AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

GenBreak: Red Teaming Text-to-Image Generation Using Large Language Models

Zilong Wang, Xiang Zheng, Xiaosen Wang, Bo Wang, Xingjun Ma

CVPR 2026 · Image generation › Text-to-image diffusion · approach: Fairness and bias in generative models

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 changedWhereMeasuredEffect
use GenBreak (SFT + GRPO with multi-reward) instead of CRTred-team LLM · Safeguarded SD 2.1 nudityTBR (%): 2.9 -> 60.8Improved
use GenBreak instead of CRTred-team LLM · Safeguarded SD 2.1 violenceTBR (%): 0.2 -> 89.7Improved
use GenBreak instead of CRTred-team LLM · Safeguarded SD 2.1 hateTBR (%): 2.9 -> 84.6Improved
use GenBreak instead of CRTred-team LLM · Safeguarded SD 2.1 nudityTCBR (%): 0.0 -> 57.9Improved
use GenBreak instead of CRTred-team LLM · Safeguarded SD 2.1 nuditylexical diversity: 0.863 -> 0.713Mixed
use GenBreak instead of vanilla RLred-team LLM · Safeguarded SD 2.1 nudityimage toxicity: 0.842 -> 0.805Hurt
transfer GenBreak prompts to Leonardo.Ai versus CRTred-team LLM · Leonardo.Ai nudityTBR (%): 0 -> 70Improved
transfer GenBreak prompts to fal.ai hate versus CRTred-team LLM · fal.ai hateTBR (%): 90 -> 75Hurt
remove the bypass reward (ablation)GenBreak RL · Safeguarded SD 2.1 nuditybypass abilityImproved
add the clean reward (ablation)GenBreak RL · Safeguarded SD 2.1 nuditystealthinessImproved
add lexical and semantic prompt diversity rewards (ablation)GenBreak RL · Safeguarded SD 2.1 nudityTCBRImproved
add image diversity reward (ablation)GenBreak RL · Safeguarded SD 2.1 nudityimage diversityMixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy