语言大模型 / LLM - arXivDaily 专题

2410.09426 2025-08-12 cs.CL cs.LG 86%

FlatQuant: Flatness Matters for LLM Quantization

Yuxuan Sun, Ruikang Liu, Haoli Bai, Han Bao, Kang Zhao, Yuening Li, Jiaxin Hu, Xianzhi Yu, Lu Hou, Chun Yuan, Xin Jiang, Wulong Liu, Jun Yao

机构 * Huawei Noah's Ark Lab（华为诺亚实验室）； Shenzhen International Graduate School, Tsinghua University（清华大学深圳国际研究生院）； The Chinese University of Hong Kong（香港中文大学）

专题命中效率与部署：LLM(title)；large language model(abstract)；language model(abstract)；post-training(abstract)

Comments 27 pages, accepted to ICML 2025

2508.07054 2025-08-12 cs.LG cs.AI 86%

Membership and Memorization in LLM Knowledge Distillation

Ziqi Zhang, Ali Shahin Shamsabadi, Hanxiao Lu, Yifeng Cai, Hamed Haddadi

机构 * Peking University（北京大学）； Brave Software（Brave软件公司）； Purdue University（普渡大学）； Imperial College London（伦敦帝国理工学院）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

2508.04451 2025-08-07 cs.LG cs.AI 86%

Automatic LLM Red Teaming

Roman Belaire, Arunesh Sinha, Pradeep Varakantham

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

2508.03487 2025-08-06 cs.SE cs.AI cs.LG 86%

BitsAI-Fix: LLM-Driven Approach for Automated Lint Error Resolution in Practice

Yuanpeng Li, Qi Long, Zhiyuan Yao, Jian Xu, Lintao Xie, Xu He, Lu Geng, Xin Han, Yueyan Chen, Wenbo Duan

机构 * ByteDance（字节跳动）； Carnegie Mellon University（卡内基梅隆大学）； Zhejiang University（浙江大学）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

2508.03148 2025-08-06 cs.LG cs.AI cs.DC 86%

Frontier: Simulating the Next Generation of LLM Inference Systems

Yicheng Feng, Xin Tan, Kin Hang Sew, Yimin Jiang, Yibo Zhu, Hong Xu

机构 * The Chinese University of Hong Kong（香港中文大学）； StepFun

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

2508.00904 2025-08-05 cs.PF cs.AI cs.AR cs.LG 86%

Forecasting LLM Inference Performance via Hardware-Agnostic Analytical Modeling

Rajeev Patwari, Ashish Sirasao, Devleena Das

机构 * Advanced Micro Devices (AMD)（先进微器件（AMD））

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

Comments 10 pages, 9 figures

2507.21170 2025-08-04 cs.CR cs.AI cs.CL 86%

OneShield -- the Next Generation of LLM Guardrails

Chad DeLuca, Anna Lisa Gentile, Shubhi Asthana, Bing Zhang, Pawan Chowdhary, Kellen Cheng, Basel Shbita, Pengyuan Li, Guang-Jie Ren, Sandeep Gopisetty

机构 * IBM Research（IBM研究院）； Princeton University（普林斯顿大学）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

2507.21276 2025-07-30 cs.AI cs.CL cs.DC 86%

LeMix: Unified Scheduling for LLM Training and Inference on Multi-GPU Systems

Yufei Li, Zexin Li, Yinglun Zhu, Cong Liu

机构 * University of California, Riverside（加州大学河滨分校）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

Comments Accepted by RTSS 2025

2507.20018 2025-07-30 cs.CY cs.AI cs.CL 86%

The Carbon Cost of Conversation, Sustainability in the Age of Language Models

Sayed Mahbub Hasan Amiri, Prasun Goswami, Md. Mainul Islam, Mohammad Shakhawat Hossen, Sayed Majhab Hasan Amiri, Naznin Akter

专题命中效率与部署：language model(title,abstract)；LLM(abstract)；large language model(abstract)；分类 cs.CL、cs.AI

