UniSAFE: A Comprehensive Benchmark for Safety Evaluation of Unified Multimodal Models

• URL: http://arxiv.org/abs/2603.17476v1
• Date: Wed, 18 Mar 2026 08:30:31 GMT
• Title: UniSAFE: A Comprehensive Benchmark for Safety Evaluation of Unified Multimodal Models
• Abstract summary: We introduce UniSAFE, the first comprehensive benchmark for system-level safety evaluation of Unified Multimodal Models (UMMs)<n>UniSAFE projects common risk scenarios across task-specific I/O configurations, enabling controlled cross-task comparisons of safety failures.<n>Our results reveal critical vulnerabilities across current UMMs, including elevated safety violations in multi-image composition and multi-turn settings.
• Score: 29.443907124394382
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Unified Multimodal Models (UMMs) offer powerful cross-modality capabilities but introduce new safety risks not observed in single-task models. Despite their emergence, existing safety benchmarks remain fragmented across tasks and modalities, limiting the comprehensive evaluation of complex system-level vulnerabilities. To address this gap, we introduce UniSAFE, the first comprehensive benchmark for system-level safety evaluation of UMMs across 7 I/O modality combinations, spanning conventional tasks and novel multimodal-context image generation settings. UniSAFE is built with a shared-target design that projects common risk scenarios across task-specific I/O configurations, enabling controlled cross-task comparisons of safety failures. Comprising 6,802 curated instances, we use UniSAFE to evaluate 15 state-of-the-art UMMs, both proprietary and open-source. Our results reveal critical vulnerabilities across current UMMs, including elevated safety violations in multi-image composition and multi-turn settings, with image-output tasks consistently more vulnerable than text-output tasks. These findings highlight the need for stronger system-level safety alignment for UMMs. Our code and data are publicly available at https://github.com/segyulee/UniSAFE

Related papers

• Does Unification Come at a Cost? Uni-SafeBench: A Safety Benchmark for Unified Multimodal Large Models [15.702553632939102]
  Unified Multimodal Large Models (UMLMs) integrate understanding and generation capabilities within a single architecture.<n>Existing safety benchmarks predominantly focus on isolated understanding or generation tasks.<n>We introduce Uni-SafeBench, a comprehensive benchmark featuring a taxonomy of six major safety categories across seven task types.<n>We develop Uni-Judger, a framework that effectively decouples contextual safety from model safety.
  arXiv  Detail & Related papers  (2026-04-01T06:45:01Z)
• Robust Multimodal Safety via Conditional Decoding [52.92816441364308]
  Multimodal large-language models (MLLMs) often experience degraded safety alignment when harmful queries exploit cross-modal interactions.<n>We propose a simple conditional decoding strategy, CASA, that utilizes internal representations of MLLMs to predict a binary safety token before response generation.
  arXiv  Detail & Related papers  (2026-03-31T23:19:50Z)
• Unsafe by Reciprocity: How Generation-Understanding Coupling Undermines Safety in Unified Multimodal Models [53.41258113970795]
  We investigate whether cross-functionality reciprocity itself constitutes a structural source of vulnerability in Unified Multimodal Models (UMMs)<n>We propose RICE: Reciprocal Interaction-based Cross-functionality Exploitation, a novel attack paradigm that explicitly exploits bidirectional interactions between understanding and generation.
  arXiv  Detail & Related papers  (2026-03-28T16:28:39Z)
• When Safe Unimodal Inputs Collide: Optimizing Reasoning Chains for Cross-Modal Safety in Multimodal Large Language Models [50.66979825532277]
  We introduce Safe-Semantics-but-Unsafe-Interpretation (SSUI), the first dataset featuring interpretable reasoning paths tailored for a cross-modal challenge.<n>A novel training framework, Safety-aware Reasoning Path Optimization (SRPO), is also designed based on the SSUI dataset.<n> Experimental results show that our SRPO-trained models achieve state-of-the-art results on key safety benchmarks.
  arXiv  Detail & Related papers  (2025-09-15T15:40:58Z)
• Align is not Enough: Multimodal Universal Jailbreak Attack against Multimodal Large Language Models [83.80177564873094]
  We propose a unified multimodal universal jailbreak attack framework.<n>We evaluate the undesirable context generation of MLLMs like LLaVA, Yi-VL, MiniGPT4, MiniGPT-v2, and InstructBLIP.<n>This study underscores the urgent need for robust safety measures in MLLMs.
  arXiv  Detail & Related papers  (2025-06-02T04:33:56Z)
• USB: A Comprehensive and Unified Safety Evaluation Benchmark for Multimodal Large Language Models [31.412080488801507]
  Unified Safety Benchmarks (USB) is one of the most comprehensive evaluation benchmarks in MLLM safety.<n>Our benchmark features high-quality queries, extensive risk categories, comprehensive modal combinations, and encompasses both vulnerability and oversensitivity evaluations.
  arXiv  Detail & Related papers  (2025-05-26T08:39:14Z)
• DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models [37.104276926258095]
  Multimodal Large Language Models (MLLMs) pose unique safety challenges due to their integration of visual and textual data.<n>We introduce textbfDREAM (textittextbfDisentangling textbfRisks to textbfEnhance Safety textbfAlignment in textbfMLLMs), a novel approach that enhances safety alignment in MLLMs through supervised fine-tuning and iterative Reinforcement Learning from AI Feedback.
  arXiv  Detail & Related papers  (2025-04-25T03:54:24Z)
• Rethinking Bottlenecks in Safety Fine-Tuning of Vision Language Models [25.606641582511106]
  We propose a novel dataset that integrates multi-image inputs with safety Chain-of-Thought (CoT) labels as fine-grained reasoning logic to improve model performance.<n>Our experiments demonstrate that fine-tuning InternVL2.5-8B with MIS significantly outperforms both powerful open-source models and API-based models in challenging multi-image tasks.
  arXiv  Detail & Related papers  (2025-01-30T17:59:45Z)
• SoK: The Security-Safety Continuum of Multimodal Foundation Models through Information Flow and Game-Theoretic Defenses [58.93030774141753]
  Multimodal foundation models (MFMs) integrate diverse data modalities to support complex and wide-ranging tasks.<n>In this paper, we unify the concepts of safety and security in the context of MFMs by identifying critical threats that arise from both model behavior and system-level interactions.
  arXiv  Detail & Related papers  (2024-11-17T23:06:20Z)
• Safe Inputs but Unsafe Output: Benchmarking Cross-modality Safety Alignment of Large Vision-Language Model [73.8765529028288]
  We introduce a novel safety alignment challenge called Safe Inputs but Unsafe Output (SIUO) to evaluate cross-modality safety alignment.<n>To empirically investigate this problem, we developed the SIUO, a cross-modality benchmark encompassing 9 critical safety domains, such as self-harm, illegal activities, and privacy violations.<n>Our findings reveal substantial safety vulnerabilities in both closed- and open-source LVLMs, underscoring the inadequacy of current models to reliably interpret and respond to complex, real-world scenarios.
  arXiv  Detail & Related papers  (2024-06-21T16:14:15Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.