语言大模型 / LLM - arXivDaily 专题

2511.15434 2025-11-20 cs.CR cs.AI 90%

Small Language Models for Phishing Website Detection: Cost, Performance, and Privacy Trade-Offs

小型语言模型在钓鱼网站检测中的应用：成本、性能与隐私的权衡

Georg Goldenits, Philip Koenig, Sebastian Raubitzek, Andreas Ekelhart

专题命中效率与部署：language model(title,abstract)；small language model(title,abstract)；LLM(abstract)；large language model(abstract)

AI总结本文研究了小型语言模型在钓鱼网站检测中的应用，探讨了其在成本、性能和隐私方面的权衡，展示了SLMs在替代外部LLM服务方面的可行性。

2409.19134 2025-11-20 cs.CR cs.CL 90%

Confidential Prompting: Privacy-preserving LLM Inference on Cloud

Caihua Li, In Gim, Lin Zhong

机构 * Yale University（耶鲁大学）

专题命中效率与部署：LLM(title,abstract)；prompting(title,abstract)；large language model(abstract)；language model(abstract)

2508.08343 2025-11-20 cs.PF cs.AI cs.CL 86%

A Data-driven ML Approach for Maximizing Performance in LLM-Adapter Serving

Ferran Agullo, Joan Oliveras, Chen Wang, Alberto Gutierrez-Torre, Olivier Tardieu, Alaa Youssef, Jordi Torres, Josep Ll. Berral

机构 * Barcelona Supercomputing Center (BSC)（巴塞罗那超级计算中心）； Universitat Politècnica de Catalunya - BarcelonaTech (UPC)（巴塞罗那技术大学）； IBM Research（IBM研究院）

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 in a computer science workshop

2511.11729 2025-11-20 cs.DC cs.LG 85%

Harli: SLO-Aware Co-location of LLM Inference and PEFT-based Finetuning on Model-as-a-Service Platforms

Ao Xu, Han Zhao, Weihao Cui, Quan Chen, Yukang Chen, Shulai Zhang, Shuang Chen, Jiemin Jiang, Zhibin Yu, Minyi Guo

机构 * Shanghai Jiao Tong University（上海交通大学）； Zhejiang University（浙江大学）； Chinese Academy of Science（中国科学院）

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

2511.15665 2025-11-20 cs.SE 85%

Quantum-Guided Test Case Minimization for LLM-Based Code Generation

Huixiang Zhang, Mahzabeen Emu

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

Comments This is a preprint version, full paper has been accepted in IEEE CASCON 2025 and will appear on lEEE Xplore

2511.15389 2025-11-20 cs.IR 85%

Unveiling Inference Scaling for Difference-Aware User Modeling in LLM Personalization

Suyu Chen, Yimeng Bai, Yulong Huang, Xiaoyan Zhao, Yang Zhang

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

2511.06247 2025-11-20 cs.DC 85%

Optimizing Long-context LLM Serving via Fine-grained Sequence Parallelism

Cong Li, Yuzhe Yang, Xuegui Zheng, Qifan Yang, Yijin Guan, Size Zheng, Li-Wen Chang, Shufan Liu, Xin Liu, Guangyu Sun

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

Comments The paper has been withdrawn to align with the internal publication policies of the affiliated organization. We plan to resubmit after obtaining the necessary approvals

2509.12527 2025-11-20 cs.LG stat.ML 83%

Selective Risk Certification for LLM Outputs via Information-Lift Statistics: PAC-Bayes, Robustness, and Skeleton Design

Sanjeda Akter, Ibne Farabi Shihab, Anuj Sharma

机构 * Department of Computer Science Iowa State University（计算机科学系爱荷华州立大学）； Department of Civil, Construction and Environmental Engineering Iowa State University（土木、建设与环境工程系爱荷华州立大学）

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

2511.12079 2025-11-20 cs.CV 82%

Point Cloud Quantization through Multimodal Prompting for 3D Understanding

Hongxuan Li, Wencheng Zhu, Huiying Xu, Xinzhong Zhu, Pengfei Zhu

专题命中效率与部署：prompting(title,abstract)；language model(abstract)

Comments Accepted by AAAI 2026. 11 pages, 7 figures

2509.19567 2025-11-20 cs.CL eess.AS 81%

Retrieval Augmented Generation based context discovery for ASR

Dimitrios Siskos, Stavros Papadopoulos, Pablo Peso Parada, Jisi Zhang, Karthikeyan Saravanan, Anastasios Drosou

机构 * Information Technologies Institute, Center for Research and Technology Hellas（信息技术研究所，希腊研究中心与技术中心）； Samsung Electronics R&D Institute UK (SRUK)（三星电子英国研发研究所（SRUK））

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；prompting(abstract)

Comments Accepted at EMNLP 2025

2511.15369 2025-11-20 cs.CV cs.AI 79%

IPTQ-ViT: Post-Training Quantization of Non-linear Functions for Integer-only Vision Transformers

