Retrieval Models Aren't Tool-Savvy: Benchmarking Tool Retrieval for Large Language Models
机构 * Shandong University（山东大学） ； Baidu Inc.（百度公司） ； Leiden University（莱顿大学）
Comments ACL 2025. Code: https://github.com/mangopy/tool-retrieval-benchmark
期刊&会议
ACL
Annual Meeting of the Association for Computational Linguistics · 会议 · Natural Language Processing
Beyond the Tip of Efficiency: Uncovering the Submerged Threats of Jailbreak Attacks in Small Language Models
机构 * Tsinghua University（清华大学） ； Zhongguancun Laboratory（中关村实验室） ； The Hong Kong University of Science and Technology (Guangzhou)（香港科技大学（广州）） ； Shandong Key Laboratory of Artificial Intelligence Security（山东省人工智能安全重点实验室）
Comments Accepted to ACL 2025 findings
APB: Accelerating Distributed Long-Context Inference by Passing Compressed Context Blocks across GPUs
机构 * NLP Group, DCST, IAI, BNRIST, Tsinghua University, Beijing, China（清华大学） ； Department of CS&T, Central South University, Changsha, China（中南大学） ； BUPT, Beijing, China（北京邮电大学） ； Pattern Recognition Center, WeChat AI, Tencent Inc.（腾讯人工智能研究院）
Comments ACL 2025 main
A Survey of LLM-based Agents in Medicine: How far are we from Baymax?
机构 * The Chinese University of Hong Kong（香港中文大学） ； Peking University（北京大学） ； City University of Hong Kong（香港城市大学） ； Massachusetts General Hospital and Harvard Medical School（麻省总医院和哈佛医学院）
Comments ACL 2025 Findings
ProcessBench: Identifying Process Errors in Mathematical Reasoning
Comments ACL 2025
P$^2$ Law: Scaling Law for Post-Training After Model Pruning
机构 * Engineering Research Center of Database and Business Intelligence, MOE, China（数据库与商业智能工程研究中心，教育部，中国） ； School of Information, Renmin University of China,Beijing, China（信息学院，中国人民大学，北京，中国） ； Key Laboratory of Data Engineering and Knowledge Engineering, MOE, China（数据工程与知识工程重点实验室，教育部，中国） ； Zhipu AI, China（智谱AI，中国）
Comments Accepted as Main of ACL2025
Reversal of Thought: Enhancing Large Language Models with Preference-Guided Reverse Reasoning Warm-up
机构 * East China Normal University（东华大学） ； Macquarie University（麦考瑞大学）
Comments Accepted by ACL 2025 Main Conference
CodeTaxo: Enhancing Taxonomy Expansion with Limited Examples via Code Language Prompts
机构 * University of Notre Dame（诺丁汉大学） ； University of Washington（华盛顿大学）
Comments Accepted by ACL2025 Findings
Less for More: Enhanced Feedback-aligned Mixed LLMs for Molecule Caption Generation and Fine-Grained NLI Evaluation
机构 * Queen Mary University of London（伦敦女王学院） ； The Alan Turing Institute（艾伦·图灵研究所）
Comments ACL25 Main
MELoRA: Mini-Ensemble Low-Rank Adapters for Parameter-Efficient Fine-Tuning
机构 * Shandong University（山东大学） ； Leiden University（莱顿大学） ； University of Amsterdam（阿姆斯特丹大学） ； Centrum Wiskunde & Informatica（数学与信息学研究中心）
Comments ACL2024
JailbreakRadar: Comprehensive Assessment of Jailbreak Attacks Against LLMs
机构 * CISPA Helmholtz Center for Information Security（CISPA海德堡信息安全研究中心）
Comments Correct typos and update new experiment results. Accepted in ACL 2025. 25 pages, 12 figures
FoodTaxo: Generating Food Taxonomies with Large Language Models
机构 * ADAPT Centre, School of Computing, Dublin City University（ADAPT中心、计算学院、都柏林城市大学） ； Lucerne School of Computer Science and Information Technology (HSLU)（卢塞恩计算机科学与信息科技学院（HSLU））
Comments To be published in ACL 2025 Industry Track. Paper website: https://foodtaxo.github.io/
Analyzing Political Bias in LLMs via Target-Oriented Sentiment Classification
机构 * Université Paris-Saclay, CEA, List（巴黎-萨克雷大学、欧洲原子能机构、List）
Comments To be published in the Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)
What Really Matters in Many-Shot Attacks? An Empirical Study of Long-Context Vulnerabilities in LLMs
机构 * Coxwave ； Seoul National University（首尔国立大学） ； Kyung Hee University（庆熙大学） ； KAIST（韩国科学技术院）
Comments Accepted by ACL 2025
Select, Read, and Write: A Multi-Agent Framework of Full-Text-based Related Work Generation
