大模型对齐与安全

M-RewardBench: Evaluating Reward Models in Multilingual Settings

Srishti Gureja, Lester James V. Miranda, Shayekh Bin Islam, Rishabh Maheshwary, Drishti Sharma, Gusti Winata, Nathan Lambert, Sebastian Ruder, Sara Hooker, Marzieh Fadaee

专题命中 偏好对齐 ：safety(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments 16 pages, 6 figures, 10 tables. Website: https://m-rewardbench.github.io/ , Updated results with latest models. Added more author information

PersLLM: A Personified Training Approach for Large Language Models

Zheni Zeng, Jiayi Chen, Huimin Chen, Yukun Yan, Yuxuan Chen, Zhenghao Liu, Zhiyuan Liu, Maosong Sun

机构 * Tsinghua University（清华大学） ； Northeastern University（东北大学） 纠错

专题命中 偏好对齐 ：DPO(abstract)；分类 cs.CL、cs.AI、cs.CY

Comments 8 pages for main text, 5 figures

A Logical Fallacy-Informed Framework for Argument Generation

Luca Mouchel, Debjit Paul, Shaobo Cui, Robert West, Antoine Bosselut, Boi Faltings

专题命中 偏好对齐 ：DPO(abstract)；分类 cs.CL、cs.AI、cs.LG

Anyprefer: An Agentic Framework for Preference Data Synthesis

Yiyang Zhou, Zhaoyang Wang, Tianle Wang, Shangyu Xing, Peng Xia, Bo Li, Kaiyuan Zheng, Zijian Zhang, Zhaorun Chen, Wenhao Zheng, Xuchao Zhang, Chetan Bansal, Weitong Zhang, Ying Wei, Mohit Bansal, Huaxiu Yao

专题命中 偏好对齐 ：alignment(abstract)；分类 cs.CL、cs.AI、cs.LG

Synthetic Data Generation & Multi-Step RL for Reasoning & Tool Use

Anna Goldie, Azalia Mirhoseini, Hao Zhou, Irene Cai, Christopher D. Manning

机构 * Department of Computer Science Stanford University（计算机科学系 斯坦福大学） 纠错

专题命中 偏好对齐 ：RLHF(abstract)；分类 cs.CL、cs.AI、cs.LG

Unintentional Unalignment: Likelihood Displacement in Direct Preference Optimization

Noam Razin, Sadhika Malladi, Adithya Bhaskar, Danqi Chen, Sanjeev Arora, Boris Hanin

机构 * Princeton Language and Intelligence, Princeton University（普林斯顿语言与智能，普林斯顿大学） 纠错

专题命中 偏好对齐 ：DPO(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Accepted to ICLR 2025; Code available at https://github.com/princeton-nlp/unintentional-unalignment

Stabilizing Reasoning in Medical LLMs with Continued Pretraining and Reasoning Preference Optimization

Wataru Kawakami, Keita Suzuki, Junichiro Iwasawa

机构 * Preferred Networks Inc.（Preferred Networks公司） ； Graduate School of Information Science and Technology, The University of Tokyo（东京大学信息科学与技术研究生院） 纠错

专题命中 偏好对齐 ：trustworthy(abstract)；分类 cs.CL、cs.AI、cs.LG

AdaViP: Aligning Multi-modal LLMs via Adaptive Vision-enhanced Preference Optimization

Jinda Lu, Jinghan Li, Yuan Gao, Junkang Wu, Jiancan Wu, Xiang Wang, Xiangnan He

机构 * University of Science and Technology of China（中国科学技术大学） 纠错

专题命中 偏好对齐 ：alignment(abstract)；DPO(abstract)

DSPO: Direct Semantic Preference Optimization for Real-World Image Super-Resolution

Miaomiao Cai, Simiao Li, Wei Li, Xudong Huang, Hanting Chen, Jie Hu, Yunhe Wang

机构 * University of Science and Technology of China（中国科学技术大学） ； Huawei Noah’s Ark Lab（华为诺亚实验室） 纠错

专题命中 偏好对齐 ：alignment(abstract)；DPO(abstract)

LLMs Meet Finance: Fine-Tuning Foundation Models for the Open FinLLM Leaderboard

Varun Rao, Youran Sun, Mahendra Kumar, Tejas Mutneja, Agastya Mukherjee, Haizhao Yang

专题命中 偏好对齐 ：DPO(abstract)；分类 cs.CL、cs.AI、cs.LG

LLM Unlearning Reveals a Stronger-Than-Expected Coreset Effect in Current Benchmarks

Soumyadeep Pal, Changsheng Wang, James Diffenderfer, Bhavya Kailkhura, Sijia Liu

专题命中 偏好对齐 ：safety(abstract)；分类 cs.CL、cs.AI、cs.LG

PaMi-VDPO: Mitigating Video Hallucinations by Prompt-Aware Multi-Instance Video Preference Learning

Xinpeng Ding, Kui Zhang, Jianhua Han, Lanqing Hong, Hang Xu, Xiaomeng Li

专题命中 偏好对齐 ：alignment(abstract)；DPO(abstract)

Direct Preference Optimization for LLM-Enhanced Recommendation Systems

Chao Sun, Yaobo Liang, Yaming Yang, Shilin Xu, Tianmeng Yang, Yunhai Tong

专题命中 偏好对齐 ：alignment(abstract)；DPO(abstract)

Comments This paper has been accepted to ICME 2025

Advancing Tool-Augmented Large Language Models: Integrating Insights from Errors in Inference Trees

