Latent Anomaly Knowledge Excavation: Unveiling Sparse Sensitive Neurons in Vision-Language Models

• URL: http://arxiv.org/abs/2604.07802v1
• Date: Thu, 09 Apr 2026 04:54:25 GMT
• Title: Latent Anomaly Knowledge Excavation: Unveiling Sparse Sensitive Neurons in Vision-Language Models
• Abstract summary: We argue that anomaly knowledge is intrinsically embedded within pre-trained models but remains latent and under-activated.<n>We propose latent anomaly knowledge excavation (LAKE), a training-free framework that identifies and elicits these critical neuronal signals.<n>Our work advocates for a paradigm shift: redefining anomaly detection as the targeted activation of latent pre-trained knowledge.
• Score: 40.6282934670186
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large-scale vision-language models (VLMs) exhibit remarkable zero-shot capabilities, yet the internal mechanisms driving their anomaly detection (AD) performance remain poorly understood. Current methods predominantly treat VLMs as black-box feature extractors, assuming that anomaly-specific knowledge must be acquired through external adapters or memory banks. In this paper, we challenge this assumption by arguing that anomaly knowledge is intrinsically embedded within pre-trained models but remains latent and under-activated. We hypothesize that this knowledge is concentrated within a sparse subset of anomaly-sensitive neurons. To validate this, we propose latent anomaly knowledge excavation (LAKE), a training-free framework that identifies and elicits these critical neuronal signals using only a minimal set of normal samples. By isolating these sensitive neurons, LAKE constructs a highly compact normality representation that integrates visual structural deviations with cross-modal semantic activations. Extensive experiments on industrial AD benchmarks demonstrate that LAKE achieves state-of-the-art performance while providing intrinsic, neuron-level interpretability. Ultimately, our work advocates for a paradigm shift: redefining anomaly detection as the targeted activation of latent pre-trained knowledge rather than the acquisition of a downstream task.

Related papers

• V-FIND: Revealing the Intrinsic Forgery Knowledge Encoded in Video Forgery Detectors [64.12551408538918]
  We find that forgery-discriminative knowledge is not uniformly distributed across the full representation space.<n>We propose a video forgery-intrinsic neuron discovery framework (V-FIND)<n>V-FIND localizes critical layers that exhibit pronounced discrepancies between real and forged videos, and then identifies latent anchor neurons that consistently carry forgery-discriminative signals.
  arXiv  Detail & Related papers  (2026-08-04T01:41:40Z)
• Neuro-Symbolic Process Anomaly Detection [2.8547732086436306]
  We propose a neuro-symbolic approach that integrates domain knowledge into neural anomaly detection.<n>We encode Declare constraints as soft logical guiderails within the learning process to distinguish between anomalous and rare but conformant behavior.<n>Our approach improves F1 scores even when as few as 10 conformant traces exist.
  arXiv  Detail & Related papers  (2026-03-27T14:30:30Z)
• Anatomy of a Lie: A Multi-Stage Diagnostic Framework for Tracing Hallucinations in Vision-Language Models [62.932580559941414]
  Vision-Language Models (VLMs) frequently "hallucinate" - generate plausible yet factually incorrect statements.<n>We propose a new paradigm for diagnosing hallucinations, recasting them from static output errors into dynamic pathologies of a model's computational cognition.
  arXiv  Detail & Related papers  (2026-03-16T17:20:38Z)
• DNA: Uncovering Universal Latent Forgery Knowledge [39.19379714306656]
  forgery detection capability is already encoded within pre-trained models.<n>DNA framework employs a coarse-to-fine excavation mechanism.<n>Hifi-Gen is a high-fidelity synthetic benchmark built upon the very latest models.
  arXiv  Detail & Related papers  (2026-01-30T03:48:30Z)
• NOBLE -- Neural Operator with Biologically-informed Latent Embeddings to Capture Experimental Variability in Biological Neuron Models [63.592664795493725]
  NOBLE is a neural operator framework that learns a mapping from a continuous frequency-modulated embedding of interpretable neuron features to the somatic voltage response induced by current injection.<n>It predicts distributions of neural dynamics accounting for the intrinsic experimental variability.<n>NOBLE is the first scaled-up deep learning framework that validates its generalization with real experimental data.
  arXiv  Detail & Related papers  (2025-06-05T01:01:18Z)
• Video Anomaly Detection via Spatio-Temporal Pseudo-Anomaly Generation : A Unified Approach [49.995833831087175]
  This work proposes a novel method for generating generic Video-temporal PAs by inpainting a masked out region of an image. In addition, we present a simple unified framework to detect real-world anomalies under the OCC setting. Our method performs on par with other existing state-of-the-art PAs generation and reconstruction based methods under the OCC setting.
  arXiv  Detail & Related papers  (2023-11-27T13:14:06Z)
• Understanding the Feature Norm for Out-of-Distribution Detection [16.40441221109391]
  A neural network trained on a classification dataset often exhibits a higher vector norm of hidden layer features for in-distribution (ID) samples. We propose a novel negative-aware norm (NAN) that can capture both the activation and deactivation tendencies of hidden layer neurons.
  arXiv  Detail & Related papers  (2023-10-09T00:17:20Z)
• Neuro-symbolic model for cantilever beams damage detection [0.0]
  We propose a neuro-symbolic model for the detection of damages in cantilever beams based on a novel cognitive architecture. The hybrid discriminative model is introduced under the name Logic Convolutional Neural Regressor.
  arXiv  Detail & Related papers  (2023-05-04T13:12:39Z)
• Generalizable Industrial Visual Anomaly Detection with Self-Induction Vision Transformer [5.116033262865781]
  We propose a self-induction vision Transformer (SIVT) for unsupervised generalizable industrial visual anomaly detection and localization. The proposed SIVT first extracts discriminatory features from pre-trained CNN as property descriptors, then reconstructs the extracted features in a self-supervisory fashion. The results reveal that the proposed method can advance state-of-the-art detection performance with an improvement of 2.8-6.3 in AUROC, and 3.3-7.6 in AP.
  arXiv  Detail & Related papers  (2022-11-22T14:56:12Z)
• Self-Supervised Masked Convolutional Transformer Block for Anomaly Detection [122.4894940892536]
  We present a novel self-supervised masked convolutional transformer block (SSMCTB) that comprises the reconstruction-based functionality at a core architectural level. In this work, we extend our previous self-supervised predictive convolutional attentive block (SSPCAB) with a 3D masked convolutional layer, a transformer for channel-wise attention, as well as a novel self-supervised objective based on Huber loss.
  arXiv  Detail & Related papers  (2022-09-25T04:56:10Z)
• A Robust Backpropagation-Free Framework for Images [47.97322346441165]
  We present an error kernel driven activation alignment algorithm for image data. EKDAA accomplishes through the introduction of locally derived error transmission kernels and error maps. Results are presented for an EKDAA trained CNN that employs a non-differentiable activation function.
  arXiv  Detail & Related papers  (2022-06-03T21:14:10Z)

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.