语言大模型 / LLM

LLMmap: Fingerprinting For Large Language Models

Dario Pasquini, Evgenios M. Kornaropoulos, Giuseppe Ateniese

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

Comments Appearing in the proceedings of the 34th USENIX Security Symposium

Evaluating the Ability of Large Language Models to Generate Verifiable Specifications in VeriFast

Wen Fan, Marilyn Rego, Xin Hu, Sanya Dod, Zhaorui Ni, Danning Xie, Jenna DiVincenzo, Lin Tan

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

Assured Automatic Programming via Large Language Models

Martin Mirchev, Andreea Costea, Abhishek Kr Singh, Abhik Roychoudhury

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

ReasoningRec: Bridging Personalized Recommendations and Human-Interpretable Explanations through LLM Reasoning

Millennium Bismay, Xiangjue Dong, James Caverlee

专题命中 推理与问题求解 ：LLM(title,abstract)；large language model(abstract,comments)；language model(abstract,comments)；分类 cs.AI

Comments Large Language Model, Recommendation, Human-Interpretable Reasoning, Personalization

Beyond Graphs: Can Large Language Models Comprehend Hypergraphs?

Yifan Feng, Chengwu Yang, Xingliang Hou, Shaoyi Du, Shihui Ying, Zongze Wu, Yue Gao

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

Autoformalization of Game Descriptions using Large Language Models

Agnieszka Mensfelt, Kostas Stathis, Vince Trencsenyi

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

Comments code: https://github.com/dicelab-rhul/game-formaliser

Are Large Language Models Capable of Generating Human-Level Narratives?

Yufei Tian, Tenghao Huang, Miri Liu, Derek Jiang, Alexander Spangher, Muhao Chen, Jonathan May, Nanyun Peng

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

Comments EMNLP 2024

DynamicRouteGPT: A Real-Time Multi-Vehicle Dynamic Navigation Framework Based on Large Language Models

Ziai Zhou, Bin Zhou, Hao Liu

专题命中 推理与问题求解 ：language model(title,abstract)；large language model(title)；分类 cs.AI

Comments This paper is 12 pages long and represents the initial draft, version 1

STEER: Assessing the Economic Rationality of Large Language Models

Narun Raman, Taylor Lundy, Samuel Amouyal, Yoav Levine, Kevin Leyton-Brown, Moshe Tennenholtz

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

Plan of Thoughts: Heuristic-Guided Problem Solving with Large Language Models

Houjun Liu

专题命中 推理与问题求解 ：language model(title,abstract)；large language model(title)；分类 cs.CL

Comments 7 pages, 2 figures

Little Red Riding Hood Goes Around the Globe:Crosslingual Story Planning and Generation with Large Language Models

Evgeniia Razumovskaia, Joshua Maynez, Annie Louis, Mirella Lapata, Shashi Narayan

专题命中 推理与问题求解 ：language model(title,abstract)；large language model(title)；分类 cs.CL

Comments Accepted to LREC-COLING 2024

Prompting Large Vision-Language Models for Compositional Reasoning

Timothy Ossowski, Ming Jiang, Junjie Hu

专题命中 推理与问题求解 ：language model(title,abstract)；prompting(title)；分类 cs.AI

HSC-GPT: A Large Language Model for Human Settlements Construction

Chen Ran, Yao Xueqi, Jiang Xuhui, Han Zhengqi, Guo Jingze, Zhang Xianyue, Lin Chunyu, Liu Chumin, Zhao Jing, Lian Zeke, Zhang Jingjing, Li Keke

专题命中 推理与问题求解 ：language model(title,abstract)；large language model(title)；分类 cs.CL

On the Planning Abilities of Large Language Models : A Critical Investigation

Karthik Valmeekam, Matthew Marquez, Sarath Sreedharan, Subbarao Kambhampati

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

Comments NeurIPS 2023 Spotlight. arXiv admin note: substantial text overlap with arXiv:2206.10498

From Complex to Simple: Unraveling the Cognitive Tree for Reasoning with Small Language Models

Junbing Yan, Chengyu Wang, Taolin Zhang, Xiaofeng He, Jun Huang, Wei Zhang

专题命中 推理与问题求解 ：language model(title,abstract)；small language model(title)；分类 cs.CL

