ahmadkaab/Trendyol-Cybersecurity-Instruction-Tuning-Dataset

Trendyol网络安全防御指令调整数据集是一个专门用于训练先进防御性安全AI助手的数据集，包含53,202条精心策划的系统/用户/助手指令调整示例，覆盖200多个专业网络安全领域。数据集由Trendyol安全团队构建，扩展了现代安全挑战的覆盖范围，包括云原生威胁、AI/ML安全、量子计算风险和高级事件响应技术。数据集结构包括系统、用户和助手三个字段，分别代表角色定义与伦理指南、现实的安全问题/场景和全面的技术响应。数据集分为训练集（90%）、验证集（5%）和测试集（5%），格式为Parquet优化列存储。数据集创建过程包括高级内容策划、多阶段处理管道、质量保证框架和主题覆盖验证。

The Trendyol Cybersecurity Defense Instruction-Tuning Dataset is a specialized dataset for training state-of-the-art defensive security AI assistants, containing 53,202 meticulously curated system/user/assistant instruction-tuning examples covering 200+ specialized cybersecurity domains. Built by the Trendyol Security Team, it expands coverage of modern security challenges including cloud-native threats, AI/ML security, quantum computing risks, and advanced incident response techniques. The dataset structure includes three fields: system, user, and assistant, representing role definition with ethical guidelines, realistic security questions/scenarios, and comprehensive technical responses, respectively. The dataset is split into train (90%), validation (5%), and test (5%) sets, stored in Parquet (optimized columnar storage) format. The dataset creation process involves advanced content curation, a multi-stage processing pipeline, a quality assurance framework, and topic coverage validation.