An Integrated Honeypot and LLM-Based Framework for near Real-Time Detection and Behavioral Analysis of Malicious Activities
Integration of deterministic preprocessing with LLM-based reasoning enables the transformation of raw honeypot logs into structured and actionable cybersecurity intelligence, reducing analyst workload while improving the explainability and reliability of intrusion analysis in near-real-time environments.