The Unreasonable Effectiveness of Model Merging for Cross-Lingual Transfer in LLMs
机构 * University of California, Los Angeles（加州大学洛杉矶分校）
Comments MRL Workshop at EMNLP 2025
期刊&会议
EMNLP
Conference on Empirical Methods in Natural Language Processing · 会议 · Natural Language Processing
GlotEval: A Test Suite for Massively Multilingual Evaluation of Large Language Models
机构 * University of Helsinki（赫尔辛基大学） ； Technical University of Darmstadt（达姆施塔特技术大学） ； University of Munich（慕尼黑大学） ； Nanjing University（南京大学） ； ELLIS Institute Finland（芬兰ELLIS研究所） ； University of Turku（图尔库大学） ； Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）
Comments EMNLP demo 2025
Contrastive Learning Using Graph Embeddings for Domain Adaptation of Language Models in the Process Industry
机构 * University of Göttingen（戈丁根大学） ； eschbach GmbH（eschbach公司）
Comments accepted to EMNLP 2025 (industry track)
Randomly Removing 50% of Dimensions in Text Embeddings has Minimal Impact on Retrieval and Classification Tasks
机构 * Data and Web Science Group, University of Mannheim（曼海姆大学数据与网络科学组）
Comments Accepted to EMNLP 2025 Main Conference (Oral), camera-ready version
OWL: Probing Cross-Lingual Recall of Memorized Texts via World Literature
Comments Accepted to EMNLP 2025 Main
Tracing Multilingual Factual Knowledge Acquisition in Pretraining
机构 * Center for Information and Language Processing, LMU Munich（信息与语言处理中心，慕尼黑大学） ； Munich Center for Machine Learning (MCML)（慕尼黑机器学习中心（MCML）） ； Bosch Center for Artificial Intelligence（博世人工智能中心） ； Sorbonne Université, CNRS, ISIR, France（索邦大学，法国国家科学研究中心，ISIR）
Comments EMNLP Findings 2025
Cross-Document Cross-Lingual NLI via RST-Enhanced Graph Fusion and Interpretability Prediction
机构 * Key Laboratory of Aerospace Information Security and Trusted Computing, Ministry of Education, School of Cyber Science and Engineering, Wuhan University（航空信息安全与可信计算重点实验室，教育部，网络安全与工程学院，武汉大学） ； Zhejiang University（浙江大学）
Comments EMNLP 2025 Main (Camera Ready)
On Relation-Specific Neurons in Large Language Models
Comments EMNLP 2025
Context Length Alone Hurts LLM Performance Despite Perfect Retrieval
机构 * University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校） ； Amazon.com Inc.（亚马逊公司） ； USC Information Sciences Institute（南加州大学信息科学研究所） ； Argonne National Laboratory（阿贡国家实验室） ； The Hebrew University of Jerusalem（耶路撒冷希伯来大学） ； University of Chicago（芝加哥大学）
Comments 18 pages (9 pages of main content), 5 figures, accepted at the Findings of EMNLP 2025
Residualized Similarity for Faithfully Explainable Authorship Verification
机构 * Department of Computer Science（计算机科学系） ； Department of Linguistics（语言学系） ； Institute for Advanced Computational Science（先进计算科学研究院） ； Stony Brook University（石溪大学）
Comments EMNLP 2025 Findings
Representation Potentials of Foundation Models for Multimodal Alignment: A Survey
机构 * Department of Electrical and Computer Engineering, Northeastern University（东北大学电气与计算机工程系） ； Khoury College of Computer Science, Northeastern University（东北大学科赫里计算机科学学院）
Journal ref The 2025 Conference on Empirical Methods in Natural Language Processing
SATER: A Self-Aware and Token-Efficient Approach to Routing and Cascading
机构 * School of Computer Science, Fudan University（复旦大学计算机学院）
Comments Accepted to EMNLP 2025 Main
Learning to vary: Teaching LMs to reproduce human linguistic variability in next-word prediction
机构 * University of Amsterdam（阿姆斯特丹大学）
Comments EMNLP UncertaiNLP Workshop 2025