Gihwan Kim, Jemin Lee, Hyungshin Kim

机构 * Chungnam National University（Chungnam 国立大学）； Electronics and Telecommunications Research Institute (ETRI)（电子与电信研究院）

专题命中效率与部署：post-training(title,abstract)；分类 cs.AI

Comments accepted in WACV 2026 (10 pages)

2511.15676 2025-11-20 cs.HC 78%

DuoZone: A User-Centric, LLM-Guided Mixed-Initiative XR Window Management System

Jing Qian, George X. Wang, Xiangyu Li, Yunge Wen, Guande Wu, Sonia Castelo Quispe, Fumeng Yang, Claudio Silva

专题命中效率与部署：LLM(title,abstract)

2511.14895 2025-11-20 eess.SP 78%

Lightweight Foundation Model for Wireless Time Series Downstream Tasks on Edge Devices

Mohammad Cheraghinia, Eli De Poorter, Jaron Fontaine, Kwang Soon Kim, Merouane Debbah, Adnan Shahid

专题命中效率与部署：foundation model(title,abstract)

Comments 6 pages, 4 figures, 5 tables. Accepted to a workshop in the IEEE GLOBECOM 2025

2508.07252 2025-11-20 cs.AR 78%

Tasa: Thermal-aware 3D-Stacked Architecture Design with Bandwidth Sharing for LLM Inference

Siyuan He, Peiran Yan, Yandong He, Youwei Zhuo, Tianyu Jia

专题命中效率与部署：LLM(title,abstract)

2505.18757 2025-11-20 cs.CV 75%

ToDRE: Effective Visual Token Pruning via Token Diversity and Task Relevance

Duo Li, Zuhao Yang, Xiaoqin Zhang, Ling Shao, Shijian Lu

机构 * CCDS, NTU, Singapore（南洋理工大学新加坡分校）； CCST, ZJUT, China（浙江工业大学计算机科学与技术学院）； Terminus AI Lab, UCAS, China（中国科学院大学人工智能实验室）

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

Comments 19 pages, 6 figures

2511.15432 2025-11-20 cs.LG cs.AI 73%

Towards Understanding Layer Contributions in Tabular In-Context Learning Models

面向表格上下文学习模型中层贡献的理解

Amir Rezaei Balef, Mykhailo Koshil, Katharina Eggensperger

机构 * University of Tübingen（图宾根大学）； TU Dortmund University（多特蒙德技术大学）； Lamarr Institute for Machine Learning and Artificial Intelligence（拉马尔机器学习与人工智能研究所）

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

AI总结本文研究了表格上下文学习模型中各层对预测的贡献，通过分析发现部分层存在冗余，为模型压缩和可解释性提升提供了新思路。

Comments Accepted at the EurIPS 2025 Workshop on AI for Tabular Data

2509.03733 2025-11-20 cs.LG cs.AI 73%

Differentiable Entropy Regularization: A Complexity-Aware Approach for Neural Optimization

Ibne Farabi Shihab, Sanjeda Akter, Anuj Sharma

机构 * Department of Computer Science Iowa State University（计算机科学系爱荷华州立大学）； Department of Civil, Construction and Environmental Engineering Iowa State University（土木、建设与环境工程系爱荷华州立大学）

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

2511.14803 2025-11-20 cs.SE cs.AI 70%

Scalable and Efficient Large-Scale Log Analysis with LLMs: An IT Software Support Case Study

Pranjal Gupta, Karan Bhukar, Harshit Kumar, Seema Nagar, Prateeti Mohapatra, Debanjana Kar

机构 * IBM

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

2508.18983 2025-11-20 cs.AI 70%

Enabling MoE on the Edge via Importance-Driven Expert Scheduling

Guoying Zhu, Meng Li, Haipeng Dai, Xuechen Liu, Weijun Wang, Keran Li, Jun xiao, Ligeng Chen, Wei Wang

机构 * State Key Laboratory for Novel Software Technology, Nanjing University（新型软件技术国家重点实验室，南京大学）； Tsinghua University（清华大学）； Honor Device Co., Ltd（荣誉设备有限公司）

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

2507.18224 2025-11-20 cs.MA cs.CL 70%

Assemble Your Crew: Automatic Multi-agent Communication Topology Design via Autoregressive Graph Generation

Shiyuan Li, Yixin Liu, Qingsong Wen, Chengqi Zhang, Shirui Pan

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

Comments Accepted as an oral presentation by AAAI 2026

2507.04416 2025-11-20 cs.CL 70%

RAT: Bridging RNN Efficiency and Attention Accuracy via Chunk-based Sequence Modeling