Comments 22 Pages, 5 Tables

2506.15690 2025-07-25 cs.LG cs.AI cs.SI stat.ME 86%

LLM Web Dynamics: Tracing Model Collapse in a Network of LLMs

Tianyu Wang, Akira Horiguchi, Lingyou Pang, Carey E. Priebe

机构 * Johns Hopkins University（约翰霍普金斯大学）； University of California, Davis（加州大学戴维斯分校）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

2501.01144 2025-07-25 cs.CL cs.LG 86%

BlockDialect: Block-wise Fine-grained Mixed Format Quantization for Energy-Efficient LLM Inference

Wonsuk Jang, Thierry Tambe

机构 * Department of Electrical Engineering, Stanford University, CA, USA（电气工程系，斯坦福大学）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.LG

Comments ICML 2025

2502.08773 2025-07-23 cs.CL cs.LG 86%

Universal Model Routing for Efficient LLM Inference

Wittawat Jitkrittum, Harikrishna Narasimhan, Ankit Singh Rawat, Jeevesh Juneja, Congchao Wang, Zifeng Wang, Alec Go, Chen-Yu Lee, Pradeep Shenoy, Rina Panigrahy, Aditya Krishna Menon, Sanjiv Kumar

机构 * Google（谷歌）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.LG

2507.14238 2025-07-22 cs.CL cs.AI cs.CY 86%

Language Models Change Facts Based on the Way You Talk

Matthew Kearney, Reuben Binns, Yarin Gal

机构 * Oxford University（牛津大学）

专题命中效率与部署：language model(title,abstract)；LLM(abstract)；large language model(abstract)；分类 cs.CL、cs.AI

2507.09019 2025-07-15 cs.LG cs.AI cs.DC 86%

On Evaluating Performance of LLM Inference Serving Systems

Amey Agrawal, Nitin Kedia, Anmol Agarwal, Jayashree Mohan, Nipun Kwatra, Souvik Kundu, Ramachandran Ramjee, Alexey Tumanov

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

2507.08877 2025-07-15 cs.LG cs.AI 86%

ODIA: Oriented Distillation for Inline Acceleration of LLM-based Function Calling

Hanlong Zhang, Jingsheng Yang, Hao Li, Yuhao He, Franck Gong

机构 * ByteDance Inc.（字节跳动公司）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

2501.00555 2025-07-15 cs.LG cs.AI stat.AP stat.ML 86%

Prune 'n Predict: Optimizing LLM Decision-making with Conformal Prediction

Harit Vishwakarma, Alan Mishler, Thomas Cook, Niccolò Dalmasso, Natraj Raman, Sumitra Ganesh

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

Journal ref 42nd International Conference on Machine Learning (ICML 2025)

2312.11462 2025-07-15 cs.LG cs.CL 86%

Cascade Speculative Drafting for Even Faster LLM Inference

Ziyi Chen, Xiaocong Yang, Jiacheng Lin, Chenkai Sun, Kevin Chen-Chuan Chang, Jie Huang

机构 * University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.LG

Comments NeurIPS 2024

2507.03871 2025-07-08 cs.LG cs.AI cs.HC 86%

Enhancing Adaptive Behavioral Interventions with LLM Inference from Participant-Described States

Karine Karine, Benjamin M. Marlin

机构 * University of Massachusetts Amherst（马萨诸塞大学阿姆赫斯特分校）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

Comments Accepted at Machine Learning for Healthcare (MLHC) 2025

2507.02595 2025-07-04 cs.CL cs.AI 86%

MPF: Aligning and Debiasing Language Models post Deployment via Multi Perspective Fusion

Xin Guan, PeiHsin Lin, Zekun Wu, Ze Wang, Ruibo Zhang, Emre Kazim, Adriano Koshiyama

机构 * University College London（伦敦大学学院）； Center for long-term AI（长期人工智能中心）

专题命中效率与部署：language model(title,abstract)；LLM(abstract)；large language model(abstract)；分类 cs.CL、cs.AI

Comments Accepted at ICML 2025 AIW Workshop

2507.01413 2025-07-03 cs.GT cs.AI cs.LG 86%

Evaluating LLM Agent Collusion in Double Auctions

Kushal Agrawal, Verona Teo, Juan J. Vazquez, Sudarsh Kunnavakkam, Vishak Srikanth, Andy Liu

