大模型推理能力

Artificial Intelligence is stupid and causal reasoning won't fix it

John Mark Bishop

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

An Ontological AI-and-Law Framework for the Autonomous Levels of AI Legal Reasoning

Lance Eliot

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Comments 22 pages, 3 figures

A Simple Approach to Case-Based Reasoning in Knowledge Bases

Rajarshi Das, Ameya Godbole, Shehzaad Dhuliawala, Manzil Zaheer, Andrew McCallum

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.CL

Better Set Representations For Relational Reasoning

Qian Huang, Horace He, Abhay Singh, Yan Zhang, Ser-Nam Lim, Austin Benson

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.LG

Comments Preprint, 17 pages

Efficient Reasoning in Regular Boardgames

Jakub Kowalski, Radosław Miernik, Maksymilian Mika, Wojciech Pawlik, Jakub Sutowicz, Marek Szykuła, Andrzej Tkaczyk

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Comments IEEE Conference on Games 2020

CNRL at SemEval-2020 Task 5: Modelling Causal Reasoning in Language with Multi-Head Self-Attention Weights based Counterfactual Detection

Rajaswa Patil, Veeky Baths

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.CL

Comments Submitted at SemEval-2020 workshop

User Memory Reasoning for Conversational Recommendation

Hu Xu, Seungwhan Moon, Honglei Liu, Bing Liu, Pararth Shah, Bing Liu, Philip S. Yu

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.CL

Reinforcement Learning with Iterative Reasoning for Merging in Dense Traffic

Maxime Bouton, Alireza Nakhaei, David Isele, Kikuo Fujimura, Mykel J. Kochenderfer

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Comments 6pages, 5 figures

Journal ref IEEE Intelligent Transportation Systems Conference (ITSC) 2020

A Novel Fuzzy Approximate Reasoning Method Based on Extended Distance Measure in SISO Fuzzy System

I. M. Son, S. I. Kwak, U. J. Han, J. H. Pak, M. Han, J. R. Pyon, U. S. Ryu

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Comments 24 pages, 1 figures

Subjective Knowledge and Reasoning about Agents in Multi-Agent Systems

Shikha Singh, Deepak Khemani

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Top k Memory Candidates in Memory Networks for Common Sense Reasoning

Vatsal Mahajan

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Comments 3 pages

Satisficing Mentalizing: Bayesian Models of Theory of Mind Reasoning in Scenarios with Different Uncertainties

Jan Pöppel, Stefan Kopp

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Comments This paper is an extended version of the paper "Satisficing Models of Bayesian Theory of Mind for Explaining Behavior of Differently Uncertain Agents" by the same authors, submitted to AAMAS 2018

Reasoning in Highly Reactive Environments

Francesco Pacenza

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Comments In Proceedings ICLP 2019, arXiv:1909.07646

Journal ref EPTCS 306, 2019, pp. 420-426

Conversational AI : Open Domain Question Answering and Commonsense Reasoning

Kinjal Basu

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Comments In Proceedings ICLP 2019, arXiv:1909.07646

Journal ref EPTCS 306, 2019, pp. 396-402

BlockPuzzle - A Challenge in Physical Reasoning and Generalization for Robot Learning

Yixiu Zhao, Ziyin Liu

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Five dimensions of reasoning in the wild

Don Perlis

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Comments minor typos corrected from AAAI version, Proceedings (Blue-Sky track) AAAI-2016, Phoenix AZ

Variable Forgetting in Reasoning about Knowledge

Kaile Su, Abdul Sattar, Guanfeng Lv, Yan Zhang

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Journal ref Journal Of Artificial Intelligence Research, Volume 35, pages 677-716, 2009

Representing and Reasoning With Probabilistic Knowledge: A Bayesian Approach

Marie desJardins

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Comments Appears in Proceedings of the Ninth Conference on Uncertainty in Artificial Intelligence (UAI1993)

Shadows and Headless Shadows: an Autobiographical Approach to Narrative Reasoning

Ladislau Boloni

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Comments arXiv admin note: substantial text overlap with arXiv:1211.5643

Real-time retrieval for case-based reasoning in interactive multiagent-based simulations

