Cover Image

Leave My Images Alone: Preventing Multi-Modal Large Language Models from Analyzing Images via Visual Prompt Injection

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Shao, Zedian, Liu, Hongbin, Hu, Yuepeng, Gong, Neil Zhenqiang
Format:	Preprint
Published:	2026
Subjects:	Computer Vision and Pattern Recognition Artificial Intelligence Cryptography and Security Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Refusing Safe Prompts for Multi-modal Large Language Models
by: Shao, Zedian, et al.
Published: (2024)

Enhancing Prompt Injection Attacks to LLMs via Poisoning Alignment
by: Shao, Zedian, et al.
Published: (2024)

A Critical Evaluation of Defenses against Prompt Injection Attacks
by: Jia, Yuqi, et al.
Published: (2025)

Jailbreaking Safeguarded Text-to-Image Models via Large Language Models
by: Jiang, Zhengyuan, et al.
Published: (2025)

SafeText: Safe Text-to-image Models via Aligning the Text Encoder
by: Hu, Yuepeng, et al.
Published: (2025)

PromptLocate: Localizing Prompt Injection Attacks
by: Jia, Yuqi, et al.
Published: (2025)

Certifiably Robust Image Watermark
by: Jiang, Zhengyuan, et al.
Published: (2024)

ObliInjection: Order-Oblivious Prompt Injection Attack to LLM Agents with Multi-source Data
by: Wang, Reachal, et al.
Published: (2025)

SecInfer: Preventing Prompt Injection via Inference-time Scaling
by: Liu, Yupei, et al.
Published: (2025)

A Transfer Attack to Image Watermarks
by: Hu, Yuepeng, et al.
Published: (2024)

Automatically Generating Visual Hallucination Test Cases for Multimodal Large Language Models
by: Liu, Zhongye, et al.
Published: (2024)

Stable Signature is Unstable: Removing Image Watermark from Diffusion Models
by: Hu, Yuepeng, et al.
Published: (2024)

Fingerprinting LLMs via Prompt Injection
by: Hu, Yuepeng, et al.
Published: (2025)

Robustness of Vision Foundation Models to Common Perturbations
by: Liu, Hongbin, et al.
Published: (2026)

Prompt Injection Attack to Tool Selection in LLM Agents
by: Shi, Jiawen, et al.
Published: (2025)

A Cross-Modal Prompt Injection Attack against Large Vision-Language Models with Image-Only Perturbation
by: Yang, Hao, et al.
Published: (2026)

Mudjacking: Patching Backdoor Vulnerabilities in Foundation Models
by: Liu, Hongbin, et al.
Published: (2024)

Securing Visually-Aware Recommender Systems: An Adversarial Image Reconstruction and Detection Framework
by: Yin, Minglei, et al.
Published: (2023)

Tracing Back the Malicious Clients in Poisoning Attacks to Federated Learning
by: Jia, Yuqi, et al.
Published: (2024)

CorruptEncoder: Data Poisoning based Backdoor Attacks to Contrastive Learning
by: Zhang, Jinghuai, et al.
Published: (2022)

DataSentinel: A Game-Theoretic Detection of Prompt Injection Attacks
by: Liu, Yupei, et al.
Published: (2025)

Model Poisoning Attacks to Federated Learning via Multi-Round Consistency
by: Xie, Yueqi, et al.
Published: (2024)

WebInject: Prompt Injection Attack to Web Agents
by: Wang, Xilong, et al.
Published: (2025)

AI-generated Image Detection: Passive or Watermark?
by: Guo, Moyang, et al.
Published: (2024)

Watermark-based Attribution of AI-Generated Content
by: Jiang, Zhengyuan, et al.
Published: (2024)

BadToken: Token-level Backdoor Attacks to Multi-modal Large Language Models
by: Yuan, Zenghui, et al.
Published: (2025)

WAInjectBench: Benchmarking Prompt Injection Detections for Web Agents
by: Liu, Yinuo, et al.
Published: (2025)

Optimization-based Prompt Injection Attack to LLM-as-a-Judge
by: Shi, Jiawen, et al.
Published: (2024)

EditTrack: Detecting and Attributing AI-assisted Image Editing
by: Jiang, Zhengyuan, et al.
Published: (2025)

Formalizing and Benchmarking Prompt Injection Attacks and Defenses
by: Liu, Yupei, et al.
Published: (2023)

Watermark Robustness and Radioactivity May Be at Odds in Federated Learning
by: Huang, Leixu, et al.
Published: (2025)

Visual Contextual Attack: Jailbreaking MLLMs with Image-Driven Context Injection
by: Miao, Ziqi, et al.
Published: (2025)

Image-based Prompt Injection: Hijacking Multimodal LLMs through Visually Embedded Adversarial Instructions
by: Nagaraja, Neha, et al.
Published: (2026)

AlignSentinel: Alignment-Aware Detection of Prompt Injection Attacks
by: Jia, Yuqi, et al.
Published: (2026)

VideoMarkBench: Benchmarking Robustness of Video Watermarking
by: Jiang, Zhengyuan, et al.
Published: (2025)

Instance-Level Data-Use Auditing of Visual ML Models
by: Huang, Zonghao, et al.
Published: (2025)

Are You Copying My Prompt? Protecting the Copyright of Vision Prompt for VPaaS via Watermark
by: Ren, Huali, et al.
Published: (2024)

MalTool: Malicious Tool Attacks on LLM Agents
by: Hu, Yuepeng, et al.
Published: (2026)

When Think-with-Image Meets Safety: What Determines Multimodal Jailbreak Robustness?
by: Tian, Yuan, et al.
Published: (2026)

Preventing Prompt Injection with Type-Directed Privilege Separation
by: Jacob, Dennis, et al.
Published: (2025)