Paper Library

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

Zhiyu Xue, Reza Abbasi-Asl, Ramtin Pedarsani

2025-06-08

2506.09067v1

AlphaSteer: Learning Refusal Steering with Principled Null-Space Constraint

Leheng Sheng, Changshuo Shen, Weixiang Zhao, Junfeng Fang, Xiaohao Liu, Zhenkai Liang, Xiang Wang, An Zhang, Tat-Seng Chua

2025-06-08

red teaming

2506.07022v1

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

Xiaoyuan Zhu, Yaowen Ye, Tianyi Qiu, Hanlin Zhu, Sijun Tan, Ajraf Mannan, Jonathan Michala, Raluca Ada Popa, Willie Neiswanger

2025-06-08

red teaming

2506.06975v3

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

Xiaoyuan Zhu, Yaowen Ye, Tianyi Qiu, Hanlin Zhu, Sijun Tan, Ajraf Mannan, Jonathan Michala, Raluca Ada Popa, Willie Neiswanger

2025-06-08

red teaming

2506.06975v2

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

Xiaoyuan Zhu, Yaowen Ye, Tianyi Qiu, Hanlin Zhu, Sijun Tan, Ajraf Mannan, Jonathan Michala, Raluca Ada Popa, Willie Neiswanger

2025-06-08

red teaming

2506.06975v1

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

Kyubyung Chae, Hyunbin Jin, Taesup Kim

2025-06-07

red teaming

2506.10020v1

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

Ruizhong Qiu, Gaotang Li, Tianxin Wei, Jingrui He, Hanghang Tong

2025-06-06

safety

2506.06444v1

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

Zarreen Reza, Alexander Mazur, Michael T. Dugdale, Robin Ray-Chaudhuri

2025-06-06

2506.05925v1

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

Lei Hsiung, Tianyu Pang, Yung-Chen Tang, Linyue Song, Tsung-Yi Ho, Pin-Yu Chen, Yaoqing Yang

2025-06-05

red teaming safety

2506.05346v1

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

Seongmin Lee, Aeree Cho, Grace C. Kim, ShengYun Peng, Mansi Phute, Duen Horng Chau

2025-06-05

safety

2506.05451v1

HoliSafe: Holistic Safety Benchmarking and Modeling with Safety Meta Token for Vision-Language Model

Youngwan Lee, Kangsan Kim, Kwanyong Park, Ilcahe Jung, Soojin Jang, Seanie Lee, Yong-Ju Lee, Sung Ju Hwang

2025-06-05

2506.04704v2

HoliSafe: Holistic Safety Benchmarking and Modeling with Safety Meta Token for Vision-Language Model

Youngwan Lee, Kangsan Kim, Kwanyong Park, Ilcahe Jung, Soojin Jang, Seanie Lee, Yong-Ju Lee, Sung Ju Hwang

2025-06-05

2506.04704v1

Adversarial Attacks on Robotic Vision Language Action Models

Eliot Krzysztof Jones, Alexander Robey, Andy Zou, Zachary Ravichandran, George J. Pappas, Hamed Hassani, Matt Fredrikson, J. Zico Kolter

2025-06-03

red teaming

2506.03350v1

It's the Thought that Counts: Evaluating the Attempts of Frontier LLMs to Persuade on Harmful Topics

Matthew Kowal, Jasper Timm, Jean-Francois Godbout, Thomas Costello, Antonio A. Arechar, Gordon Pennycook, David Rand, Adam Gleave, Kellin Pelrine

2025-06-03

red teaming

2506.02873v1

From Prompts to Protection: Large Language Model-Enabled In-Context Learning for Smart Public Safety UAV

Yousef Emami, Hao Zhou, Miguel Gutierrez Gaitan, Kai Li, Luis Almeida, Zhu Han

2025-06-03

2506.02649v1

IndoSafety: Culturally Grounded Safety for LLMs in Indonesian Languages

Muhammad Falensi Azmi, Muhammad Dehan Al Kautsar, Alfan Farizki Wicaksono, Fajri Koto

2025-06-03

safety

2506.02573v1

BitBypass: A New Direction in Jailbreaking Aligned Large Language Models with Bitstream Camouflage

Kalyan Nakka, Nitesh Saxena

2025-06-03

red teaming

2506.02479v1

Should LLM Safety Be More Than Refusing Harmful Instructions?

Utsav Maskey, Mark Dras, Usman Naseem

2025-06-03

safety

2506.02442v2

Should LLM Safety Be More Than Refusing Harmful Instructions?

Utsav Maskey, Mark Dras, Usman Naseem

2025-06-03

safety

2506.02442v1

AnswerCarefully: A Dataset for Improving the Safety of Japanese LLM Output

Hisami Suzuki, Satoru Katsumata, Takashi Kodama, Tetsuro Takahashi, Kouta Nakayama, Satoshi Sekine

2025-06-03

safety

2506.02372v1

Evaluating LLM Agent Adherence to Hierarchical Safety Principles: A Lightweight Benchmark for Probing Foundational Controllability Components

Ram Potham

2025-06-03

safety

2506.02357v1

June 09 - June 15, 2025

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

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

HoliSafe: Holistic Safety Benchmarking and Modeling with Safety Meta Token for Vision-Language Model

HoliSafe: Holistic Safety Benchmarking and Modeling with Safety Meta Token for Vision-Language Model

Adversarial Attacks on Robotic Vision Language Action Models

It's the Thought that Counts: Evaluating the Attempts of Frontier LLMs to Persuade on Harmful Topics

From Prompts to Protection: Large Language Model-Enabled In-Context Learning for Smart Public Safety UAV

IndoSafety: Culturally Grounded Safety for LLMs in Indonesian Languages

BitBypass: A New Direction in Jailbreaking Aligned Large Language Models with Bitstream Camouflage

Should LLM Safety Be More Than Refusing Harmful Instructions?

Should LLM Safety Be More Than Refusing Harmful Instructions?

AnswerCarefully: A Dataset for Improving the Safety of Japanese LLM Output

Evaluating LLM Agent Adherence to Hierarchical Safety Principles: A Lightweight Benchmark for Probing Foundational Controllability Components