Exploring Visual Vulnerabilities via Multi-Loss Adversarial Search for Jailbreaking Vision-Language Models

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Hao, Shuyang, Hooi, Bryan, Liu, Jun, Chang, Kai-Wei, Huang, Zi, Cai, Yujun
Natura:	Preprint
Pubblicazione:	2024
Soggetti:	Computer Vision and Pattern Recognition
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Making Every Step Effective: Jailbreaking Large Vision-Language Models Through Hierarchical KV Equalization
di: Hao, Shuyang, et al.
Pubblicazione: (2025)

Tit-for-Tat: Safeguarding Large Vision-Language Models Against Jailbreak Attacks via Adversarial Defense
di: Hao, Shuyang, et al.
Pubblicazione: (2025)

Text Speaks Louder than Vision: ASCII Art Reveals Textual Biases in Vision-Language Models
di: Wang, Zhaochen, et al.
Pubblicazione: (2025)

How does Watermarking Affect Visual Language Models in Document Understanding?
di: Xu, Chunxue, et al.
Pubblicazione: (2025)

Jailbreak Vision Language Models via Bi-Modal Adversarial Prompt
di: Ying, Zonghao, et al.
Pubblicazione: (2024)

Seeing is Believing: Mitigating Hallucination in Large Vision-Language Models via CLIP-Guided Decoding
di: Deng, Ailin, et al.
Pubblicazione: (2024)

JaiLIP: Jailbreaking Vision-Language Models via Loss Guided Image Perturbation
di: Mia, Md Jueal, et al.
Pubblicazione: (2025)

Cure or Poison? Embedding Instructions Visually Alters Hallucination in Vision-Language Models
di: Wang, Zhaochen, et al.
Pubblicazione: (2025)

MAA: Meticulous Adversarial Attack against Vision-Language Pre-trained Models
di: Zhang, Peng-Fei, et al.
Pubblicazione: (2025)

Words or Vision: Do Vision-Language Models Have Blind Faith in Text?
di: Deng, Ailin, et al.
Pubblicazione: (2025)

Vision Language Models Map Logos to Text via Semantic Entanglement in the Visual Projector
di: Li, Sifan, et al.
Pubblicazione: (2025)

Lost in Edits? A $λ$-Compass for AIGC Provenance
di: You, Wenhao, et al.
Pubblicazione: (2025)

Jailbreaking Vision-Language Models Through the Visual Modality
di: Azulay, Aharon, et al.
Pubblicazione: (2026)

Revealing Physical-World Semantic Vulnerabilities: Universal Adversarial Patches for Infrared Vision-Language Models
di: Hu, Chengyin, et al.
Pubblicazione: (2026)

GRPO-TTA: Test-Time Visual Tuning for Vision-Language Models via GRPO-Driven Reinforcement Learning
di: Li, Yujun, et al.
Pubblicazione: (2026)

Jailbreak Large Vision-Language Models Through Multi-Modal Linkage
di: Wang, Yu, et al.
Pubblicazione: (2024)

Images are Achilles' Heel of Alignment: Exploiting Visual Vulnerabilities for Jailbreaking Multimodal Large Language Models
di: Li, Yifan, et al.
Pubblicazione: (2024)

FrameMind: Frame-Interleaved Video Reasoning via Reinforcement Learning
di: Ge, Haonan, et al.
Pubblicazione: (2025)

Securing Vision-Language Models with a Robust Encoder Against Jailbreak and Adversarial Attacks
di: Hossain, Md Zarif, et al.
Pubblicazione: (2024)

Jailbreaks on Vision Language Model via Multimodal Reasoning
di: Noheria, Aarush, et al.
Pubblicazione: (2026)

Activation Steering Meets Preference Optimization: Defense Against Jailbreaks in Vision Language Models
di: Wu, Sihao, et al.
Pubblicazione: (2025)

Seeing Through Deception: Uncovering Misleading Creator Intent in Multimodal News with Vision-Language Models
di: Wu, Jiaying, et al.
Pubblicazione: (2025)

When Text and Images Don't Mix: Bias-Correcting Language-Image Similarity Scores for Anomaly Detection
di: Goodge, Adam, et al.
Pubblicazione: (2024)

Universal Adversarial Perturbations for Vision-Language Pre-trained Models
di: Zhang, Peng-Fei, et al.
Pubblicazione: (2024)

Visual Adversarial Attack on Vision-Language Models for Autonomous Driving
di: Zhang, Tianyuan, et al.
Pubblicazione: (2024)

Adversarial Prompt Distillation for Vision-Language Models
di: Luo, Lin, et al.
Pubblicazione: (2024)

JailbreakZoo: Survey, Landscapes, and Horizons in Jailbreaking Large Language and Vision-Language Models
di: Jin, Haibo, et al.
Pubblicazione: (2024)

SemVink: Advancing VLMs' Semantic Understanding of Optical Illusions via Visual Global Thinking
di: Li, Sifan, et al.
Pubblicazione: (2025)

Dataset Distillation via Vision-Language Category Prototype
di: Zou, Yawen, et al.
Pubblicazione: (2025)

VRSA: Jailbreaking Multimodal Large Language Models through Visual Reasoning Sequential Attack
di: Zhao, Shiji, et al.
Pubblicazione: (2025)

Benchmarking Large Vision-Language Models via Directed Scene Graph for Comprehensive Image Captioning
di: Lu, Fan, et al.
Pubblicazione: (2024)

Principled Steering via Null-space Projection for Jailbreak Defense in Vision-Language Models
di: Zhu, Xingyu, et al.
Pubblicazione: (2026)

SegDebias: Test-Time Bias Mitigation for ViT-Based CLIP via Segmentation
di: Wu, Fangyu, et al.
Pubblicazione: (2025)

Spatio-Temporal Foundation Models: Vision, Challenges, and Opportunities
di: Goodge, Adam, et al.
Pubblicazione: (2025)

PruneVid: Visual Token Pruning for Efficient Video Large Language Models
di: Huang, Xiaohu, et al.
Pubblicazione: (2024)

LLMs are Good Action Recognizers
di: Qu, Haoxuan, et al.
Pubblicazione: (2024)

TagAlign: Improving Vision-Language Alignment with Multi-Tag Classification
di: Liu, Qinying, et al.
Pubblicazione: (2023)

MMT-ARD: Multimodal Multi-Teacher Adversarial Distillation for Robust Vision-Language Models
di: Li, Yuqi, et al.
Pubblicazione: (2025)

TAIJI: Textual Anchoring for Immunizing Jailbreak Images in Vision Language Models
di: Yin, Xiangyu, et al.
Pubblicazione: (2025)

When Lighting Deceives: Exposing Vision-Language Models' Illumination Vulnerability Through Illumination Transformation Attack
di: Liu, Hanqing, et al.
Pubblicazione: (2025)