Pierre De Loor, Romain Bénard, Chevaillier Pierre

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.AI

Journal ref Expert Systems with Applications 38, 5 (2011) 5145-5153

A Hybrid Reasoning Model for Indirect Answers

Nancy Green, Sandra Carberry

专题命中 其他推理 ：reasoning(title,abstract)；分类 cs.CL

Comments To appear in Proc. of ACL-94. 8 pages, uuencoded compressed Postscript file; extract with Unix uudecode and uncompress. Contact Author for latex version

When AI Agents Disagree Like Humans: Reasoning Trace Analysis for Human-AI Collaborative Moderation

当AI代理像人类一样产生分歧：用于人机协作 moderation 的推理轨迹分析

Michał Wawer, Jarosław A. Chudziak

专题命中 其他推理 ：reasoning(title,abstract)

AI总结 本文研究AI代理在仇恨言论 moderation 中的分歧模式，通过推理轨迹分析发现分歧结构比幅度更能预测人类判断需求，提出从共识寻求转向不确定性揭示的多代理设计。

Comments Accepted to the ICLR 2026 Workshop on "From Human Cognition to AI Reasoning: Models, Methods, and Applications (HCAIR)

Not All Tokens Are Equal: Inflation-Aware Routing for Agentic LLM Systems

并非所有token都平等：面向智能体大语言模型系统的感知通胀路由

Heming Fu, Shan Lin, Qianqian Xie, Guojun Xiong

机构 * Stony Brook University（石溪大学） ； Wuhan University（武汉大学） ； Shanghai Jiao Tong University（上海交通大学） 纠错

专题命中 其他推理 ：CoT(abstract,abstract_cn)；reasoning(abstract)；分类 cs.CL、cs.AI

AI总结 针对智能体LLM系统的token通胀问题，提出四阶段路由框架InflationAgent，通过测量通胀、引入CBE信号、最大化SER并采用新鲜升级策略，在GSM8K上优于FrugalGPT。

Evaluating LLM-Based Goal Extraction in Requirements Engineering: Prompting Strategies and Their Limitations

评估基于大语言模型的目标提取在需求工程中的应用：提示策略及其局限性

Anna Arnaudo, Riccardo Coppola, Maurizio Morisio, Flavio Giobergia, Andrea Bioddo, Angelo Bongiorno, Luca Dadone

机构 * Department of Control and Computer Engineering（控制与计算机工程系） 纠错

专题命中 其他推理 ：CoT(abstract,abstract_cn)；chain-of-thought(abstract)；分类 cs.CL、cs.AI

AI总结 本文探讨了通过三个阶段自动提取功能目标以实现目标导向的需求工程，提出基于工程提示的LLM链，实验表明反馈循环机制在零样本学习中表现更优，但提示策略仍是性能限制因素。

Comments 11 pages, 1 figure. This contribution will be published in the conference proceedings of EASE 2026 Conference (https://conf.researchr.org/home/ease-2026/prompt-se-2026)

OneReason Technical Report

OneReason 技术报告

OneRec Team, Biao Yang, Boyang Ding, Chenglong Chu, Dunju Zang, Fei Pan, Han Li, Hao Jiang, Honghui Bao, Huanjie Wang, Jian Liang, Jiangxia Cao, Jiao Ou, Jiaxin Deng, Jinghao Zhang, Kun Gai, Lu Ren, Peiru Du, Pengfei Zheng, Rongzhou Zhang, Ruiming Tang, Shiyao Wang, Siyang Mao, Siyuan Lou, Teng Shi, Wei Yuan, Wenlong Xu, Xingchen Liu, Xingmei Wang, Xinqi Jin, Yan Sun, Yan Wang, Yifei Hu, Yingzhi He, Yufei Ye, Yuhao Wang, Yunhao Zhou, Yuqin Dai, Zhao Liu, Zhipeng Wei, Zhixin Ling, Ziming Li, Zixing Zhang, Ziyuan Liu, An Zhang, Changxin Lao, Chaoyi Ma, Chengru Song, Defu Lian, Fan Yang, Guowang Zhang, Hao Peng, Jiayao Shen, Jie Chen, Jun Xu, Junmin Chen, Kun Zhang, Kuo Cai, Mingxing Wen, Minmao Wang, Minxuan Lv, Qi Zhang, Qiang Luo, Sheng Yu, Shijie Li, Shijie Yi, Shuang Yang, Shugui Liu, Shuni Chen, Tinghai Zhang, Tingting Gao, Xiang Wang, Xiangyu Wu, Xiangyu Zhao, Xiao Lv, Xiaoyou Zhou, Xuming Wang, Yong Du, Zejian Zhang, Zhaojie Liu, Zhiyang Zhang, Zhuang Zhuang, Ziqi Wang, Ziyi Zhao

