Comprehensive Evaluation and Analysis for NSFW Concept Erasure in Text-to-Image Diffusion Models

2025/05/21 by Die Chen, Chen, Die, Zhiwen Li +13
Computer Science · #Advanced Text Analysis Techniques #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Sentiment Analysis and Opinion Mining #Text and Document Classification Technologies

paper · pdf · doi:10.48550/arxiv.2505.15450

openalex publication_date 2025/05/21 · openalex created_date 2025/10/19 · openalex updated_date 2026/07/28

Abstract

Text-to-image diffusion models have gained widespread application across various domains, demonstrating remarkable creative potential. However, the strong generalization capabilities of diffusion models can inadvertently lead to the generation of not-safe-for-work (NSFW) content, posing significant risks to their safe deployment. While several concept erasure methods have been proposed to mitigate the issue associated with NSFW content, a comprehensive evaluation of their effectiveness across various scenarios remains absent. To bridge this gap, we introduce a full-pipeline toolkit specifically designed for concept erasure and conduct the first systematic study of NSFW concept erasure methods. By examining the interplay between the underlying mechanisms and empirical observations, we provide in-depth insights and practical guidance for the effective application of concept erasure methods in various real-world scenarios, with the aim of advancing the understanding of content safety in diffusion models and establishing a solid foundation for future research and development in this critical area.

Citations

• Responsible Diffusion Models via Constraining Text Embeddings within Safe Regions
• ACE: Anti-Editing Concept Erasure in Text-to-Image Models
• Meta-Unlearning on Diffusion Models: Preventing Relearning Unlearned Concepts
• Holistic Unlearning Benchmark: A Multi-Faceted Evaluation for Text-to-Image Diffusion Model Unlearning
• TextToon: Real-Time Text Toonify Head Avatar from Single Video
• Growth Inhibitors for Suppressing Inappropriate Image Concepts in Diffusion Models
• Defensive Unlearning with Adversarial Training for Robust Concept Erasure in Diffusion Models
• Evaluating Text-to-Visual Generation with Image-to-Text Generation
• MACE: Mass Concept Erasure in Diffusion Models
• UnlearnCanvas: Stylized Image Dataset for Enhanced Machine Unlearning Evaluation in Diffusion Models
• Erasing Undesirable Influence in Diffusion Models
• One-Dimensional Adapter to Rule Them All: Concepts, Diffusion Models and Erasing Applications
• SalUn: Empowering Machine Unlearning via Gradient-based Weight Saliency in Both Image Classification and Generation
• To Generate or Not? Safety-Driven Unlearned Diffusion Models Are Still Easy To Generate Unsafe Images ... For Now
• Ring-A-Bell! How Reliable are Concept Removal Methods for Diffusion Models?
• Prompting4Debugging: Red-Teaming Text-to-Image Diffusion Models by Finding Problematic Prompts
• Unified Concept Editing in Diffusion Models
• On the Trustworthiness Landscape of State-of-the-art Generative Models: A Survey and Outlook
• Unsafe Diffusion: On the Generation of Unsafe Images and Hateful Memes From Text-To-Image Models
• ImageReward: Learning and Evaluating Human Preferences for Text-to-Image Generation
• Forget-Me-Not: Learning to Forget in Text-to-Image Diffusion Models
• Ablating Concepts in Text-to-Image Diffusion Models
• Erasing Concepts from Diffusion Models
• Safe Latent Diffusion: Mitigating Inappropriate Degeneration in Diffusion Models
• LAION-5B: An open large-scale dataset for training next generation image-text models
• Red-Teaming the Stable Diffusion Safety Filter
• Classifier-Free Diffusion Guidance
• Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding
• High-Resolution Image Synthesis with Latent Diffusion Models
• High-Resolution Image Synthesis with Latent Diffusion Models
• GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models
• LoRA: Low-Rank Adaptation of Large Language Models
• Learning Transferable Visual Models From Natural Language Supervision
• Improving Image Captioning with Better Use of Captions
• Denoising Diffusion Probabilistic Models
• DeepFakes: a New Threat to Face Recognition? Assessment and Detection
• Datasheets for Datasets
• The Unreasonable Effectiveness of Deep Features as a Perceptual Metric
• Deep Unsupervised Learning using Nonequilibrium Thermodynamics
• Microsoft COCO: Common Objects in Context

Cited by

• What Lurks Within? Concept Auditing for Shared Diffusion Models at Scale

Related

• Comprehensive Assessment and Analysis for NSFW Content Erasure in Text-to-Image Diffusion Models
• Pruning for Robust Concept Erasing in Diffusion Models
• Reliable and Efficient Concept Erasure of Text-to-Image Diffusion Models
• CE-SDWV: Effective and Efficient Concept Erasure for Text-to-Image Diffusion Models via a Semantic-Driven Word Vocabulary
• TRCE: Towards Reliable Malicious Concept Erasure in Text-to-Image Diffusion Models