大模型推理能力 - arXivDaily 专题

2504.14395 2025-04-22 cs.CV cs.AI cs.MA 85%

Hydra: An Agentic Reasoning Approach for Enhancing Adversarial Robustness and Mitigating Hallucinations in Vision-Language Models

Chung-En, Yu, Hsuan-Chih, Chen, Brian Jalaian, Nathaniel D. Bastian

机构 * University of West Florida（西弗吉尼亚大学）； New York University（纽约大学）； United States Military Academy（美国军事学院）

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

2504.09586 2025-04-15 cs.CL 85%

Short-Path Prompting in LLMs: Analyzing Reasoning Instability and Solutions for Robust Performance

Zuoli Tang, Junjie Ou, Kaiqin Hu, Chunwei Wu, Zhaoxin Huan, Chilin Fu, Xiaolu Zhang, Jun Zhou, Chenliang Li

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

Comments Under review

2503.12721 2025-04-15 cs.AI 85%

Can Reasoning Models Reason about Hardware? An Agentic HLS Perspective

Luca Collini, Andrew Hennessee, Ramesh Karri, Siddharth Garg

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

Comments 7 pages, submitted for peer review

2504.01700 2025-04-03 cs.HC cs.AI cs.RO 85%

Reasoning LLMs for User-Aware Multimodal Conversational Agents

Hamed Rahimi, Jeanne Cattoni, Meriem Beghili, Mouad Abrini, Mahdi Khoramshahi, Maribel Pino, Mohamed Chetouani

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

2503.11926 2025-03-18 cs.AI 85%

Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation

Bowen Baker, Joost Huizinga, Leo Gao, Zehao Dou, Melody Y. Guan, Aleksander Madry, Wojciech Zaremba, Jakub Pachocki, David Farhi

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

2406.16144 2025-03-18 cs.CL 85%

Chain-of-Probe: Examining the Necessity and Accuracy of CoT Step-by-Step

Zezhong Wang, Xingshan Zeng, Weiwen Liu, Yufei Wang, Liangyou Li, Yasheng Wang, Lifeng Shang, Xin Jiang, Qun Liu, Kam-Fai Wong

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

Comments Accepted by Findings of NAACL 2025

2503.08308 2025-03-12 cs.AI 85%

Seeing and Reasoning with Confidence: Supercharging Multimodal LLMs with an Uncertainty-Aware Agentic Framework

Zhuo Zhi, Chen Feng, Adam Daneshmend, Mine Orlu, Andreas Demosthenous, Lu Yin, Da Li, Ziquan Liu, Miguel R. D. Rodrigues

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

2502.18239 2025-02-26 cs.LG 85%

Unveiling and Causalizing CoT: A Causal Pespective

Jiarun Fu, Lizhong Ding, Hao Li, Pengqi Li, Qiuning Wei, Xu Chen

专题命中其他推理：CoT(title,abstract)；reasoning(abstract)；chain-of-thought(abstract)；分类 cs.LG

2412.13171 2024-12-18 cs.CL 85%

Compressed Chain of Thought: Efficient Reasoning Through Dense Representations

Jeffrey Cheng, Benjamin Van Durme

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

2410.16924 2024-10-23 cs.AI 85%

SleepCoT: A Lightweight Personalized Sleep Health Model via Chain-of-Thought Distillation

Huimin Zheng, Xiaofeng Xing, Xiangmin Xu

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

2311.05922 2024-03-11 cs.CL 85%

Chain of Thought with Explicit Evidence Reasoning for Few-shot Relation Extraction

Xilai Ma, Jing Li, Min Zhang

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

Comments Findings of EMNLP 2023

2310.19750 2023-10-31 cs.CL 85%

Chain-of-Thought Embeddings for Stance Detection on Social Media

Joseph Gatto, Omar Sharif, Sarah Masud Preum

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

Comments Accepted at EMNLP-2023, 8 pages

2310.17342 2023-10-27 cs.CL 85%

ACT-SQL: In-Context Learning for Text-to-SQL with Automatically-Generated Chain-of-Thought

Hanchong Zhang, Ruisheng Cao, Lu Chen, Hongshen Xu, Kai Yu

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

2310.04959 2023-10-10 cs.CL 85%

Towards Better Chain-of-Thought Prompting Strategies: A Survey

Zihan Yu, Liang He, Zhen Wu, Xinyu Dai, Jiajun Chen

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

2305.01879 2023-09-01 cs.CL 85%

SCOTT: Self-Consistent Chain-of-Thought Distillation

Peifeng Wang, Zhengyang Wang, Zheng Li, Yifan Gao, Bing Yin, Xiang Ren

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

Comments 11 pages, 8 figures. Accepted to ACL 2023

2608.01598 2026-08-04 cs.CL cs.AI 新提交 85%

PICTURE: Enhancing Theory-of-Mind in Large Language Models by Revealing, Not Hiding, Characters' Lack of Knowledge

