# SPG: Sparse-Projected Guides with Sparse Autoencoders for Zero-Shot Anomaly Detection

arXiv: https://arxiv.org/abs/2604.02871 · HTML: https://mlsanity.com/paper/2604.02871/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Sparse autoencoder training and evaluation, one of 13 directions in Mechanistic interpretability.
- Builds on 14 of its 20 references that are in the map.
- First posted on arXiv 3 Apr 2026.

## Placed in

- Mechanistic interpretability / [Sparse autoencoder training and evaluation](https://mlsanity.com/mechinterp/sparse-autoencoder-training-and-evaluation/)

## Builds on (most cited first, 14 of 14)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Sigmoid Loss for Language Image Pre-Training](https://mlsanity.com/paper/2303.15343/) (2023)
- [Towards Total Recall in Industrial Anomaly Detection](https://mlsanity.com/paper/2106.08265/) (2021)
- [Scaling and evaluating sparse autoencoders](https://mlsanity.com/paper/2406.04093/) (2024)
- [SPot-the-Difference Self-Supervised Pre-training for Anomaly Detection and Segmentation](https://mlsanity.com/paper/2207.14315/) (2022)
- [WinCLIP: Zero-/Few-Shot Anomaly Classification and Segmentation](https://mlsanity.com/paper/2303.14814/) (2023)
- [AnomalyCLIP: Object-agnostic Prompt Learning for Zero-shot Anomaly Detection](https://mlsanity.com/paper/2310.18961/) (2023)
- [k-Sparse Autoencoders](https://mlsanity.com/paper/1312.5663/) (2013)
- [AA-CLIP: Enhancing Zero-shot Anomaly Detection via Anomaly-Aware CLIP](https://mlsanity.com/paper/2503.06661/) (2025)
- [AnomalyDINO: Boosting Patch-based Few-shot Anomaly Detection with DINOv2](https://mlsanity.com/paper/2405.14529/) (2024)
- [VCP-CLIP: A visual context prompting model for zero-shot anomaly segmentation](https://mlsanity.com/paper/2407.12276/) (2024)
- [AdaptCLIP: Adapting CLIP for Universal Visual Anomaly Detection](https://mlsanity.com/paper/2505.09926/) (2025)
- [Supervised Anomaly Detection for Complex Industrial Images](https://mlsanity.com/paper/2405.04953/) (2024)

## Built on by (most cited first, 1 of 1)

- [When Does Unsupervised Learning Succeed or Fail? A PoS Perspective on Reconstruction-Based Anomaly Detection](https://mlsanity.com/paper/2609.28832/) (2026)

## Abstract (arXiv)

We study zero-shot anomaly detection and segmentation using frozen foundation model features, where all learnable parameters are trained only on a labeled auxiliary dataset and deployed to unseen target categories without any target-domain adaptation. Existing prompt-based approaches use handcrafted or learned prompt embeddings as reference vectors for normal/anomalous states. We propose Sparse-Projected Guides (SPG), a prompt-free framework that learns sparse guide coefficients in the Sparse Autoencoder (SAE) latent space, which generate normal/anomaly guide vectors via the SAE dictionary. SPG employs a two stage learning strategy on the labeled auxiliary dataset: (i) train an SAE on patch-token features, and (ii) optimize only guide coefficients using auxiliary pixel-level masks while freezing the backbone and SAE. On MVTec AD and VisA under cross-dataset zero-shot settings, SPG achieves competitive image-level detection and strong pixel-level segmentation; with DINOv3, SPG attains the highest pixellevel AUROC among the compared methods. We also report SPG instantiated with OpenCLIP (ViT-L/14@336px) to align the backbone with CLIP-based baselines. Moreover, the learned guide coefficients trace decisions back to a small set of dictionary atoms, revealing category-general and category-specific factors.