Paper Library

AdversariaL attacK sAfety aLIgnment(ALKALI): Safeguarding LLMs through GRACE: Geometric Representation-Aware Contrastive Enhancement- Introducing Adversarial Vulnerability Quality Index (AVQI)

Danush Khanna, Krishna Kumar, Basab Ghosh, Vinija Jain, Vasu Sharma, Aman Chadha, Amitava Das

2025-06-10

red teaming safety

2506.08885v1

Design Patterns for Securing LLM Agents against Prompt Injections

Luca Beurer-Kellner, Beat Buesser, Ana-Maria Creţu, Edoardo Debenedetti, Daniel Dobos, Daniel Fabian, Marc Fischer, David Froelicher, Kathrin Grosse, Daniel Naeff, Ezinwanne Ozoani, Andrew Paverd, Florian Tramèr, Václav Volhejn

2025-06-10

2506.08837v3

Evaluation empirique de la sécurisation et de l'alignement de ChatGPT et Gemini: analyse comparative des vulnérabilités par expérimentations de jailbreaks

Rafaël Nouailles

2025-06-10

red teaming

2506.10029v1

AsFT: Anchoring Safety During LLM Fine-Tuning Within Narrow Safety Basin

Shuo Yang, Qihui Zhang, Yuyang Liu, Yue Huang, Xiaojun Jia, Kunpeng Ning, Jiayu Yao, Jigang Wang, Hailiang Dai, Yibing Song, Li Yuan

2025-06-10

safety

2506.08473v2

AsFT: Anchoring Safety During LLM Fine-Tuning Within Narrow Safety Basin

Shuo Yang, Qihui Zhang, Yuyang Liu, Yue Huang, Xiaojun Jia, Kunpeng Ning, Jiayu Yao, Jigang Wang, Hailiang Dai, Yibing Song, Li Yuan

2025-06-10

safety

2506.08473v1

RSafe: Incentivizing proactive reasoning to build robust and adaptive LLM safeguards

Jingnan Zheng, Xiangtian Ji, Yijun Lu, Chenhang Cui, Weixiang Zhao, Gelei Deng, Zhenkai Liang, An Zhang, Tat-Seng Chua

2025-06-09

red teaming

2506.07736v2

RSafe: Incentivizing proactive reasoning to build robust and adaptive LLM safeguards

Jingnan Zheng, Xiangtian Ji, Yijun Lu, Chenhang Cui, Weixiang Zhao, Gelei Deng, Zhenkai Liang, An Zhang, Tat-Seng Chua

2025-06-09

red teaming

2506.07736v1

LLMs Caught in the Crossfire: Malware Requests and Jailbreak Challenges

Haoyang Li, Huan Gao, Zhiyuan Zhao, Zhiyu Lin, Junyu Gao, Xuelong Li

2025-06-09

red teaming

2506.10022v1

Evaluating LLMs Robustness in Less Resourced Languages with Proxy Models

Maciej Chrabąszcz, Katarzyna Lorenc, Karolina Seweryn

2025-06-09

red teaming

2506.07645v1

TwinBreak: Jailbreaking LLM Security Alignments based on Twin Prompts

Torsten Krauß, Hamid Dashtbani, Alexandra Dmitrienko

2025-06-09

red teaming

2506.07596v1

When Style Breaks Safety: Defending Language Models Against Superficial Style Alignment

Yuxin Xiao, Sana Tonekaboni, Walter Gerych, Vinith Suriyakumar, Marzyeh Ghassemi

2025-06-09

red teaming

2506.07452v1

Beyond Jailbreaks: Revealing Stealthier and Broader LLM Security Risks Stemming from Alignment Failures

Yukai Zhou, Sibei Yang, Wenjie Wang

2025-06-09

red teaming

2506.07402v1

June 09 - June 15, 2025

AdversariaL attacK sAfety aLIgnment(ALKALI): Safeguarding LLMs through GRACE: Geometric Representation-Aware Contrastive Enhancement- Introducing Adversarial Vulnerability Quality Index (AVQI)

Design Patterns for Securing LLM Agents against Prompt Injections

Evaluation empirique de la sécurisation et de l'alignement de ChatGPT et Gemini: analyse comparative des vulnérabilités par expérimentations de jailbreaks

AsFT: Anchoring Safety During LLM Fine-Tuning Within Narrow Safety Basin

AsFT: Anchoring Safety During LLM Fine-Tuning Within Narrow Safety Basin

RSafe: Incentivizing proactive reasoning to build robust and adaptive LLM safeguards

RSafe: Incentivizing proactive reasoning to build robust and adaptive LLM safeguards

LLMs Caught in the Crossfire: Malware Requests and Jailbreak Challenges

Evaluating LLMs Robustness in Less Resourced Languages with Proxy Models

TwinBreak: Jailbreaking LLM Security Alignments based on Twin Prompts

When Style Breaks Safety: Defending Language Models Against Superficial Style Alignment

Beyond Jailbreaks: Revealing Stealthier and Broader LLM Security Risks Stemming from Alignment Failures

June 02 - June 08, 2025

Enhancing the Safety of Medical Vision-Language Models by Synthetic Demonstrations

AlphaSteer: Learning Refusal Steering with Principled Null-Space Constraint

Auditing Black-Box LLM APIs with a Rank-Based Uniformity Test

Auditing Black-Box LLM APIs with a Rank-Based Uniformity Test

Auditing Black-Box LLM APIs with a Rank-Based Uniformity Test

From Threat to Tool: Leveraging Refusal-Aware Injection Attacks for Safety Alignment

Saffron-1: Towards an Inference Scaling Paradigm for LLM Safety Assurance

Small Models, Big Support: A Local LLM Framework for Teacher-Centric Content Creation and Assessment using RAG and CAG

To Protect the LLM Agent Against the Prompt Injection Attack with Polymorphic Prompt

Why LLM Safety Guardrails Collapse After Fine-tuning: A Similarity Analysis Between Alignment and Fine-tuning Datasets

Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safety

Sentinel: SOTA model to protect against prompt injections