大模型推理能力 - arXivDaily 专题

2509.17380 2025-09-23 cs.AI 83%

Correlation or Causation: Analyzing the Causal Structures of LLM and LRM Reasoning Process

Zhizhang FU, Guangsheng Bao, Hongbo Zhang, Chenkai Hu, Yue Zhang

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

2503.17287 2025-09-23 cs.CL 83%

FastCuRL: Curriculum Reinforcement Learning with Stage-wise Context Scaling for Efficient Training R1-like Reasoning Models

Mingyang Song, Mao Zheng, Zheng Li, Wenjie Yang, Xuan Luo, Yue Pan, Feng Zhang

机构 * Tencent Hunyuan（腾讯文英）

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

Comments Accepted by EMNLP 2025

2507.20936 2025-09-23 cs.LG cs.AI cs.CL 82%

Dissecting Persona-Driven Reasoning in Language Models via Activation Patching

Ansh Poonia, Maeghal Jain

机构 * Independent（独立研究者）

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

Comments EMNLP (Findings) 2025

2507.14063 2025-09-23 cs.CL 79%

Collaborative Rational Speech Act: Pragmatic Reasoning for Multi-Turn Dialog

Lautaro Estienne, Gabriel Ben Zenou, Nona Naderi, Jackie Cheung, Pablo Piantanida

机构 * International Laboratory on Learning Systems (ILLS)（国际学习系统实验室）； Laboratoire Interdisciplinaire des Sciences du Numérique (LISN)（跨学科数字科学实验室）； Mila - Quebec AI Institute（魁北克人工智能研究所）； CNRS（法国国家科学研究中心）； CentraleSupélec（中央超导研究所）； Université Paris-Saclay（巴黎萨克雷大学）； McGill University（麦吉尔大学）

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

2508.18635 2025-09-23 cs.LG 79%

STRATA-TS: Selective Knowledge Transfer for Urban Time Series Forecasting with Retrieval-Guided Reasoning

Yue Jiang, Chenxi Liu, Yile Chen, Qin Chao, Shuai Liu, Cheng Long, Gao Cong

机构 * Nanyang Technological University（南洋理工大学）

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

2509.09712 2025-09-23 cs.CL cs.AI 79%

The Thinking Therapist: Training Large Language Models to Deliver Acceptance and Commitment Therapy using Supervised Fine-Tuning and Odds Ratio Policy Optimization

Talha Tahir

机构 * Department of Psychiatry University of Toronto（精神病学系多伦多大学）

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

2509.18076 2025-09-23 cs.AI 77%

Improving Large Language Models Function Calling and Interpretability via Guided-Structured Templates

Hy Dang, Tianyi Liu, Zhuofeng Wu, Jingfeng Yang, Haoming Jiang, Tao Yang, Pei Chen, Zhengyang Wang, Helen Wang, Huasheng Li, Bing Yin, Meng Jiang

机构 * University of Notre Dame（诺丁汉大学）； Amazon（亚马逊）

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

Comments Accepted to EMNLP 2025 Main Conference

2509.17917 2025-09-23 cs.AI 70%

Orcust: Stepwise-Feedback Reinforcement Learning for GUI Agent

Junyu Lu, Songxin Zhang, Zejian Xie, Zhuoyang Song, Jiaxing Zhang

机构 * Lionrock AI Lab（狮岩人工智能实验室）； China Merchants Research Institute of Advanced Technology（中国商人先进科技研究院）

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

2502.00919 2025-09-23 cs.CL cs.AI cs.LG 67%

Attention Sinks: A 'Catch, Tag, Release' Mechanism for Embeddings

Stephen Zhang, Mustafa Khan, Vardan Papyan

机构 * University of Toronto（多伦多大学）； Vector Institute（向量研究所）

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

2509.17765 2025-09-23 cs.CL cs.AI cs.CV eess.AS 62%

Qwen3-Omni Technical Report

Jin Xu, Zhifang Guo, Hangrui Hu, Yunfei Chu, Xiong Wang, Jinzheng He, Yuxuan Wang, Xian Shi, Ting He, Xinfa Zhu, Yuanjun Lv, Yongqi Wang, Dake Guo, He Wang, Linhan Ma, Pei Zhang, Xinyu Zhang, Hongkun Hao, Zishan Guo, Baosong Yang, Bin Zhang, Ziyang Ma, Xipin Wei, Shuai Bai, Keqin Chen, Xuejing Liu, Peng Wang, Mingkun Yang, Dayiheng Liu, Xingzhang Ren, Bo Zheng, Rui Men, Fan Zhou, Bowen Yu, Jianxin Yang, Le Yu, Jingren Zhou, Junyang Lin

机构 * Qwen Team（通义实验室）

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

Comments https://github.com/QwenLM/Qwen3-Omni

2509.17730 2025-09-23 cs.LG cs.CL 62%

ConfClip: Confidence-Weighted and Clipped Reward for Reinforcement Learning in LLMs

Bonan Zhang, Zhongqi Chen, Bowen Song, Qinya Li, Fan Wu, Guihai Chen

机构 * Shanghai Jiao Tong University Ant Group（上海交通大学-蚂蚁集团）

专题命中其他推理：reasoning(abstract)；分类 cs.CL、cs.LG

2509.16633 2025-09-23 cs.CV cs.AI cs.CL 62%

When Big Models Train Small Ones: Label-Free Model Parity Alignment for Efficient Visual Question Answering using Small VLMs

