大模型对齐与安全 - arXivDaily 专题

2510.16540 2025-10-21 cs.CV cs.AI 79%

Enhancing Compositional Reasoning in CLIP via Reconstruction and Alignment of Text Descriptions

Jihoon Kwon, Kyle Min, Jy-yong Sohn

机构 * Seoul National University（首尔国立大学）； Oracle（Oracle公司）； Yonsei University（延世大学）； Intel Labs（英特尔实验室）

专题命中其他安全：alignment(title,abstract)；分类 cs.AI

Comments Accepted at NeurIPS 2025 (poster). This is the camera-ready version

2510.17598 2025-10-21 cs.AI cs.CL cs.LG 78%

Reasoning Distillation and Structural Alignment for Improved Code Generation

Amir Jalilifard, Anderson de Rezende Rocha, Marcos Medeiros Raimundo

机构 * Universidade Estadual de Campinas（坎皮纳斯州立大学）

专题命中其他安全：alignment(title)；分类 cs.CL、cs.AI、cs.LG

2410.00003 2025-10-21 cs.CV 78%

Large Language Model-Guided Semantic Alignment for Human Activity Recognition

Hua Yan, Heng Tan, Yi Ding, Pengfei Zhou, Vinod Namboodiri, Yu Yang

机构 * Lehigh University（莱维理工大学）； University of Texas at Dallas（德克萨斯大学达拉斯分校）； University of Pittsburgh（匹兹堡大学）

专题命中其他安全：alignment(title)；safety(abstract)

2502.11018 2025-10-21 cs.CL cs.AI 76%

GRIFFIN: Effective Token Alignment for Faster Speculative Decoding

Shijing Hu, Jingyang Li, Xingyu Xie, Zhihui Lu, Kim-Chuan Toh, Pan Zhou

机构 * Fudan University（复旦大学）； National University of Singapore（新加坡国立大学）； Singapore Management University（新加坡管理学院）

专题命中其他安全：alignment(title)；分类 cs.CL、cs.AI

2503.03313 2025-10-21 cs.LG cs.CL 62%

LLM as GNN: Graph Vocabulary Learning for Text-Attributed Graph Foundation Models

Xi Zhu, Haochen Xue, Ziwei Zhao, Wujiang Xu, Jingyuan Huang, Minghao Guo, Qifan Wang, Kaixiong Zhou, Imran Razzak, Yongfeng Zhang

机构 * Rutgers University（罗格斯大学）； Mohamed bin Zayed University of Artificial Intelligence (MBZUAI)（穆罕默德·本·扎耶德人工智能大学）； University of Science and Technology of China（中国科学技术大学）； Meta AI ； North Carolina State University（北卡罗来纳州立大学）

专题命中其他安全：alignment(abstract)；分类 cs.CL、cs.LG

2412.11704 2025-10-21 cs.CL cs.AI 62%

Adapting Chat Language Models Using Only Target Unlabeled Language Data

Atsuki Yamaguchi, Terufumi Morishita, Aline Villavicencio, Nikolaos Aletras

机构 * University of Sheffield（谢菲尔德大学）； Hitachi, Ltd.（日立株式会社）； University of Exeter（埃克塞特大学）

专题命中其他安全：safety(abstract)；分类 cs.CL、cs.AI

Comments Accepted to TMLR

2410.10807 2025-10-21 cs.LG cs.AI stat.ML 62%

HardNet: Hard-Constrained Neural Networks with Universal Approximation Guarantees

Youngjae Min, Navid Azizan

机构 * Massachusetts Institute of Technology（麻省理工学院）

专题命中其他安全：safety(abstract)；分类 cs.AI、cs.LG

2510.16068 2025-10-21 cs.CY cs.AI 62%

Co-Designing Interdisciplinary Design Projects with AI

Wei Ting Liow, Sumbul Khan, Lay Kee Ang

专题命中其他安全：alignment(abstract)；分类 cs.AI、cs.CY

Comments to be published in IEEE TALE 2025

2510.15996 2025-10-21 cs.LG cs.AI 62%

Using Kolmogorov-Smirnov Distance for Measuring Distribution Shift in Machine Learning

Ozan K. Tonguz, Federico Taschin

机构 * Carnegie Mellon University, USA（卡内基梅隆大学，美国）； KTH Royal Institute of Technology, Sweden（皇家理工学院）

专题命中其他安全：safety(abstract)；分类 cs.AI、cs.LG

2510.16952 2025-10-21 cs.HC cs.CL 57%

Real-Time World Crafting: Generating Structured Game Behaviors from Natural Language with Large Language Models