机构 * stanford（斯坦福大学）； caltech（加州理工学院）； yale（耶鲁大学）； cmu（卡内基梅隆大学）； arb（Arb研究）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

2412.12386 2025-07-01 cs.CL cs.LG 86%

Interpretable LLM-based Table Question Answering

Giang Nguyen, Ivan Brugere, Shubham Sharma, Sanjay Kariyappa, Anh Totti Nguyen, Freddy Lecue

机构 * Auburn University（阿伯拉姆大学）； J.P. Morgan（摩根大通）； NVIDIA（英伟达）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.LG

Comments Published in Transactions on Machine Learning Research (TMLR) in 06/2025. Reviews at: https://openreview.net/forum?id=2eTsZBoU2W

2506.22396 2025-06-30 cs.CL cs.AI 86%

QuickSilver -- Speeding up LLM Inference through Dynamic Token Halting, KV Skipping, Contextual Token Fusion, and Adaptive Matryoshka Quantization

Danush Khanna, Aditya Kumar Guru, Srivarshinee Sridhar, Zidan Ahmed, Rubhav Bahirwani, Meetu Malhotra, Vinija Jain, Aman Chadha, Amitava Das, Kripabandhu Ghosh

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

Comments Preprint. Under submission

2502.15294 2025-06-30 cs.CL cs.AI 86%

Round Attention: A Novel Round-Level Attention Mechanism to Accelerate LLM Inference

Yaohua Tang, Zhicheng Hu, Kun Cheng, Fan Mo, Qiheng Lv, Hua Wang, Zhi Chen

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.AI

2408.00523 2025-06-26 cs.CR cs.AI cs.LG 86%

Fuzz-Testing Meets LLM-Based Agents: An Automated and Efficient Framework for Jailbreaking Text-To-Image Generation Models

Yingkai Dong, Xiangtao Meng, Ning Yu, Zheng Li, Shanqing Guo

机构 * School of Cyber Science and Technology, Shandong University（山东大学网络科学与技术学院）； Netflix Eyeline Studios ； State Key Laboratory of Cryptography and Digital Economy Security, Shandong University（山东大学密码与数字经济安全国家重点实验室）； Shandong Key Laboratory of Artificial Intelligence Security, Shandong University（山东省人工智能安全重点实验室）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

2506.18129 2025-06-24 cs.CL cs.AI 86%

$ϕ^{\infty}$: Clause Purification, Embedding Realignment, and the Total Suppression of the Em Dash in Autoregressive Language Models

Bugra Kilictas, Faruk Alpay

机构 * Bahcesehir University（巴切希尔大学）

专题命中效率与部署：language model(title,abstract)；large language model(abstract)；foundation model(abstract)；分类 cs.CL、cs.AI

Comments 16 pages, 3 figures

2506.13514 2025-06-17 cs.CL cs.LG cs.NA math.NA 86%

TensorSLM: Energy-efficient Embedding Compression of Sub-billion Parameter Language Models on Low-end Devices

Mingxue Xu, Yao Lei Xu, Danilo P. Mandic

机构 * Imperial College London（伦敦帝国理工学院）

专题命中效率与部署：language model(title,abstract)；large language model(abstract)；small language model(abstract)；分类 cs.CL、cs.LG

Comments ICML 2025 Workshop on Tiny Titans: The next wave of On-Device Learning for Foundational Models (TTODLer-FM)

2506.12204 2025-06-17 cs.LG cs.AI cs.OS 86%

Semantic Scheduling for LLM Inference

Wenyue Hua, Dujian Ding, Yile Gu, Yujie Ren, Kai Mei, Minghua Ma, William Yang Wang

机构 * University of California, Santa Barbara（加州大学圣芭芭拉分校）； University of British Columbia（不列颠哥伦比亚大学）； University of Washington, Seattle（华盛顿大学（西雅图））； EPFL（苏黎世联邦理工学院）； Rutgers University, New Brunswick（罗格斯大学新 Brunswick分校）； Microsoft（微软公司）

