paper-with-me

홈 › Papers

Unveiling Safety Vulnerabilities of Large Language Models

2023-11-07 · George Kour, Marcel Zalmanovici, Naama Zwerdling, Esther Goldbraich, Ora Nova Fandina, Ateret Anaby-Tavor, Orna Raz, Eitan Farchi

As large language models become more prevalent, their possible harmful or inappropriate responses are a cause for concern. This paper introduces a unique dataset containing adversarial examples in the form of questions, which we call AttaQ, designed to provoke such harmful or inappropriate responses. We assess the efficacy of our dataset by analyzing the vulnerabilities of various models when subjected to it. Additionally, we introduce a novel automatic approach for identifying and naming vulnerable semantic regions - input semantic areas for which the model is likely to produce harmful outputs. This is achieved through the application of specialized clustering techniques that consider both the semantic similarity of the input attacks and the harmfulness of the model's responses. Automatically identifying vulnerable semantic regions enhances the evaluation of model weaknesses, facilitating targeted improvements to its safety mechanisms and overall reliability.

📄 PDF Abstract BibTeX arXiv:2311.04124

Code (0)

등록된 구현이 없습니다.

Tasks

Semantic SimilaritySemantic Textual Similarity

Similar Papers 제목 키워드 기반

Exploring the Adversarial Vulnerabilities of Vision-Language-Action Models in Robotics

2024-11-18 · Taowen Wang, Cheng Han, James Chenhao Liang, Wenhao Yang 외

Recently in robotics, Vision-Language-Action (VLA) models have emerged as a transformative approach, enabling robots to execute complex tasks by integrating visual and linguistic inputs within an end-to-end learning fram…

Vision-Language-Action

LLMs in Web Development: Evaluating LLM-Generated PHP Code Unveiling Vulnerabilities and Limitations

2024-04-21 · Rebeka Tóth, Tamas Bisztray, László Erdodi

This study evaluates the security of web application code generated by Large Language Models, analyzing 2,500 GPT-4 generated PHP websites. These were deployed in Docker containers and tested for vulnerabilities using a …

Beyond Text: Unveiling Privacy Vulnerabilities in Multi-modal Retrieval-Augmented Generation

2025-05-20 · Jiankun Zhang, Shenglai Zeng, Jie Ren, Tianqi Zheng 외

Multimodal Retrieval-Augmented Generation (MRAG) systems enhance LMMs by integrating external multimodal databases, but introduce unexplored privacy vulnerabilities. While text-based RAG privacy risks have been studied, …

Privacy PreservingRAGRetrievalRetrieval-augmented Generation

Hidden in the Noise: Unveiling Backdoors in Audio LLMs Alignment through Latent Acoustic Pattern Triggers

2025-08-04 · Liang Lin, Miao Yu, Kaiwen Luo, Yibo Zhang 외 arxiv

As Audio Large Language Models (ALLMs) emerge as powerful tools for speech processing, their safety implications demand urgent attention. While considerable research has explored textual and vision safety, audio's distin…

ProjLens: Unveiling the Role of Projectors in Multimodal Model Safety

2026-04-21 · Kun Wang, Cheng Qian, Miao Yu, Lilan Peng 외 arxiv

Multimodal Large Language Models (MLLMs) have achieved remarkable success in cross-modal understanding and generation, yet their deployment is threatened by critical safety vulnerabilities. While prior works have demonst…