Austin Drake, Hang Dong

机构 * University of Exeter Department of Computer Science（埃克塞特大学计算机科学系）

专题命中其他安全：alignment(abstract)；分类 cs.CL

Comments 16 pages, 11 figures (including appendix). To be presented at the 5th Wordplay @ EMNLP workshop (2025)

2510.16885 2025-10-21 cs.LG 57%

UniGTE: Unified Graph-Text Encoding for Zero-Shot Generalization across Graph Tasks and Domains

Duo Wang, Yuan Zuo, Guangyue Lu, Junjie Wu

机构 * MIIT Key Laboratory of Data Intelligence and Management, Beihang University（信息科技部数据智能与管理重点实验室，北京航空航天大学）

专题命中其他安全：alignment(abstract)；分类 cs.LG

Journal ref NeurIPS 2025

2510.16772 2025-10-21 cs.CV cs.AI 57%

Region in Context: Text-condition Image editing with Human-like semantic reasoning

Thuy Phuong Vu, Dinh-Cuong Hoang, Minhhuy Le, Phan Xuan Tan

机构 * Greenwich Vietnam FPT University（越南格林威治FPT大学）

专题命中其他安全：alignment(abstract)；分类 cs.AI

2507.18067 2025-10-21 cs.LG cs.CE 57%

Multiscale Neural PDE Surrogates for Prediction and Downscaling: Application to Ocean Currents

Abdessamad El-Kabid, Loubna Benabbou, Redouane Lguensat, Alex Hernández-García

机构 * Mila – Québec AI Institute（魁北克人工智能研究所）； École polytechnique Palaiseau, France（巴黎理工大学Palaiseau分校）； Université du Québec à Rimouski Lévis, Québec, Canada（魁北克 Rimouski 大学 Lévis 分校）； Institut Pierre-Simon Laplace, IRD Sorbonne Université Paris, France（皮埃尔-西蒙·拉普拉斯研究所，IRD 巴黎索邦大学）； Université de Montréal Montréal, Québec, Canada（蒙特利尔大学）

专题命中其他安全：safety(abstract)；分类 cs.LG

Comments Tackling Climate Change with Machine Learning: workshop at NeurIPS 2025

2504.11521 2025-10-21 cs.LG cs.RO 57%

LANGTRAJ: Diffusion Model and Dataset for Language-Conditioned Trajectory Simulation

Wei-Jer Chang, Wei Zhan, Masayoshi Tomizuka, Manmohan Chandraker, Francesco Pittaluga

机构 * UC Berkeley（加州大学伯克利分校）； NEC Labs America（NEC美国实验室）； UC San Diego（加州大学圣地亚哥分校）

专题命中其他安全：safety(abstract)；分类 cs.LG

Comments ICCV 2025

2509.25748 2025-10-21 cs.CV cs.AI 57%

Dolphin v1.0 Technical Report

Taohan Weng, Kaibing Hu, Henan Liu, Siya Liu, Xiaoyang Liu, Zhenyu Liu, Jiren Ren, Boyan Wang, Boyang Wang, Yiyu Wang, Yalun Wu, Chaoran Yan, Kaiwen Yan, Jinze Yu, Chi Zhang, Duo Zhang, Haoyun Zheng, Xiaoqing Guo, Jacques Souquet, Hongcheng Guo, Anjie Le

专题命中其他安全：alignment(abstract)；分类 cs.AI

2508.19131 2025-10-21 cs.RO cs.AI cs.CV 57%

ZeST: an LLM-based Zero-Shot Traversability Navigation for Unknown Environments

Shreya Gummadi, Mateus V. Gasparino, Gianluca Capezzuto, Marcelo Becker, Girish Chowdhary

机构 * Field Robotics Engineering and Science Hub (FRESH), Illinois Autonomous Farm, University of Illinois at Urbana-Champaign (UIUC), IL（伊利诺伊大学厄巴纳-香槟分校）； Mobile Robotics Group, São Carlos School of Engineering, University of São Paulo (EESC-USP), São Carlos, SP, Brazil（圣保罗大学）

专题命中其他安全：safety(abstract)；分类 cs.AI

2505.18752 2025-10-21 cs.CL 57%

Unifying Attention Heads and Task Vectors via Hidden State Geometry in In-Context Learning

Haolin Yang, Hakaze Cho, Yiqiao Zhong, Naoya Inoue

机构 * University of Chicago（芝加哥大学）； JAIST ； University of Wisconsin - Madison（威斯康星大学麦迪逊分校）； RIKEN（日本研究机构）

