语言大模型 / LLM - arXivDaily 专题

2210.09132 2022-10-18 cs.CL 74%

Pseudo-OOD training for robust language models

Dhanasekar Sundararaman, Nikhil Mehta, Lawrence Carin

专题命中其他LLM ：language model(title)；分类 cs.CL

Comments Work in progress

2205.10183 2022-10-06 cs.CL 74%

Prototypical Calibration for Few-shot Learning of Language Models

Zhixiong Han, Yaru Hao, Li Dong, Yutao Sun, Furu Wei

专题命中其他LLM ：language model(title)；分类 cs.CL

2110.07483 2022-08-02 cs.CL 74%

On the Pitfalls of Analyzing Individual Neurons in Language Models

Omer Antverg, Yonatan Belinkov

专题命中其他LLM ：language model(title)；分类 cs.CL

Comments ICLR 2022 Main Conference

2206.07988 2022-06-17 cs.AI 74%

PreCogIIITH at HinglishEval : Leveraging Code-Mixing Metrics & Language Model Embeddings To Estimate Code-Mix Quality

Prashant Kodali, Tanmay Sachan, Akshay Goindani, Anmol Goel, Naman Ahuja, Manish Shrivastava, Ponnurangam Kumaraguru

专题命中其他LLM ：language model(title)；分类 cs.AI

2203.09904 2022-03-21 cs.CL 74%

Do Multilingual Language Models Capture Differing Moral Norms?

Katharina Hämmerl, Björn Deiseroth, Patrick Schramowski, Jindřich Libovický, Alexander Fraser, Kristian Kersting

专题命中其他LLM ：language model(title)；分类 cs.CL

2109.12393 2021-09-28 cs.CL 74%

Sorting through the noise: Testing robustness of information processing in pre-trained language models

Lalchand Pandia, Allyson Ettinger

专题命中其他LLM ：language model(title)；分类 cs.CL

Comments Accepted at EMNLP 2021

2106.01536 2021-09-21 cs.CL 74%

BERT meets LIWC: Exploring State-of-the-Art Language Models for Predicting Communication Behavior in Couples' Conflict Interactions

Jacopo Biggiogera, George Boateng, Peter Hilpert, Matthew Vowels, Guy Bodenmann, Mona Neysari, Fridtjof Nussbeck, Tobias Kowatsch

专题命中其他LLM ：language model(title)；分类 cs.CL

Comments 5 pages. Accepted at the 2nd Workshop on Social Affective Multimodal Interaction for Health (SAMIH) at ICMI 2021

2108.13654 2021-09-01 cs.CL 74%

Discretized Integrated Gradients for Explaining Language Models

Soumya Sanyal, Xiang Ren

专题命中其他LLM ：language model(title)；分类 cs.CL

Comments Accepted in EMNLP 2021

2106.01541 2021-06-04 cs.CL 74%

MPC-BERT: A Pre-Trained Language Model for Multi-Party Conversation Understanding

Jia-Chen Gu, Chongyang Tao, Zhen-Hua Ling, Can Xu, Xiubo Geng, Daxin Jiang

专题命中其他LLM ：language model(title)；分类 cs.CL

Comments Accepted by ACL 2021

2104.08017 2021-04-19 cs.SE cs.CL 74%

BERT2Code: Can Pretrained Language Models be Leveraged for Code Search?

Abdullah Al Ishtiaq, Masum Hasan, Md. Mahim Anjum Haque, Kazi Sajeed Mehrab, Tanveer Muttaqueen, Tahmid Hasan, Anindya Iqbal, Rifat Shahriyar

专题命中其他LLM ：language model(title)；分类 cs.CL

Comments Submitted to ICANN2021

2101.05494 2021-01-15 cs.CL 74%

Hostility Detection in Hindi leveraging Pre-Trained Language Models

Ojasv Kamal, Adarsh Kumar, Tejas Vaidhya

专题命中其他LLM ：language model(title)；分类 cs.CL

2011.02541 2020-11-06 cs.CL 74%

MTLB-STRUCT @PARSEME 2020: Capturing Unseen Multiword Expressions Using Multi-task Learning and Pre-trained Masked Language Models

Shiva Taslimipoor, Sara Bahaadini, Ekaterina Kochmar

专题命中其他LLM ：language model(title)；分类 cs.CL

Comments accepted for publication at MWE-LEX 2020 Workshop at COLING

1910.11559 2020-08-12 cs.CL cs.SD eess.AS 74%

SpeechBERT: An Audio-and-text Jointly Learned Language Model for End-to-end Spoken Question Answering

