AI Paper Atlas

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

Obliviate: Erasing Concepts from Autoregressive Image Generation Models

Hossein Shakibania, Jonas Henry Grebe, Tobias Braun, Ege Aktemur, Saleh Aslani, Mehmet Görkem Yiğit, Marcus Rohrbach

ECCV 2026 · Image generation › Text-to-image diffusion · approach: Concept erasure in diffusion models · 2 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
use KL distribution matching instead of cross entropy token supervision (ablation)Obliviate (Liquid) · Explicit content RABCDR: 5.26 -> 3.15Improved
use KL distribution matching instead of cross entropy token supervision (ablation)Obliviate (Liquid) · MJHQ-30K utilityFID: 23.24 -> 15.41Improved
use comprehensive concept prompt instead of simple prompt for explicit content (ablation)Obliviate (Emu3-Gen) · Explicit contentCDR: 27.78 -> 4.76Improved
use comprehensive concept prompt instead of simple prompt for explicit content (ablation)Obliviate (Liquid) · Explicit contentCDR: 6.36 -> 3.74Improved
use comprehensive descriptive prompt instead of simple brand name (ablation)Obliviate (Emu3-Gen) · Coca-Cola unbrandingCDR: 4.14 -> 91.19Hurt
use comprehensive descriptive prompt instead of simple brand name (ablation)Obliviate (Janus-Pro) · Coca-Cola unbrandingCDR: 0.18 -> 19.60Hurt
increase negative guidance scale eta from 1.0 to 2.0 (ablation)Obliviate (Emu3-Gen) · Explicit content RABCDR: 11.57 -> 5.26Improved
increase negative guidance scale eta from 1.0 to 2.0 (ablation)Obliviate (Emu3-Gen) · MJHQ-30K utilityFID: 15.33 -> 25.62Hurt
increase negative guidance scale eta from 1.0 to 10.0 (ablation)Obliviate (Janus-Pro) · Explicit content T2I-RiskyPromptCDR: 33.48 -> 18.11Improved
Obliviate versus supervised fine-tuning for brand erasureSFT (Janus-Pro) · Coca-Cola unbrandingCDR: 32.19 -> 0.18Improved
Obliviate versus EAR on Janus-Pro explicit contentEAR · MJHQ-30K utilityFID: 31.63 -> 12.31Improved
Obliviate versus negative prompting for gory contentOriginal model (Liquid) · Gory content T2I-RiskyPromptCDR: 87.65 -> 39.06Improved

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