Imagen de Portada

Beyond the Safety Tax: Mitigating Unsafe Text-to-Image Generation via External Safety Rectification

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Meng, Xiangtao, Dong, Yingkai, Yu, Ning, Wang, Li, Li, Zheng, Guo, Shanqing
Formato:	Preprint
Publicado:	2025
Materias:	Computer Vision and Pattern Recognition Artificial Intelligence Cryptography and Security
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Beyond Known Fakes: Generalized Detection of AI-Generated Images via Post-hoc Distribution Alignment
por: Wang, Li, et al.
Publicado: (2025)

Fuzz-Testing Meets LLM-Based Agents: An Automated and Efficient Framework for Jailbreaking Text-To-Image Generation Models
por: Dong, Yingkai, et al.
Publicado: (2024)

AVA: Inconspicuous Attribute Variation-based Adversarial Attack bypassing DeepFake Detection
por: Meng, Xiangtao, et al.
Publicado: (2023)

VidLeaks: Membership Inference Attacks Against Text-to-Video Models
por: Wang, Li, et al.
Publicado: (2026)

DCMI: A Differential Calibration Membership Inference Attack Against Retrieval-Augmented Generation
por: Gao, Xinyu, et al.
Publicado: (2025)

UnsafeBench: Benchmarking Image Safety Classifiers on Real-World and AI-Generated Images
por: Qu, Yiting, et al.
Publicado: (2024)

GuardTrace-VL: Detecting Unsafe Multimodel Reasoning via Iterative Safety Supervision
por: Xiang, Yuxiao, et al.
Publicado: (2025)

PromptGuard: Soft Prompt-Guided Unsafe Content Moderation for Text-to-Image Models
por: Yuan, Lingzhi, et al.
Publicado: (2025)

SurrogatePrompt: Bypassing the Safety Filter of Text-to-Image Models via Substitution
por: Ba, Zhongjie, et al.
Publicado: (2023)

Beyond Text Prompts: Precise Concept Erasure through Text-Image Collaboration
por: Li, Jun, et al.
Publicado: (2026)

DAVSP: Safety Alignment for Large Vision-Language Models via Deep Aligned Visual Safety Prompt
por: Zhang, Yitong, et al.
Publicado: (2025)

On the Proactive Generation of Unsafe Images From Text-To-Image Models Using Benign Prompts
por: Wu, Yixin, et al.
Publicado: (2023)

FaceSwapGuard: Safeguarding Facial Privacy from DeepFake Threats through Identity Obfuscation
por: Wang, Li, et al.
Publicado: (2025)

Membership Inference Attack Against Masked Image Modeling
por: Li, Zheng, et al.
Publicado: (2024)

T2I-RiskyPrompt: A Benchmark for Safety Evaluation, Attack, and Defense on Text-to-Image Model
por: Zhang, Chenyu, et al.
Publicado: (2025)

SafeGuider: Robust and Practical Content Safety Control for Text-to-Image Models
por: Qi, Peigui, et al.
Publicado: (2025)

On the Generation and Mitigation of Harmful Geometry in Image-to-3D Models
por: Liu, Yule, et al.
Publicado: (2026)

From Defender to Devil? Unintended Risk Interactions Induced by LLM Defenses
por: Meng, Xiangtao, et al.
Publicado: (2025)

Membership Inference on Text-to-Image Diffusion Models via Conditional Likelihood Discrepancy
por: Zhai, Shengfang, et al.
Publicado: (2024)

The VLLM Safety Paradox: Dual Ease in Jailbreak Attack and Defense
por: Guo, Yangyang, et al.
Publicado: (2024)

When Understanding Becomes a Risk: Authenticity and Safety Risks in the Emerging Image Generation Paradigm
por: Leng, Ye, et al.
Publicado: (2026)

Contextual Image Attack: How Visual Context Exposes Multimodal Safety Vulnerabilities
por: Xiong, Yuan, et al.
Publicado: (2025)

The Structural Safety Generalization Problem
por: Broomfield, Julius, et al.
Publicado: (2025)

Pushing the Limits of Safety: A Technical Report on the ATLAS Challenge 2025
por: Ying, Zonghao, et al.
Publicado: (2025)

DREAM: Scalable Red Teaming for Text-to-Image Generative Systems via Distribution Modeling
por: Li, Boheng, et al.
Publicado: (2025)

One Prompt to Verify Your Models: Black-Box Text-to-Image Models Verification via Non-Transferable Adversarial Attacks
por: Guo, Ji, et al.
Publicado: (2024)

T2VSafetyBench: Evaluating the Safety of Text-to-Video Generative Models
por: Miao, Yibo, et al.
Publicado: (2024)

DiffZOO: A Purely Query-Based Black-Box Attack for Red-teaming Text-to-Image Generative Model via Zeroth Order Optimization
por: Dang, Pucheng, et al.
Publicado: (2024)

ID-Cloak: Crafting Identity-Specific Cloaks Against Personalized Text-to-Image Generation
por: Teng, Qianrui, et al.
Publicado: (2025)

Towards Understanding Unsafe Video Generation
por: Pang, Yan, et al.
Publicado: (2024)

Defenses at Odds: Measuring and Explaining Defense Conflicts in Large Language Models
por: Meng, Xiangtao, et al.
Publicado: (2026)

When Memory Becomes a Vulnerability: Towards Multi-turn Jailbreak Attacks against Text-to-Image Generation Systems
por: Zhao, Shiqian, et al.
Publicado: (2025)

Scaling Exposes the Trigger: Input-Level Backdoor Detection in Text-to-Image Diffusion Models via Cross-Attention Scaling
por: Li, Zida, et al.
Publicado: (2026)

Responsible Diffusion: A Comprehensive Survey on Safety, Ethics, and Trust in Diffusion Models
por: Wei, Kang, et al.
Publicado: (2025)

Unsafe LLM-Based Search: Quantitative Analysis and Mitigation of Safety Risks in AI Web Search
por: Luo, Zeren, et al.
Publicado: (2025)

Unveiling and Mitigating Memorization in Text-to-image Diffusion Models through Cross Attention
por: Ren, Jie, et al.
Publicado: (2024)

A Survey of Safety on Large Vision-Language Models: Attacks, Defenses and Evaluations
por: Ye, Mang, et al.
Publicado: (2025)

SafeGRPO: Self-Rewarded Multimodal Safety Alignment via Rule-Governed Policy Optimization
por: Rong, Xuankun, et al.
Publicado: (2025)

SafeGen: Mitigating Sexually Explicit Content Generation in Text-to-Image Models
por: Li, Xinfeng, et al.
Publicado: (2024)

Mitigating Sexual Content Generation via Embedding Distortion in Text-conditioned Diffusion Models
por: Ahn, Jaesin, et al.
Publicado: (2025)