Yung-Sung Chuang, Chi-Liang Liu, Hung-Yi Lee, Lin-shan Lee

专题命中其他LLM ：language model(title)；分类 cs.CL

Comments Interspeech 2020

2007.13306 2020-07-28 cs.CL cs.CY 74%

Public Sentiment Toward Solar Energy: Opinion Mining of Twitter Using a Transformer-Based Language Model

Serena Y. Kim, Koushik Ganesan, Princess Dickens, Soumya Panda

专题命中其他LLM ：language model(title)；分类 cs.CL

1809.01201 2019-04-03 cond-mat.dis-nn cs.CL cs.FL 74%

Random Language Model

E. DeGiuli

专题命中其他LLM ：language model(title)；分类 cs.CL

Comments 5 pages + 3 pages SI

Journal ref Phys. Rev. Lett. 122, 128301 (2019)

1607.03895 2016-07-15 cs.CL physics.soc-ph 74%

Tie-breaker: Using language models to quantify gender bias in sports journalism

Liye Fu, Cristian Danescu-Niculescu-Mizil, Lillian Lee

专题命中其他LLM ：language model(title)；分类 cs.CL

Comments Best paper award at the IJCAI workshop on NLP Meets Journalism; 5 pages, 2 figures; data and other info available at http://www.cs.cornell.edu/~liye/tennis.html

cs/0402055 2010-06-01 cs.CL 74%

Lexical Base as a Compressed Language Model of the World (on the material of the Ukrainian language)

Solomiya Buk

专题命中其他LLM ：language model(title)；分类 cs.CL

Comments 8 pages, 2 tables

Journal ref Psychology of Language and Communication. 2009, vol. 13, no. 2, pp. 35-44

cmp-lg/9711003 2009-11-30 cmp-lg cs.CL 74%

Probabilistic Parsing Using Left Corner Language Models

Christopher D. Manning, Bob Carpenter

专题命中其他LLM ：language model(title)；分类 cs.CL

Comments 12 pages, uses iwpt97.sty

Journal ref Proceedings of the Fifth International Workshop on Parsing Technologies, MIT, Boston MA, 1997

2608.10279 2026-08-25 cs.CR cs.AI cs.CL 版本更新 73%

Withholding the Completing Chunk: Exact Release-Boundary Equivalence for Production Streaming Guardrails

withhold the Completing Chunk: 面向流式大语言模型输出的确定性配对完成护栏

Christopher M. Frost

专题命中其他LLM ：LLM(summary_cn)；分类 cs.CL、cs.AI

AI总结本研究提出流式LLM输出的确定性配对完成护栏，通过扫描前缀扣留配对完成块，实验验证其效果，表明它是小型固定策略的精确发布边界后盾。

Comments 9 pages, 2 figures, 4 tables. Substantially revised version

2608.19875 2026-08-21 cs.CL cs.AI 新提交 73%

A knowledge-guided agentic framework for mitigating patient-context ambiguity in health queries

一种用于缓解医疗查询中患者上下文歧义的知识引导智能体框架

Mahyar Abbasian, Saba A. Farahani, Arshia Ilaty, Hung Cao, Ramesh Jain, Amir M. Rahmani

专题命中其他LLM ：language model(abstract)；prompting(abstract)；分类 cs.CL、cs.AI

AI总结该研究提出知识引导智能体框架，通过解析医疗查询、构建假设、识别缺失上下文并提问，缓解患者上下文歧义，在诊断检索和饮食安全分类任务中显著提升了多种语言模型的性能。

Comments 48 pages, 3 figures, 6 tables, journal

2608.19220 2026-08-21 cs.CL cs.AI 新提交 73%

Can Conversational AI loosen Us-Versus-Them Boundaries? The Effects of Common, Dual, and Separate Identity Framings on Pro-Immigrant Intergroup Helping

