LLM Log Anomaly Detection: Study Reveals Performance, Efficiency, and Robustness Insights

A new empirical study published on arXiv investigates the performance, efficiency, and robustness of Large Language Models (LLMs) when applied to log anomaly detection. The research systematically analyzes various adaptation strategies, model architectures, parameter scales, and quantization settings across three public log datasets. Key findings indicate that adaptation strategies significantly impact detection effectiveness, while model scaling offers varied gains depending on the dataset. The study also highlights that models with similar accuracy can have different computational costs, and low-bit quantization generally maintains detection performance. AI

IMPACT Provides practical insights for deploying LLMs in log anomaly detection, focusing on efficiency and robustness beyond simple accuracy.

RANK_REASON The item is an academic paper published on arXiv detailing an empirical study of LLM performance on a specific task. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• alphaXiv
• arXiv
• Bibliographic Explorer
• CatalyzeX
• Connected Papers
• CORE Recommender
• DagsHub
• Gotit.pub
• Hugging Face
• IArxiv Recommender
• Influence Flower
• Large Language Models
• Litmaps
• ScienceCast
• scite Smart Citations

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →