语言大模型 / LLM - arXivDaily 专题

2503.01141 2025-04-02 cs.CL cs.AI 79%

How Well do LLMs Compress Their Own Chain-of-Thought? A Token Complexity Approach

Ayeong Lee, Ethan Che, Tianyi Peng

机构 * Columbia Business School（哥伦比亚大学商学院）

专题命中推理与问题求解：large language model(abstract)；language model(abstract)；prompting(abstract)；分类 cs.CL、cs.AI

2503.21464 2025-03-28 cs.CL cs.AI cs.PF 79%

Harnessing Chain-of-Thought Metadata for Task Routing and Adversarial Prompt Detection

Ryan Marinelli, Josef Pichlmeier, Tamas Bisztray

机构 * University of Oslo（奥斯陆大学）； BMW Group（宝马集团）； Ludwig-Maximilians-Universität（慕尼黑大学）

专题命中推理与问题求解：large language model(abstract)；language model(abstract)；prompting(abstract)；分类 cs.CL、cs.AI

2502.09056 2025-03-28 cs.CL cs.AI 79%

Adapting Language-Specific LLMs to a Reasoning Model in One Day via Model Merging -- An Open Recipe

Kunat Pipatanakul, Pittawat Taveekitworachai, Potsawee Manakul, Kasima Tharnpipitchai

机构 * SCB 10X R&D SCBX Group（SCB 10X研发中心SCBX集团）

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

Comments 9 pages

2404.14963 2025-03-28 cs.CL cs.AI 79%

Achieving >97% on GSM8K: Deeply Understanding the Problems Makes LLMs Better Solvers for Math Word Problems

Qihuang Zhong, Kang Wang, Ziyang Xu, Juhua Liu, Liang Ding, Bo Du

机构 * Wuhan University（武汉大学）； University of Sydney（悉尼大学）

专题命中推理与问题求解：large language model(abstract)；language model(abstract)；prompting(abstract)；分类 cs.CL、cs.AI

Comments The article has been accepted by Frontiers of Computer Science (FCS), with the DOI: { 10.1007/s11704-025-41102-z }

2503.19988 2025-03-27 cs.LG cs.AI cs.DB 79%

ExCoT: Optimizing Reasoning for Text-to-SQL with Execution Feedback

Bohan Zhai, Canwen Xu, Yuxiong He, Zhewei Yao

机构 * Snowflake Inc.（雪花公司）

专题命中推理与问题求解：large language model(abstract)；language model(abstract)；preference optimization(abstract)；分类 cs.AI、cs.LG

2502.06759 2025-03-21 cs.CL cs.AI cs.DB 79%

Rationalization Models for Text-to-SQL

Gaetano Rossiello, Nhan Pham, Michael Glass, Junkyu Lee, Dharmashankar Subramanian

机构 * IBM T.J. Watson Research Center（IBM T.J.沃森研究中心）

专题命中推理与问题求解：large language model(abstract)；language model(abstract)；small language model(abstract)；分类 cs.CL、cs.AI

Comments Published at ICLR 2025 Workshop on Reasoning and Planning for LLMs

2503.15272 2025-03-20 cs.CL cs.AI 79%

MAMM-Refine: A Recipe for Improving Faithfulness in Generation with Multi-Agent Collaboration

David Wan, Justin Chih-Yao Chen, Elias Stengel-Eskin, Mohit Bansal

机构 * UNC Chapel Hill（北卡罗来纳大学教堂山分校）

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

Comments NAACL 2025, 18 pages. Code: https://github.com/meetdavidwan/mammrefine

2503.14484 2025-03-19 cs.MA cs.AI cs.CL 79%

Gricean Norms as a Basis for Effective Collaboration

Fardin Saad, Pradeep K. Murukannaiah, Munindar P. Singh

机构 * North Carolina State University（北卡罗来纳州立大学）； Delft University of Technology（代尔夫特理工大学）

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

Comments Accepted to AAMAS 2025. 8 pages (excl. references), 9 figures/tables. (Appendix: 5 pages, 6 figures/tables). Code available at: https://github.com/fardinsaad/Gricean-Norms

2411.19488 2025-03-18 cs.CV cs.AI cs.LG 79%

Interleaved-Modal Chain-of-Thought

Jun Gao, Yongqi Li, Ziqiang Cao, Wenjie Li

机构 * School of Computer Science and Technology, Soochow University（苏州大学计算机科学与技术学院）； Department of Computer Science, The Hong Kong Polytechnic University（香港理工大学计算机科学系）