对话式AI能否弱化“我们与他们”的界限？共同、双重及分离身份框架对支持移民的群体间帮助的影响

Oluwadamilola Jeboda, John F. Dovidio, Jonas R. Kunst

专题命中其他LLM ：LLM(abstract,abstract_cn)；分类 cs.CL、cs.AI

AI总结该研究通过与GPT-4o开展五轮对话的预先注册实验，发现强调共同或双重身份的AI对话可弱化群体间界限，提升行动意愿，且效应在各调节变量中基本一致。

2608.12387 2026-08-14 cs.CL cs.AI 新提交 73%

Query Timing Produces Opposite Positional Biases Between LLMs and Humans

查询时机在大型语言模型（LLMs）与人类之间产生相反的位置偏差

Jasin Cekinmez, Addison J. Wu, Thomas L. Griffiths

机构 * Princeton University（普林斯顿大学）

专题命中其他LLM ：large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

AI总结该研究探究查询时机对LLMs和人类位置偏差的影响，发现二者存在差异，且新LLMs的位置偏差比前代更显著。

Comments Entropic Award (Top 3 Paper), ICBINB @ ICLR 2026

2501.14844 2026-08-14 cs.CL cs.AI cs.MA 版本更新 73%

Unmasking Conversational Bias in AI Multiagent Systems

揭示AI多智能体系统中的对话偏见

Erica Coppolillo, Giuseppe Manco, Luca Maria Aiello

专题命中其他LLM ：large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

AI总结该研究针对AI多智能体系统的对话偏见检测空白，提出对话型LLMs多智能体系统偏见量化框架，发现现有问卷式方法无法检测到该场景下的偏见，呼吁开发更复杂的偏见检测与缓解工具包。

2608.11197 2026-08-12 cs.LG cs.CL 新提交 73%

Beyond a Bag of Features: Set-Level Instability in Sparse Autoencoders

超越特征袋：稀疏自编码器中的集合级不稳定性

Nikolai Bolik, Lennart Stöpler, Artur Andrzejak

机构 * Heidelberg University（海德堡大学）

专题命中其他LLM ：LLM(abstract,abstract_cn)；分类 cs.CL、cs.LG

AI总结该研究以稀疏自编码器（SAE）隐集合重叠度为相似度度量，发现SAE特征不通过简单特征袋语义组合，其激活集合与人类概念判断存在显著不匹配。

2608.08881 2026-08-11 cs.AI cs.CL 新提交 73%

Theory-Guided Deception Detection: A RAG-Based Artificial Intelligence Exploration

理论引导的欺骗检测：基于检索增强生成（RAG）的人工智能探索

David M. Markowitz, Timothy R. Levine

专题命中其他LLM ：large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

AI总结该研究基于欺骗理论开发7个RAG模型，对比4个大语言模型与基线模型的欺骗判断，发现RAG模型反应偏差更低但准确率无显著差异，理论引导AI判断当前参数下不可靠，优化后或有潜力。

2608.06779 2026-08-10 q-bio.QM cs.AI cs.CL 新提交 73%

Genotypic Triggers: Exposing Pharmacogenomic Blind Spots via Host-Specific Backdoors in Generative Antimicrobial Peptide Models

基因型触发器：通过生成抗菌肽模型中宿主特定后门暴露药物基因组学盲区

Doniyorkhon Obidov, Xiaolong Guo, Yonghui Li, Kaichen Yang

专题命中其他LLM ：large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

AI总结该研究提出基因型触发器后门攻击，操控抗菌肽生成模型使其针对特定HLA等位基因携带者生成高免疫原性风险肽，同时保留抗菌效力与低毒性，以暴露药物基因组学盲区。

2607.13246 2026-08-04 cs.LG cs.AI 版本更新 73%

Reassessing Muon for Matrix Factorization

重新评估用于矩阵分解的Muon优化器

Ali Parviz, Gal Mishne, Alex Cloninger

机构 * Halicioğlu Data Science Institute, UC San Diego（加州大学圣地亚哥分校哈利西奥卢数据科学研究所）； Department of Mathematics, UC San Diego（加州大学圣地亚哥分校数学系）

专题命中其他LLM ：large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

AI总结研究在低秩矩阵分解问题上重新评估Muon优化器，通过与精心调整的自适应基线对比，发现Muon在此设置下不始终优于AdamW，一些优势对超参数选择敏感，更细致地展现谱感知正交化何时有益，主张在受控问题上评估优化器。

2602.02288 2026-08-04 cs.LG cs.AI 版本更新 73%

AROpt: An Optimization Method for Autoregressive Time Series Forecasting

AROpt: 一种用于自回归时间序列预测的优化方法

Zheng Li, Jerry Cheng, Huanying Gu

机构 * Department of Computer Science（计算机科学系）； New York Institute of Technology（纽约理工学院）

