语言大模型 / LLM

2508.01932 2025-08-05 cs.CV cs.AI 57%

Proactive Disentangled Modeling of Trigger-Object Pairings for Backdoor Defense

Kyle Stein, Andrew A. Mahyari, Guillermo Francia, Eman El-Sheikh

机构 * Department of Intelligent Systems and Robotics, University of West Florida（智能系统与机器人系，西佛罗里达大学）； Florida Institute For Human and Machine Cognition (IHMC)（佛罗里达人类与机器认知研究所）； Center for Cybersecurity, University of West Florida（网络安全中心，西佛罗里达大学）

专题命中评测与基准：language model(abstract)；分类 cs.AI

Journal ref Computers, Materials & Continua, 2025

2508.02082 2025-08-05 cs.CV 50%

S-RRG-Bench: Structured Radiology Report Generation with Fine-Grained Evaluation Framework

Yingshu Li, Yunyi Liu, Zhanyu Wang, Xinyu Liang, Lingqiao Liu, Lei Wang, Luping Zhou

机构 * University of Sydney（悉尼大学）； University of Wollongong（沃林戈大学）； University of Adelaide（阿德莱德大学）； First Clinical Medical College, Guangzhou University of Chinese Medicine（广州中医药大学第一临床学院）

专题命中评测与基准：LLM(abstract)

2508.01921 2025-08-05 cs.CV 50%

InspectVLM: Unified in Theory, Unreliable in Practice

Conor Wallace, Isaac Corley, Jonathan Lwowski

机构 * Zeitview

专题命中评测与基准：language model(abstract)

Comments Accepted to 2025 ICCV VISION Workshop

2504.13490 2025-08-05 cs.CV 50%

Early Timestep Zero-Shot Candidate Selection for Instruction-Guided Image Editing

Joowon Kim, Ziseok Lee, Donghyeon Cho, Sanghyun Jo, Yeonsung Jung, Kyungsu Kim, Eunho Yang

机构 * KAIST（韩国科学技术院）； Seoul National University（首尔国立大学）； OGQ

专题命中评测与基准：language model(abstract)

Comments Accepted to ICCV 2025

2412.18645 2025-08-05 cs.CV 50%

Scendi Score: Prompt-Aware Diversity Evaluation via Schur Complement of CLIP Embeddings

Azim Ospanov, Mohammad Jalali, Farzan Farnia

机构 * The Chinese University of Hong Kong, Department of Computer Science & Engineering（香港中文大学计算机科学与工程系）

专题命中评测与基准：LLM(abstract)

2307.15640 2025-08-05 cs.CV cs.MM 50%

CLIP Brings Better Features to Visual Aesthetics Learners

Liwu Xu, Jinjin Xu, Yuzhe Yang, Xilu Wang, Yijie Huang, Yaqian Li

机构 * OPPO AI Center（OPPO人工智能中心）； University of Surrey（塞弗尔大学）

专题命中评测与基准：language model(abstract)

Comments ICME 2025 Oral

2508.02668 2025-08-05 cs.LG 90%

LOST: Low-rank and Sparse Pre-training for Large Language Models

Jiaxi Li, Lu Yin, Li Shen, Jinjin Xu, Liwu Xu, Tianjin Huang, Wenwu Wang, Shiwei Liu, Xilu Wang

机构 * University of Surrey（塞维利亚大学）； University of Oxford（牛津大学）； Sun Yat-sen University（中山大学）； Bytedance（字节跳动）； Alibaba Group（阿里巴巴集团）； University of Exeter（埃克塞特大学）

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

2508.02128 2025-08-05 cs.LG cs.AI 90%

Amber Pruner: Leveraging N:M Activation Sparsity for Efficient Prefill in Large Language Models

Tai An, Ruwu Cai, Yanzhe Zhang, Yang Liu, Hao Chen, Pengcheng Xie, Sheng Chang, Yiwu Yao, Gongyi Wang

