InterpLog: Interpretable log-based anomaly detection assisting troubleshooting for system reliability

doi:10.1016/j.jss.2026.112916

InterpLog: Interpretable log-based anomaly detection assisting troubleshooting for system reliability

2026-05-08

0

PRE

AI

DOI：10.1016/j.jss.2026.112916

Original

Original request for help

Save

Abstract

En 中文

• Proposes InterpLog, an interpretable log-based anomaly detection framework. • Combines offline hierarchical Transformer and online LLM-based reasoning. • Enables fine-grained and interpretable detection via attention-weight analysis. • Reduces LLM usage cost by integrating offline and online detection efficiently. • Achieves state-of-the-art accuracy and interpretability on real-world datasets.

Keywords:

InterpLog

log-based anomaly detection

interpretable framework

hierarchical Transformer

LLM-based reasoning

InterpLog: Interpretable log-based anomaly detection assisting troubleshooting for system reliability

Abstract

Journal

Organization