机构 * Gaoling School of Artificial Intelligence, Renmin University of China, Beijing, China（中国人民大学北京校区人工智能学院） ； Engineering Research Center of Next-Generation Intelligent Search and Recommendation, Ministry of Education（教育部下一代智能搜索与推荐工程研究中心） ； Beijing Key Laboratory of Research on Large Models and Intelligent Governance（北京大型模型与智能治理重点实验室）
Comments Accepted by ACL 2025 (Findings)
Inconsistent Tokenizations Cause Language Models to be Perplexed by Japanese Grammar
机构 * University of Tokyo（东京大学）
Comments In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics, 2025
DocMEdit: Towards Document-Level Model Editing
机构 * School of Computer Science and Technology, Beijing Institute of Technology, Beijing, China（北京理工大学计算机科学与技术学院） ； School of Computer Science and Engineering, Beihang University, Beijing, China（北京航空航天大学计算机科学与工程学院）
Comments Accepted by ACL 2025 findings
Automated CAD Modeling Sequence Generation from Text Descriptions via Transformer-Based Large Language Models
机构 * Shenzhen Institute for Advanced Study, University of Electronic Science and Technology of China（深圳先进研究院长江大学） ； i4AI Ltd（i4AI公司）
Comments Accepted by ACL 2025 Main Conference
Frictional Agent Alignment Framework: Slow Down and Don't Break Things
机构 * Department of Computer Science, Colorado State University（计算机科学系，科罗拉多州立大学）
Comments 48 pages (main paper: 10 pages incl. Limitations and Acknowledgments; references: 6 pages; appendix: 32 pages), 9 figures, 12 tables, appearing in Proceedings of ACL 2025, Vienna, Austria
gec-metrics: A Unified Library for Grammatical Error Correction Evaluation
机构 * Nara Institute of Science and Technology（奈良科学技術研究所）
Comments Accepted at ACL 2025 System Demonstration Track, 11 pages, 9 figures
ChartLens: Fine-grained Visual Attribution in Charts
机构 * University of Maryland（马里兰大学） ； Adobe Research（Adobe研究）
Comments ACL 2025 (Main)
Towards Reliable Large Audio Language Model
机构 * X-LANCE Lab, School of Computer Science, MoE Key Lab of Artificial Intelligence, Shanghai Jiao Tong University（X-LANCE实验室，计算机科学学院，人工智能国家重点实验室，上海交通大学） ； ByteDance（字节跳动）
Comments ACL 2025 Findings
Eta-WavLM: Efficient Speaker Identity Removal in Self-Supervised Speech Representations Using a Simple Linear Equation
机构 * Università degli studi di Torino（托斯纳大学） ； Cerence Inc（Cerence公司）
Comments Full paper accepted at ACL 2025
Unveiling Dual Quality in Product Reviews: An NLP-Based Approach
机构 * National Information Processing Institute（国家信息处理研究所）
Comments Accepted for ACL 2025 Industry Track
Self-Critique Guided Iterative Reasoning for Multi-hop Question Answering
机构 * Harbin Institute of Technology（哈尔滨工业大学） ； Peng Cheng Laboratory（鹏城实验室）
Comments ACL 2025 Findings
CCHall: A Novel Benchmark for Joint Cross-Lingual and Cross-Modal Hallucinations Detection in Large Language Models
机构 * School of Computer Science and Engineering, Central South University, China（中南大学计算机科学与工程学院） ； National University of Singapore, Singapore（新加坡国立大学）
Comments Accepted at ACL 2025 Main Conference
ASPO: Adaptive Sentence-Level Preference Optimization for Fine-Grained Multimodal Reasoning
机构 * Northwestern Polytechnical University, School of Automation, Xi’an, China（西北工业大学自动化学院，西安，中国） ； Northwestern Polytechnical University, School of Cybersecurity, Xi’an, China（西北工业大学网络安全学院，西安，中国） ； Binjiang Institute of Artificial Intelligence, ZJUT, Hangzhou, China（浙大宁波理工学院滨海人工智能研究院，杭州，中国） ； Alibaba Group, Hangzhou, China（阿里巴巴集团，杭州，中国）
Comments Accepted by ACL 2025 findings
Sci-LoRA: Mixture of Scientific LoRAs for Cross-Domain Lay Paraphrasing
机构 * Virginia Tech（弗吉尼亚理工大学）
Comments 18 pages, 3 figures, ACL 2025 Findings
Writing Like the Best: Exemplar-Based Expository Text Generation
机构 * University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
Comments Accepted to ACL 2025. Camera-ready version
Inference Compute-Optimal Video Vision Language Models
机构 * MIT（麻省理工学院） ； Georgia Tech（佐治亚理工学院） ； Meta ； UC Davis（加州大学戴维斯分校）
Comments Annual Meeting of the Association for Computational Linguistics (ACL), 2025
