ParaICL: Towards Parallel In-Context Learning
机构 * Shanda AI Research Institute（上海沙达人工智能研究院） ； Nanyang Technological University（南洋理工大学） ； Salesforce Research（Salesforce 研究）
Comments Accepted by NAACL 2025
期刊&会议
NAACL
North American Chapter of the Association for Computational Linguistics · 会议 · Natural Language Processing
On the Vulnerability of Text Sanitization
Comments NAACL 2025
Can't Hide Behind the API: Stealing Black-Box Commercial Embedding Models
Comments NAACL Findings 2025
NeMo-Inspector: A Visualization Tool for LLM Generation Analysis
机构 * NVIDIA Corporation（英伟达公司）
Comments Presented at the NAACL 2025 conference
CORG: Generating Answers from Complex, Interrelated Contexts
机构 * KAIST AI（韩国科学技术院人工智能实验室） ； Adobe Research（Adobe研究院）
Comments published at Findings of NAACL 2025
Are LLM-Judges Robust to Expressions of Uncertainty? Investigating the effect of Epistemic Markers on LLM-based Evaluation
Comments NAACL 2025 Oral (21 pages, 6 figures, 15 tables)
TaeBench: Improving Quality of Toxic Adversarial Examples
机构 * AWS Bedrock Science（AWS Bedrock科学）
Comments Accepted for publication in NAACL 2025. The official version will be available in the ACL Anthology
Black-Box Visual Prompt Engineering for Mitigating Object Hallucination in Large Vision Language Models
Comments NAACL 2025
HateSieve: A Contrastive Learning Framework for Detecting and Segmenting Hateful Content in Multimodal Memes
机构 * University of Ottawa（渥太华大学） ； American University（美国美国大学）
Comments Accepted at NAACL 2025 Findings; camera-ready version
Journal ref Findings Assoc. Comput. Linguistics: NAACL 2025, 5201-5215 (2025)
Uncovering Bias in Large Vision-Language Models at Scale with Counterfactuals
机构 * Intel Labs（英特尔实验室） ； National Research Council Canada（加拿大国家研究理事会）
Comments Accepted to NAACL 2025 main track (oral)
Does Generative AI speak Nigerian-Pidgin?: Issues about Representativeness and Bias for Multilingualism in LLMs
机构 * Mila - Quebec AI Institute McGill University, Canada（魁北克AI研究所麦吉尔大学，加拿大） ； LT3, IDLab, Universiteit Gent（LT3，IDLab，根特大学） ； Bloomberg（彭博） ； University of Trento（特伦托大学）
Comments Accepted to NAACL 2025 (findings), please cite ACL anthology reference on https://aclanthology.org/2025.findings-naacl.85/
Local Prompt Optimization
机构 * Microsoft（微软）
Comments Accepted as Oral at NAACL 2025 (Main Conference)
MICE for CATs: Model-Internal Confidence Estimation for Calibrating Agents with Tools
机构 * CMU LTI（卡内基梅隆大学语言技术研究所） ； Microsoft（微软公司）
Comments Accepted at NAACL 2025. Code: https://github.com/microsoft/mice_for_cats
A Bayesian Optimization Approach to Machine Translation Reranking
机构 * University of Cambridge（剑桥大学） ； Karlsruhe Institute of Technology（卡尔斯鲁厄理工学院） ； ETH Zürich（苏黎世联邦理工学院）
Comments NAACL 2025 camera ready
Unleashing Multi-Hop Reasoning Potential in Large Language Models through Repetition of Misordered Context
机构 * Department of Electrical and Computer Engineering, Seoul National University（电子与计算机工程系，首尔国立大学） ； AIIS, ASRI, INMC, ISRC, and IPAI, Seoul National University（人工智能研究所、人工智能研究机构、智能网络中心、智能系统研究所以及人工智能与公共政策研究所，首尔国立大学）
Comments NAACL 2025 Findings
AdaCAD: Adaptively Decoding to Balance Conflicts between Contextual and Parametric Knowledge
机构 * UNC Chapel Hill（北卡罗来纳大学教堂山分校）
Comments NAACL 2025 (main conference), Code: https://github.com/HanNight/AdaCAD
Correcting Negative Bias in Large Language Models through Negative Attention Score Alignment
机构 * Department of Electrical and Computer Engineering, Seoul National University（电子与计算机工程系，首尔国立大学） ； Samsung SDS, Korea（三星SDS，韩国） ； AIIS, ASRI, INMC, ISRC, and IPAI, Seoul National University（人工智能研究所、人工智能研究机构、智能网络中心、智能系统研究室和人工智能政策研究所，首尔国立大学）
Comments NAACL 2025 Oral
A Practical Analysis of Human Alignment with *PO
机构 * University of Southern California, Information Sciences Institute（南加州大学信息科学研究所） ； Microsoft（微软公司）
Comments Accepted to NAACL 2025 findings papers. 9 pages, 7 figures, 3 tables
Pula: Training Large Language Models for Setswana
机构 * Data Science for Social Impact University of Pretoria School of Computing Clemson University（数据科学与社会影响 乌得勒支大学 学术 computing 勃林茅伊大学） ； Data Science for Social Impact Department of Computer Science University of Pretoria（数据科学与社会影响 计算机科学系 乌得勒支大学）
Comments NAACL 2025. 10 pages, 5 tables, 1 figure
When2Call: When (not) to Call Tools
机构 * Harvard University（哈佛大学） ； NVIDIA（NVIDIA公司）
Comments NAACL 2025
Latent Factor Models Meets Instructions: Goal-conditioned Latent Factor Discovery without Task Supervision
机构 * University of California, San Diego（加州大学圣地亚哥分校） ； Allen Institute for AI（AI研究院）
Comments NAACL 2025
From Single to Multi: How LLMs Hallucinate in Multi-Document Summarization
机构 * University of California Irvine（加州大学伊文斯分校） ； Megagon Labs（Megagon实验室）
Comments NAACL 2025 - Findings
AdTEC: A Unified Benchmark for Evaluating Text Quality in Search Engine Advertising
机构 * CyberAgent ； Nara Institute of Science and Technology（奈良科学技術研究所）
Comments Accepted to NAACL 2025
RAG LLMs are Not Safer: A Safety Analysis of Retrieval-Augmented Generation for Large Language Models
机构 * Bloomberg AI（布莱尔人工智能） ； University of Maryland（马里兰大学） ； Johns Hopkins University（约翰霍普金斯大学）
Comments NAACL 2025
AMR-RE: Abstract Meaning Representations for Retrieval-Based In-Context Learning in Relation Extraction
Comments Accepted to NAACL 2025 SRW
PicPersona-TOD : A Dataset for Personalizing Utterance Style in Task-Oriented Dialogue with Image Persona
机构 * Graduate School of Artificial Intelligence, POSTECH, Republic of Korea（人工智能研究生院，POSTECH，韩国） ； Department of Computer Science and Engineering, POSTECH, Republic of Korea（计算机科学与工程系，POSTECH，韩国）
Comments Accepted in NAACL 2025 main
PatientDx: Merging Large Language Models for Protecting Data-Privacy in Healthcare
Journal ref Workshop CL4Health @ NAACL 2025, May 2025, Albuquerque, New Mexico, United States
Synthetic Lyrics Detection Across Languages and Genres
机构 * Deezer Research（Deepz研究）
Comments Published in the TrustNLP Workshop at NAACL 2025
Evaluating Text Style Transfer Evaluation: Are There Any Reliable Metrics?
Comments Accepted at NAACL SRW 2025
ChunkRAG: Novel LLM-Chunk Filtering Method for RAG Systems
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)
