Image de couverture de livre

Image-based Prompt Injection: Hijacking Multimodal LLMs through Visually Embedded Adversarial Instructions

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Nagaraja, Neha, Zhang, Lan, Wang, Zhilong, Zhang, Bo, Patil, Pawan
Format:	Preprint
Publié:	2026
Sujets:	Computer Vision and Pattern Recognition Artificial Intelligence Cryptography and Security
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

To Protect the LLM Agent Against the Prompt Injection Attack with Polymorphic Prompt
par: Wang, Zhilong, et autres
Publié: (2025)

Natias: Neuron Attribution based Transferable Image Adversarial Steganography
par: Fan, Zexin, et autres
Publié: (2024)

Semantic Router: On the Feasibility of Hijacking MLLMs via a Single Adversarial Perturbation
par: Li, Changyue, et autres
Publié: (2025)

Iteratively Prompting Multimodal LLMs to Reproduce Natural and AI-Generated Images
par: Naseh, Ali, et autres
Publié: (2024)

One Prompt to Verify Your Models: Black-Box Text-to-Image Models Verification via Non-Transferable Adversarial Attacks
par: Guo, Ji, et autres
Publié: (2024)

A Cross-Modal Prompt Injection Attack against Large Vision-Language Models with Image-Only Perturbation
par: Yang, Hao, et autres
Publié: (2026)

Jailbreak Vision Language Models via Bi-Modal Adversarial Prompt
par: Ying, Zonghao, et autres
Publié: (2024)

Hijack-GAN: Unintended-Use of Pretrained, Black-Box GANs
par: Wang, Hui-Po, et autres
Publié: (2020)

Improving Adversarial Transferability of Vision-Language Pre-training Models through Collaborative Multimodal Interaction
par: Fu, Jiyuan, et autres
Publié: (2024)

Boosting Adversarial Transferability with Spatial Adversarial Alignment
par: Chen, Zhaoyu, et autres
Publié: (2025)

Probabilistic Modeling of Jailbreak on Multimodal LLMs: From Quantification to Application
par: Xu, Wenzhuo, et autres
Publié: (2025)

SLICE: Semantic Latent Injection via Compartmentalized Embedding for Image Watermarking
par: Gao, Zheng, et autres
Publié: (2026)

Beyond Text Prompts: Precise Concept Erasure through Text-Image Collaboration
par: Li, Jun, et autres
Publié: (2026)

BadBone: Backdoor Attacks Against Backbone Models in Visual Prompt Learning
par: Yang, Ziqing, et autres
Publié: (2026)

VLATTACK: Multimodal Adversarial Attacks on Vision-Language Tasks via Pre-trained Models
par: Yin, Ziyi, et autres
Publié: (2023)

DAVSP: Safety Alignment for Large Vision-Language Models via Deep Aligned Visual Safety Prompt
par: Zhang, Yitong, et autres
Publié: (2025)

Modeling Electromagnetic Signal Injection Attacks on Camera-based Smart Systems: Applications and Mitigation
par: Zhang, Youqian, et autres
Publié: (2024)

Visual Contextual Attack: Jailbreaking MLLMs with Image-Driven Context Injection
par: Miao, Ziqi, et autres
Publié: (2025)

Patronus: Safeguarding Text-to-Image Models against White-Box Adversaries
par: Li, Xinfeng, et autres
Publié: (2025)

BodhiPromptShield: Pre-Inference Prompt Mediation for Suppressing Privacy Propagation in LLM/VLM Agents
par: Ma, Bo, et autres
Publié: (2026)

Hidden Tail: Adversarial Image Causing Stealthy Resource Consumption in Vision-Language Models
par: Zhang, Rui, et autres
Publié: (2025)

Environmental Injection Attacks against GUI Agents in Realistic Dynamic Environments
par: Zhang, Yitong, et autres
Publié: (2025)

Generating Image Adversarial Examples by Embedding Digital Watermarks
par: Xiang, Yuexin, et autres
Publié: (2020)

Controllable Adversarial Makeup for Privacy via Text-Guided Diffusion
par: Kwon, Youngjin, et autres
Publié: (2025)

ZeroPur: Succinct Training-Free Adversarial Purification
par: Liu, Erhu, et autres
Publié: (2024)

Vulnerabilities in AI-generated Image Detection: The Challenge of Adversarial Attacks
par: Diao, Yunfeng, et autres
Publié: (2024)

Unrecognizable Yet Identifiable: Image Distortion with Preserved Embeddings
par: Zakharov, Dmytro, et autres
Publié: (2024)

Understanding Impacts of Electromagnetic Signal Injection Attacks on Object Detection
par: Zhang, Youqian, et autres
Publié: (2024)

Enhancing Adversarial Attacks: The Similar Target Method
par: Zhang, Shuo, et autres
Publié: (2023)

SnatchML: Hijacking ML models without Training Access
par: Ghorbel, Mahmoud, et autres
Publié: (2024)

Leave My Images Alone: Preventing Multi-Modal Large Language Models from Analyzing Images via Visual Prompt Injection
par: Shao, Zedian, et autres
Publié: (2026)

Stealthy Multi-Task Adversarial Attacks
par: Guo, Jiacheng, et autres
Publié: (2024)

Are You Copying My Prompt? Protecting the Copyright of Vision Prompt for VPaaS via Watermark
par: Ren, Huali, et autres
Publié: (2024)

SafeRedir: Prompt Embedding Redirection for Robust Unlearning in Image Generation Models
par: Liu, Renyang, et autres
Publié: (2026)

OpenRT: An Open-Source Red Teaming Framework for Multimodal LLMs
par: Wang, Xin, et autres
Publié: (2026)

TPatch: A Triggered Physical Adversarial Patch
par: Zhu, Wenjun, et autres
Publié: (2023)

Anomaly Unveiled: Securing Image Classification against Adversarial Patch Attacks
par: Chattopadhyay, Nandish, et autres
Publié: (2024)

Time Traveling to Defend Against Adversarial Example Attacks in Image Classification
par: Etim, Anthony, et autres
Publié: (2024)

R-TPT: Improving Adversarial Robustness of Vision-Language Models through Test-Time Prompt Tuning
par: Sheng, Lijun, et autres
Publié: (2025)

Transferability of Adversarial Attacks in Video-based MLLMs: A Cross-modal Image-to-Video Approach
par: Huang, Linhao, et autres
Publié: (2025)