机构 * OneRec Team（OneRec团队） 纠错

专题命中 其他推理 ：reasoning(abstract)；chain-of-thought(abstract)；CoT(abstract)；分类 cs.CL、cs.AI

AI总结 针对生成式推荐模型中推理能力难以激活的问题，提出 OneReason 方法，通过增强感知和认知能力实现有效推理。

Comments Work in progress

TOOLCAD: Exploring Tool-Using Large Language Models in Text-to-CAD Generation with Reinforcement Learning

TOOLCAD: 探索用于文本到CAD生成的工具使用大型语言模型与强化学习

Yifei Gong, Xing Wu, Wenda Liu, Kang Tu

机构 * School of Computer Engineering & Science, Shanghai University（上海大学计算机工程与科学学院） 纠错

专题命中 其他推理 ：CoT(abstract,abstract_cn)；reasoning(abstract)；分类 cs.CL、cs.AI

AI总结 本文提出TOOLCAD框架，利用LLM作为工具使用代理进行文本到CAD生成，并通过交互式CAD建模环境和强化学习提升模型性能。

Comments ACL2026

Rectifying LLM Thought from Lens of Optimization

从优化角度纠正大语言模型的思维

Junnan Liu, Hongwei Liu, Songyang Zhang, Kai Chen

机构 * Monash University（莫纳什大学） ； Shanghai Artificial Intelligence Laboratory（上海人工智能实验室） 纠错

专题命中 其他推理 ：reasoning(abstract)；chain-of-thought(abstract)；CoT(abstract)；分类 cs.CL、cs.AI

AI总结 本文通过优化视角分析LLM推理过程，提出RePro方法改进推理性能，减少过度思考等亚优行为。

Comments Accepted by ICLR 2026

Finding and Reactivating Post-Trained LLMs' Hidden Safety Mechanisms

发现并重新激活已训练大语言模型的隐藏安全机制

Mingjie Li, Wai Man Si, Michael Backes, Yang Zhang, Yisen Wang

机构 * Cranberry-Lemon University（蔓越莓柠檬大学） 纠错

专题命中 其他推理 ：reasoning(abstract)；chain-of-thought(abstract)；CoT(abstract)；分类 cs.CL、cs.AI

AI总结 本文研究了大语言模型在后训练过程中安全机制的退化原因，提出SafeReAct方法通过LoRA适配器恢复被抑制的安全行为，提升模型在有害提示下的安全性而不影响推理性能。

Debiasing Large Language Models toward Social Factors in Online Behavior Analytics through Prompt Knowledge Tuning

通过提示知识微调去偏大型语言模型在在线行为分析中的社会因素

Hossein Salemi, Jitin Krishnan, Hemant Purohit

机构 * George Mason University（乔治梅森大学） 纠错

专题命中 其他推理 ：reasoning(abstract)；chain-of-thought(abstract)；CoT(abstract)；分类 cs.CL、cs.AI

AI总结 本文研究了在社交情境中通过引入用户目标推断 dispositional 因果性和消息上下文推断 situational 因果性对 LLM 性能的影响，提出了一种通过增强指令提示来减少社会归因偏见的方法，提升零样本分类任务的性能。

Comments This is a preprint of the accepted paper for publication in IEEE Transactions on Computational Social Systems

Knowledge Distillation for Large Language Models

为大型语言模型进行知识蒸馏

Alejandro Paredes La Torre, Barbara Flores, Diego Rodriguez

专题命中 其他推理 ：reasoning(abstract)；chain-of-thought(abstract)；CoT(abstract)；分类 cs.CL、cs.AI

AI总结 本文提出一种高效的压缩框架，结合引导的思维链强化学习，通过知识蒸馏和链式思维提示提升模型效率，实现更小规模的模型部署。

Comments Code and data are available at: https://github.com/AlejandroParedesLT/knowledge_distillLLM