paper-with-me

홈 › Papers

Visual Exclusivity Attacks: Automatic Multimodal Red Teaming via Agentic Planning

2026-02-05 · Yunbei Zhang, Yingqiang Ge, Weijie Xu, Yuhui Xu, Jihun Hamm, Chandan K. Reddy arxiv

Current multimodal red teaming treats images as wrappers for malicious payloads via typography or adversarial noise. These attacks are structurally brittle, as standard defenses neutralize them once the payload is exposed. We introduce Visual Exclusivity (VE), a more resilient Image-as-Basis threat where harm emerges only through reasoning over visual content such as technical schematics. To systematically exploit VE, we propose Multimodal Multi-turn Agentic Planning (MM-Plan), a framework that reframes jailbreaking from turn-by-turn reaction to global plan synthesis. MM-Plan trains an attacker planner to synthesize comprehensive, multi-turn strategies, optimized via Group Relative Policy Optimization (GRPO), enabling self-discovery of effective strategies without human supervision. To rigorously benchmark this reasoning-dependent threat, we introduce VE-Safety, a human-curated dataset filling a critical gap in evaluating high-risk technical visual understanding. MM-Plan achieves 46.3% attack success rate against Claude 4.5 Sonnet and 13.8% against GPT-5, outperforming baselines by 2--5x where existing methods largely fail. These findings reveal that frontier models remain vulnerable to agentic multimodal attacks, exposing a critical gap in current safety alignment. Warning: This paper contains potentially harmful content.

📄 PDF Abstract BibTeX arXiv:2603.20198

Code (0)

등록된 구현이 없습니다.

Tasks

Red Teaming

Similar Papers 제목 키워드 기반

ARMs: Adaptive Red-Teaming Agent against Multimodal Models with Plug-and-Play Attacks

2025-10-03 · Zhaorun Chen, Xun Liu, Mintong Kang, Jiawei Zhang 외 arxiv

As vision-language models (VLMs) gain prominence, their multimodal interfaces also introduce new safety vulnerabilities, making the safety evaluation challenging and critical. Existing red-teaming efforts are either rest…

Red-teaming the Multimodal Reasoning: Jailbreaking Vision-Language Models via Cross-modal Entanglement Attacks

2026-02-09 · Yu Yan, Sheng Sun, Shengjia Cheng, Teli Liu 외 arxiv

Vision-Language Models (VLMs) with multimodal reasoning capabilities are high-value attack targets, given their potential for handling complex multimodal harmful tasks. Mainstream black-box jailbreak attacks on VLMs work…

Multimodal Reasoning

Against The Achilles' Heel: A Survey on Red Teaming for Generative Models

2024-03-31 · Lizhi Lin, Honglin Mu, Zenan Zhai, Minghan Wang 외

Generative models are rapidly gaining popularity and being integrated into everyday applications, raising concerns over their safe use as various vulnerabilities are exposed. In light of this, the field of red teaming is…

Red TeamingSurvey

Red Teaming GPT-4V: Are GPT-4V Safe Against Uni/Multi-Modal Jailbreak Attacks?

2024-04-04 · Shuo Chen, Zhen Han, Bailan He, Zifeng Ding 외

Various jailbreak attacks have been proposed to red-team Large Language Models (LLMs) and revealed the vulnerable safeguards of LLMs. Besides, some methods are not limited to the textual modality and extend the jailbreak…

Red Teaming

BELT: Old-School Backdoor Attacks can Evade the State-of-the-Art Defense with Backdoor Exclusivity Lifting

2023-12-08 · Huming Qiu, Junjie Sun, Mi Zhang, Xudong Pan 외

Deep neural networks (DNNs) are susceptible to backdoor attacks, where malicious functionality is embedded to allow attackers to trigger incorrect classifications. Old-school backdoor attacks use strong trigger features …