Xiuying Wei, Anunay Yadav, Razvan Pascanu, Caglar Gulcehre

机构 * CLAIRE, EPFL（EPFL的CLAIRE）； Google DeepMind（谷歌DeepMind）

专题命中效率与部署：language model(abstract)；SFT(abstract)；分类 cs.CL

Comments Accepted by NeurIPS 2025

2511.15462 2025-11-20 cs.CY cs.AI 57%

Insights from the ICLR Peer Review and Rebuttal Process

从ICLR同行评审与反驳流程中的洞察

Amir Hossein Kargaran, Nafiseh Nikeghbal, Jing Yang, Nedjma Ousidhoum

专题命中效率与部署：LLM(abstract)；分类 cs.AI

AI总结本文通过分析ICLR 2024和2025的同行评审流程，揭示了评分变化的主要影响因素及反驳策略对论文结果的提升作用。

2511.15107 2025-11-20 cs.SE cs.AI 57%

Effective Code Membership Inference for Code Completion Models via Adversarial Prompts

Yuan Jiang, Zehao Li, Shan Huang, Christoph Treude, Xiaohong Su, Tiantian Wang

机构 * Harbin Institute of Technology（哈尔滨理工大学）； Singapore Management University（新加坡国立大学）

专题命中效率与部署：language model(abstract)；分类 cs.AI

2511.15499 2025-11-20 cs.CV 50%

Learning to Expand Images for Efficient Visual Autoregressive Modeling

学习扩展图像以实现高效的视觉自回归建模

Ruiqing Yang, Kaixin Zhang, Zheng Zhang, Shan You, Tao Huang

机构 * University of Electronic Science and Technology of China（电子科技大学）； School of Computer Science and Engineering, Central South University（中南大学计算机科学与工程学院）； Xidian University（西安电子科技大学）； SenseTime Research（商汤科技研究院）； Shanghai Jiao Tong University（上海交通大学）

专题命中效率与部署：language model(abstract)

AI总结本文提出EAR模型，通过中心向外的生成方式提升视觉自回归模型的效率与生成质量，实现保真度与效率的最佳平衡。

Comments 16 pages, 18 figures, includes appendix with additional visualizations, submitted as arXiv preprint

2511.15316 2025-11-20 cs.CV 50%

What Your Features Reveal: Data-Efficient Black-Box Feature Inversion Attack for Split DNNs

Zhihan Ren, Lijun He, Jiaxi Liang, Xinzhu Fu, Haixia Bi, Fan Li

机构 * Xi’an Jiaotong University（西安交通大学）

专题命中效率与部署：language model(abstract)

2511.15312 2025-11-20 cs.CV 50%

A Multimodal Transformer Approach for UAV Detection and Aerial Object Recognition Using Radar, Audio, and Video Data

Mauro Larrat, Claudomiro Sales

机构 * Institute of Exact and Natural Sciences（精确与自然科学研究所）； Federal University of Pará（巴西亚马逊联邦大学）

专题命中效率与部署：prompting(abstract)

Comments 23 pages, 7 figures

1. 效率与部署 26 篇

Small Language Models for Phishing Website Detection: Cost, Performance, and Privacy Trade-Offs

Confidential Prompting: Privacy-preserving LLM Inference on Cloud

A Data-driven ML Approach for Maximizing Performance in LLM-Adapter Serving

Harli: SLO-Aware Co-location of LLM Inference and PEFT-based Finetuning on Model-as-a-Service Platforms

Quantum-Guided Test Case Minimization for LLM-Based Code Generation

Unveiling Inference Scaling for Difference-Aware User Modeling in LLM Personalization

Optimizing Long-context LLM Serving via Fine-grained Sequence Parallelism

Selective Risk Certification for LLM Outputs via Information-Lift Statistics: PAC-Bayes, Robustness, and Skeleton Design

Point Cloud Quantization through Multimodal Prompting for 3D Understanding

Retrieval Augmented Generation based context discovery for ASR

IPTQ-ViT: Post-Training Quantization of Non-linear Functions for Integer-only Vision Transformers

DuoZone: A User-Centric, LLM-Guided Mixed-Initiative XR Window Management System

Lightweight Foundation Model for Wireless Time Series Downstream Tasks on Edge Devices

Tasa: Thermal-aware 3D-Stacked Architecture Design with Bandwidth Sharing for LLM Inference

ToDRE: Effective Visual Token Pruning via Token Diversity and Task Relevance

Towards Understanding Layer Contributions in Tabular In-Context Learning Models

Differentiable Entropy Regularization: A Complexity-Aware Approach for Neural Optimization

Scalable and Efficient Large-Scale Log Analysis with LLMs: An IT Software Support Case Study

Enabling MoE on the Edge via Importance-Driven Expert Scheduling

Assemble Your Crew: Automatic Multi-agent Communication Topology Design via Autoregressive Graph Generation

RAT: Bridging RNN Efficiency and Attention Accuracy via Chunk-based Sequence Modeling

Insights from the ICLR Peer Review and Rebuttal Process

Effective Code Membership Inference for Code Completion Models via Adversarial Prompts

Learning to Expand Images for Efficient Visual Autoregressive Modeling

What Your Features Reveal: Data-Efficient Black-Box Feature Inversion Attack for Split DNNs

A Multimodal Transformer Approach for UAV Detection and Aerial Object Recognition Using Radar, Audio, and Video Data