Paper Library

CoP: Agentic Red-teaming for Large Language Models using Composition of Principles

Chen Xiong, Pin-Yu Chen, Tsung-Yi Ho

2025-06-01

red teaming

2506.00781v1

SafeSteer: Interpretable Safety Steering with Refusal-Evasion in LLMs

Shaona Ghosh, Amrita Bhattacharjee, Yftah Ziser, Christopher Parisien

2025-06-01

safety

2506.04250v1

SafeTuneBed: A Toolkit for Benchmarking LLM Safety Alignment in Fine-Tuning

Saad Hossain, Samanvay Vajpayee, Sirisha Rambhatla

2025-05-31

safety

2506.00676v1

AgentAuditor: Human-Level Safety and Security Evaluation for LLM Agents

Hanjun Luo, Shenyu Dai, Chiming Ni, Xinfeng Li, Guibin Zhang, Kun Wang, Tongliang Liu, Hanan Salam

2025-05-31

safety

2506.00641v1

Con Instruction: Universal Jailbreaking of Multimodal Large Language Models via Non-Textual Modalities

Jiahui Geng, Thy Thy Tran, Preslav Nakov, Iryna Gurevych

2025-05-31

red teaming

2506.00548v1

Wide Reflective Equilibrium in LLM Alignment: Bridging Moral Epistemology and AI Safety

Matthew Brophy

2025-05-31

safety

2506.00415v1

TRIDENT: Enhancing Large Language Model Safety with Tri-Dimensional Diversified Red-Teaming Data Synthesis

Xiaorui Wu, Xiaofeng Mao, Fei Li, Xin Zhang, Xuanhong Li, Chong Teng, Donghong Ji, Zhuang Li

2025-05-30

red teaming

2505.24672v1

Benchmarking Large Language Models for Cryptanalysis and Mismatched-Generalization

Utsav Maskey, Chencheng Zhu, Usman Naseem

2025-05-30

red teaming

2505.24621v1

AMIA: Automatic Masking and Joint Intention Analysis Makes LVLMs Robust Jailbreak Defenders

Yuqi Zhang, Yuchun Miao, Zuchao Li, Liang Ding

2025-05-30

2505.24519v1

Model Unlearning via Sparse Autoencoder Subspace Guided Projections

Xu Wang, Zihao Li, Benyou Wang, Yan Hu, Difan Zou

2025-05-30

2505.24428v1

From Hallucinations to Jailbreaks: Rethinking the Vulnerability of Large Foundation Models

Haibo Jin, Peiyan Zhang, Peiran Wang, Man Luo, Haohan Wang

2025-05-30

red teaming

2505.24232v1

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap

Wenhan Yang, Spencer Stice, Ali Payani, Baharan Mirzasoleiman

2025-05-30

safety

2505.24208v1

The State of Multilingual LLM Safety Research: From Measuring the Language Gap to Mitigating It

Zheng-Xin Yong, Beyza Ermis, Marzieh Fadaee, Stephen H. Bach, Julia Kreutzer

2025-05-30

safety

2505.24119v1

Understanding Refusal in Language Models with Sparse Autoencoders

Wei Jie Yeo, Nirmalendu Prakash, Clement Neo, Roy Ka-Wei Lee, Erik Cambria, Ranjan Satapathy

2025-05-29

red teaming

2505.23556v1

Adaptive Jailbreaking Strategies Based on the Semantic Understanding Capabilities of Large Language Models

Mingyu Yu, Wei Wang, Yanjie Wei, Sujuan Qin

2025-05-29

red teaming

2505.23404v1

Operationalizing CaMeL: Strengthening LLM Defenses for Enterprise Deployment

Krti Tallam, Emma Miller

2025-05-28

safety

2505.22852v1

GeneBreaker: Jailbreak Attacks against DNA Language Models with Pathogenicity Guidance

Zaixi Zhang, Zhenghong Zhou, Ruofan Jin, Le Cong, Mengdi Wang

2025-05-28

2505.23839v1

JailBound: Jailbreaking Internal Safety Boundaries of Vision-Language Models

Jiaxin Song, Yixu Wang, Jie Li, Rui Yu, Yan Teng, Xingjun Ma, Yingchun Wang

2025-05-26

red teaming

2505.19610v2

May 26 - June 01, 2025

CoP: Agentic Red-teaming for Large Language Models using Composition of Principles

SafeSteer: Interpretable Safety Steering with Refusal-Evasion in LLMs

SafeTuneBed: A Toolkit for Benchmarking LLM Safety Alignment in Fine-Tuning

AgentAuditor: Human-Level Safety and Security Evaluation for LLM Agents

Con Instruction: Universal Jailbreaking of Multimodal Large Language Models via Non-Textual Modalities

Wide Reflective Equilibrium in LLM Alignment: Bridging Moral Epistemology and AI Safety

TRIDENT: Enhancing Large Language Model Safety with Tri-Dimensional Diversified Red-Teaming Data Synthesis

Benchmarking Large Language Models for Cryptanalysis and Mismatched-Generalization

AMIA: Automatic Masking and Joint Intention Analysis Makes LVLMs Robust Jailbreak Defenders

Model Unlearning via Sparse Autoencoder Subspace Guided Projections

From Hallucinations to Jailbreaks: Rethinking the Vulnerability of Large Foundation Models

Bootstrapping LLM Robustness for VLM Safety via Reducing the Pretraining Modality Gap

The State of Multilingual LLM Safety Research: From Measuring the Language Gap to Mitigating It

Understanding Refusal in Language Models with Sparse Autoencoders

Adaptive Jailbreaking Strategies Based on the Semantic Understanding Capabilities of Large Language Models

Operationalizing CaMeL: Strengthening LLM Defenses for Enterprise Deployment

GeneBreaker: Jailbreak Attacks against DNA Language Models with Pathogenicity Guidance

JailBound: Jailbreaking Internal Safety Boundaries of Vision-Language Models

May 19 - May 25, 2025

Understanding and Mitigating Overrefusal in LLMs from an Unveiling Perspective of Safety Decision Boundary