CyberXP Agent: AI-Powered Cyber Threat Assessment

The CyberXP Agent is an AI model developed by Abaryan, built upon the meta-llama/Llama-3.2-1B-Instruct base model. It has been fine-tuned using Supervised Fine-Tuning (SFT) on the AlicanKiraz0/Cybersecurity-Dataset-v1 dataset, specifically designed to assist cybersecurity teams with threat assessment and incident response workflows.

Key Capabilities

• Cyber Threat Assessment: Provides AI-powered insights for detecting and responding to cyber threats.
• Incident Response Workflow Integration: Designed with real-world incident response procedures in mind.
• Textual Reasoning: Assesses threats based on textual analysis and retrieval, offering advisory recommendations.

Limitations and Responsible Use

This model is intended for research and educational purposes only in cybersecurity AI. It should not be used as an autonomous incident-response system or for real-world security decisions without expert human review and validation. Limitations include potential gaps in covering emerging threats due to dataset scope and the model's inability to execute or monitor real system activity. Users are advised to maintain human oversight and follow internal security policies.

How to Get Started

An interactive Hugging Face Space is available for easy access and experimentation: https://huggingface.co/spaces/abaryan/CyberXP_AGENT_Llama_3.2