Look But Don't Touch with Sparse Autoencoders for Unlearning in Diffusion Models

• URL: http://arxiv.org/abs/2606.31699v2
• Date: Wed, 01 Jul 2026 18:42:22 GMT
• Title: Look But Don't Touch with Sparse Autoencoders for Unlearning in Diffusion Models
• Abstract summary: Sparse autoencoders (SAEs) have been proposed as interpretable tools for concept-level manipulation.<n>We show that while SAEs reliably detect and localize semantic concepts within diffusion model activations, direct intervention in their latent space frequently induces out-of-distribution activations.<n>We use SAE activations purely as semantic detectors to identify image regions containing the target object, and replace those patch embeddings with the ones that do not contain it.
• Score: 11.366541829206199
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Sparse autoencoders (SAEs) have recently been proposed as interpretable tools for concept-level manipulation, under the assumption that isolated features can serve as controllable intervention points. In this work, we systematically evaluate this assumption in the context of object erasure and steering in diffusion models. We show that while SAEs reliably detect and localize semantic concepts within diffusion model activations, direct intervention in their latent space frequently induces out-of-distribution activations, resulting in severe visual artifacts. To disentangle detection from intervention, we use SAE activations purely as semantic detectors to identify image regions containing the target object, and replace those patch embeddings with the ones that do not contain it. This detection-based replacement preserves the diffusion model's activation statistics and produces significantly cleaner erasure results than latent steering. Our findings reveal a fundamental gap between concept detection and concept intervention in diffusion models: monosemantic or sparse features are not inherently suitable as control knobs for steering. These results position SAEs as powerful interpretability tools for analyzing generative models, but highlight important limitations when used for direct manipulation, such as unlearning.

Related papers

• Actionable Hallucination Detection: Translating Latent Uncertainty into Agentic Critique [1.1172382217477128]
  Large Language Models (LLMs) deployed as AI agents frequently exhibit user specification-grounding failures.<n>Existing detection methods fail to provide actionable, real-time correction as they either do not localize the hallucinations, or incur prohibitive inference latency.<n>We introduce the Latent Critic, a lightweight low-rank adapter (LoRA) that operates concurrently with a frozen base LLM's generation.
  arXiv  Detail & Related papers  (2026-08-11T03:26:16Z)
• Scaling Pretrained Representations Enables Label-Free Out-of-Distribution Detection Without Fine-Tuning [5.008779702997125]
  We show that modern pretrained models encode sufficient geometric structure for accurate label-free out-of-distribution detection.<n>Our results suggest that label-free OOD detection depends strongly on the geometry exposed by frozen pretrained backbones.
  arXiv  Detail & Related papers  (2026-05-07T03:45:52Z)
• CASL: Concept-Aligned Sparse Latents for Interpreting Diffusion Models [45.90361318326864]
  Internal activations of diffusion models encode rich semantic information, but interpreting such representations remains challenging.<n>We introduce CASL (Concept-Aligned Sparse Latents), a supervised framework that aligns sparse latent dimensions of diffusion models with semantic concepts.<n>Unlike editing methods, CASL-Steer is used solely as a causal probe to reveal how concept-aligned latents influence generated content.
  arXiv  Detail & Related papers  (2026-01-21T20:14:17Z)
• Did Models Sufficient Learn? Attribution-Guided Training via Subset-Selected Counterfactual Augmentation [61.248535801314375]
  Subset-Selected Counterfactual Augmentation (SS-CA)<n>We develop Counterfactual LIMA to identify minimal spatial region sets whose removal can selectively alter model predictions.<n>Experiments show that SS-CA improves generalization on in-distribution (ID) test data and achieves superior performance on out-of-distribution (OOD) benchmarks.
  arXiv  Detail & Related papers  (2025-11-15T08:39:22Z)
• LLM-Assisted Semantic Guidance for Sparsely Annotated Remote Sensing Object Detection [25.9348571356454]
  LLM-assisted semantic guidance framework tailored for sparsely annotated remote sensing object detection.<n>Dense Pseudo-Label Assignment mechanism adaptively assigns pseudo-labels for both unlabeled and sparsely labeled data.<n> Adaptive Hard-Negative Reweighting Module to stabilize the supervised learning branch.
  arXiv  Detail & Related papers  (2025-09-21T08:05:43Z)
• Orthogonal Subspace Decomposition for Generalizable AI-Generated Image Detection [58.87142367781417]
  A naively trained detector tends to favor overfitting to the limited and monotonous fake patterns, causing the feature space to become highly constrained and low-ranked.<n>One potential remedy is incorporating the pre-trained knowledge within the vision foundation models to expand the feature space.<n>By freezing the principal components and adapting only the remained components, we preserve the pre-trained knowledge while learning fake patterns.
  arXiv  Detail & Related papers  (2024-11-23T19:10:32Z)
• Unsupervised Model Diagnosis [49.36194740479798]
  This paper proposes Unsupervised Model Diagnosis (UMO) to produce semantic counterfactual explanations without any user guidance. Our approach identifies and visualizes changes in semantics, and then matches these changes to attributes from wide-ranging text sources.
  arXiv  Detail & Related papers  (2024-10-08T17:59:03Z)
• Diffusion-Based Particle-DETR for BEV Perception [94.88305708174796]
  Bird-Eye-View (BEV) is one of the most widely-used scene representations for visual perception in Autonomous Vehicles (AVs) Recent diffusion-based methods offer a promising approach to uncertainty modeling for visual perception but fail to effectively detect small objects in the large coverage of the BEV. Here, we address this problem by combining the diffusion paradigm with current state-of-the-art 3D object detectors in BEV.
  arXiv  Detail & Related papers  (2023-12-18T09:52:14Z)
• Weakly-supervised Contrastive Learning for Unsupervised Object Discovery [52.696041556640516]
  Unsupervised object discovery is promising due to its ability to discover objects in a generic manner. We design a semantic-guided self-supervised learning model to extract high-level semantic features from images. We introduce Principal Component Analysis (PCA) to localize object regions.
  arXiv  Detail & Related papers  (2023-07-07T04:03:48Z)
• Feature Alignment by Uncertainty and Self-Training for Source-Free Unsupervised Domain Adaptation [1.6498361958317636]
  Most unsupervised domain adaptation (UDA) methods assume that labeled source images are available during model adaptation. We propose a source-free UDA method that uses only a pre-trained source model and unlabeled target images. Our method captures the aleatoric uncertainty by incorporating data augmentation and trains the feature generator with two consistency objectives.
  arXiv  Detail & Related papers  (2022-08-31T14:28:36Z)
• DAAIN: Detection of Anomalous and Adversarial Input using Normalizing Flows [52.31831255787147]
  We introduce a novel technique, DAAIN, to detect out-of-distribution (OOD) inputs and adversarial attacks (AA) Our approach monitors the inner workings of a neural network and learns a density estimator of the activation distribution. Our model can be trained on a single GPU making it compute efficient and deployable without requiring specialized accelerators.
  arXiv  Detail & Related papers  (2021-05-30T22:07:13Z)
• Unsupervised Anomaly Detection with Adversarial Mirrored AutoEncoders [51.691585766702744]
  We propose a variant of Adversarial Autoencoder which uses a mirrored Wasserstein loss in the discriminator to enforce better semantic-level reconstruction. We put forward an alternative measure of anomaly score to replace the reconstruction-based metric. Our method outperforms the current state-of-the-art methods for anomaly detection on several OOD detection benchmarks.
  arXiv  Detail & Related papers  (2020-03-24T08:26:58Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.