专题命中其他安全：alignment(abstract)；分类 cs.CL

Comments 52 pages, 70 figures, 24 tables, NeurIPS 2025

2504.10906 2025-10-21 cs.CL 57%

Understanding LLMs' Cross-Lingual Context Retrieval: How Good It Is And Where It Comes From

Changjiang Gao, Hankun Lin, Xin Huang, Xue Han, Junlan Feng, Chao Deng, Jiajun Chen, Shujian Huang

机构 * National Key Laboratory for Novel Software Technology, Nanjing University（新型软件技术国家重点实验室，南京大学）； China Mobile Research Beijing, China（中国移动北京研究院）

专题命中其他安全：alignment(abstract)；分类 cs.CL

2411.04863 2025-10-21 cs.LG q-bio.BM 57%

OneProt: Towards Multi-Modal Protein Foundation Models

Klemens Flöge, Srisruthi Udayakumar, Johanna Sommer, Marie Piraud, Stefan Kesselheim, Vincent Fortuin, Stephan Günneman, Karel J van der Weg, Holger Gohlke, Erinc Merdivan, Alina Bazarova

专题命中其他安全：alignment(abstract)；分类 cs.LG

Comments 34 pages, 7 figures, 11 tables

2408.08821 2025-10-21 cs.IR cs.AI 57%

EasyRec: Simple yet Effective Language Models for Recommendation

Xubin Ren, Chao Huang

机构 * The University of Hong Kong（香港大学）

专题命中其他安全：alignment(abstract)；分类 cs.AI

Comments Published as an EMNLP'25 main paper

2510.17617 2025-10-21 cs.HC cs.CV 50%

ImaGGen: Zero-Shot Generation of Co-Speech Semantic Gestures Grounded in Language and Image Input

Hendric Voss, Stefan Kopp

机构 * Social Cognitive Systems Group, Bielefeld University（比勒菲尔德大学社会认知系统小组）

专题命中其他安全：alignment(abstract)

2510.17188 2025-10-21 cs.CV 50%

HIDISC: A Hyperbolic Framework for Domain Generalization with Generalized Category Discovery

Vaibhav Rathore, Divyam Gupta, Biplab Banerjee

机构 * Indian Institute of Technology Bombay（印度理工学院孟买）

专题命中其他安全：alignment(abstract)

Comments Accpeted at NeurIPS (2025) Main Conference

2510.16803 2025-10-21 cs.IR 50%

An Efficient Framework for Whole-Page Reranking via Single-Modal Supervision

Zishuai Zhang, Sihao Yu, Wenyi Xie, Ying Nie, Junfeng Wang, Zhiming Zheng, Dawei Yin, Hainan Zhang

专题命中其他安全：alignment(abstract)

1. 其他安全 23 篇

Enhancing Compositional Reasoning in CLIP via Reconstruction and Alignment of Text Descriptions

Reasoning Distillation and Structural Alignment for Improved Code Generation

Large Language Model-Guided Semantic Alignment for Human Activity Recognition

GRIFFIN: Effective Token Alignment for Faster Speculative Decoding

LLM as GNN: Graph Vocabulary Learning for Text-Attributed Graph Foundation Models

Adapting Chat Language Models Using Only Target Unlabeled Language Data

HardNet: Hard-Constrained Neural Networks with Universal Approximation Guarantees

Co-Designing Interdisciplinary Design Projects with AI

Using Kolmogorov-Smirnov Distance for Measuring Distribution Shift in Machine Learning

Real-Time World Crafting: Generating Structured Game Behaviors from Natural Language with Large Language Models

UniGTE: Unified Graph-Text Encoding for Zero-Shot Generalization across Graph Tasks and Domains

Region in Context: Text-condition Image editing with Human-like semantic reasoning

Multiscale Neural PDE Surrogates for Prediction and Downscaling: Application to Ocean Currents

LANGTRAJ: Diffusion Model and Dataset for Language-Conditioned Trajectory Simulation

Dolphin v1.0 Technical Report

ZeST: an LLM-based Zero-Shot Traversability Navigation for Unknown Environments

Unifying Attention Heads and Task Vectors via Hidden State Geometry in In-Context Learning

Understanding LLMs' Cross-Lingual Context Retrieval: How Good It Is And Where It Comes From

OneProt: Towards Multi-Modal Protein Foundation Models

EasyRec: Simple yet Effective Language Models for Recommendation

ImaGGen: Zero-Shot Generation of Co-Speech Semantic Gestures Grounded in Language and Image Input

HIDISC: A Hyperbolic Framework for Domain Generalization with Generalized Category Discovery

An Efficient Framework for Whole-Page Reranking via Single-Modal Supervision