CyberStrike SFT 120K: Offensive Cybersecurity Instruction-Response Pairs

121,422 expert-level instruction-response pairs for offensive cybersecurity tasks. The dataset was created by author 'oyildirim' and is described as the largest open-source offensive cybersecurity SFT dataset. It was last updated on June 17, 2026.

Use Cases

• Fine-tuning language models to respond to offensive security prompts based on expert-level instruction-response pairs.
• Training AI assistants for penetration testing and ethical hacking scenarios based on the described security domains.
• Benchmarking LLM performance on cybersecurity tasks using the described 15 security generators.
• Developing red team training tools that leverage structured security question-answer data.

Strengths

• Contains 121,422 expert-level instruction-response pairs.
• Spans 15 distinct security generators, suggesting broad topic coverage.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but other scale metrics like file size and formats are unknown.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

huggingface

Freshness

Last updated 2026-06-17 23:21:03; freshness should be verified.