hypnonyx/traffico
hypnonyx/traffico is a 0.3 billion parameter language model, fine-tuned from Google's Gemma 2.7B, specialized in cybersecurity. It analyzes TCP/IP network traffic to detect cyberattacks and maps network flow patterns to the MITRE ATT&CK framework. This model excels at classifying network flows as normal or malicious and providing ATT&CK-mapped threat classifications for security teams.
Loading preview...
Overview
hypnonyx/traffico is a specialized language model, fine-tuned from Google's Gemma 2.7B, designed for cybersecurity applications. It focuses on analyzing TCP/IP network traffic to identify cyberattacks and correlate them with the MITRE ATT&CK framework. The model was trained using a synthetic dataset derived from real-world network traffic (CIC-IDS2017 + UNSW-NB15) and enriched with MITRE ATT&CK techniques, enabling it to classify network flows and provide threat classifications.
Key Capabilities
- Network Intrusion Detection: Classifies network flows as benign or malicious in real-time.
- Threat Intelligence: Maps detected attacks to specific MITRE ATT&CK techniques and tactics.
- Security Monitoring: Analyzes TCP/IP flows from network sensors and IDS systems.
- Incident Response: Helps understand adversary behavior patterns from network telemetry.
- ATT&CK Mapping: Provides classifications for various attack types, including DoS, DDoS, PortScan, Brute Force, Infiltration, Botnet, and Web attacks, linking them to MITRE ATT&CK categories like Reconnaissance, Initial Access, Lateral Movement, and Impact.
Good For
- Security teams needing to understand adversary tactics from network behavior.
- Researchers studying attack-to-technique mappings in security datasets.
- Integrating into existing security monitoring and incident response workflows.
Limitations
- The model is not exhaustive and may not cover all possible or novel adversary behaviors.
- It is provided "as is" and requires validation against specific security requirements.
- Comprehensive defensive coverage is not guaranteed solely by using this model.