专题命中其他LLM ：large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

AI总结本文提出一种新方法，通过强制预测误差随预测时间增加而增长，提升自回归时间序列预测的准确性与灵活性，实验表明在多个基准上取得新突破。

Comments 16 pages, 5 figures, 6 tables

2307.14544 2026-07-28 cs.CL cs.AI 版本更新 73%

Speed Reading Tool Powered by Artificial Intelligence for Students with ADHD, Dyslexia, and Short Attention Span

由人工智能驱动的针对患有注意力缺陷多动障碍、阅读障碍和注意力持续时间短的学生的快速阅读工具

Megat Irfan Zackry Bin Ismail, Ahmad Nazran bin Yusri, Muhammad Hafizzul Bin Abdul Manap, Muhammad Muizzuddin Bin Kamarozaman

专题命中其他LLM ：large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

AI总结针对阅读障碍、多动症及注意力不集中学生，提出用云托管大语言模型与自适应排版结合的人工智能工具，通过向Gemini API请求实时生成摘要，接受多种文件输入，优化可读性，经Flask框架部署，显著提升阅读速度与理解能力。

2607.18256 2026-07-22 cs.AI cs.LG 新提交 73%

PEARL: Solver-in-the-Loop Interactive Optimization Modeling from Natural Language

PEARL：基于自然语言的循环求解器交互式优化建模

Hongliang Lu, Zhong Li, Yuxuan Chen, Yuan Lan, Fan Zhang, Zaiwen Wen

专题命中其他LLM ：large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

AI总结研究旨在解决优化建模问题，提出PEARL系统，通过在循环中使用Python执行和数学编程求解器，学习优化策略，在多轮工具集成设置下提升求解率，在优化建模任务中表现优于强大基线。

1. 其他LLM 12251 篇

Pseudo-OOD training for robust language models

Prototypical Calibration for Few-shot Learning of Language Models

On the Pitfalls of Analyzing Individual Neurons in Language Models

PreCogIIITH at HinglishEval : Leveraging Code-Mixing Metrics & Language Model Embeddings To Estimate Code-Mix Quality

Do Multilingual Language Models Capture Differing Moral Norms?

Sorting through the noise: Testing robustness of information processing in pre-trained language models

BERT meets LIWC: Exploring State-of-the-Art Language Models for Predicting Communication Behavior in Couples' Conflict Interactions

Discretized Integrated Gradients for Explaining Language Models

MPC-BERT: A Pre-Trained Language Model for Multi-Party Conversation Understanding

BERT2Code: Can Pretrained Language Models be Leveraged for Code Search?

Hostility Detection in Hindi leveraging Pre-Trained Language Models

MTLB-STRUCT @PARSEME 2020: Capturing Unseen Multiword Expressions Using Multi-task Learning and Pre-trained Masked Language Models

SpeechBERT: An Audio-and-text Jointly Learned Language Model for End-to-end Spoken Question Answering

Public Sentiment Toward Solar Energy: Opinion Mining of Twitter Using a Transformer-Based Language Model

Random Language Model

Tie-breaker: Using language models to quantify gender bias in sports journalism

Lexical Base as a Compressed Language Model of the World (on the material of the Ukrainian language)

Probabilistic Parsing Using Left Corner Language Models

Withholding the Completing Chunk: Exact Release-Boundary Equivalence for Production Streaming Guardrails

A knowledge-guided agentic framework for mitigating patient-context ambiguity in health queries

Can Conversational AI loosen Us-Versus-Them Boundaries? The Effects of Common, Dual, and Separate Identity Framings on Pro-Immigrant Intergroup Helping

Query Timing Produces Opposite Positional Biases Between LLMs and Humans

Unmasking Conversational Bias in AI Multiagent Systems

Beyond a Bag of Features: Set-Level Instability in Sparse Autoencoders

Theory-Guided Deception Detection: A RAG-Based Artificial Intelligence Exploration

Genotypic Triggers: Exposing Pharmacogenomic Blind Spots via Host-Specific Backdoors in Generative Antimicrobial Peptide Models

Reassessing Muon for Matrix Factorization

AROpt: An Optimization Method for Autoregressive Time Series Forecasting

Speed Reading Tool Powered by Artificial Intelligence for Students with ADHD, Dyslexia, and Short Attention Span

PEARL: Solver-in-the-Loop Interactive Optimization Modeling from Natural Language