AI Paper Atlas

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

Beautiful Images, Toxic Words: Understanding and Addressing Offensive Text in Generated Images

Aditya Kumar, Tom Blanchard, Adam Dziedzic, Franziska Boenisch

AAAI 2026 · Image generation › Text-to-image diffusion · approach: Text-driven synthesis and semantic guidance · 0 citations (OpenAlex)

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
restrict fine-tuning to text-generation layers instead of all attention layers (ablation)NSFW-Intervention (SD3) · ToxicBench NSFW words (test)delta NGramLD (NSFW): +0.49 -> +3.63Improved
restrict fine-tuning to text-generation layers instead of all attention layers (ablation)NSFW-Intervention (SDXL) · ToxicBench NSFW words (test)delta NGramLD (NSFW): 1.97 -> 5.69Improved
restrict fine-tuning to text-generation layers instead of all attention layers (ablation)NSFW-Intervention (DeepFloyd IF) · ToxicBench NSFW words (test)delta NGramLD (NSFW): 0.19 -> 4.72Improved
fine-tune the diffusion backbone rather than the CLIP text encoder (ablation)NSFW-Intervention (SD3) · ToxicBench NSFW words (test)NGramLD NSFW minus benign delta: 1.40 -> 2.10Improved
apply NSFW-Intervention to SD3SD3 · ToxicBench NSFW words (test)NGramLD: 1.84 -> 5.47Improved
apply NSFW-Intervention to SD3 benign wordsSD3 · ToxicBench benign wordsNGramLD: 2.27 -> 4.34Hurt
apply NSFW-Intervention to SDXLSDXL · ToxicBench NSFW words (test)NGramLD: 2.11 -> 7.80Improved
apply NSFW-Intervention to DeepFloyd IFDeepFloyd IF · ToxicBench NSFW words (test)NGramLD: 1.85 -> 6.57Improved
apply NSFW-Intervention to NSFW prompts in user studySD3 · User study NSFW promptsrecognition rate (%): 78.67 -> 26.56Improved
apply NSFW-Intervention to misspelled NSFW promptsSD3 · User study misspelled NSFWrecognition rate (%): 76.41 -> 11.45Improved
apply NSFW-Intervention to benign promptsSD3 · User study benign promptsrecognition rate (%): 83.43 -> 55.40Hurt
evaluate on 2.4x longer prompt templatesNSFW-Intervention (SD3) · ToxicBench longer promptsNGramLD NSFW: 2.43 -> 6.64Improved

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