Abhirama Subramanyam Penamakuri, Navlika Singh, Piyush Arora, Anand Mishra

机构 * Indian Institute of Technology Jodhpur（印度理工学院朱诺尔）

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

Comments Accepted to EMNLP (Main) 2025

2509.16297 2025-09-23 cs.CY cs.AI cs.CL 62%

How Large Language Models are Designed to Hallucinate

Richard Ackermann, Simeon Emanuilov

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

Comments 23 pages, 2 tables, 2 figures

2505.23323 2025-09-23 cs.CL 57%

Neither Stochastic Parroting nor AGI: LLMs Solve Tasks through Context-Directed Extrapolation from Training Data Priors

Harish Tayyar Madabushi, Melissa Torgbi, Claire Bonial

机构 * University of Bath（巴斯大学）； DEVCOM Army Research Laboratory（陆军研究实验室）

专题命中其他推理：reasoning(abstract)；分类 cs.CL

2504.05652 2025-09-23 cs.CR cs.CL 57%

Sugar-Coated Poison: Benign Generation Unlocks LLM Jailbreaking

Yu-Hang Wu, Yu-Jie Xiong, Hao Zhang, Jia-Chen Zhang, Zheng Zhou

机构 * Shanghai University of Engineering Science（上海工程技术大学）； Institute of Computing Technology, Chinese Academy of Sciences（中国科学院计算技术研究所）

专题命中其他推理：reasoning(abstract)；分类 cs.CL

Comments Accepted by EMNLP2025

2509.17757 2025-09-23 cs.CV cs.MA 50%

Multi-Agent Amodal Completion: Direct Synthesis with Fine-Grained Semantic Guidance

Hongxing Fan, Lipeng Wang, Haohua Chen, Zehuan Huang, Jiangtao Wu, Lu Sheng

机构 * School of Computer Science and Engineering（计算机科学与工程学院）； Beihang University（北京航空航天大学）； School of Software（软件学院）

专题命中其他推理：reasoning(abstract)

2506.08283 2025-09-23 cs.IR 50%

Serendipitous Recommendation with Multimodal LLM

Haoting Wang, Jianling Wang, Hao Li, Fangjun Yi, Mengyu Fu, Youwei Zhang, Yifan Liu, Liang Liu, Minmin Chen, Ed H. Chi, Lichan Hong, Haokai Lu

专题命中其他推理：chain-of-thought(abstract)

Comments Accepted by 2025 Recsys EARL Workshop

2509.16970 2025-09-23 cs.CV 50%

LLM-Assisted Semantic Guidance for Sparsely Annotated Remote Sensing Object Detection

Wei Liao, Chunyan Xu, Chenxu Wang, Zhen Cui

机构 * Nanjing University of Science and Technology（南京理工大学）； Beijing Normal University（北京师范大学）

专题命中其他推理：reasoning(abstract)

2509.16895 2025-09-23 cs.IR 50%

Temporal-Aware User Behaviour Simulation with Large Language Models for Recommender Systems

Xinye Wanyan, Danula Hettiachchi, Chenglong Ma, Ziqi Xu, Jeffrey Chan

专题命中其他推理：reasoning(abstract)

1. 其他推理 19 篇

Correlation or Causation: Analyzing the Causal Structures of LLM and LRM Reasoning Process

FastCuRL: Curriculum Reinforcement Learning with Stage-wise Context Scaling for Efficient Training R1-like Reasoning Models

Dissecting Persona-Driven Reasoning in Language Models via Activation Patching

Collaborative Rational Speech Act: Pragmatic Reasoning for Multi-Turn Dialog

STRATA-TS: Selective Knowledge Transfer for Urban Time Series Forecasting with Retrieval-Guided Reasoning

The Thinking Therapist: Training Large Language Models to Deliver Acceptance and Commitment Therapy using Supervised Fine-Tuning and Odds Ratio Policy Optimization

Improving Large Language Models Function Calling and Interpretability via Guided-Structured Templates

Orcust: Stepwise-Feedback Reinforcement Learning for GUI Agent

Attention Sinks: A 'Catch, Tag, Release' Mechanism for Embeddings

Qwen3-Omni Technical Report

ConfClip: Confidence-Weighted and Clipped Reward for Reinforcement Learning in LLMs

When Big Models Train Small Ones: Label-Free Model Parity Alignment for Efficient Visual Question Answering using Small VLMs

How Large Language Models are Designed to Hallucinate

Neither Stochastic Parroting nor AGI: LLMs Solve Tasks through Context-Directed Extrapolation from Training Data Priors

Sugar-Coated Poison: Benign Generation Unlocks LLM Jailbreaking

Multi-Agent Amodal Completion: Direct Synthesis with Fine-Grained Semantic Guidance

Serendipitous Recommendation with Multimodal LLM

LLM-Assisted Semantic Guidance for Sparsely Annotated Remote Sensing Object Detection

Temporal-Aware User Behaviour Simulation with Large Language Models for Recommender Systems