Comments emnlp 2023

BRAINTEASER: Lateral Thinking Puzzles for Large Language Models

Yifan Jiang, Filip Ilievski, Kaixin Ma, Zhivar Sourati

专题命中 推理与问题求解 ：language model(title,abstract)；large language model(title)；分类 cs.CL

ALCUNA: Large Language Models Meet New Knowledge

Xunjian Yin, Baizhou Huang, Xiaojun Wan

专题命中 推理与问题求解 ：language model(title,abstract)；large language model(title)；分类 cs.CL

Comments EMNLP 2023

Learning Horn Envelopes via Queries from Large Language Models

Sophie Blum, Raoul Koudijs, Ana Ozaki, Samia Touileb

专题命中 推理与问题求解 ：language model(title,abstract)；large language model(title)；分类 cs.LG

Comments 35 pages, 2 figures; manuscript accepted for publication in the International Journal of Approximate Reasoning (IJAR)

Robot Task Planning Based on Large Language Model Representing Knowledge with Directed Graph Structures

Yue Zhen, Sheng Bi, Lu Xing-tong, Pan Wei-qin, Shi Hai-peng, Chen Zi-rui, Fang Yi-shu

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

PertMind: Eliciting Emergent Biological Reasoning in LLM via Reinforcement Learning on Cellular Perturbation Data

PertMind：通过细胞扰动数据的强化学习在大语言模型中引出涌现生物推理能力

Zhenchao Tang, Xiaogang Xu, Tianxu Lv, Jiahui Guan, Jiale Zhou, Haohuai He, Zhi Song, Hanbo Huang, Jiehui Huang, Jiafei Wu, Zhe Liu

专题命中 推理与问题求解 ：LLM(title)；large language model(abstract)；language model(abstract)；post-training(abstract)

AI总结 本研究提出PertMind模型，结合可信轨迹监督初始化与多层面强化信号，以细胞扰动图谱为强化学习环境训练，实现生物推理能力的涌现，且可迁移至多类生物任务。

Comments Project page: https://shapsider.github.io/PertMind/ Code: https://github.com/shapsider/PertMind Model: https://huggingface.co/tzcfly/PertMind

Prior Audit-Repair Context Shifts LLM Verifier Thresholds Toward Leniency

先验审计-修复上下文调整大语言模型验证器阈值以趋向宽松

Parsa Mazaheri, Kasra Mazaheri

机构 * University of California, Santa Cruz（加州大学圣克鲁兹分校） ； Massachusetts Institute of Technology（麻省理工学院） 纠错

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

AI总结 该研究发现，LLM检查器上下文中的审计→修复事件会降低误报率，调整了验证器阈值趋向宽松，且该效应与负性不对称性预测相反，不随推理启用而消失。

Comments 12 pages, 2 figures, 4 tables. Code and analysis artefacts: https://github.com/parsa-mz/crtitxer

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 推理与问题求解 ：LLM(title,summary_cn)；language model(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。

Measuring in-context algorithmic reasoning in language models against an exact Bayes-optimal standard

基于精确贝叶斯最优标准测量语言模型的上下文内算法推理能力

Hector Zenil, Luan Ozelim

机构 * Oxford Immune Algorithmics（牛津免疫算法学公司） ； Oxford University Innovation（牛津大学创新公司） ； London Institute for Healthcare Engineering（伦敦医疗工程研究所） ； King’s College London（伦敦国王学院） 纠错

专题命中 推理与问题求解 ：language model(title,abstract)；large language model(abstract)；post-training(abstract)；分类 cs.AI、cs.LG

AI总结 本研究提出F-ICL基准，以精确贝叶斯最优标准测量语言模型的上下文内算法推理能力，发现多数模型分布与最优标准存在差距，该差距与准确率无关且不受模型规模缩小，基准已开放。

A Comprehensive FP8 Training Recipe for Reasoning-Enhanced Language Models

用于推理增强型语言模型的综合FP8训练方案

Wenjun Wang, Shuo Cai, Congkai Xie, Mingfa Feng, Yiming Zhang, Zhen Li, Kejing Yang, Ming Li, Jiannong Cao, Hongxia Yang

机构 * The Hong Kong Polytechnic University（香港理工大学） 纠错

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

AI总结 本文提出整合持续预训练与监督微调的端到端FP8训练方案，通过混合粒度量化实现高效无损失训练，效率较BF16提升显著，将发布代码推动大规模模型训练普及。

Comments This paper has been withdrawn by the authors due to a significant bug discovered in our data processing pipeline. This bug affects the validity of the experimental results, and we can no longer stand by the conclusions presented

