cs.LGNov 5, 2026

Evaluating Language Model Features for Zero-Shot Time-Series Anomaly Detection Under Isolation Assumption

Authors: Romain Hermary, Nesryne Mejri, Djamila Aouada

Organizations: University of Luxembourg

Abstract

Time-Series Anomaly Detection (TSAD) is a fundamentally challenging task: it is unsupervised, requires high local precision capabilities, and suffers from limited labelled data. Recent work has explored adapting Large Language Models (LLMs) to time-series via cross-modal transfer. However, existing approaches are poorly suited to Multivariate TSAD (MTSAD), as tokenisation often obscures fine-grained temporal and variables structures, and most methods optimise auxiliary objectives such as forecasting or reconstruction. We posit that these surrogate objectives become unreliable as model capacity increases. Highly expressive models risk reconstructing or forecasting anomalies, undermining the foundations of both strategies. Instead, we propose to use LLMs purely as feature extractors and directly analyse their induced feature spaces. In this work, investigate the ability of LLMs to provide useful representations for MTSADs in a zero-shot setting through classification. To this end, we study the separability of the extracted features, under the assumption that anomalous features should be isolated from normal ones. We analyse their temporal and contextual discriminative ability using simple step-wise tokenisation schemes, with one time-step per token, which preserve fine-grained local information. Their usability for MTSAD is assessed using an unsupervised Isolation Forest, enabling a direct evaluation of the resulting representation quality. We further investigate whether the observed performance can be attributed to the language-related pre-training of the LLMs or arises independently of it. Our results provide empirical insight into the ability of LLMs to represent multivariate time-series for anomaly detection and highlight broader applications of cross-modal foundation models.

Figures & tables

Appendix figures & tables31 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Exploring Zero-Shot Foundation Models for Multivariate Time Series Anomaly Detection

    Jul 14, 2026Martin Uray, Saverio Messineo, Roland Kwitt +1Industrial Anomaly DetectionZero-Shot Anomaly Detection

  2. Structured Frequency-Domain Evidence for LLM-Based Time-Series Anomaly Detection

    Aug 25, 2026Jungwook Seo, Sangwon Son, Minjeong Kim +3LLM-Based Time Series Anomaly DetectionTime-Series Anomaly Detection

  3. LEARN-TS: LLM-Enhanced Alignment and Reconstruction with Normality Guidance for Multivariate Time-Series Anomaly Detection

    Sep 30, 2026Jahyeob Koo, Kio Yun, Byoungmo Koo +1LLM-Based Time Series Anomaly DetectionAnomaly Localization