Apppublic

AYI-NEDJIMI/CyberSec-Leaderboard

sourceHugging Faceapache-2.0updated 7mo agoView on Hugging Face

0likes

App README

CyberSec Leaderboard

Evaluate and rank Large Language Models on real-world cybersecurity knowledge using the CyberSec-Bench benchmark.

Features

—Leaderboard: Ranked table of models with scores across 5 cybersecurity categories
—Benchmark Info: Details about the CyberSec-Bench dataset (200 questions, 5 categories, bilingual EN/FR)
—Submit Model: Submit your model for evaluation on CyberSec-Bench
—About: Methodology and author information

Benchmark Categories

Category	Questions	Description
Compliance	40	ISO 27001, GDPR/RGPD, NIS2, DORA, AI Act
Offensive Security	40	MITRE ATT&CK, OWASP, penetration testing
Defensive Security	40	SOC operations, incident response, threat detection
Digital Forensics	40	Evidence collection, memory forensics, malware analysis
Cloud Security	40	Azure, AWS, GCP, Zero Trust, DevSecOps

Dataset

CyberSec-Bench on Hugging Face

Author

Built by AYI-NEDJIMI