Cybersecurity Q&A V2: 2.6 Million Examples for LLM Fine-Tuning

2,621,468 question-answer examples for training large language models on cybersecurity topics. The dataset aggregates records from the NIST NVD CVE database and a detailed CVE analysis dataset, covering vulnerabilities, attack techniques, and defensive strategies. It was created by user 'rezaduty' and last updated on Hugging Face in June 2026.

Use Cases

• Fine-tuning LLMs for cybersecurity question-answering based on the described Q&A structure.
• Training models to analyze Common Vulnerabilities and Exposures (CVEs) based on the inclusion of NVD records.
• Developing systems for automated vulnerability detection and remediation suggestion based on the coverage of attack techniques and defensive strategies.

Strengths

• Contains 2,621,468 examples, providing a large-scale resource for model training.
• Sources data from the authoritative NIST NVD CVE database, covering records from 2002 to 2025.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file format details are unknown, which may limit suitability assessment.

Provenance

Source

Primarily the NIST National Vulnerability Database (NVD) CVE records and the AlicanKiraz0/All-CVE-Records-Training-Dataset.

Collection Method

Likely aggregated and formatted for question-answering tasks from existing CVE databases.

Time Range

CVE records from 2002 to 2025.

Freshness

Last updated 2026-06-04 03:09:24; freshness should be verified.