大模型推理能力 - arXivDaily 专题

2407.20564 2024-07-31 cs.CL 89%

CLR-Fact: Evaluating the Complex Logical Reasoning Capability of Large Language Models over Factual Knowledge

Tianshi Zheng, Jiaxin Bai, Yicheng Wang, Tianqing Fang, Yue Guo, Yauwai Yim, Yangqiu Song

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；chain-of-thought(abstract)；分类 cs.CL

Comments 9 pages

2406.10625 2024-07-02 cs.CL 89%

On the Hardness of Faithful Chain-of-Thought Reasoning in Large Language Models

Sree Harsha Tanneru, Dan Ley, Chirag Agarwal, Himabindu Lakkaraju

专题命中逻辑推理：reasoning(title,abstract)；chain-of-thought(title,abstract)；CoT(abstract)；分类 cs.CL

2205.08794 2022-12-12 cs.CL 89%

LogiGAN: Learning Logical Reasoning via Adversarial Pre-training

Xinyu Pi, Wanjun Zhong, Yan Gao, Nan Duan, Jian-Guang Lou

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；verifier(abstract)；分类 cs.CL

Comments Accepted by NeurIPS 2022

2308.16118 2024-05-02 cs.CL cs.AI 89%

Response: Emergent analogical reasoning in large language models

Damian Hodel, Jevin West

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI

Comments Response to publication in Nature Human Behaviour titled "Emergent analogical reasoning in large language models," (Webb, Holyoak, and Lu, 2023, arXiv:2212.09196). 14 pages

2607.23019 2026-07-28 cs.AI cs.LO cs.SC 新提交 89%

Reason Popper-ly: Patching In-Context Reasoning with Inductive Logic Programming

合理地运用波普尔方法：用归纳逻辑编程修补上下文推理

Zirong Chen, Meiyi Ma

机构 * Vanderbilt University（范德堡大学）

专题命中逻辑推理：reasoning(title,abstract)；CoT(abstract,abstract_cn)；chain-of-thought(abstract)；verifier(abstract)

AI总结研究针对思维链提示中间步骤逻辑不合理问题，提出Reason Popper-ly框架，用归纳逻辑编程学习规则作在线验证器校正步骤，在CLUTRR基准测试中提升了模型准确率，还产生细粒度错误分类法，优于完全外部符号管道。

Comments Accepted at the 20th Conference on Neurosymbolic Learning and Reasoning

2312.03633 2026-07-15 cs.CL cs.AI cs.LG 89%

Exploring the Reversal Curse and Other Deductive Logical Reasoning in BERT and GPT-Based Large Language Models

探索BERT和GPT类大语言模型中的反向诅咒及其他演绎逻辑推理

Da Wu, Jingye Yang, Kai Wang

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI、cs.LG

AI总结本研究发现BERT对反向诅咒免疫，而GPT在处理多集合逻辑推理时存在困难，揭示了两种模型在简单与复杂逻辑推理中的差异。

Comments Final revision. To appear in Patterns

2506.15787 2026-05-12 cs.AI cs.CL cs.LG 89%

SLR: Automated Synthesis for Scalable Logical Reasoning

SLR：可扩展逻辑推理中的自动化合成

Lukas Helff, Ahmad Omar, Felix Friedrich, Antonia Wüst, Hikaru Shindo, Rupert Mitchell, Tim Woydt, Patrick Schramowski, Wolfgang Stammer, Kristian Kersting

机构 * TU Darmstadt（图宾根大学）； DFKI（德意志联邦防务学院）； Meta FAIR ； CERTAIN, Germany（德国CERTAIN）； Lab1141 ； MPI-Inf（马克斯·普朗克研究所）

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI、cs.LG

AI总结 SLR通过可扩展逻辑推理系统自动化生成指令提示、验证程序和潜在真实规则，创建SLR-Bench基准测试，展示LLM在逻辑推理中的表现及课程学习提升效果。

2605.07334 2026-05-11 cs.CV 89%

RCoT-Seg: Reinforced Chain-of-Thought for Video Reasoning and Segmentation

RCoT-Seg：强化链式推理用于视频推理与分割

Junwei Wen, Deshui Miao, Guangming Lu, Xin Li, Wenjie Pei

机构 * Harbin Institute of Technology, Shenzhen（哈尔滨工业大学（深圳））； Pengcheng Laboratory（鹏城实验室）； Pazhou Lab (Huangpu)（琶洲实验室（黄埔））

专题命中逻辑推理：reasoning(title,abstract)；chain-of-thought(title)；CoT(abstract,abstract_cn)