PICTURE：通过揭示而非隐藏角色的知识缺失来增强大语言模型的心智理论能力

Eojin Jeon, SangKeun Lee

机构 * Korea University（高丽大学）

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

AI总结本研究针对大语言模型心智理论推理中事件隐藏导致的性能下降问题，提出PICTURE提示方法，通过在思维链中明确角色知识缺失，使模型在错误信念任务上性能提升7.3%

2603.12277 2026-06-30 cs.CL cs.AI cs.CR 85%

Prompt Injection as Role Confusion

提示注入作为角色混淆

Charles Ye, Jasmine Cui, Dylan Hadfield-Menell

机构 * Massachusetts Institute of Technology（麻省理工学院）

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

AI总结本文通过角色探测和CoT伪造攻击，揭示提示注入源于LLM对文本来源的角色感知混淆，并提出角色混淆程度可预测攻击成功率。

Comments ICML 2026

2607.18618 2026-07-22 cs.CL cs.AI cs.LG 新提交 85%

LatentMT: Machine Translation with Latent Reasoning

LatentMT：具有潜在推理的机器翻译

Wei-Rui Chen, Samar M. Magdy, Chiyu Zhang, Wenhui Zhu, Zhipeng Wang, Muhammad Abdul-Mageed

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

AI总结研究提出LatentMT，将潜在推理循环语言模型用于机器翻译，采用小型主干模型经轻量级训练，在多语言翻译方向表现出色，循环推理早期提升质量后期饱和，且训练和推理计算效率高，为紧凑高效的机器翻译提供新路径。

2604.22280 2026-07-17 cs.CV 版本更新 85%

Beyond Chain-of-Thought: Rewrite as a Universal Interface for Generative Multimodal Embeddings

超越思维链：重写作为生成式多模态嵌入的通用接口

Peixi Wu, Ke Mei, Feipeng Ma, Bosong Chai, Zhibin Lan, Chenxi Zhao, Shannan Yan, Jie Chen, Zhangchi Hu, Yansong Peng, Bo Lin, Junjie Zhou, Dacheng Yin, Tianyi Wang, Fengyun Rao, Jing Lyu, Hebei Li, Xiaoyan Sun

机构 * WeChat Vision, Tencent Inc.（腾讯微信视觉部）； Zhejiang University（浙江大学）； Tsinghua University（清华大学）； Institute of Artificial Intelligence, Hefei Comprehensive National Science Center（合肥综合性国家科学中心人工智能研究院）

专题命中其他推理：chain-of-thought(title,abstract)；reasoning(abstract)；CoT(abstract)

AI总结针对思维链推理在检索中产生冗余和语义歧义的问题，提出重写驱动的多模态嵌入框架RIME，联合优化生成与嵌入，并通过跨模态对齐和精炼强化学习实现高效准确的检索。

Comments Accepted by ACMMM 2026

2601.05175 2026-03-24 cs.CV 85%

VideoAuto-R1: Video Auto Reasoning via Thinking Once, Answering Twice

VideoAuto-R1：通过一次推理、两次回答实现视频自动推理

Shuming Liu, Mingchen Zhuge, Changsheng Zhao, Jun Chen, Lemeng Wu, Zechun Liu, Chenchen Zhu, Zhipeng Cai, Chong Zhou, Haozhe Liu, Ernie Chang, Saksham Suri, Hongyu Xu, Qi Qian, Wei Wen, Balakrishnan Varadarajan, Zhuang Liu, Hu Xu, Florian Bordes, Raghuraman Krishnamoorthi, Bernard Ghanem, Vikas Chandra, Yunyang Xiong

机构 * Meta AI ； King Abdullah University of Science and Technology (KAUST)（卡布斯大学）； Princeton University（普林斯顿大学）

专题命中其他推理：reasoning(title,abstract)；chain-of-thought(abstract)；CoT(abstract)

AI总结本文提出VideoAuto-R1框架，通过一次推理两次回答策略提升视频理解效率，实现准确率和效率的双重提升。

Comments Accepted to CVPR 2026. Project page: https://ivul-kaust.github.io/projects/videoauto-r1/

2603.09877 2026-03-11 cs.CV 85%

InternVL-U: Democratizing Unified Multimodal Models for Understanding, Reasoning, Generation and Editing

InternVL-U: 使统一多模态模型在理解、推理、生成和编辑方面更加普及

Changyao Tian, Danni Yang, Guanzhou Chen, Erfei Cui, Zhaokai Wang, Yuchen Duan, Penghao Yin, Sitao Chen, Ganlin Yang, Mingxin Liu, Zirun Zhu, Ziqian Fan, Leyao Gu, Haomin Wang, Qi Wei, Jinhui Yin, Xue Yang, Zhihang Zhong, Qi Qin, Yi Xin, Bin Fu, Yihao Liu, Jiaye Ge, Qipeng Guo, Gen Luo, Hongsheng Li, Yu Qiao, Kai Chen, Hongjie Zhang

