Synthetic CVE Reports for Instruction Tuning

471 synthetic instruction-tuning pairs generated for cybersecurity threat intelligence. The data is derived from the NIST National Vulnerability Database (NVD). The author, organization, and last update date are unknown.

Use Cases

• Fine-tune language models for CVE report generation based on the synthetic instruction pairs.
• Train models to understand and summarize cybersecurity vulnerabilities based on the NVD-derived content.
• Develop AI assistants for cybersecurity analysts using the structured instruction-response format.
• Benchmark model performance on technical text generation tasks in the cybersecurity domain.

Strengths

• Contains 471 distinct instruction-tuning pairs, providing a defined corpus for model training.
• Derived from the authoritative NIST NVD, suggesting a foundation in real-world vulnerability data.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

NIST National Vulnerability Database (NVD)

Collection Method

Synthetic generation (likely using Llama 3.1) to create instruction-tuning pairs.