# SAEmnesia: Erasing Concepts in Diffusion Models with Supervised Sparse Autoencoders

arXiv: https://arxiv.org/abs/2509.21379 · HTML: https://mlsanity.com/paper/2509.21379/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Placed in Robustness, privacy and adversarial aspects, one of 14 directions in Diffusion models.
- Builds on 49 of its 61 references that are in the map.
- First posted on arXiv 23 Sep 2025.

## Placed in

- Diffusion models / [Robustness, privacy and adversarial aspects](https://mlsanity.com/diffusion/robustness-privacy-and-adversarial-aspects/)
- LLM safety and security / [Vision-language and text-to-image safety](https://mlsanity.com/safety/vision-language-and-text-to-image-safety/)
- Mechanistic interpretability / [Interpreting and steering diffusion models](https://mlsanity.com/mechinterp/interpreting-and-steering-diffusion-models/)

## Builds on (most cited first, 30 of 49)

- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [Going Deeper with Convolutions](https://mlsanity.com/paper/1409.4842/) (2014)
- [Erasing Concepts from Diffusion Models](https://mlsanity.com/paper/2303.07345/) (2023)
- [Safe Latent Diffusion: Mitigating Inappropriate Degeneration in Diffusion Models](https://mlsanity.com/paper/2211.05105/) (2022)
- [Mechanistic Interpretability for AI Safety -- A Review](https://mlsanity.com/paper/2404.14082/) (2024)
- [Unified Concept Editing in Diffusion Models](https://mlsanity.com/paper/2308.14761/) (2023)
- [SalUn: Empowering Machine Unlearning via Gradient-based Weight Saliency in Both Image Classification and Generation](https://mlsanity.com/paper/2310.12508/) (2023)
- [Diffusion Models already have a Semantic Latent Space](https://mlsanity.com/paper/2210.10960/) (2022)
- [Ablating Concepts in Text-to-Image Diffusion Models](https://mlsanity.com/paper/2303.13516/) (2023)
- [MACE: Mass Concept Erasure in Diffusion Models](https://mlsanity.com/paper/2403.06135/) (2024)
- [Forget-Me-Not: Learning to Forget in Text-to-Image Diffusion Models](https://mlsanity.com/paper/2303.17591/) (2023)
- [Ring-A-Bell! How Reliable are Concept Removal Methods for Diffusion Models?](https://mlsanity.com/paper/2310.10012/) (2023)
- [What the DAAM: Interpreting Stable Diffusion Using Cross Attention](https://mlsanity.com/paper/2210.04885/) (2022)
- [k-Sparse Autoencoders](https://mlsanity.com/paper/1312.5663/) (2013)
- [To Generate or Not? Safety-Driven Unlearned Diffusion Models Are Still Easy To Generate Unsafe Images ... For Now](https://mlsanity.com/paper/2310.11868/) (2023)
- [Selective Amnesia: A Continual Learning Approach to Forgetting in Deep Generative Models](https://mlsanity.com/paper/2305.10120/) (2023)
- [One-Dimensional Adapter to Rule Them All: Concepts, Diffusion Models and Erasing Applications](https://mlsanity.com/paper/2312.16145/) (2023)
- [Defensive Unlearning with Adversarial Training for Robust Concept Erasure in Diffusion Models](https://mlsanity.com/paper/2405.15234/) (2024)
- [Reliable and Efficient Concept Erasure of Text-to-Image Diffusion Models](https://mlsanity.com/paper/2407.12383/) (2024)
- [Understanding the Latent Space of Diffusion Models through the Lens of Riemannian Geometry](https://mlsanity.com/paper/2307.12868/) (2023)
- [Receler: Reliable Concept Erasing of Text-to-Image Diffusion Models via Lightweight Erasers](https://mlsanity.com/paper/2311.17717/) (2023)
- [Interpreting and Editing Vision-Language Representations to Mitigate Hallucinations](https://mlsanity.com/paper/2410.02762/) (2024)
- [EraseAnything: Enabling Concept Erasure in Rectified Flow Transformers](https://mlsanity.com/paper/2412.20413/) (2024)
- [Discover-then-Name: Task-Agnostic Concept Bottlenecks via Automated Concept Discovery](https://mlsanity.com/paper/2407.14499/) (2024)
- [SAeUron: Interpretable Concept Unlearning in Diffusion Models with Sparse Autoencoders](https://mlsanity.com/paper/2501.18052/) (2025)
- [Get What You Want, Not What You Don't: Image Content Suppression for Text-to-Image Diffusion Models](https://mlsanity.com/paper/2402.05375/) (2024)
- [Erasing Undesirable Influence in Diffusion Models](https://mlsanity.com/paper/2401.05779/) (2024)
- [ConceptPrune: Concept Editing in Diffusion Models via Skilled Neuron Pruning](https://mlsanity.com/paper/2405.19237/) (2024)
- [Localizing and Editing Knowledge in Text-to-Image Generative Models](https://mlsanity.com/paper/2310.13730/) (2023)
- [Scissorhands: Scrub Data Influence via Connection Sensitivity in Networks](https://mlsanity.com/paper/2401.06187/) (2024)

## Built on by (most cited first, 10 of 10)

- [ActivationReasoning: Logical Reasoning in Latent Activation Spaces](https://mlsanity.com/paper/2510.18184/) (2025)
- [Disentangled Sparse Representations for Concept-Separated Diffusion Unlearning](https://mlsanity.com/paper/2605.12122/) (2026)
- [Where Concept Erasure Should Occur: Concept-Layer Alignment in Text-to-Video Diffusion Models](https://mlsanity.com/paper/2605.25941/) (2026)
- [Look But Don't Touch with Sparse Autoencoders for Unlearning in Diffusion Models](https://mlsanity.com/paper/2606.31699/) (2026)
- [PEAK: Precise and Persistent Concept Erasure via k-Sparse Autoencoders](https://mlsanity.com/paper/2608.10985/) (2026)
- [Supervised sparse auto-encoders for interpretable and compositional representations](https://mlsanity.com/paper/2602.00924/) (2026)
- [Robust and Generalizable Safety Steering for Text-to-Image Diffusion Transformers](https://mlsanity.com/paper/2605.30049/) (2026)
- [TORINO: Token Reduction via Interpretable Concept Overlap in Vision-Language Models](https://mlsanity.com/paper/2607.04593/) (2026)
- [ClasSAE: Class-Aligned Sparse Autoencoders via Differentiable Feature-Class Affinity](https://mlsanity.com/paper/2610.04020/) (2026)
- [SharedSAE: a unified sparse dictionary across u-net blocks and denoising time for diffusion models](https://mlsanity.com/paper/2610.13997/) (2026)

## Abstract (arXiv)

Concept unlearning in diffusion models is hampered by feature splitting, where concepts are distributed across many latent features, making their removal challenging and computationally expensive. We introduce SAEmnesia, a supervised sparse autoencoder framework that overcomes this by enforcing one-to-one concept-neuron mappings. By systematically labeling concepts during training, our method achieves feature centralization, binding each concept to a single, interpretable neuron. This enables highly targeted and efficient concept erasure. Compared to the state-of-the-art sparse autoencoder-based unlearning approach, SAEmnesia reduces hyperparameter search by 96.67% and achieves a 9.22% improvement on the UnlearnCanvas benchmark for objects. Our method also shows superior scalability in sequential unlearning, improving accuracy by 28.4% when removing nine objects, establishing a step forward for precise and controllable concept erasure. Moreover, SAEmnesia effectively suppresses nudity on the I2P benchmark and remains robust to adversarial attacks. Source code available at https://github.com/EIDOSLAB/SAEmnesia.