专题命中推理与问题求解：large language model(abstract)；language model(abstract)；prompting(abstract)；分类 cs.AI、cs.LG

Comments CVPR 2025 Main Conference

2502.17017 2025-03-17 cs.CL cs.IT cs.LG math.IT math.LO 79%

Quantifying Logical Consistency in Transformers via Query-Key Alignment

Eduard Tulchinskii, Anastasia Voznyuk, Laida Kushnareva, Andrei Andriiainen, Irina Piontkovskaya, Evgeny Burnaev, Serguei Barannikov

机构 * Skolkovo Institute of Science and Technology（斯科尔科沃科学技术研究所）； Moscow Institute of Physics and Technology（莫斯科物理技术学院）； CNRS（法国国家科学研究中心）； Université Paris Cité（巴黎西岱大学）； Artificial Intelligence Research Institute (AIRI)（人工智能研究所（AIRI））

专题命中推理与问题求解：large language model(abstract)；language model(abstract)；prompting(abstract)；分类 cs.CL、cs.LG

2503.05212 2025-03-10 cs.CL cs.AI 79%

Knowledge Updating? No More Model Editing! Just Selective Contextual Reasoning

Guoxiu He, Xin Song, Aixin Sun

机构 * School of Economics and Management, East China Normal University（华东师范大学经济与管理学部）； College of Computing and Data Science, Nanyang Technological University（南洋理工大学计算与数据科学学院）

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

2503.03979 2025-03-07 cs.CL cs.AI cs.HC 79%

ReasonGraph: Visualisation of Reasoning Paths

Zongqian Li, Ehsan Shareghi, Nigel Collier

机构 * University of Cambridge（剑桥大学）； Monash University（莫纳什大学）

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

2403.11807 2025-03-07 cs.AI cs.CL 79%

How Far Are We on the Decision-Making of LLMs? Evaluating LLMs' Gaming Ability in Multi-Agent Environments

Jen-tse Huang, Eric John Li, Man Ho Lam, Tian Liang, Wenxuan Wang, Youliang Yuan, Wenxiang Jiao, Xing Wang, Zhaopeng Tu, Michael R. Lyu

机构 * The Chinese University of Hong Kong（香港中文大学）； Tencent AI Lab（腾讯AI实验室）； The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳））； Tsinghua University（清华大学）

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

Comments Accepted to ICLR 2025; 11 pages of main text; 26 pages of appendices; Included models: GPT-3.5-{0613, 1106, 0125}, GPT-4-0125, GPT-4o-0806, Gemini-{1.0, 1.5)-Pro, LLaMA-3.1-{7, 70, 405}B, Mixtral-8x{7, 22}B, Qwen-2-72B

2410.06209 2025-03-07 cs.LG cs.AI cs.LO 79%

LeanAgent: Lifelong Learning for Formal Theorem Proving

Adarsh Kumarappan, Mo Tiwari, Peiyang Song, Robert Joseph George, Chaowei Xiao, Anima Anandkumar

机构 * California Institute of Technology（加州理工学院）； Stanford University（斯坦福大学）； University of Wisconsin, Madison（威斯康星大学麦迪逊分校）

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

2410.05459 2025-03-06 cs.LG cs.CL stat.ML 79%

From Sparse Dependence to Sparse Attention: Unveiling How Chain-of-Thought Enhances Transformer Sample Efficiency

Kaiyue Wen, Huaqing Zhang, Hongzhou Lin, Jingzhao Zhang

机构 * Stanford University（斯坦福大学）； IIIS, Tsinghua University（清华大学信息科学技术学院（IIIS））； Amazon（亚马逊公司）； Shanghai AI Lab（上海人工智能实验室）； Shanghai Qizhi Institute（上海智慧研究院）

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.LG

Comments 43 pages,11 figures

2503.01923 2025-03-05 cs.CL cs.AI 79%

Output Length Effect on DeepSeek-R1's Safety in Forced Thinking

Xuying Li, Zhuo Li, Yuji Kosuga, Victor Bian

机构 * HydroX AI

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

2503.01880 2025-03-05 cs.CL cs.AI cs.SI 79%

BEYONDWORDS is All You Need: Agentic Generative AI based Social Media Themes Extractor

Mohammed-Khalil Ghali, Abdelrahman Farrag, Sarah Lam, Daehan Won

机构 * School of Systems Science and Industrial Engineering, State University of New York at Binghamton（纽约州立大学宾汉姆顿分校系统科学与工业工程学院）