机构 * Shanghai AI Laboratory（上海人工智能实验室）； Fudan University（复旦大学）； University of Science and Technology of China（中国科学技术大学）； Shanghai Jiao Tong University（上海交通大学）； South China University of Technology（华南理工大学）； Nanjing University（南京大学）； Xiamen University（厦门大学）； CUHK MMLab（香港中文大学MMLab）

专题命中其他推理：reasoning(title,abstract)；chain-of-thought(abstract)；CoT(abstract)

AI总结 InternVL-U通过轻量级统一多模态模型，在保持强大生成能力的同时，实现了在理解、推理、生成和编辑方面的高效平衡。

Comments technical report, 61 pages, https://github.com/OpenGVLab/InternVL-U

2602.06034 2026-02-26 cs.CV 85%

V-Retrver: Evidence-Driven Agentic Reasoning for Universal Multimodal Retrieval

V-Retrver: 以证据驱动的代理推理用于通用多模态检索

Dongyang Chen, Chaoyang Wang, Dezhao Su, Xi Xiao, Zeyu Zhang, Jing Xiong, Qing Li, Yuzhang Shang, Shichao Kan

机构 * Tsinghua University（清华大学）； University of Central Florida（中央佛罗里达大学）； Fudan University（复旦大学）； The Australian National University（澳大利亚国立大学）； The University of Hong Kong（香港大学）； Pengcheng Laboratory（鹏城实验室）； Central South University（中南大学）

专题命中其他推理：reasoning(title,abstract)；chain-of-thought(abstract)；CoT(abstract)

AI总结 V-Retrver通过引入证据驱动的代理推理机制，提升多模态检索的准确率和推理可靠性。

Comments Project page: https://github.com/chendy25/V-Retrver

2504.10829 2026-02-06 cs.CV 85%

LayoutCoT: Unleashing the Deep Reasoning Potential of Large Language Models for Layout Generation

LayoutCoT: 解锁大型语言模型在布局生成中的深度推理潜力

Hengyu Shi, Junhao Su, Tianyang Han, Junfeng Luo, Jialin Gao

机构 * Meituan（美团）

专题命中其他推理：reasoning(title,abstract)；chain-of-thought(abstract)；CoT(abstract)

AI总结 LayoutCoT通过结合RAG和CoT技术，利用LLMs的推理能力实现无需训练的高质量布局生成。

2504.14642 2025-12-16 cs.CV 85%

Relation-R1: Progressively Cognitive Chain-of-Thought Guided Reinforcement Learning for Unified Relation Comprehension

Relation-R1: 逐步认知链式思维引导的强化学习用于统一关系理解

Lin Li, Wei Chen, Jiahui Li, Kwang-Ting Cheng, Long Chen

专题命中其他推理：chain-of-thought(title,abstract)；reasoning(abstract)；CoT(abstract)

AI总结 Relation-R1通过整合认知链式思维引导的强化学习，实现了对二元和N元关系的统一理解，提升了视觉-语义定位能力。

Comments AAAI 2026

2503.13939 2025-10-28 cs.CV 85%

Med-R1: Reinforcement Learning for Generalizable Medical Reasoning in Vision-Language Models

Yuxiang Lai, Jike Zhong, Ming Li, Shitian Zhao, Yuheng Li, Konstantinos Psounis, Xiaofeng Yang

机构 * Department of Computer Science and Informatics, Emory University（计算机科学与信息学系，埃默里大学）； Department of Computer Science and Department of Electrical and Computer Engineering, University of Southern California（计算机科学系和电气与计算机工程系，南加州大学）； Department of Computer Science, University of Tokyo（计算机科学系，东京大学）； Department of Computer Science, Johns Hopkins University（计算机科学系，约翰霍普金斯大学）； Department of Biomedical Engineering, Georgia Institute of Technology and Emory University（生物医学工程系，佐治亚理工学院和埃默里大学）

专题命中其他推理：reasoning(title,abstract)；chain-of-thought(abstract)；CoT(abstract)

2508.12587 2025-09-25 cs.CV 85%

Multimodal Chain of Continuous Thought for Latent-Space Reasoning in Vision-Language Models

Tan-Hanh Pham, Chris Ngo

机构 * Harvard Medical School, Harvard University（哈佛医学院、哈佛大学）； Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital（阿提尼乌拉A.马丁努斯生物医学成像中心、麻省总医院）； Knovel Engineering Lab, Singapore（Knovel工程实验室、新加坡）

专题命中其他推理：reasoning(title,abstract)；chain-of-thought(abstract)；CoT(abstract)