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

Comments 18 pages, 3 figures

2502.19002 2025-06-16 cs.LG cs.AI math.OC stat.ML 86%

The Sharpness Disparity Principle in Transformers for Accelerating Language Model Pre-Training

Jinbo Wang, Mingze Wang, Zhanpeng Zhou, Junchi Yan, Weinan E, Lei Wu

机构 * School of Mathematical Sciences, Peking University（北京大学数学科学学院）； Center for Machine Learning Research, Peking University（北京大学机器学习研究中心）； Shanghai Jiao Tong University（上海交通大学）； AI for Science Institute, Beijing, China（北京人工智能科学研究院）

专题命中效率与部署：language model(title,abstract)；LLM(abstract)；large language model(abstract)；分类 cs.AI、cs.LG

Comments 21 pages, accepted by ICML 2025

2506.09452 2025-06-12 cs.LG cs.CL cs.CR cs.IT math.IT 86%

Learning Obfuscations Of LLM Embedding Sequences: Stained Glass Transform

Jay Roberts, Kyle Mylonakis, Sidhartha Roy, Kaan Kale

机构 * Protopia AI

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.CL、cs.LG

Comments Submitted to IEEE S&P 2026

2506.07330 2025-06-10 cs.LG cs.AI cs.CR 86%

JavelinGuard: Low-Cost Transformer Architectures for LLM Security

Yash Datta, Sharath Rajasekar

专题命中效率与部署：LLM(title,abstract)；large language model(abstract)；language model(abstract)；分类 cs.AI、cs.LG

Comments 16 pages, 1 Figure and 5 Tables

1. 效率与部署 22280 篇

FlatQuant: Flatness Matters for LLM Quantization

Membership and Memorization in LLM Knowledge Distillation

Automatic LLM Red Teaming

BitsAI-Fix: LLM-Driven Approach for Automated Lint Error Resolution in Practice

Frontier: Simulating the Next Generation of LLM Inference Systems

Forecasting LLM Inference Performance via Hardware-Agnostic Analytical Modeling

OneShield -- the Next Generation of LLM Guardrails

LeMix: Unified Scheduling for LLM Training and Inference on Multi-GPU Systems

The Carbon Cost of Conversation, Sustainability in the Age of Language Models

LLM Web Dynamics: Tracing Model Collapse in a Network of LLMs

BlockDialect: Block-wise Fine-grained Mixed Format Quantization for Energy-Efficient LLM Inference

Universal Model Routing for Efficient LLM Inference

Language Models Change Facts Based on the Way You Talk

On Evaluating Performance of LLM Inference Serving Systems

ODIA: Oriented Distillation for Inline Acceleration of LLM-based Function Calling

Prune 'n Predict: Optimizing LLM Decision-making with Conformal Prediction

Cascade Speculative Drafting for Even Faster LLM Inference

Enhancing Adaptive Behavioral Interventions with LLM Inference from Participant-Described States

MPF: Aligning and Debiasing Language Models post Deployment via Multi Perspective Fusion

Evaluating LLM Agent Collusion in Double Auctions

Interpretable LLM-based Table Question Answering

QuickSilver -- Speeding up LLM Inference through Dynamic Token Halting, KV Skipping, Contextual Token Fusion, and Adaptive Matryoshka Quantization

Round Attention: A Novel Round-Level Attention Mechanism to Accelerate LLM Inference

Fuzz-Testing Meets LLM-Based Agents: An Automated and Efficient Framework for Jailbreaking Text-To-Image Generation Models

$ϕ^{\infty}$: Clause Purification, Embedding Realignment, and the Total Suppression of the Em Dash in Autoregressive Language Models

TensorSLM: Energy-efficient Embedding Compression of Sub-billion Parameter Language Models on Low-end Devices

Semantic Scheduling for LLM Inference

The Sharpness Disparity Principle in Transformers for Accelerating Language Model Pre-Training

Learning Obfuscations Of LLM Embedding Sequences: Stained Glass Transform

JavelinGuard: Low-Cost Transformer Architectures for LLM Security