机构 * ia.ac.cn（中国科学院）

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

2508.01268 2025-08-05 cs.AI cs.CR 89%

Win-k: Improved Membership Inference Attacks on Small Language Models

Roya Arkhmammadova, Hosein Madadi Tamar, M. Emre Gursoy

机构 * Department of Computer Engineering, Koç University（计算机工程系，科奇大学）

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

2502.05213 2025-08-05 cs.CR cs.AI 89%

DERMARK: A Dynamic, Efficient and Robust Multi-bit Watermark for Large Language Models

Qihao Lin, Chen Tang, Lan zhang, Junyang zhang, Xiangyang Li

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

Comments 8 pages, 7 figures

2306.15933 2025-08-05 cs.CL cs.AI cs.LG 88%

You Can Generate It Again: Data-to-Text Generation with Verification and Correction Prompting

Xuan Ren, Zeyu Zhang, Lingqiao Liu

机构 * AIML, University of Adelaide（人工智能实验室，阿德莱德大学）

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

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

2508.01744 2025-08-05 cs.LG 85%

AGFT: An Adaptive GPU Frequency Tuner for Real-Time LLM Inference Optimization

Zicong Ye, Kunming Zhang, Guoming Tang

机构 * The Hong Kong University of Science（香港科技大学）

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

2508.01625 2025-08-05 cs.LG cs.AI 85%

EAC-MoE: Expert-Selection Aware Compressor for Mixture-of-Experts Large Language Models

Yuanteng Chen, Yuantian Shao, Peisong Wang, Jian Cheng

机构 * Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）； School of Artificial Intelligence, University of Chinese Academy of Sciences（中国科学院大学人工智能学院）； Nanjing University of Science and Technology（南京理工大学）； AIRIA

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

Comments 22 pages, 13 figures. ACL 2025

2508.01989 2025-08-05 cs.DC 85%

Prefill-Decode Aggregation or Disaggregation? Unifying Both for Goodput-Optimized LLM Serving

Chao Wang, Pengfei Zuo, Zhangyu Chen, Yunkai Liang, Zhou Yu, Ming-Chang Yang

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

Comments 17 pages, 19 figures

2508.01558 2025-08-05 cs.CV 85%

EvoVLMA: Evolutionary Vision-Language Model Adaptation

Kun Ding, Ying Wang, Shiming Xiang

机构 * Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）

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

Comments This paper has been accepted by ACM Multimedia 2025 (ACM MM 2025)

2508.01279 2025-08-05 cs.HC 85%

ViseGPT: Towards Better Alignment of LLM-generated Data Wrangling Scripts and User Prompts

Jiajun Zhu, Xinyu Cheng, Zhongsu Luo, Yunfan Zhou, Xinhuan Shu, Di Weng, Yingcai Wu

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

Comments Accepted at Annual ACM Symposium on User Interface Software and Technology (UIST'25), September 28-October 1, 2025, Busan, Republic of Korea

2412.15450 2025-08-05 cs.CL 83%

Fietje: An open, efficient LLM for Dutch

Bram Vanroy

机构 * KU Leuven（库勒芬大学）； Dutch Language Institute（荷兰语言学院）

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

Journal ref Computational Linguistics in the Netherlands Journal,14 (2025), 473-504

2508.02193 2025-08-05 cs.CL cs.LG 81%

Seed Diffusion: A Large-Scale Diffusion Language Model with High-Speed Inference

Yuxuan Song, Zheng Zhang, Cheng Luo, Pengyang Gao, Fan Xia, Hao Luo, Zheng Li, Yuehang Yang, Hongli Yu, Xingwei Qu, Yuwei Fu, Jing Su, Ge Zhang, Wenhao Huang, Mingxuan Wang, Lin Yan, Xiaoying Jia, Jingjing Liu, Wei-Ying Ma, Ya-Qin Zhang, Yonghui Wu, Hao Zhou

