AnomalyGen is the first automated framework that leverages large language models to generate high-quality, semantically meaningful log sequences specifically for anomaly detection tasks.

• Addresses critical dataset scarcity in log-based anomaly detection research
• Overcomes limitations of existing datasets: incomplete coverage, artificial patterns, and poor semantics
• Generates realistic anomaly scenarios with complete semantic context
• Enables more robust and reliable security monitoring systems

This research significantly advances security capabilities by providing high-fidelity synthetic data for training detection systems, potentially improving threat detection accuracy in production environments.

AnomalyGen: An Automated Semantic Log Sequence Generation Framework with LLM for Anomaly Detection