From Input Perception to Predictive Insight: Modeling Model Blind Spots Before They Become Errors
机构 * University of Sheffield（谢菲尔德大学） ； University of Exeter（埃克塞特大学） ； The Alan Turing Institute（艾伦·图灵研究所） ； UFRN, Brazil（巴西UFRN）
Comments EMNLP 2025
期刊&会议
EMNLP
Conference on Empirical Methods in Natural Language Processing · 会议 · Natural Language Processing
Tokenization and Representation Biases in Multilingual Models on Dialectal NLP Tasks
机构 * SUPSI, IDSIA（SUPSI和IDSIA）
Comments Accepted in EMNLP-2025 Main conference
Do Before You Judge: Self-Reference as a Pathway to Better LLM Evaluation
机构 * Department of Computer Science and Information Engineering（计算机科学与信息工程系） ； National Taiwan University（国立台湾大学） ； Institute of Information Science, Academia Sinica（学术院信息研究所） ； AI Research Center (AINTU)（人工智能研究中心）
Comments Accepted as a long findings paper at EMNLP 2025
Mahānāma: A Unique Testbed for Literary Entity Discovery and Linking
机构 * Indian Institute of Technology, Kharagpur（印度理工学院，克拉格浦尔） ； Kyoto University（京都大学） ； BharatGen（巴哈特基因）
Comments Accepted to EMNLP 2025. This is the authors' version. The official version will appear in the ACL Anthology
EnAnchored-X2X: English-Anchored Optimization for Many-to-Many Translation
Comments Accepted to EMNLP 2025
CHURRO: Making History Readable with an Open-Weight Large Vision-Language Model for High-Accuracy, Low-Cost Historical Text Recognition
机构 * Stanford University（斯坦福大学）
Comments EMNLP 2025
Personality Vector: Modulating Personality of Large Language Models by Model Merging
机构 * Department of Human-Artificial Intelligence Interaction, Sungkyunkwan University（人类-人工智能交互系，成均馆大学） ； Department of Immersive Media Engineering, Sungkyunkwan University（沉浸式媒体工程系，成均馆大学）
Comments EMNLP 2025
Learning Contextual Retrieval for Robust Conversational Search
Comments EMNLP 2025 main conference
Unmasking Fake Careers: Detecting Machine-Generated Career Trajectories via Multi-layer Heterogeneous Graphs
Comments Accepted at EMNLP 2025 Main
AutoSpec: An Agentic Framework for Automatically Drafting Patent Specification
机构 * Columbia University（哥伦比亚大学）
Comments EMNLP Findings 2025
SteinerSQL: Graph-Guided Mathematical Reasoning for Text-to-SQL Generation
机构 * Vanderbilt University（范德比大学） ； Zhengzhou University（郑州大学）
Comments Accept in Non-archival EMNLP 2025 MathNLP
Evaluating Language Translation Models by Playing Telephone
机构 * Stony Brook University（石溪大学）
Comments Accepted to EMNLP 2025 Main Conference as a long paper
GuessingGame: Measuring the Informativeness of Open-Ended Questions in Large Language Models
机构 * University of Cincinnati（辛辛那提大学）
Comments EMNLP 2025, 17 pages, 2 figures
Uncertainty in Semantic Language Modeling with PIXELS
机构 * Department of Artificial Intelligence, Bernoulli Institute, University of Groningen（人工智能系、伯努利研究所、 Groningen大学）
Comments 9 pages, 6 figures, UncertaiNLP 2025 Workshop @ EMNLP Camera Ready
STARQA: A Question Answering Dataset for Complex Analytical Reasoning over Structured Databases
机构 * Bloomberg（贝莱德） ； Yardi School of Artificial Intelligence, Indian Institute of Technology, Delhi（雅迪人工智能学院，印度理工学院，德里）
Comments Accepted to EMNLP 2025 long paper
Representation-based Broad Hallucination Detectors Fail to Generalize Out of Distribution
机构 * Samsung AI Research Center（三星人工智能研究中心）
Comments Accepted in EMNLP 2025 Findings