机构 * ByteDance Seed（字节跳动种子）； Institute for AI Industry Research (AIR)（人工智能产业研究院）； Tsinghua University（清华大学）

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

Comments Demo is available at https://studio.seed.ai/exp/seed_diffusion/; Project page is https://seed.bytedance.com/seed_diffusion

2508.01647 2025-08-05 cs.CR cs.AI cs.CL 81%

DUP: Detection-guided Unlearning for Backdoor Purification in Language Models

Man Hu, Yahui Ding, Yatao Yang, Liangyu Chen, Yanhao Jia, Shuai Zhao

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

2508.00838 2025-08-05 cs.DL cs.AI cs.CL 81%

The Attribution Crisis in LLM Search Results

Ilan Strauss, Jangho Yang, Tim O'Reilly, Sruly Rosenblat, Isobel Moure

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

2508.01540 2025-08-05 cs.CV cs.AI 79%

MagicVL-2B: Empowering Vision-Language Models on Mobile Devices with Lightweight Visual Encoders via Curriculum Learning

Yi Liu, Xiao Xu, Zeyu Xu, Meng Zhang, Yibo Li, Haoyu Chen, Junkang Zhang, Qiang Wang, Jifa Sun, Siling Lin, Shengxun Cheng, Lingshu Zhang, Kang Wang

机构 * Project Leader（项目负责人）； Corresponding Author（通讯作者）

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

2508.01261 2025-08-05 cs.AI 79%

Unifying Mixture of Experts and Multi-Head Latent Attention for Efficient Language Models

Sushant Mehta, Raj Dandekar, Rajat Dandekar, Sreedath Panat

机构 * Vizuara AI Labs（Vizuara AI实验室）

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

2508.01667 2025-08-05 cs.CV 78%

Rein++: Efficient Generalization and Adaptation for Semantic Segmentation with Vision Foundation Models

Zhixiang Wei, Xiaoxiao Ma, Ruishen Yan, Tao Tu, Huaian Chen, Jinjin Zheng, Yi Jin, Enhong Chen

机构 * University of Science and Technology of China（中国科学技术大学）

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

2508.01548 2025-08-05 cs.CV 78%

A Glimpse to Compress: Dynamic Visual Token Pruning for Large Vision-Language Models

Quan-Sheng Zeng, Yunheng Li, Qilong Wang, Peng-Tao Jiang, Zuxuan Wu, Ming-Ming Cheng, Qibin Hou

机构 * VCIP, CS, Nankai University（VCIP计算机科学系，南开大学）； Shanghai Innovation Institute（上海创新研究院）； Tianjin University（天津大学）； vivo Mobile Communication Co., Ltd（vivo移动通信有限公司）

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

Comments 15 pages, 10 figures. Project page: https://github.com/HVision-NKU/GlimpsePrune

2508.01236 2025-08-05 cs.CV 78%

Mitigating Information Loss under High Pruning Rates for Efficient Large Vision Language Models

Mingyu Fu, Wei Suo, Ji Ma, Lin Yuanbo Wu, Peng Wang, Yanning Zhang

机构 * Northwestern Polytechnical University（西北工业大学）； National Engineering Laboratory for Integrated Aero-Space-Ground-Ocean Big Data Application Technology（集成空天地海大数据应用技术国家工程实验室）； Swansea University（斯旺西大学）

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

Comments accepted by ACM MM 2025

2508.02523 2025-08-05 cs.CR 75%

Transportation Cyber Incident Awareness through Generative AI-Based Incident Analysis and Retrieval-Augmented Question-Answering Systems

Ostonya Thomas, Muhaimin Bin Munir, Jean-Michel Tine, Mizanur Rahman, Yuchen Cai, Khandakar Ashrafi Akbar, Md Nahiyan Uddin, Latifur Khan, Trayce Hockstad, Mashrur Chowdhury

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

Comments This paper has been submitted to the Transportation Research Board (TRB) for consideration for presentation at the 2026 Annual Meeting