2508.04416 2025-09-04 cs.CV 85%

Thinking With Videos: Multimodal Tool-Augmented Reinforcement Learning for Long Video Reasoning

Haoji Zhang, Xin Gu, Jiawen Li, Chixiang Ma, Sule Bai, Chubin Zhang, Bowen Zhang, Zhichao Zhou, Dongliang He, Yansong Tang

专题命中其他推理：reasoning(title,abstract)；chain-of-thought(abstract)；CoT(abstract)

2505.21425 2025-08-13 cs.SE 85%

GUARD:Dual-Agent based Backdoor Defense on Chain-of-Thought in Neural Code Generation

Naizhu Jin, Zhong Li, Tian Zhang, Qingkai Zeng

专题命中其他推理：chain-of-thought(title,abstract)；reasoning(abstract)；CoT(abstract)

Comments Accepted by SEKE 2025

2507.02859 2025-07-04 cs.CV 85%

Bootstrapping Grounded Chain-of-Thought in Multimodal LLMs for Data-Efficient Model Adaptation

Jiaer Xia, Bingkui Tong, Yuhang Zang, Rui Shao, Kaiyang Zhou

机构 * Hong Kong Baptist University（香港 Baptist 大学）； Sichuan University（四川大学）； Shanghai AI Lab（上海人工智能实验室）； Harbin Institute of Technology (Shenzhen)（哈尔滨工业大学（深圳））

专题命中其他推理：chain-of-thought(title,abstract)；reasoning(abstract)；CoT(abstract)

Comments Accepted by ICCV2025

2504.20769 2025-04-30 cs.CL cs.AI cs.LG 85%

Chain-of-Defensive-Thought: Structured Reasoning Elicits Robustness in Large Language Models against Reference Corruption

Wenxiao Wang, Parsa Hosseini, Soheil Feizi

机构 * Department of Computer Science, University of Maryland, College Park, Maryland, USA（计算机科学系，马里兰大学，College Park，马里兰，美国）

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

1. 其他推理 5817 篇

Hydra: An Agentic Reasoning Approach for Enhancing Adversarial Robustness and Mitigating Hallucinations in Vision-Language Models

Short-Path Prompting in LLMs: Analyzing Reasoning Instability and Solutions for Robust Performance

Can Reasoning Models Reason about Hardware? An Agentic HLS Perspective

Reasoning LLMs for User-Aware Multimodal Conversational Agents

Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation

Chain-of-Probe: Examining the Necessity and Accuracy of CoT Step-by-Step

Seeing and Reasoning with Confidence: Supercharging Multimodal LLMs with an Uncertainty-Aware Agentic Framework

Unveiling and Causalizing CoT: A Causal Pespective

Compressed Chain of Thought: Efficient Reasoning Through Dense Representations

SleepCoT: A Lightweight Personalized Sleep Health Model via Chain-of-Thought Distillation

Chain of Thought with Explicit Evidence Reasoning for Few-shot Relation Extraction

Chain-of-Thought Embeddings for Stance Detection on Social Media

ACT-SQL: In-Context Learning for Text-to-SQL with Automatically-Generated Chain-of-Thought

Towards Better Chain-of-Thought Prompting Strategies: A Survey

SCOTT: Self-Consistent Chain-of-Thought Distillation

PICTURE: Enhancing Theory-of-Mind in Large Language Models by Revealing, Not Hiding, Characters' Lack of Knowledge

Prompt Injection as Role Confusion

LatentMT: Machine Translation with Latent Reasoning

Beyond Chain-of-Thought: Rewrite as a Universal Interface for Generative Multimodal Embeddings

VideoAuto-R1: Video Auto Reasoning via Thinking Once, Answering Twice

InternVL-U: Democratizing Unified Multimodal Models for Understanding, Reasoning, Generation and Editing

V-Retrver: Evidence-Driven Agentic Reasoning for Universal Multimodal Retrieval

LayoutCoT: Unleashing the Deep Reasoning Potential of Large Language Models for Layout Generation

Relation-R1: Progressively Cognitive Chain-of-Thought Guided Reinforcement Learning for Unified Relation Comprehension

Med-R1: Reinforcement Learning for Generalizable Medical Reasoning in Vision-Language Models

Multimodal Chain of Continuous Thought for Latent-Space Reasoning in Vision-Language Models

Thinking With Videos: Multimodal Tool-Augmented Reinforcement Learning for Long Video Reasoning

GUARD:Dual-Agent based Backdoor Defense on Chain-of-Thought in Neural Code Generation

Bootstrapping Grounded Chain-of-Thought in Multimodal LLMs for Data-Efficient Model Adaptation

Chain-of-Defensive-Thought: Structured Reasoning Elicits Robustness in Large Language Models against Reference Corruption