Sijia Chen, Yibo Wang, Yi-Feng Wu, Qing-Guo Chen, Zhao Xu, Weihua Luo, Kaifu Zhang, Lijun Zhang

专题命中 偏好对齐 ：DPO(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Accepted by NeurIPS 2024

Bayesian WeakS-to-Strong from Text Classification to Generation

Ziyun Cui, Ziyang Zhang, Guangzhi Sun, Wen Wu, Chao Zhang

专题命中 偏好对齐 ：alignment(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Accepted by ICLR2025

Rejected Dialects: Biases Against African American Language in Reward Models

Joel Mire, Zubin Trivadi Aysola, Daniel Chechelnitsky, Nicholas Deas, Chrysoula Zerva, Maarten Sap

专题命中 偏好对齐 ：alignment(abstract)；分类 cs.CL、cs.AI、cs.CY

Comments Accepted to NAACL Findings 2025

Enhancing Multi-Step Reasoning Abilities of Language Models through Direct Q-Function Optimization

Kaixuan Ji, Guanlin Liu, Ning Dai, Qingping Yang, Renjie Zheng, Zheng Wu, Chen Dun, Quanquan Gu, Lin Yan

专题命中 偏好对齐 ：DPO(abstract)；分类 cs.CL、cs.AI、cs.LG

Self-Generated Critiques Boost Reward Modeling for Language Models

Yue Yu, Zhengxing Chen, Aston Zhang, Liang Tan, Chenguang Zhu, Richard Yuanzhe Pang, Yundi Qian, Xuewei Wang, Suchin Gururangan, Chao Zhang, Melanie Kambadur, Dhruv Mahajan, Rui Hou

专题命中 偏好对齐 ：RLHF(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Accepted to NAACL 2025 (Main Conference)

Journal ref NAACL 2025

PerPO: Perceptual Preference Optimization via Discriminative Rewarding

Zining Zhu, Liang Zhao, Kangheng Lin, Jinze Yang, En Yu, Chenglong Liu, Haoran Wei, Jianjian Sun, Zheng Ge, Xiangyu Zhang

专题命中 偏好对齐 ：alignment(abstract)；分类 cs.CL、cs.AI、cs.LG

On Teacher Hacking in Language Model Distillation

Daniil Tiapkin, Daniele Calandriello, Johan Ferret, Sarah Perrin, Nino Vieillard, Alexandre Ramé, Mathieu Blondel

专题命中 偏好对齐 ：RLHF(abstract)；分类 cs.CL、cs.AI、cs.LG

Eliciting Language Model Behaviors with Investigator Agents

Xiang Lisa Li, Neil Chowdhury, Daniel D. Johnson, Tatsunori Hashimoto, Percy Liang, Sarah Schwettmann, Jacob Steinhardt

专题命中 偏好对齐 ：DPO(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments 20 pages, 7 figures

Mitigating Hallucinated Translations in Large Language Models with Hallucination-focused Preference Optimization

Zilu Tang, Rajen Chatterjee, Sarthak Garg

专题命中 偏好对齐 ：safety(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments NAACL 2025 Main Conference Long paper (9 pages)

Journal ref NAACL 2025

Aligning with Human Judgement: The Role of Pairwise Preference in Large Language Model Evaluators

Yinhong Liu, Han Zhou, Zhijiang Guo, Ehsan Shareghi, Ivan Vulić, Anna Korhonen, Nigel Collier

专题命中 偏好对齐 ：RLHF(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments This paper has been accepted by COLM 2024

Critical Tokens Matter: Token-Level Contrastive Estimation Enhances LLM's Reasoning Capability

Zicheng Lin, Tian Liang, Jiahao Xu, Qiuzhi Lin, Xing Wang, Ruilin Luo, Chufan Shi, Siheng Li, Yujiu Yang, Zhaopeng Tu

专题命中 偏好对齐 ：DPO(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Work in progress

AGFSync: Leveraging AI-Generated Feedback for Preference Optimization in Text-to-Image Generation

Jingkun An, Yinghao Zhu, Zongjian Li, Enshen Zhou, Haoran Feng, Xijie Huang, Bohua Chen, Yemin Shi, Chengwei Pan

专题命中 偏好对齐 ：alignment(abstract)；DPO(abstract)

Comments Accepted by AAAI-2025

Offline Reinforcement Learning for LLM Multi-Step Reasoning

Huaijie Wang, Shibo Hao, Hanze Dong, Shenao Zhang, Yilin Bao, Ziran Yang, Yi Wu

专题命中 偏好对齐 ：DPO(abstract)；分类 cs.CL、cs.AI、cs.LG

A Post-Training Enhanced Optimization Approach for Small Language Models

Keke Zhai

专题命中 偏好对齐 ：alignment(abstract)；分类 cs.CL、cs.AI、cs.LG

Pipeline Analysis for Developing Instruct LLMs in Low-Resource Languages: A Case Study on Basque

Ander Corral, Ixak Sarasua, Xabier Saralegi

专题命中 偏好对齐 ：alignment(abstract)；分类 cs.CL、cs.AI、cs.LG

Why Does ChatGPT "Delve" So Much? Exploring the Sources of Lexical Overrepresentation in Large Language Models

Tom S. Juzek, Zina B. Ward

专题命中 偏好对齐 ：RLHF(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments 15 pages, 8 figures, The 31st International Conference on Computational Linguistics (COLING 2025)

VickreyFeedback: Cost-efficient Data Construction for Reinforcement Learning from Human Feedback

Guoxi Zhang, Jiuding Duan

专题命中 偏好对齐 ：RLHF(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments 16 pages, 5 figures