专题命中推理与问题求解：LLM(abstract)；language model(abstract)；prompting(abstract)；分类 cs.CL、cs.AI

2410.05218 2025-03-05 cs.LG cs.CL stat.ML 79%

Density estimation with LLMs: a geometric investigation of in-context learning trajectories

Toni J. B. Liu, Nicolas Boullé, Raphaël Sarfati, Christopher J. Earls

机构 * Cornell University（康奈尔大学）； Imperial College London（伦敦帝国学院）

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.LG

2407.11418 2025-03-04 cs.DB cs.AI cs.CL 79%

Semantic Operators: A Declarative Model for Rich, AI-based Data Processing

Liana Patel, Siddharth Jha, Melissa Pan, Harshit Gupta, Parth Asawa, Carlos Guestrin, Matei Zaharia

机构 * Stanford University（斯坦福大学）； UC Berkeley（加州大学伯克利分校）

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

2502.21290 2025-03-03 cs.AI cs.LG q-bio.QM 79%

Contextualizing biological perturbation experiments through language

Menghua Wu, Russell Littman, Jacob Levine, Lin Qiu, Tommaso Biancalani, David Richmond, Jan-Christian Huetter

机构 * Massachusetts Institute of Technology（麻省理工学院）； Genentech（基因泰克公司）； Meta AI

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

Comments The Thirteenth International Conference on Learning Representations (2025)

2502.21212 2025-03-03 cs.LG cs.AI 79%

Transformers Learn to Implement Multi-step Gradient Descent with Chain of Thought

Jianhao Huang, Zixuan Wang, Jason D. Lee

专题命中推理与问题求解：large language model(abstract)；language model(abstract)；prompting(abstract)；分类 cs.AI、cs.LG

Comments ICLR 2025 Spotlight

2502.16606 2025-02-27 cs.CL cs.AI 79%

Reasoning about Affordances: Causal and Compositional Reasoning in LLMs

Magnus F. Gjerde, Vanessa Cheung, David Lagnado

专题命中推理与问题求解：large language model(abstract)；language model(abstract)；prompting(abstract)；分类 cs.CL、cs.AI

Comments 21 pages, 7 figures, 3 tables

2501.11613 2025-02-25 cs.CL cs.AI cs.ET cs.HC cs.PL 79%

Conversation Routines: A Prompt Engineering Framework for Task-Oriented Dialog Systems

Giorgio Robino

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

Comments Minor typos revision

2502.16171 2025-02-25 cs.CL cs.AI 79%

EPERM: An Evidence Path Enhanced Reasoning Model for Knowledge Graph Question and Answering

Xiao Long, Liansheng Zhuang, Aodi Li, Minghong Yao, Shafei Wang

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

2502.13723 2025-02-20 cs.CL cs.AI 79%

Direct Value Optimization: Improving Chain-of-Thought Reasoning in LLMs with Refined Values

Hongbo Zhang, Han Cui, Guangsheng Bao, Linyi Yang, Jun Wang, Yue Zhang

机构 * Zhejiang University（浙江大学）； School of Engineering, Westlake University（西湖大学工程学院）； Institute of Advanced Technology, Westlake Institute for Advanced Study（西湖高等研究院先进技术研究所）； University College London（伦敦大学学院）

专题命中推理与问题求解：large language model(abstract)；language model(abstract)；preference optimization(abstract)；分类 cs.CL、cs.AI

Comments preprint

2502.10458 2025-02-18 cs.LG cs.AI 79%

I Think, Therefore I Diffuse: Enabling Multimodal In-Context Reasoning in Diffusion Models

Zhenxing Mi, Kuan-Chieh Wang, Guocheng Qian, Hanrong Ye, Runtao Liu, Sergey Tulyakov, Kfir Aberman, Dan Xu

机构 * The Hong Kong University of Science and Technology（香港科技大学）

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

Comments Project page: https://mizhenxing.github.io/ThinkDiff, 19 pages, 14 figures

2502.10411 2025-02-18 cs.CY cs.AI cs.CL cs.IR cs.MA 79%

TrueReason: An Exemplar Personalised Learning System Integrating Reasoning with Foundational Models

Sahan Bulathwela, Daniel Van Niekerk, Jarrod Shipton, Maria Perez-Ortiz, Benjamin Rosman, John Shawe-Taylor

机构 * University College London（伦敦大学学院）； University of the Witwatersrand（威特沃特斯兰德大学）

