Paper Library

DeRAG: Black-box Adversarial Attacks on Multiple Retrieval-Augmented Generation Applications via Prompt Injection

Jerry Wang, Fang Yu

2025-07-20

red teaming

2507.15042v1

AlphaAlign: Incentivizing Safety Alignment with Extremely Simplified Reinforcement Learning

Yi Zhang, An Zhang, XiuYu Zhang, Leheng Sheng, Yuxin Chen, Zhenkai Liang, Xiang Wang

2025-07-20

2507.14987v1

Automated Safety Evaluations Across 20 Large Language Models: The Aymara LLM Risk and Responsibility Matrix

Juan Manuel Contreras

2025-07-19

safety

2507.14719v1

Innocence in the Crossfire: Roles of Skip Connections in Jailbreaking Visual Language Models

Palash Nandi, Maithili Joshi, Tanmoy Chakraborty

2025-07-18

red teaming

2507.13761v1

TopicAttack: An Indirect Prompt Injection Attack via Topic Transition

Yulin Chen, Haoran Li, Yuexin Li, Yue Liu, Yangqiu Song, Bryan Hooi

2025-07-18

red teaming

2507.13686v2

Paper Summary Attack: Jailbreaking LLMs through LLM Safety Papers

Liang Lin, Zhihao Xu, Xuehai Tang, Shi Liu, Biyu Zhou, Fuqing Zhu, Jizhong Han, Songlin Hu

2025-07-17

red teaming safety

2507.13474v1

Prompt Injection 2.0: Hybrid AI Threats

Jeremy McHugh, Kristina Šekrst, Jon Cefalu

2025-07-17

red teaming

2507.13169v1

Exploiting Jailbreaking Vulnerabilities in Generative AI to Bypass Ethical Safeguards for Facilitating Phishing Attacks

Rina Mishra, Gaurav Varshney

2025-07-16

red teaming

2507.12185v1

LLMs Encode Harmfulness and Refusal Separately

Jiachen Zhao, Jing Huang, Zhengxuan Wu, David Bau, Weiyan Shi

2025-07-16

red teaming

2507.11878v1

Jailbreak-Tuning: Models Efficiently Learn Jailbreak Susceptibility

Brendan Murphy, Dillon Bowen, Shahrad Mohammadzadeh, Julius Broomfield, Adam Gleave, Kellin Pelrine

2025-07-15

red teaming

2507.11630v1

The Devil behind the mask: An emergent safety vulnerability of Diffusion LLMs

Zichen Wen, Jiashu Qu, Dongrui Liu, Zhiyuan Liu, Ruixi Wu, Yicun Yang, Xiangqi Jin, Haoyun Xu, Xuyang Liu, Weijia Li, Chaochao Lu, Jing Shao, Conghui He, Linfeng Zhang

2025-07-15

red teaming safety

2507.11097v1

ARMOR: Aligning Secure and Safe Large Language Models via Meticulous Reasoning

Zhengyue Zhao, Yingzi Ma, Somesh Jha, Marco Pavone, Chaowei Xiao

2025-07-14

red teaming

2507.11500v1

July 14 - July 20, 2025

DeRAG: Black-box Adversarial Attacks on Multiple Retrieval-Augmented Generation Applications via Prompt Injection

AlphaAlign: Incentivizing Safety Alignment with Extremely Simplified Reinforcement Learning

Automated Safety Evaluations Across 20 Large Language Models: The Aymara LLM Risk and Responsibility Matrix

Innocence in the Crossfire: Roles of Skip Connections in Jailbreaking Visual Language Models

TopicAttack: An Indirect Prompt Injection Attack via Topic Transition

Paper Summary Attack: Jailbreaking LLMs through LLM Safety Papers

Prompt Injection 2.0: Hybrid AI Threats

Exploiting Jailbreaking Vulnerabilities in Generative AI to Bypass Ethical Safeguards for Facilitating Phishing Attacks

LLMs Encode Harmfulness and Refusal Separately

Jailbreak-Tuning: Models Efficiently Learn Jailbreak Susceptibility

The Devil behind the mask: An emergent safety vulnerability of Diffusion LLMs

ARMOR: Aligning Secure and Safe Large Language Models via Meticulous Reasoning

July 07 - July 13, 2025

Measuring What Matters: A Framework for Evaluating Safety Risks in Real-World LLM Applications

SEALGuard: Safeguarding the Multilingual Conversations in Southeast Asian Languages for LLM Software Systems

SEALGuard: Safeguarding the Multilingual Conversations in Southeast Asian Languages for LLM Software Systems

SEALGuard: Safeguarding the Multilingual Conversations in Southeast Asian Languages for LLM Software Systems

A Dynamic Stackelberg Game Framework for Agentic AI Defense Against LLM Jailbreaking

Defending Against Prompt Injection With a Few DefensiveTokens

GuardVal: Dynamic Large Language Model Jailbreak Evaluation for Comprehensive Safety Testing

May I have your Attention? Breaking Fine-Tuning based Prompt Injection Defenses using Architecture-Aware Attacks

An attention-aware GNN-based input defender against multi-turn jailbreak on LLMs

On the Robustness of Verbal Confidence of LLMs in Adversarial Attacks

Foundation Model Self-Play: Open-Ended Strategy Innovation via Foundation Models

The bitter lesson of misuse detection