DCE-LLM: Dead Code Elimination with Large Language Models
Comments Accepted by regular paper in NAACL 2025, with 13 pages, 5 figures
期刊&会议
NAACL
North American Chapter of the Association for Computational Linguistics · 会议 · Natural Language Processing
Transferable Post-training via Inverse Value Learning
机构 * Chinese Information Processing Laboratory Institute of Software, Chinese Academy of Sciences（中国科学院软件研究所信息处理实验室） ； University of Chinese Academy of Sciences（中国科学院大学） ； Alibaba Group（阿里巴巴集团）
Comments NAACL 2025 Camera Ready
UniToMBench: Integrating Perspective-Taking to Improve Theory of Mind in LLMs
机构 * Algoverse AI Research（Algoverse AI研究院）
Comments Accepted at Conference of the North American Chapter of the Association for Computational Linguistics, Student Research Workshop 2025 (NAACL SRW 2025)
From Annotation to Adaptation: Metrics, Synthetic Data, and Aspect Extraction for Aspect-Based Sentiment Analysis with Large Language Models
机构 * Saint Mary’s University（圣玛丽大学） ； York University（约克大学）
Comments Accepted to NAACL SRW 2025
Unmasking Implicit Bias: Evaluating Persona-Prompted LLM Responses in Power-Disparate Social Scenarios
Comments NAACL 2025; 10 pages of main text. For interactive visualisation, see https://inc0mple.github.io/Implicit_Bias_Interactive_Data_Viz
Intrinsic Bias is Predicted by Pretraining Data and Correlates with Downstream Performance in Vision-Language Encoders
机构 * Carnegie Mellon University（卡内基梅隆大学） ； University of Washington（华盛顿大学）
Comments Accepted to NAACL Main, 2025
Can Large Language Models Invent Algorithms to Improve Themselves?: Algorithm Discovery for Recursive Self-Improvement through Reinforcement Learning
机构 * NEC Corporation（日本电报电话株式会社）
Comments Accepted at NAACL 2025 (main)
COVID-19 Literature Knowledge Graph Construction and Drug Repurposing Report Generation
Comments 12 pages, Accepted by Proceedings of 2021 Annual Conference of the North American Chapter of the Association for Computational Linguistics System Demonstrations, for resources see http://blender.cs.illinois.edu/covid19/, for video see http://159.89.180.81/demo/covid/Covid-KG_DemoVideo.mp4, for slides see https://eaglew.github.io/files/Covid-KG_DemoVideo_with_ethics.pdf
Eliciting Critical Reasoning in Retrieval-Augmented Language Models via Contrastive Explanations
机构 * Idiap Research Institute（Idiap研究机构） ； Department of Computer Science, University of Manchester（曼彻斯特大学计算机科学系） ； National Biomarker Centre (NBC), CRUK Manchester Institute（国家生物标志物中心（NBC），CRUK曼彻斯特研究所）
Journal ref 2025.naacl-long.557
RONA: Pragmatically Diverse Image Captioning with Coherence Relations
机构 * The Pennsylvania State University（宾夕法尼亚州立大学） ； National Institute of Technology, Tiruchirappalli（特里奇里帕利理工学院）
Comments Accepted in the NAACL Fourth Workshop on Intelligent and Interactive Writing Assistants (In2Writing), Albuquerque, New Mexico, May 2025, https://in2writing.glitch.me
Preserving Cultural Identity with Context-Aware Translation Through Multi-Agent AI Systems
机构 * Shahjalal University of Science and Technology（沙赫拉尔科学与技术大学） ； University of Oklahoma（俄克拉荷马大学）
Comments Accepted in NAACL 2025 Workshop on Language Models for Underserved Communities (https://openreview.net/forum?id=RiCfefEHII)
Language Models are Crossword Solvers
机构 * Indian Statistical Institute, Kolkata（印度统计研究所，加尔各答）
Comments 10 pages, 3 figures, 6 Appendix. Accepted to NAACL 2025
Journal ref NAACL 2025 (Volume 1: Long Papers), pages 2074-2090
Mitigating Hallucination in Abstractive Summarization with Domain-Conditional Mutual Information
机构 * Graduate School of Data Science, Seoul National University（数据科学研究生院，首尔国立大学）
Comments Accepted by Findings of NAACL 2024
DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models
机构 * Alibaba Group（阿里巴巴集团） ； Zhejiang University（浙江大学） ； M-A-P ； The Chinese University of Hong Kong（香港中文大学）
Comments [NAACL 2025] The first four authors contribute equally, 23 pages, repo at https://github.com/Kizna1ver/DREAM
