paper-with-me

홈 › Papers

MLLMGuard: A Multi-dimensional Safety Evaluation Suite for Multimodal Large Language Models

2024-06-11 · Tianle Gu, Zeyang Zhou, Kexin Huang, Dandan Liang, Yixu Wang, Haiquan Zhao, Yuanqi Yao, Xingge Qiao, Keqing Wang, Yujiu Yang, Yan Teng, Yu Qiao, Yingchun Wang

Powered by remarkable advancements in Large Language Models (LLMs), Multimodal Large Language Models (MLLMs) demonstrate impressive capabilities in manifold tasks. However, the practical application scenarios of MLLMs are intricate, exposing them to potential malicious instructions and thereby posing safety risks. While current benchmarks do incorporate certain safety considerations, they often lack comprehensive coverage and fail to exhibit the necessary rigor and robustness. For instance, the common practice of employing GPT-4V as both the evaluator and a model to be evaluated lacks credibility, as it tends to exhibit a bias toward its own responses. In this paper, we present MLLMGuard, a multidimensional safety evaluation suite for MLLMs, including a bilingual image-text evaluation dataset, inference utilities, and a lightweight evaluator. MLLMGuard's assessment comprehensively covers two languages (English and Chinese) and five important safety dimensions (Privacy, Bias, Toxicity, Truthfulness, and Legality), each with corresponding rich subtasks. Focusing on these dimensions, our evaluation dataset is primarily sourced from platforms such as social media, and it integrates text-based and image-based red teaming techniques with meticulous annotation by human experts. This can prevent inaccurate evaluation caused by data leakage when using open-source datasets and ensures the quality and challenging nature of our benchmark. Additionally, a fully automated lightweight evaluator termed GuardRank is developed, which achieves significantly higher evaluation accuracy than GPT-4. Our evaluation results across 13 advanced models indicate that MLLMs still have a substantial journey ahead before they can be considered safe and responsible.

📄 PDF Abstract BibTeX arXiv:2406.07594

Code (1)

Carol-gutianle/MLLMGuard 공식 구현 pytorch

Tasks

Red Teaming

Methods 이 논문이 사용한 방법론

Residual Connection 설명 없음
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
BPE Byte Pair Encoding, or BPE, is a subword segmentation algorithm that encodes rare and unknown words as sequences of subword units. The intuition is that various word…
Label Smoothing Label Smoothing is a regularization technique that introduces noise for the labels. This accounts for the fact that datasets may have mistakes in them, so maximizing the…
Adam 설명 없음
Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.

Similar Papers 제목 키워드 기반

SDEval: Safety Dynamic Evaluation for Multimodal Large Language Models

2025-08-08 · Hanqing Wang, Yuan Tian, Mingyu Liu, Zhenhao Zhang 외 arxiv

In the rapidly evolving landscape of Multimodal Large Language Models (MLLMs), the safety concerns of their outputs have earned significant attention. Although numerous datasets have been proposed, they may become outdat…

LinguaSafe: A Comprehensive Multilingual Safety Benchmark for Large Language Models

2025-08-18 · Zhiyuan Ning, Tianle Gu, Jiaxin Song, Shixin Hong 외 arxiv

The widespread adoption and increasing prominence of large language models (LLMs) in global technologies necessitate a rigorous focus on ensuring their safety across a diverse range of linguistic and cultural contexts. T…

SkillAudit: From Fixed-Suite Benchmarking to Skill-Centered Assessment

2026-06-21 · Dexu Yu, Youhua Li, Zhaoyang Guan, Xianhao Lin 외 arxiv

Agent skills have become a practical way to extend large language model agents, but the growing skill ecosystem still lacks a reliable way to judge whether a skill is worth deploying. Existing evaluation methods remain l…

OutSafe-Bench: A Benchmark for Multimodal Offensive Content Detection in Large Language Models

2025-11-13 · Yuping Yan, Yuhan Xie, Yuanshuai Li, Yingchao Yu 외 arxiv

Since Multimodal Large Language Models (MLLMs) are increasingly being integrated into everyday tools and intelligent agents, growing concerns have arisen regarding their possible output of unsafe contents, ranging from t…

How Brittle is Agent Safety? Rethinking Agent Risk under Intent Concealment and Task Complexity

2025-11-11 · Zihan Ma, Dongsheng Zhu, Shudong Liu, Taolin Zhang 외 arxiv

Current safety evaluations for LLM-driven agents primarily focus on atomic harms, failing to address sophisticated threats where malicious intent is concealed or diluted within complex tasks. We address this gap with a t…