona State University(亚利桑那州立大学)
SPOT: Span-level Pause-of-Thought for Efficient and Interpretable Latent Reasoning in Large Language Models
School of New Media and Communication, Tianjin University(新媒体与传播学院,天津大学);School of Computer Science and Engineering, Beihang University(计算机科学与工程学院,北航大学)
Scalable Qualitative Coding with LLMs: Chain-of-Thought Reasoning Matches Human Performance in Some Hermeneutic Tasks
Center for Complex Networks and Systems Research(复杂网络与系统研究中心);Luddy School of Informatics, Computer Science, and Engineering(勒迪信息学、计算机科学与工程学院);Department of Sociology(社会学系);Indiana University(印第安纳大学)
Separate First, Fuse Later: Mitigating Cross-Modal Interference in Audio-Visual LLMs Reasoning with Modality-Specific Chain-of-Thought
Tianjin Key Laboratory of Cognitive Computing(天津认知计算实验室);Tianjin University(天津大学);Huiyan Technology Company, Ltd.(慧颜科技有限公司);Chinese Academy of Sciences(中国科学院);Tencent(腾讯)
Multi-Agent Multimodal Large Language Model Framework for Automated Interpretation of Fuel Efficiency Analytics in Public Transportation
SDU Center for Energy Informatics, The Maersk Mc-Kinney Moller Institute, Faculty of Engineering, University of Southern Denmark(南丹麦大学工程学院马士基麦克-金尼穆勒研究所SDU能源信息学中心);Institute for Technologies and Management of Digital Transformation, University of Wuppertal(伍珀塔尔大学数字化转型技术与管理研究所)