AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

SAEmnesia: Erasing Concepts in Diffusion Models with Supervised Sparse Autoencoders

Enrico Cassano, Riccardo Renzulli, Marco Nurisso, Mirko Zaffaroni, Alan Perotti, Marco Grangetto

ICML 2026 · Image generation › Fast diffusion sampling · approach: Diffusion models as perception tools · 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
add supervised concept-assignment and decorrelation losses to the SAE (SAEmnesia vs unsupervised SAeUron)SAE on SD v1.5 · UnlearnCanvas objectsaverage of UA, IRA, CRA: 82.29 -> 91.51Improved
supervised one-to-one concept-latent mapping removes the latent-combination searchSAEmnesia · UnlearnCanvas objectsevaluations needed per concept: 210 -> 7Improved
supervised training for feature centralizationSAEmnesia · Flowers concept scoremax feature score: 0.0166 -> 0.0404Improved
apply supervised SAE to sequential unlearning of nine objectsSAEmnesia · UnlearnCanvas sequential, 9 objectsUA: 64 -> 92.4Improved
supervised SAE under white-box UnlearnDiffAtk attackSAEmnesia · UnlearnCanvas objectsUA after attack: 34.20 -> 57.50Improved
supervised SAE under black-box Ring-A-Bell attackSAEmnesia · UnlearnCanvas objectsUA after attack: 79.50 -> 97.00Improved
supervised SAE on nudity removalSAEmnesia · I2P nudity, SD v1.4NudeNet detections: 18 -> 9Improved
uniform steering multipliers across objects with supervised SAESAEmnesia · UnlearnCanvas objectsUA, IRA, CRA across multipliersImproved
using only the top-scoring latent for k-NN classification (ablation)SAEmnesia · 20 UnlearnCanvas objectsk-NN accuracyNo clear effect
compare with SalUn on object unlearning onlySAEmnesia · UnlearnCanvas objectsaverage of UA, IRA, CRA: 94.28 (SalUn) vs 91.51Hurt

Later papers that compared with this method

1 later accepted paper reports SAEmnesia as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 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