Professional pretraining corpus for defensive security LLMs. Rebuilt 2026-07-14.
Records: 186,104 | Avg length: 377 chars | Dedup SHA256 | Split 95/5
This dataset merges 12 Kaggle sources into one high-quality text field optimized for causal LM pretraining:
jeffborschowa/malwarebazaar-threat-intelligence-csv
oriolakolawole/ransomware-and-goodware-pe-header
joebeachcapital/tunadromd-malware-detection
atharvasoundankar/global-cybersecurity-threats-2015-2024… See the full description on the dataset page:
https://huggingface.co/datasets/11-47/GOT_Defense_1.