Intent-Aware Schema Generation And Refinement For Literature Review Tables
机构 * New York University（纽约大学） ； AI2 ； Northwestern University（西北大学）
Comments To Appear at EMNLP Findings 2025
Compound AI Systems Optimization: A Survey of Methods, Challenges, and Future Directions
机构 * National Taiwan University（国立台湾大学）
Comments Accepted to EMNLP 2025 (Main)
LLM-Hanabi: Evaluating Multi-Agent Gameplays with Theory-of-Mind and Rationale Inference in Imperfect Information Collaboration Game
机构 * Department of Computer Science and Engineering, HKUST, Hong Kong SAR, China（计算机科学与工程系，香港科技大学，香港特别行政区，中国）
Comments EMNLP 2025 Wordplay
Instability in Downstream Task Performance During LLM Pretraining
机构 * Nara Institute of Science and Technology（那拉科学与技术研究所） ； Tohoku University（东北大学） ； Research and Development Center for Large Language Models, National Institute of Informatics（大型语言模型研究与开发中心，信息学国家研究所）
Comments Accepted to EMNLP 2025 Findings
QuantAgents: Towards Multi-agent Financial System via Simulated Trading
机构 * South China University of Technology（南方科技大学） ； Pazhou Lab（琶洲实验室） ； Foshan University（佛山大学）
Comments This paper has been accepted by EMNLP 2025
Language Model Based Text-to-Audio Generation: Anti-Causally Aligned Collaborative Residual Transformers
机构 * The Hong Kong Polytechnic University（香港理工大学） ； Alibaba Group（阿里巴巴集团）
Comments Accepted to EMNLP 2025
GenQuest: An LLM-based Text Adventure Game for Language Learners
机构 * Hosei University（立命馆大学） ； King’s College London（伦敦大学国王学院） ； Kindai University（_kindai大学） ； Tokyo Uni. of Science（东京科学大学） ； University of Sheffield（谢菲尔德大学）
Comments Workshop on Wordplay: When Language Meets Games, EMNLP 2025
On the Role of Unobserved Sequences on Sample-based Uncertainty Quantification for LLMs
机构 * National Institute of Advanced Industrial Science and Technology (AIST)（国家先进工业科学与技术研究院）
Comments Accepted to UncertaiNLP workshop of EMNLP 2025
Good Intentions Beyond ACL: Who Does NLP for Social Good, and Where?
机构 * Northwestern University（西北大学） ； University of California, Los Angeles（加州大学洛杉矶分校） ； University of California, Davis（加州大学戴维斯分校）
Comments EMNLP 2025
Time Is Effort: Estimating Human Post-Editing Time for Grammar Error Correction Tool Evaluation
机构 * University of Waterloo（滑铁卢大学） ； Vector Institute（向量研究所） ； Scribendi Inc.（Scribendi公司）
Comments Accepted for publication in the 4th HCI+NLP Workshop (Fourth Workshop on Bridging Human-Computer Interaction and Natural Language Processing; part of EMNLP 2025)
SliceMoE: Routing Embedding Slices Instead of Tokens for Fine-Grained and Balanced Transformer Scaling
Comments EMNLP 2025 Main, 8 pages, 9 figures
Internal states before wait modulate reasoning patterns
机构 * Independent（独立研究者） ； Northeastern University（东北大学）
Comments Accepted to EMNLP Findings 2025
Exploring Chain-of-Thought Reasoning for Steerable Pluralistic Alignment
机构 * Columbia University（哥伦比亚大学） ； Barnard College（巴纳德学院）
Comments ACL EMNLP 2025
Zero-Shot Fine-Grained Image Classification Using Large Vision-Language Models
机构 * Department of Computer Science（计算机科学系） ； Virginia Tech（弗吉尼亚理工大学）
Comments Accepted to EMNLP 2025 Findings
A Survey of LLM-Based Applications in Programming Education: Balancing Automation and Human Oversight
Comments 2025 EMNLP HCI+NLP Workshop Short Paper
Backdoor-Powered Prompt Injection Attacks Nullify Defense Methods
Comments EMNLP 2025 Findings
Format Inertia: A Failure Mechanism of LLMs in Medical Pre-Consultation
Comments EMNLP 2025 Industry Track
