Sparsity as a Key: Unlocking New Insights from Latent Structures for Out-of-Distribution Detection

• URL: http://arxiv.org/abs/2604.26409v1
• Date: Wed, 29 Apr 2026 08:23:38 GMT
• Title: Sparsity as a Key: Unlocking New Insights from Latent Structures for Out-of-Distribution Detection
• Abstract summary: We present the first application of SAEs to the ViT [] token for out-of-distribution (OOD) detection.<n>Our findings demonstrate that the sparse, disentangled features revealed by SAEs can serve as a powerful, interpretable tool for robust OOD detection in vision models.
• Score: 4.496267803613718
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Sparse Autoencoders (SAEs) have demonstrated significant success in interpreting Large Language Models (LLMs) by decomposing dense representations into sparse, semantic components. However, their potential for analyzing Vision Transformers (ViTs) remains largely under-explored. In this work, we present the first application of SAEs to the ViT [CLS] token for out-of-distribution (OOD) detection, addressing the limitation of existing methods that rely on entangled feature representations. We propose a novel framework utilizing a Top-k SAE to disentangle the dense [CLS] features into a structured latent space. Through this analysis, we reveal that in-distribution (ID) data exhibits consistent, class-specific activation patterns, which we formalize as Class Activation Profiles (CAPs). Our study uncovers a key structural invariant: while ID samples preserve a stable pattern within CAPs, OOD samples systematically disrupt this structure. Leveraging this insight, we introduce a scoring function based on the divergence of core energy profiles to quantify the deviation from ideal activation profiles. Our method achieves strong results on the FPR95 metric, critical for safety-sensitive applications across multiple benchmarks, while also achieving competitive AUROC. Overall, our findings demonstrate that the sparse, disentangled features revealed by SAEs can serve as a powerful, interpretable tool for robust OOD detection in vision models.

Related papers

• Latent Performance Profiling of Large Language Models [47.009623327601226]
  We introduce Latent Performance Profiling (LPP) -- a framework that derives task-agnostic diagnostics from hidden activations and output distributions.<n>Unlike static accuracy scores, LPP provides stable, architecture-sensitive signatures across models of similar size.<n>We show that models with similar benchmark scores can exhibit contrasting latent profiles, such as differences in entropy or adaptability.
  arXiv  Detail & Related papers  (2026-05-28T14:41:26Z)
• StepVAR: Structure-Texture Guided Pruning for Visual Autoregressive Models [98.72926158261937]
  We propose a training-free token pruning framework for Visual AutoRegressive models.<n>We employ a lightweight high-pass filter to capture local texture details, while leveraging Principal Component Analysis (PCA) to preserve global structural information.<n>To maintain valid next-scale prediction under sparse tokens, we introduce a nearest neighbor feature propagation strategy.
  arXiv  Detail & Related papers  (2026-03-02T11:35:05Z)
• Matching-Based Few-Shot Semantic Segmentation Models Are Interpretable by Design [8.993770750003673]
  Few-Shot Semantic (FSS) models achieve strong performance in segmenting novel classes with minimal labeled examples.<n>This paper introduces the first dedicated method for interpreting matching-based FSS models.<n>Our Affinity Explainer approach extracts attribution maps that highlight which pixels in support images contribute most to query segmentation predictions.
  arXiv  Detail & Related papers  (2025-11-22T19:22:10Z)
• Analysis of Variational Sparse Autoencoders [1.675385127117872]
  We investigate whether incorporating variational methods into SAE architectures can improve feature organization and interpretability.<n>We introduce the Variational Sparse Autoencoder (vSAE), which replaces deterministic ReLU gating with sampling from learned Gaussian posteriors.<n>Our findings suggest that naive application of variational methods to SAEs does not improve feature organization or interpretability.
  arXiv  Detail & Related papers  (2025-09-26T23:09:56Z)
• Dense SAE Latents Are Features, Not Bugs [86.50389855919292]
  We show that dense latents serve functional roles in language model computation.<n>We identify classes tied to position tracking, context binding, entropy regulation, letter-specific output signals, part-of-speech, and principal component reconstruction.
  arXiv  Detail & Related papers  (2025-06-18T17:59:35Z)
• "Principal Components" Enable A New Language of Images [79.45806370905775]
  We introduce a novel visual tokenization framework that embeds a provable PCA-like structure into the latent token space.<n>Our approach achieves state-of-the-art reconstruction performance and enables better interpretability to align with the human vision system.
  arXiv  Detail & Related papers  (2025-03-11T17:59:41Z)
• Clustering Properties of Self-Supervised Learning [14.756786256090704]
  Self-supervised learning (SSL) methods have proven remarkably effective at capturing semantically rich representations with strong clustering properties.<n>We propose a novel positive-feedback SSL method, termed Representation Self-Assignment (ReSA), which leverages the model's clustering properties to promote learning in a self-guided manner.
  arXiv  Detail & Related papers  (2025-01-30T16:05:35Z)
• TopoFR: A Closer Look at Topology Alignment on Face Recognition [58.45515807380505]
  We propose TopoFR, a novel FR model that leverages a topological structure alignment strategy called PTSA and a hard sample mining strategy named SDE.<n> PTSA uses persistent homology to align the topological structures of the input and latent spaces, effectively preserving the structure information and improving the generalization performance of FR model.<n> Experimental results on popular face benchmarks demonstrate the superiority of our TopoFR over the state-of-the-art methods.
  arXiv  Detail & Related papers  (2024-10-14T14:58:30Z)
• Latent Causal Probing: A Formal Perspective on Probing with Causal Models of Data [3.376269351435396]
  We develop a formal perspective on probing using structural causal models (SCM) We extend a recent study of LMs in the context of a synthetic grid-world navigation task. Our techniques provide robust empirical evidence for the ability of LMs to induce the latent concepts underlying text.
  arXiv  Detail & Related papers  (2024-07-18T17:59:27Z)
• Be Your Own Neighborhood: Detecting Adversarial Example by the Neighborhood Relations Built on Self-Supervised Learning [64.78972193105443]
  This paper presents a novel AE detection framework, named trustworthy for predictions. performs the detection by distinguishing the AE's abnormal relation with its augmented versions. An off-the-shelf Self-Supervised Learning (SSL) model is used to extract the representation and predict the label.
  arXiv  Detail & Related papers  (2022-08-31T08:18:44Z)
• Structure-Aware Feature Generation for Zero-Shot Learning [108.76968151682621]
  We introduce a novel structure-aware feature generation scheme, termed as SA-GAN, to account for the topological structure in learning both the latent space and the generative networks. Our method significantly enhances the generalization capability on unseen-classes and consequently improve the classification performance.
  arXiv  Detail & Related papers  (2021-08-16T11:52:08Z)

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.