AI总结本文提出RCoT-Seg框架，通过分离时间推理与空间感知，改进视频分割中关键帧选择与定位，提升多目标场景下的鲁棒性与精度。

Comments 21 pages

2509.25851 2026-01-30 cs.CV 89%

MuSLR: Multimodal Symbolic Logical Reasoning

MuSLR：多模态符号逻辑推理

Jundong Xu, Hao Fei, Yuhui Zhang, Liangming Pan, Qijun Huang, Qian Liu, Preslav Nakov, Min-Yen Kan, William Yang Wang, Mong-Li Lee, Wynne Hsu

机构 * National University of Singapore（新加坡国立大学）； Stanford University（斯坦福大学）； Peking University（北京大学）； UniMelb（墨尔本大学）； University of Auckland（奥克兰大学）； MBZUAI（穆斯林人工智能研究所）； University of California, Santa Barbara（加州大学圣芭芭拉分校）

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；chain-of-thought(abstract)

AI总结 MuSLR提出了一种多模态符号逻辑推理基准，通过形式逻辑规则提升VLMs的推理能力，显著提升链式推理性能及复杂逻辑处理效果。

Comments Accepted by NeurIPS 2025

2509.18571 2025-09-24 cs.CV 89%

Live-E2T: Real-time Threat Monitoring in Video via Deduplicated Event Reasoning and Chain-of-Thought

Yuhan Wang, Cheng Liu, Zihan Zhao, Weichao Wu

机构 * School of Mechatronical Engineering, Beijing Institute of Technology（机械工程学院，北京理工大学）

专题命中逻辑推理：reasoning(title,abstract)；chain-of-thought(title,abstract)；logical reasoning(abstract)

2505.14615 2025-09-23 cs.AI cs.CL cs.LG cs.LO 89%

SATBench: Benchmarking LLMs' Logical Reasoning via Automated Puzzle Generation from SAT Formulas

Anjiang Wei, Yuheng Wu, Yingjia Wan, Tarun Suresh, Huanmi Tan, Zhanke Zhou, Sanmi Koyejo, Ke Wang, Alex Aiken

机构 * Stanford University（斯坦福大学）； UCLA（加州大学洛杉矶分校）； UIUC（伊利诺伊大学香槟分校）； CMU（卡内基梅隆大学）； Nanjing University（南京大学）

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI、cs.LG

2508.03366 2025-08-06 cs.AI cs.CL cs.LG cs.SC 89%

A Comparative Study of Neurosymbolic AI Approaches to Interpretable Logical Reasoning

Michael K. Chen

机构 * Raffles Institution（拉夫斯学院）

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Accepted to NeSy 2025

2502.01100 2025-07-16 cs.AI cs.CL cs.LG 89%

ZebraLogic: On the Scaling Limits of LLMs for Logical Reasoning

Bill Yuchen Lin, Ronan Le Bras, Kyle Richardson, Ashish Sabharwal, Radha Poovendran, Peter Clark, Yejin Choi

机构 * Allen Institute for AI（艾伦人工智能研究所）； University of Washington（华盛顿大学）； Stanford University（斯坦福大学）

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Accepted to ICML 2025

2504.20835 2025-04-30 cs.SD eess.AS 89%

Enhancing Non-Core Language Instruction-Following in Speech LLMs via Semi-Implicit Cross-Lingual CoT Reasoning

Hongfei Xue, Yufeng Tang, Hexin Liu, Jun Zhang, Xuelong Geng, Lei Xie

机构 * Northwestern Polytechnical University（西北工业大学）； ByteDance Inc.（字节跳动公司）； Nanyang Technological University（南洋理工大学）

专题命中逻辑推理：reasoning(title,abstract)；CoT(title,abstract)；chain-of-thought(abstract)

Comments 10 pages, 6 figures, Submitted to ACM MM 2025

2411.14215 2024-11-22 cs.CL cs.AI cs.LG 89%

Evaluating the Robustness of Analogical Reasoning in Large Language Models

Martha Lewis, Melanie Mitchell

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI、cs.LG

Comments 31 pages, 13 figures. arXiv admin note: text overlap with arXiv:2402.08955

2309.13339 2024-03-27 cs.CL cs.AI cs.LG cs.SC 89%

Enhancing Zero-Shot Chain-of-Thought Reasoning in Large Language Models through Logic

Xufeng Zhao, Mengdi Li, Wenhao Lu, Cornelius Weber, Jae Hee Lee, Kun Chu, Stefan Wermter

专题命中逻辑推理：reasoning(title,abstract)；chain-of-thought(title,abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Accepted in COLING 2024. Code see https://github.com/xf-zhao/LoT

2308.08614 2023-08-21 cs.LG cs.AI cs.CL 89%

Boosting Logical Reasoning in Large Language Models through a New Framework: The Graph of Thought

