AI Paper Atlas

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

Responsible Text-to-Image Diffusion: Interpretable and Linearly Controllable Semantics for Fair and Safe Generation

Sayedmoslem Shokrolahi, Jae-Mo Kang, Il-Min Kim

ICML 2026 · Image generation › Text-to-image diffusion · approach: Concept erasure in diffusion 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
add learned gender concept vectors with equal probabilityPixArt-alpha · WinoBias genderdeviation ratio: 0.82 -> 0.05Improved
add learned concept vectors to FLUXFLUX.1-dev · WinoBias racedeviation ratio: 0.84 -> 0.08Improved
add concept vectors for gender and race jointlyPixArt-alpha · WinoBias gender x racedeviation ratio: 0.73 -> 0.02Improved
add anti-unsafe concept vectorsSDXL · I2P unsafe imagespercentage unsafe: 56 -> 14Improved
add anti-unsafe concept vectorsFLUX.1-dev · I2P unsafe imagespercentage unsafe: 64 -> 20Improved
use concept-alignment loss instead of simple reconstruction loss (ablation)concept vector learning · spurious attribute 'old' leakageCLIP delta old: 1.02 -> 0.61Improved
use concept-alignment loss instead of simple reconstruction loss (ablation)concept vector learning · target attribute 'female'CLIP delta female: 1.28 -> 1.45Improved
restrict intervention to Top-5 heads instead of all heads in selected layers (ablation)PixArt-alpha · WinoBiasCLIP score: 31.62 -> 34.01Improved
extend intervention to all layers instead of selected layers (ablation)PixArt-alpha · WinoBiasCLIP score: 31.62 -> 29.61Hurt
keep Top-5 heads in the intervention instead of excluding them (ablation)PixArt-alpha · WinoBiasdeviation ratio: 0.02 -> 0.01Improved
increase training images beyond about 3,000 (1,500 for SDXL) (ablation)concept vector learning · doctor promptdeviation ratioNo clear effect
vary the target-removed prompt template (five templates) (ablation)concept vector learning · a photo of a doctordeviation ratioNo clear effect

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