Adversarial smuggling attacks encode harmful content into human-readable visuals that evade MLLM detection, achieving over 90% attack success rates on models like GPT-5 and Qwen3-VL via the new SmuggleBench benchmark.
Visual- roleplay: Universal jailbreak attack on multimodal large language mod- els via role-playing image character
9 Pith papers cite this work. Polarity classification is still indexing.
9
Pith papers citing it
citation-role summary
background 4
citation-polarity summary
roles
background 4polarities
background 4representative citing papers
A narrative survey that catalogs fifty papers on diffusion-based adversarial techniques across text, vision, and vision-language models, proposes a six-class taxonomy of diffusion roles plus a unified five-dimension evaluation framework, and releases a companion catalog.
MLingualFC benchmark finds flowchart jailbreaks succeed at high rates for Latin-script languages but much lower rates for Punjabi in multilingual VLMs, pointing to language-dependent safety gaps.
CAVI framework uses character-guided token pruning, orthogonal feature modulation, and modality-adaptive role steering to resolve modality-role interference in multimodal RPAs.
DACO curates a 15,000-concept dictionary from 400K image-caption pairs and uses it to initialize an SAE that enables granular, concept-specific steering of MLLM activations, raising safety scores on MM-SafetyBench and JailBreakV while preserving general capabilities.
A neuron-targeted safety tuning method for VLLMs reduces attack success rates from ~20-30% to ~4-6% on average across ten languages while using less than 0.03% of parameters.
A comprehensive survey that taxonomizes safety threats to large models and agents, reviews defenses and benchmarks, and outlines open challenges.
citing papers explorer
-
Making MLLMs Blind: Adversarial Smuggling Attacks in MLLM Content Moderation
Adversarial smuggling attacks encode harmful content into human-readable visuals that evade MLLM detection, achieving over 90% attack success rates on models like GPT-5 and Qwen3-VL via the new SmuggleBench benchmark.
-
Adversarial Diffusion Across Modalities: A Fusion Survey of Attacks, Defenses, and Evaluation for Text, Vision, and Vision-Language Models
A narrative survey that catalogs fifty papers on diffusion-based adversarial techniques across text, vision, and vision-language models, proposes a six-class taxonomy of diffusion roles plus a unified five-dimension evaluation framework, and releases a companion catalog.
-
MLingualFC: Evaluating Jailbreak Vulnerabilities in Multilingual Vision-Language Models
MLingualFC benchmark finds flowchart jailbreaks succeed at high rates for Latin-script languages but much lower rates for Punjabi in multilingual VLMs, pointing to language-dependent safety gaps.
-
Through the Lens of Character: Resolving Modality-Role Interference in Multimodal Role-Playing Agent
CAVI framework uses character-guided token pruning, orthogonal feature modulation, and modality-adaptive role steering to resolve modality-role interference in multimodal RPAs.
-
Dictionary-Aligned Concept Control for Safeguarding Multimodal LLMs
DACO curates a 15,000-concept dictionary from 400K image-caption pairs and uses it to initialize an SAE that enables granular, concept-specific steering of MLLM activations, raising safety scores on MM-SafetyBench and JailBreakV while preserving general capabilities.
-
Targeted Interpretable Safety Neuron Enhancement for Multilingual Vision-Language Large Models
A neuron-targeted safety tuning method for VLLMs reduces attack success rates from ~20-30% to ~4-6% on average across ten languages while using less than 0.03% of parameters.
-
Safety at Scale: A Comprehensive Survey of Large Model and Agent Safety
A comprehensive survey that taxonomizes safety threats to large models and agents, reviews defenses and benchmarks, and outlines open challenges.
- Localization then Neutralization: Gradient-guided Token Suppression against Visual Prompt Injection Attack
- RedDiffuser: Auditing Multimodal Safety Failures in Vision-Language Models via Reinforced Diffusion