Bin Lei, pei-Hung Lin, Chunhua Liao, Caiwen Ding

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI、cs.LG

2305.17626 2023-06-06 cs.AI cs.CL cs.LG 89%

In-Context Analogical Reasoning with Pre-Trained Language Models

Xiaoyang Hu, Shane Storks, Richard L. Lewis, Joyce Chai

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI、cs.LG

2212.09282 2023-06-06 cs.CL cs.AI cs.LG 89%

APOLLO: A Simple Approach for Adaptive Pretraining of Language Models for Logical Reasoning

Soumya Sanyal, Yichong Xu, Shuohang Wang, Ziyi Yang, Reid Pryzant, Wenhao Yu, Chenguang Zhu, Xiang Ren

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Accepted at ACL 2023, code available at https://github.com/INK-USC/APOLLO

2305.03742 2023-05-09 cs.AI cs.CL cs.LG 89%

Improved Logical Reasoning of Language Models via Differentiable Symbolic Programming

Hanlin Zhang, Jiani Huang, Ziyang Li, Mayur Naik, Eric Xing

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI、cs.LG

Comments ACL 2023 Findings

2210.00312 2023-03-02 cs.CL cs.AI cs.CV cs.LG cs.MM 89%

Multimodal Analogical Reasoning over Knowledge Graphs

Ningyu Zhang, Lei Li, Xiang Chen, Xiaozhuan Liang, Shumin Deng, Huajun Chen

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Accepted by ICLR 2023. The project website is https://zjunlp.github.io/project/MKG_Analogy/introduction.html

2002.04326 2020-08-25 cs.CL cs.AI cs.LG 89%

ReClor: A Reading Comprehension Dataset Requiring Logical Reasoning

Weihao Yu, Zihang Jiang, Yanfei Dong, Jiashi Feng

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI、cs.LG

Comments ICLR 2020 paper. Project page: http://whyu.me/reclor/

2601.09446 2026-01-15 cs.CL cs.AI 89%

Improving Symbolic Translation of Language Models for Logical Reasoning

提升语言模型的符号翻译以增强逻辑推理

Ramya Keerthy Thatikonda, Jiuzhou Han, Wray Buntine, Ehsan Shareghi

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI

AI总结本文通过微调和增量推理框架，提升小型语言模型在逻辑推理中的符号翻译能力，减少错误率并提高推理性能。

Comments The Third workshop of NeusymBridge @AAAI 2026 (Bridging Neurons and Symbols for NLP and Knowledge Graph Reasoning)

2503.11207 2025-06-05 cs.AI cs.LG 89%

Can Large Reasoning Models do Analogical Reasoning under Perceptual Uncertainty?

Giacomo Camposampiero, Michael Hersche, Roger Wattenhofer, Abu Sebastian, Abbas Rahimi

机构 * IBM Research - Zurich（IBM瑞士研究实验室）； ETH Zürich（苏黎世联邦理工学院）

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.AI、cs.LG

Comments Accepted at the 19th International Conference on Neural-Symbolic Learning and Reasoning (NeSy) 2025

2209.08750 2022-09-20 cs.AI cs.LG 89%

Knowledge-based Analogical Reasoning in Neuro-symbolic Latent Spaces

Vishwa Shah, Aditya Sharma, Gautam Shroff, Lovekesh Vig, Tirtharaj Dash, Ashwin Srinivasan

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.AI、cs.LG

Comments 13 pages, 4 figures, Accepted at 16th International Workshop on Neural-Symbolic Learning and Reasoning as part of the 2nd International Joint Conference on Learning & Reasoning (IJCLR 2022)

2608.08168 2026-08-11 cs.CL 新提交 89%

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders

思考与非思考：基于稀疏自编码器的大语言模型推理机制可解释性研究

Bo Cheng, Qiaolin Lu, Yi Chang, Yuan Wu

专题命中逻辑推理：reasoning(title,abstract)；chain-of-thought(abstract,abstract_cn)；CoT(abstract,abstract_cn)；分类 cs.CL

AI总结本研究基于稀疏自编码器分析DeepSeek-R1-Distill-Qwen-7B的推理机制，揭示思考与非思考模式的神经差异，为大语言模型推理可解释性提供新见解。

2606.17312 2026-06-17 cs.AI 新提交 89%

Quantifying Consistency in LLM Logical Reasoning via Structural Uncertainty

通过结构不确定性量化LLM逻辑推理中的一致性

Baishali Chaudhury, Mengdie Flora Wang, Hyunji Hayley Park, Rahul Ghosh, Sungmin Hong, Jae Oh Woo