PoisonedParrot: Subtle Data Poisoning Attacks to Elicit Copyright-Infringing Content from Large Language Models
机构 * University of Maryland, College Park（马里兰大学College Park分校） ； University of California, Santa Barbara（加州大学圣巴巴拉分校） ； Capital One（Capital One公司）
Comments 18 pages, 18 figures. Accepted at NAACL 2025
TurboFuzzLLM: Turbocharging Mutation-based Fuzzing for Effectively Jailbreaking Large Language Models in Practice
机构 * Amazon Web Services, USA（亚马逊网络服务（美国））
Comments Oral presentation at NAACL 2025 industry track
Evaluating Morphological Compositional Generalization in Large Language Models
机构 * EPFL（瑞士联邦理工学院） ； Idiap Research Institute（Idiap研究机构） ； Duke University（杜克大学） ； Brandeis University（布兰迪斯大学） ； LMU Munich（慕尼黑大学） ； University of Cambridge（剑桥大学） ； Johns Hopkins University（约翰霍普金斯大学） ； Università della Svizzera Italiana（瑞士意大利大学） ； New York University（纽约大学）
Comments Accepted to NAACL 2025
Beyond the Safety Bundle: Auditing the Helpful and Harmless Dataset
机构 * McGill University（麦吉尔大学） ； Mila - Quebec AI Institute（魁北克人工智能研究所）
Comments NAACL Main Conference 2025 - Accepted as an Oral
Unmasking Database Vulnerabilities: Zero-Knowledge Schema Inference Attacks in Text-to-SQL Systems
机构 * Department of Information Systems and Cyber Security（信息系统与网络安全系） ； The University of Texas at San Antonio（德克萨斯大学圣安东尼奥分校）
Comments Accepted to NAACL 2025 Findings
What Goes Into a LM Acceptability Judgment? Rethinking the Impact of Frequency and Length
机构 * Carnegie Mellon University（卡内基梅隆大学） ； New York University（纽约大学）
Comments Accepted to NAACL 2025 (Main Conference)
PromptRefine: Enhancing Few-Shot Performance on Low-Resource Indic Languages with Example Selection from Related Example Banks
机构 * University of Maryland, College Park（马里兰大学 College Park 分校） ； Adobe Research（Adobe 研究）
Comments Accepted at NAACL 2025
BanTH: A Multi-label Hate Speech Detection Dataset for Transliterated Bangla
机构 * Research and Development, Penta Global Limited（研发部，Penta全球有限公司） ； Islamic University of Technology（伊斯兰技术大学） ； CCDS Lab, Independent University（CCDS实验室，独立大学）
Comments Published in NAACL Findings 2025
Data Poisoning for In-context Learning
Comments NAACL 2025
Probing Audio-Generation Capabilities of Text-Based Language Models
机构 * Algoverse AI Research（Algoverse AI研究）
Comments Accepted at Conference of the North American Chapter of the Association for Computational Linguistics 2025, Student Research Workshop (NAACL SRW)
Anticipating Future with Large Language Model for Simultaneous Machine Translation
机构 * Carnegie Mellon University（卡内基梅隆大学） ； NVIDIA（英伟达）
Comments NAACL 2025 Main
MIDAS: Multi-level Intent, Domain, And Slot Knowledge Distillation for Multi-turn NLU
机构 * The University of Sydney（悉尼大学） ； Kyung Hee University（庆熙大学） ； The University of Melbourne（墨尔本大学）
Comments Accepted by NAACL 2025
Verify-in-the-Graph: Entity Disambiguation Enhancement for Complex Claim Verification with Interactive Graph Representation
机构 * Viettel Artificial Intelligence and Data Services Center（越南 Viettel 人工智能与数据服务中心） ； Viettel Group（越南 Viettel 集团）
Comments Published at NAACL 2025 Main Conference
Journal ref https://aclanthology.org/2025.naacl-long.268
Ensuring Safe and High-Quality Outputs: A Guideline Library Approach for Language Models
机构 * XiaMen University（厦门大学） ； Yanshan University（雁山大学） ； IDEA Research（IDEA研究） ； Inner Mongolia University（内蒙古大学） ； Microsoft（微软） ； Microsoft Research Asia（微软亚洲研究院）
Comments Accepted to NAACL 2024 main conference
Journal ref Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)
HARP: Hesitation-Aware Reframing in Transformer Inference Pass
Comments Accepted to NAACL 2025 main (long)
Patent-CR: A Dataset for Patent Claim Revision
机构 * University of Cambridge（剑桥大学）
Comments Accepted to NAACL 2025. 15 pages, 6 tables, 3 figures
Journal ref Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), pages 2300-2314
