Research map / Adversarial robustness and security / Attacks on text and graphs
JailbreakDiffBench: A Comprehensive Benchmark for Jailbreaking Diffusion Models
Xiaolong Jin, Zixuan Weng, Hanxi Guo, Chenlong Yin, Siyuan Cheng, Guangyu Shen, Xiangyu Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add in-context examples to GPT-4o-mini prompt checker (ablation) | GPT-4o-mini · JDBench prompts | accuracy: 0.8540 -> 0.8720 | Improved |
| add in-context examples to GPT-4o prompt checker (ablation) | GPT-4o · JDBench prompts | FPR: 0.5481 -> 0.2869 | Improved |
| add in-context examples to GPT-4o prompt checker (ablation) | GPT-4o · JDBench prompts | accuracy: 0.6184 -> 0.7691 | Improved |
| add in-context examples to GPT-4o-mini (ablation) | GPT-4o-mini · Copyright category | recall: 0.562 -> 0.342 | Hurt |
| use LLM checker GPT-4o instead of Multihead Detector for images | Multihead Detector · image checking | accuracy: 0.7571 -> 0.9203 | Improved |
| use rule-based keyword matcher vs BERT classifier | NSFW-Word-Matcher · JDBench prompts | F1: 0.2635 -> 0.7991 | Hurt |
| use larger GPT-4o instead of GPT-4o-mini as prompt checker | GPT-4o-mini · JDBench prompts | FPR: 0.1376 -> 0.5481 | Hurt |
| black-box PGJ attack vs white-box RingABell against checkers | RingABell · SD-v1.5 | BPR: 0.50% -> 51.75% | Improved |
| bypassing checkers does not translate to harmful images | PGJ · SD-v1.5 | ASR: 51.75% -> 5.75% | Hurt |
| newer SD-3.5-Large-Turbo vs SD-v1.5 vulnerability | SD-v1.5 · DACA attack | BPR: 57.00% -> 56.25% | No clear effect |
Closest papers by meaning
- JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models (NeurIPS 2024 · similarity 0.90)
- DiffusionGuard: A Robust Defense Against Malicious Diffusion-based Image Editing (ICLR 2025 · similarity 0.88)
- T2V-OptJail: Discrete Prompt Optimization for Text-to-Video Jailbreak Attacks (NeurIPS 2025 · similarity 0.88)
- Why Jailbreaks Succeed in Diffusion Language Models: An Energy Landscape Analysis (NeurIPS 2026 · similarity 0.88)
- Elijah: Eliminating Backdoors Injected in Diffusion Models via Distribution Shift (AAAI 2024 · similarity 0.88)
- DiffHammer: Rethinking the Robustness of Diffusion-Based Adversarial Purification (NeurIPS 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