机构 * AWS Generative AI Innovation Center（AWS生成式AI创新中心）

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.AI

AI总结提出结构不确定性框架，通过自偏好排序的稳定性评估LLM推理一致性，在逻辑和数学任务中与答案分散度互补，提升不可靠实例识别。

Comments Published at ICLR 2026 Workshop on Logical Reasoning of Large Language Models. Accepted as best paper

2406.13803 2025-11-21 cs.CL 89%

LLMs as Models for Analogical Reasoning

大型语言模型作为类比推理模型

Sam Musker, Alex Duchnowski, Raphaël Millière, Ellie Pavlick

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL

AI总结本文通过设计新的类比推理任务，探讨了大型语言模型在类比推理中的表现，发现其在某些条件下能匹配人类，但对特定变体反应不同，表明LLMs可能在某些领域提供可能的解释，但无法完全替代人类的类比能力。

Comments The title has been changed from Semantic Structure-Mapping in LLM and Human Analogical Reasoning to LLMs as Models for Analogical Reasoning to improve clarity and accuracy

Journal ref Journal of Memory and Language 145 (2025) 104676

2506.15211 2025-06-19 cs.CL 89%

ProtoReasoning: Prototypes as the Foundation for Generalizable Reasoning in LLMs

Feng He, Zijun Chen, Xinnian Liang, Tingting Ma, Yunqi Qiu, Shuangzhi Wu, Junchi Yan

机构 * Shanghai Jiao Tong University（上海交通大学）

专题命中逻辑推理：reasoning(title,abstract)；chain-of-thought(abstract)；CoT(abstract)；planning(abstract)

2607.21412 2026-08-12 cs.AI cs.CL cs.SE 版本更新 88%

Euclid-MCP: A Model Context Protocol Server for Deterministic Logical Reasoning via Prolog

欧几里得-MCP：一个通过Prolog进行确定性逻辑推理的模型上下文协议服务器

Bartolomeo Bogliolo

专题命中逻辑推理：reasoning(title,abstract)；logical reasoning(title,abstract)；分类 cs.CL、cs.AI

AI总结研究针对大型语言模型多步逻辑推理不可靠问题，提出开源的欧几里得-MCP服务器，通过引入欧几里得-IR及支持特定循环的工具接口实现确定性逻辑推理，经评估在处理大问题时效果优于LLMs，可作稳定推理基础。

1. 逻辑推理 3002 篇

CLR-Fact: Evaluating the Complex Logical Reasoning Capability of Large Language Models over Factual Knowledge

On the Hardness of Faithful Chain-of-Thought Reasoning in Large Language Models

LogiGAN: Learning Logical Reasoning via Adversarial Pre-training

Response: Emergent analogical reasoning in large language models

Reason Popper-ly: Patching In-Context Reasoning with Inductive Logic Programming

Exploring the Reversal Curse and Other Deductive Logical Reasoning in BERT and GPT-Based Large Language Models

SLR: Automated Synthesis for Scalable Logical Reasoning

RCoT-Seg: Reinforced Chain-of-Thought for Video Reasoning and Segmentation

MuSLR: Multimodal Symbolic Logical Reasoning

Live-E2T: Real-time Threat Monitoring in Video via Deduplicated Event Reasoning and Chain-of-Thought

SATBench: Benchmarking LLMs' Logical Reasoning via Automated Puzzle Generation from SAT Formulas

A Comparative Study of Neurosymbolic AI Approaches to Interpretable Logical Reasoning

ZebraLogic: On the Scaling Limits of LLMs for Logical Reasoning

Enhancing Non-Core Language Instruction-Following in Speech LLMs via Semi-Implicit Cross-Lingual CoT Reasoning

Evaluating the Robustness of Analogical Reasoning in Large Language Models

Enhancing Zero-Shot Chain-of-Thought Reasoning in Large Language Models through Logic

Boosting Logical Reasoning in Large Language Models through a New Framework: The Graph of Thought

In-Context Analogical Reasoning with Pre-Trained Language Models

APOLLO: A Simple Approach for Adaptive Pretraining of Language Models for Logical Reasoning

Improved Logical Reasoning of Language Models via Differentiable Symbolic Programming

Multimodal Analogical Reasoning over Knowledge Graphs

ReClor: A Reading Comprehension Dataset Requiring Logical Reasoning

Improving Symbolic Translation of Language Models for Logical Reasoning

Can Large Reasoning Models do Analogical Reasoning under Perceptual Uncertainty?

Knowledge-based Analogical Reasoning in Neuro-symbolic Latent Spaces

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders

Quantifying Consistency in LLM Logical Reasoning via Structural Uncertainty

LLMs as Models for Analogical Reasoning

ProtoReasoning: Prototypes as the Foundation for Generalizable Reasoning in LLMs

Euclid-MCP: A Model Context Protocol Server for Deterministic Logical Reasoning via Prolog