Cognitive-Level Adaptive Generation via Capability-Aware Retrieval and Style Adaptation
机构 * Zhejiang University（浙江大学）
Comments Accepted to Findings of EMNLP 2026
Quantifying Compositionality of Classic and State-of-the-Art Embeddings
机构 * University of Oxford（牛津大学） ； University of Bristol（布里斯托大学） ； University of Amsterdam（阿姆斯特丹大学）
Comments Findings of the Association for Computational Linguistics: EMNLP 2025
DRES: Fake news detection by dynamic representation and ensemble selection
机构 * École de Technologie Supérieure (ÉTS-Montréal)（École de Technologie Supérieure（蒙特利尔）） ； Universidade Tecnológica Federal do Paraná (UTFPR)（技术联邦大学（帕拉纳州）） ； Universidade Federal de Pernambuco (UFPE)（佩鲁纳博联邦大学）
Comments Accepted as oral presentation at EMNLP 2025
Benchmarking Contextual and Paralinguistic Reasoning in Speech-LLMs: A Case Study with In-the-Wild Data
机构 * Institute of Infocomm Research (I 2 R)（信息通信研究所）
Comments Accepted in EMNLP Findings 2025
PromptSculptor: Multi-Agent Based Text-to-Image Prompt Optimization
机构 * University of Connecticut（康涅狄格大学） ； Central University of Finance and Economics（中央财经大学） ； Baylor University（贝勒大学）
Comments Accepted to EMNLP 2025 System Demonstration Track
LoSiA: Efficient High-Rank Fine-Tuning via Subnet Localization and Optimization
机构 * Tsinghua University Department of Computer Science and Technology（清华大学计算机科学与技术系）
Comments Accepted to EMNLP 2025 (Oral); 20 pages, 12 figures
The Medium Is Not the Message: Deconfounding Document Embeddings via Linear Concept Erasure
机构 * ETH Zurich（苏黎世联邦理工学院） ； University of Zurich（苏黎世大学） ； Center for Data Science, New York University（纽约大学数据科学中心）
Comments Accepted to EMNLP 2025 (Main)
Why Do Some Inputs Break Low-Bit LLM Quantization?
Comments EMNLP 2025
How Well Can Reasoning Models Identify and Recover from Unhelpful Thoughts?
Comments Findings of EMNLP 2025
RadialRouter: Structured Representation for Efficient and Robust Large Language Models Routing
机构 * Department of Automation, Tsinghua University（自动化系，清华大学） ； Beijing National Research Center for Information Science and Technology（北京信息科学与技术国家研究中心）
Comments EMNLP 2025 Findings
Threading the Needle: Reweaving Chain-of-Thought Reasoning to Explain Human Label Variation
Comments Accepted by EMNLP 2025 Main (Oral), 25 pages, 7 figures
LASER: Stratified Selective Sampling for Instruction Tuning with Dedicated Scoring Strategy
机构 * ScaDS.AI & TU Dresden（ScaDS.AI与德累斯顿技术大学） ； Fraunhofer IIS（弗劳恩霍夫研究所）
Comments Accepted at Findings of the Association for Computational Linguistics: EMNLP 2025
Playpen: An Environment for Exploring Learning Through Conversational Interaction
机构 * University of Potsdam（波恩大学） ； University of Trento（特伦特大学） ； Saarland University（萨尔大学） ； ETH Zurich（苏黎世联邦理工学院） ； Amsterdam UMC（阿姆斯特丹医疗中心） ； Free University of Bozen Bolzano（博兹纳自由大学） ； University of Amsterdam（阿姆斯特丹大学） ； Heriot-Watt University（赫里奥特-瓦特大学） ； DFKI（德意志联邦研究所） ； UCL（伦敦大学学院） ； University of Edinburgh（爱丁堡大学）
Comments Accepted at EMNLP 2025 (Main) Source code: https://github.com/lm-playpen/playpen Please send correspodence to: lm-playschool@googlegroups.com
Modeling Subjectivity in Cognitive Appraisal with Language Models
机构 * Queen Mary University of London（伦敦大学玛丽女王学院） ； King’s College London（伦敦大学国王学院） ； The University of Texas at Austin（德克萨斯大学奥斯汀分校） ； The Alan Turing Institute（艾伦·图灵研究所）
Comments EMNLP 2025 Findings