2508.02215 2025-08-05 cs.LG cs.AI cs.CL 75%

LeanK: Learnable K Cache Channel Pruning for Efficient Decoding

Yike Zhang, Zhiyuan He, Huiqiang Jiang, Chengruidong Zhang, Yuqing Yang, Jianyong Wang, Lili Qiu

机构 * Tsinghua University（清华大学）； Microsoft Research（微软研究院）

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

2508.01092 2025-08-05 cs.HC 75%

DescribePro: Collaborative Audio Description with Human-AI Interaction

Maryam Cheema, Sina Elahimanesh, Samuel Martin, Pooyan Fazli, Hasti Seifi

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

Comments ASSETS 25 19 pages, 8 figures

2505.02331 2025-08-05 cs.CV cs.SD 75%

VAEmo: Efficient Representation Learning for Visual-Audio Emotion with Knowledge Injection

Hao Cheng, Zhiwei Zhao, Yichao He, Zhenzhen Hu, Jia Li, Meng Wang, Richang Hong

机构 * Hefei University of Technology（合肥工业大学）

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

Comments Source code and pre-trained models will be available at https://github.com/MSA-LMC/VAEmo

1. 评测与基准 84 篇

Proactive Disentangled Modeling of Trigger-Object Pairings for Backdoor Defense

S-RRG-Bench: Structured Radiology Report Generation with Fine-Grained Evaluation Framework

InspectVLM: Unified in Theory, Unreliable in Practice

Early Timestep Zero-Shot Candidate Selection for Instruction-Guided Image Editing

Scendi Score: Prompt-Aware Diversity Evaluation via Schur Complement of CLIP Embeddings

CLIP Brings Better Features to Visual Aesthetics Learners

2. 效率与部署 44 篇

LOST: Low-rank and Sparse Pre-training for Large Language Models

Amber Pruner: Leveraging N:M Activation Sparsity for Efficient Prefill in Large Language Models

Win-k: Improved Membership Inference Attacks on Small Language Models

DERMARK: A Dynamic, Efficient and Robust Multi-bit Watermark for Large Language Models

You Can Generate It Again: Data-to-Text Generation with Verification and Correction Prompting

Forecasting LLM Inference Performance via Hardware-Agnostic Analytical Modeling

AGFT: An Adaptive GPU Frequency Tuner for Real-Time LLM Inference Optimization

EAC-MoE: Expert-Selection Aware Compressor for Mixture-of-Experts Large Language Models

Prefill-Decode Aggregation or Disaggregation? Unifying Both for Goodput-Optimized LLM Serving

EvoVLMA: Evolutionary Vision-Language Model Adaptation

ViseGPT: Towards Better Alignment of LLM-generated Data Wrangling Scripts and User Prompts

Fietje: An open, efficient LLM for Dutch

Seed Diffusion: A Large-Scale Diffusion Language Model with High-Speed Inference

DUP: Detection-guided Unlearning for Backdoor Purification in Language Models

The Attribution Crisis in LLM Search Results

MagicVL-2B: Empowering Vision-Language Models on Mobile Devices with Lightweight Visual Encoders via Curriculum Learning

Unifying Mixture of Experts and Multi-Head Latent Attention for Efficient Language Models

Rein++: Efficient Generalization and Adaptation for Semantic Segmentation with Vision Foundation Models

A Glimpse to Compress: Dynamic Visual Token Pruning for Large Vision-Language Models

Mitigating Information Loss under High Pruning Rates for Efficient Large Vision Language Models

Transportation Cyber Incident Awareness through Generative AI-Based Incident Analysis and Retrieval-Augmented Question-Answering Systems

LeanK: Learnable K Cache Channel Pruning for Efficient Decoding

DescribePro: Collaborative Audio Description with Human-AI Interaction

VAEmo: Efficient Representation Learning for Visual-Audio Emotion with Knowledge Injection