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

Comments To be published as a book chapter

2502.09621 2025-02-14 cs.CV cs.AI cs.CL 79%

MME-CoT: Benchmarking Chain-of-Thought in Large Multimodal Models for Reasoning Quality, Robustness, and Efficiency

Dongzhi Jiang, Renrui Zhang, Ziyu Guo, Yanwei Li, Yu Qi, Xinyan Chen, Liuhui Wang, Jianhan Jin, Claire Guo, Shen Yan, Bo Zhang, Chaoyou Fu, Peng Gao, Hongsheng Li

专题命中推理与问题求解：large language model(abstract)；language model(abstract)；prompting(abstract)；分类 cs.CL、cs.AI

Comments Project Page: https://mmecot.github.io/

2502.06873 2025-02-12 cs.CL cs.AI 79%

Multimodal Cognitive Reframing Therapy via Multi-hop Psychotherapeutic Reasoning

Subin Kim, Hoonrae Kim, Heejin Do, Gary Geunbae Lee

机构 * Graduate School of Artificial Intelligence, POSTECH（浦项科技大学人工智能研究生院）； Department of Computer Science and Engineering, POSTECH（浦项科技大学计算机科学与工程系）

专题命中推理与问题求解：large language model(abstract)；language model(abstract)；prompting(abstract)；分类 cs.CL、cs.AI

Comments NAACL 2025 Main

2409.19020 2025-02-11 cs.CL cs.LG 79%

DiaSynth: Synthetic Dialogue Generation Framework for Low Resource Dialogue Applications

Sathya Krishnan Suresh, Wu Mengjun, Tushar Pranav, Eng Siong Chng

机构 * Nanyang Technological University（南洋理工大学）； Singapore Institute of Technology（新加坡理工学院）

专题命中推理与问题求解：LLM(abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.LG

Comments 13 pages, 1 figure

Journal ref NAACL 2025

1. 推理与问题求解 19037 篇

How Well do LLMs Compress Their Own Chain-of-Thought? A Token Complexity Approach

Harnessing Chain-of-Thought Metadata for Task Routing and Adversarial Prompt Detection

Adapting Language-Specific LLMs to a Reasoning Model in One Day via Model Merging -- An Open Recipe

Achieving >97% on GSM8K: Deeply Understanding the Problems Makes LLMs Better Solvers for Math Word Problems

ExCoT: Optimizing Reasoning for Text-to-SQL with Execution Feedback

Rationalization Models for Text-to-SQL

MAMM-Refine: A Recipe for Improving Faithfulness in Generation with Multi-Agent Collaboration

Gricean Norms as a Basis for Effective Collaboration

Interleaved-Modal Chain-of-Thought

Quantifying Logical Consistency in Transformers via Query-Key Alignment

Knowledge Updating? No More Model Editing! Just Selective Contextual Reasoning

ReasonGraph: Visualisation of Reasoning Paths

How Far Are We on the Decision-Making of LLMs? Evaluating LLMs' Gaming Ability in Multi-Agent Environments

LeanAgent: Lifelong Learning for Formal Theorem Proving

From Sparse Dependence to Sparse Attention: Unveiling How Chain-of-Thought Enhances Transformer Sample Efficiency

Output Length Effect on DeepSeek-R1's Safety in Forced Thinking

BEYONDWORDS is All You Need: Agentic Generative AI based Social Media Themes Extractor

Density estimation with LLMs: a geometric investigation of in-context learning trajectories

Semantic Operators: A Declarative Model for Rich, AI-based Data Processing

Contextualizing biological perturbation experiments through language

Transformers Learn to Implement Multi-step Gradient Descent with Chain of Thought

Reasoning about Affordances: Causal and Compositional Reasoning in LLMs

Conversation Routines: A Prompt Engineering Framework for Task-Oriented Dialog Systems

EPERM: An Evidence Path Enhanced Reasoning Model for Knowledge Graph Question and Answering

Direct Value Optimization: Improving Chain-of-Thought Reasoning in LLMs with Refined Values

I Think, Therefore I Diffuse: Enabling Multimodal In-Context Reasoning in Diffusion Models

TrueReason: An Exemplar Personalised Learning System Integrating Reasoning with Foundational Models

MME-CoT: Benchmarking Chain-of-Thought in Large Multimodal Models for Reasoning Quality, Robustness, and Efficiency

Multimodal Cognitive Reframing Therapy via Multi-hop Psychotherapeutic Reasoning

DiaSynth: Synthetic Dialogue Generation Framework for Low Resource Dialogue Applications