license: apache-2.0
language:
- en
task_categories: - text-generation
tags: - security
- cybersecurity
- red-team
- pentest
- instruction-tuning
- sft
pretty_name: RedSec Security SFT v1
size_categories: - 10K<n<100K
configs: - config_name: default
data_files:- split: train
path: train.jsonl - split: validation
path: valid.jsonl - split: test
path: test.jsonl
- split: train
RedSec Security SFT v1
A chat-formatted supervised fine-tuning dataset for security-focused language models, intended for authorized red-team, penetration-testing, and defensive use. Each record is a {"messages": [...]} conversation with an optional system turn, a user turn, and an assistant turn.
| Split | Rows |
|---|---|
| train | 55,459 |
| validation | 1,155 |
| test | 1,155 |
| total | 57,769 |
Sources and attribution
This is a derivative work. It combines, reformats, deduplicates, and splits two Apache-2.0 datasets. All credit for the underlying content goes to their authors.
- Trendyol/Trendyol-Cybersecurity-Instruction-Tuning-Dataset (Apache-2.0): system/user/assistant cyber-defense instruction data.
- CyberNative/Code_Vulnerability_Security_DPO (Apache-2.0): vulnerability and secure-code pairs. The
questionfield was used as the user turn and thechosenfield as the assistant turn.
This dataset is released under Apache-2.0 to match its sources, with their license and attribution notices preserved here.
Changes from the sources
- Mapped both sources to a unified chat format,
{"messages": [...]}. - Kept system prompts where present.
- Dropped rows with empty user or assistant content.
- Deduplicated by SHA-256 of the message content.
- Split into train, validation, and test (about 96/2/2).
Intended use
Fine-tuning security assistants for authorized security testing and research. Not for unauthorized access or harmful use.
Reproducibility
Regenerable from the project builder that produced it (data/build_v1.py): load both sources, map to chat format, dedup, and split.