Think Short, Defer Smart, Act, and Repeat: Calibrated Reasoning and Uncertainty-Aware Deferral for Edge LLM Agents

短思考、智能延迟、行动与循环：面向边缘大语言模型智能体的校准推理与感知不确定性的延迟机制

Amirmohammad Farzaneh, Osvaldo Simeone

专题命中 推理与问题求解 ：LLM(title,summary_cn)；分类 cs.AI、cs.LG

AI总结 该研究提出TSDS框架，结合收敛探测器与困惑度延迟规则，经LTT流程校准后，可在边缘LLM智能体上减少43%-73%的思考计算量，同时保证奖励与云端调用率，在多类基准任务上表现优异。

A Cost-Effective Multimodal LLM Reasoning Framework for Question Answering over Irregular Clinical Time Series

一种用于不规则临床时间序列问答的经济高效多模态大语言模型推理框架

Frank Nie, Ethan B Liu, Yuan Zhu, Wei Fan, Jindong Han

机构 * Shandong University（山东大学） ； University of Auckland（奥克兰大学） 纠错

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

AI总结 针对不规则临床时间序列问答，提出ClinPRISM框架。该框架含不规则感知多尺度编码器、时间证据蒸馏器和渐进对齐策略，通过构建训练数据，在40亿参数大语言模型主干上实现最优性能，兼具低时间序列令牌数和低推理延迟。

Silent Failures in Quantized LLM Reasoning: A Taxonomy-Based Analysis of Hollow Convergence and Failure Mode Shifts

量化语言模型推理中的无声失败：基于分类法的空洞收敛和失败模式转移分析

Renuka Oladri, Mohan Vamsi Varadaraju Priya, Jerry Wu

专题命中 推理与问题求解 ：LLM(title)；large language model(abstract)；language model(abstract)；post-training(abstract)

AI总结 研究量化大语言模型推理中的无声失败，用六类失败分类法对五个模型在三种精度和四个基准下的思维链输出分类，发现空洞收敛等问题有精度和基准依赖性，且无法从文本特征可靠检测，揭示了标准评估管道未捕捉到的失败模式。

Discourse-Aware Policy Analysis with Argumentation: A Hybrid LLM-Symbolic Framework for Disaster Governance

基于论证的话语感知政策分析：灾害治理的混合大语言模型-符号框架

Stylianos Loukas Vasileiou, Olga Derendiaeva

机构 * New Mexico State University（新墨西哥州立大学） ； Sun Yat-Sen University（中山大学） 纠错

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

AI总结 研究灾害治理中政策文件推理隐含问题，提出混合大语言模型-符号框架Apaf，通过分类论点、生成框架介导关系子类型进行关键话语分析，发布新数据集，论证图准确、可解释且跨辖区稳定。

A Model Can Help Itself: Reward-Free Self-Training for LLM Reasoning

模型可以自我帮助：无需奖励的自我训练用于大语言模型推理

Mengqi Li, Lei Zhao, Anthony Man-Cho So, Ruoyu Sun, Xiao Li

机构 * The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳）） ； Shanghai Jiao Tong University（上海交通大学） 纠错

专题命中 推理与问题求解 ：LLM(title)；post-training(abstract,abstract_cn)；language model(abstract)；分类 cs.AI、cs.LG

AI总结 本文提出SePT方法，通过自我生成和训练自我生成响应交替进行，无需外部奖励提升大语言模型推理性能，实验表明在多个数学推理基准上有效。

Bridging Symbolic Control and Neural Reasoning in LLM Agents -- The Structured Cognitive Loop

连接符号控制与神经推理的LLM代理——结构认知循环

Myung Ho Kim

机构 * JEI University（JEI大学） 纠错

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

AI总结 本文提出结构认知循环(SCL)架构，通过分离检索、认知、控制、行动和记忆模块，解决LLM代理的推理与执行纠缠、记忆波动和行动序列失控问题，实现零策略违规、防止冗余调用并保持决策可追溯性。

Comments This update clarifies the theoretical architecture by separating Regulation as the Soft Symbolic Control layer from Control as a deterministic runtime engine, while adding explicit discussion of how the current implementation should be interpreted in light of that distinction