# HiProbe-VAD: Video Anomaly Detection via Hidden States Probing in Tuning-Free Multimodal LLMs

arXiv: https://arxiv.org/abs/2507.17394 · HTML: https://mlsanity.com/paper/2507.17394/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Builds on 52 of its 71 references that are in the map.
- First posted on arXiv 23 Jul 2025.

## Builds on (most cited first, 30 of 52)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [LLaVA-OneVision: Easy Visual Task Transfer](https://mlsanity.com/paper/2408.03326/) (2024)
- [MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models](https://mlsanity.com/paper/2304.10592/) (2023)
- [InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Tasks](https://mlsanity.com/paper/2312.14238/) (2023)
- [Video-LLaMA: An Instruction-tuned Audio-Visual Language Model for Video Understanding](https://mlsanity.com/paper/2306.02858/) (2023)
- [Understanding intermediate layers using linear classifier probes](https://mlsanity.com/paper/1610.01644/) (2016)
- [ImageBind: One Embedding Space To Bind Them All](https://mlsanity.com/paper/2305.05665/) (2023)
- [How Far Are We to GPT-4V? Closing the Gap to Commercial Multimodal Models with Open-Source Suites](https://mlsanity.com/paper/2404.16821/) (2024)
- [Real-world Anomaly Detection in Surveillance Videos](https://mlsanity.com/paper/1801.04264/) (2018)
- [DeepSeek-VL2: Mixture-of-Experts Vision-Language Models for Advanced Multimodal Understanding](https://mlsanity.com/paper/2412.10302/) (2024)
- [Future Frame Prediction for Anomaly Detection -- A New Baseline](https://mlsanity.com/paper/1712.09867/) (2017)
- [Layer by Layer: Uncovering Hidden Representations in Language Models](https://mlsanity.com/paper/2502.02013/) (2025)
- [Learning Temporal Regularity in Video Sequences](https://mlsanity.com/paper/1604.04574/) (2016)
- [LLMs Know More Than They Show: On the Intrinsic Representation of LLM Hallucinations](https://mlsanity.com/paper/2410.02707/) (2024)
- [Not only Look, but also Listen: Learning Multimodal Violence Detection under Weak Supervision](https://mlsanity.com/paper/2007.04687/) (2020)
- [Weakly-supervised Video Anomaly Detection with Robust Temporal Feature Magnitude Learning](https://mlsanity.com/paper/2101.10030/) (2021)
- [MIST: Multiple Instance Self-Training Framework for Video Anomaly Detection](https://mlsanity.com/paper/2104.01633/) (2021)
- [A Hybrid Video Anomaly Detection Framework via Memory-Augmented Flow Reconstruction and Flow-Guided Frame Prediction](https://mlsanity.com/paper/2108.06852/) (2021)
- [Generative Cooperative Learning for Unsupervised Video Anomaly Detection](https://mlsanity.com/paper/2203.03962/) (2022)
- [Not All Layers of LLMs Are Necessary During Inference](https://mlsanity.com/paper/2403.02181/) (2024)
- [Dual Memory Units with Uncertainty Regulation for Weakly Supervised Video Anomaly Detection](https://mlsanity.com/paper/2302.05160/) (2023)
- [MGFN: Magnitude-Contrastive Glance-and-Focus Network for Weakly-Supervised Video Anomaly Detection](https://mlsanity.com/paper/2211.15098/) (2022)
- [Anomaly Detection in Autonomous Driving: A Survey](https://mlsanity.com/paper/2204.07974/) (2022)
- [Harnessing Large Language Models for Training-free Video Anomaly Detection](https://mlsanity.com/paper/2404.01014/) (2024)
- [Learning Normal Dynamics in Videos with Meta Prototype Network](https://mlsanity.com/paper/2104.06689/) (2021)
- [Holmes-VAU: Towards Long-term Video Anomaly Understanding at Any Granularity](https://mlsanity.com/paper/2412.06171/) (2024)

## Built on by (most cited first, 9 of 9)

- [Unlocking Vision-Language Models for Video Anomaly Detection via Fine-Grained Prompting](https://mlsanity.com/paper/2510.02155/) (2025)
- [Steering and Rectifying Latent Representation Manifolds in Frozen Multi-modal LLMs for Video Anomaly Detection](https://mlsanity.com/paper/2602.24021/) (2026)
- [SphereVAD: Training-Free Video Anomaly Detection via Geodesic Inference on the Unit Hypersphere](https://mlsanity.com/paper/2605.08003/) (2026)
- [Fine-Grained Human Pose Editing Assessment via Layer-Selective MLLMs](https://mlsanity.com/paper/2601.10369/) (2026)
- [COPRA: Conditional Parameter Adaptation with Reinforcement Learning for Video Anomaly Detection](https://mlsanity.com/paper/2605.15325/) (2026)
- [Structured Evidence Selection for Weakly Supervised Video Anomaly Detection](https://mlsanity.com/paper/2607.10298/) (2026)
- [Beyond Hazard Resemblance: Contrastive Event Adjudication for Training-Free Video Anomaly Detection](https://mlsanity.com/paper/2608.09908/) (2026)
- [HUGIN: Enhancing Vision-Language Planning for Autonomous Logistics Sorting](https://mlsanity.com/paper/2608.11692/) (2026)
- [RoMod: Temporal Routing Modulation via Mixture-of-Experts for Video Anomaly Detection](https://mlsanity.com/paper/2610.05131/) (2026)

## Abstract (arXiv)

Video Anomaly Detection (VAD) aims to identify and locate deviations from normal patterns in video sequences. Traditional methods often struggle with substantial computational demands and a reliance on extensive labeled datasets, thereby restricting their practical applicability. To address these constraints, we propose HiProbe-VAD, a novel framework that leverages pre-trained Multimodal Large Language Models (MLLMs) for VAD without requiring fine-tuning. In this paper, we discover that the intermediate hidden states of MLLMs contain information-rich representations, exhibiting higher sensitivity and linear separability for anomalies compared to the output layer. To capitalize on this, we propose a Dynamic Layer Saliency Probing (DLSP) mechanism that intelligently identifies and extracts the most informative hidden states from the optimal intermediate layer during the MLLMs reasoning. Then a lightweight anomaly scorer and temporal localization module efficiently detects anomalies using these extracted hidden states and finally generate explanations. Experiments on the UCF-Crime and XD-Violence datasets demonstrate that HiProbe-VAD outperforms existing training-free and most traditional approaches. Furthermore, our framework exhibits remarkable cross-model generalization capabilities in different MLLMs without any tuning, unlocking the potential of pre-trained MLLMs for video anomaly detection and paving the way for more practical and scalable solutions.