MME-Safety：面向多模态大语言模型安全评估的细粒度基准

Computer Science > Computation and Language arXiv:2609.20850 (cs) [Submitted on 8 Aug 2026 (v1), last revised 22 Sep 2026 (this version, v2)] Title:MME-Safety: A Fine-grained Benchmark for Safety Evaluation of MLLMs Authors:Yilian Shi, Yueming Lyu, Haoxiang Tan, Linzhuang Zou, Qihao Wang, Guihua Yu, Chenyang Si, Caifeng Shan View a PDF of the paper titled MME-Safety: A Fine-grained Benchmark for Safety Evaluation of MLLMs, by Yilian Shi and 7 other authors View PDF HTML (experimental) Abstract:While Multimodal Large Language Models (MLLMs) show remarkable advancements, their cross-modal capabilities introduce complex vulnerabilities that easily bypass unimodal filters. Existing benchmarks lack fine-grained intent-related annotations and rely on unidimensional metrics, hindering comprehensive robustness evaluation. To address this, we propose MME-Safety, a rigorously verified benchmark featuring a unique four-dimensional annotation schema that categorizes risk scenarios, harm severity, and modality-specific stealth levels. Furthermore, we introduce a hierarchical evaluation framework to assess fundamental response reliability, actual risk exposure, and the structural integrity of defensive behaviors. Extensive zero-shot evaluations across 17 state-of-the-art MLLMs provide a comprehensive safety profile of current multimodal systems. Our analysis systematically investigates cross-modal input configurations and uncovers safety implications associated with Chain-of-Thought (CoT) reasoning. These multifaceted findings underscore the urgent need for robust, reasoning-aware safety alignment in the multimodal landscape. Subjects: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV) Cite as: arXiv:2609.20850 [cs.CL] (or arXiv:2609.20850v2 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2609.20850 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Yilian Shi [view email] [v1] Sat, 8 Aug 2026 02:49:39 UTC (23,806 KB) [v2] Tue, 22 Sep 2026 12:08:57 UTC (23,806 KB) Full-text links: Access Paper: View a PDF of the paper titled MME-Safety: A Fine-grained Benchmark for Safety Evaluation of MLLMs, by Yilian Shi and 7 other authors View PDF HTML (experimental) TeX Source view license Current browse context: cs.CL new | recent | 2026-09 Change to browse by: cs cs.CV References & Citations NASA ADS Google Scholar Semantic Scholar export BibTeX citation Loading... BibTeX formatted citation × loading... Data provided by: Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv (What is alphaXiv?) Links to Code Toggle CatalyzeX Code Finder for Papers (What is CatalyzeX?) DagsHub Toggle DagsHub (What is DagsHub?) GotitPub Toggle Gotit.pub (What is GotitPub?) Huggingface Toggle Hugging Face (What is Huggingface?) ScienceCast Toggle ScienceCast (What is ScienceCast?) Demos Demos Replicate Toggle Replicate (What is Replicate?) Spaces Toggle Hugging Face Spaces (What is Spaces?) Spaces Toggle TXYZ.AI (What is TXYZ.AI?) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower (What are Influence Flowers?) Core recommender toggle CORE Recommender (What is CORE?) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs. Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)