datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Trendyol-Cybersecurity-Instruction-Tuning-Dataset
Trendyol Cybersecurity Defense Instruction-Tuning Dataset (v2.0)
🚀 TL;DR
53,202 meticulously curated system/user/assistant instruction-tuning examples covering 200+ specialized cybersecurity domains. Built by the Trendyol Security Team for training state-of-the-art defensive security AI assistants. Expanded from 21K to 53K rows with comprehensive coverage of modern security challenges including cloud-native threats, AI/ML security, quantum computing risks… See the full description on the dataset page: https://huggingface.co/datasets/Trendyol/Trendyol-Cybersecurity-Instruction-Tuning-Dataset.cyber-security
Cybersecurity AI Knowledge Base — PhD-Level Dataset
Overview
This is the most comprehensive cybersecurity knowledge base ever assembled for AI training. It covers all domains of cybersecurity at PhD-level depth — from offensive red teaming and bug bounty exploitation to defensive SOC operations, digital forensics, and cutting-edge AI/LLM security.
Size: 16 GB | Files: 507 | Domains: 30+ | Sources: 15+ platforms
Purpose
Train the world's most… See the full description on the dataset page: https://huggingface.co/datasets/Vyber07/cyber-security.cyberstrike-sft-120k
CyberStrike SFT 120K
The largest open-source offensive cybersecurity SFT dataset
121,422 expert-level red team instruction-response pairs across 15 security generators
Quick Start •
Why CyberStrike •
Domains •
Data Format •
Training Guide •
Benchmarks •
Contributing •
License
Why CyberStrike?
Most LLMs refuse or give surface-level answers to offensive security questions. Security professionals —… See the full description on the dataset page: https://huggingface.co/datasets/oyildirim/cyberstrike-sft-120k.CyberSecurity-1M
CyberSecurity-1M
A large-scale, multi-source cybersecurity knowledge dataset containing 1.19M records across 16 categories, collected exclusively for academic, non-commercial research purposes. Last updated: 2026-05-27.
Disclaimer: This dataset is provided for academic research only. All content is aggregated from publicly available sources. The views, opinions, and information expressed in the dataset content do not represent the views or positions of the research team. The… See the full description on the dataset page: https://huggingface.co/datasets/WhitzardAgent/CyberSecurity-1M.cybersecurity_qa
Cybersecurity QA
This dataset contains instruction–response pairs focused on cybersecurity concepts.It can be used for instruction-tuned fine-tuning of LLMs
Dataset Structure
Format: JSONL (.jsonl)
Each line is a JSON object with fields:
instruction: the task or question
input: optional extra context (empty string in this dataset)
output: the expected answer
Example:
{"instruction": "What is cybersecurity's primary purpose?", "input": "", "output":… See the full description on the dataset page: https://huggingface.co/datasets/mariiazhiv/cybersecurity_qa.cybersec-master-dataset
Cybersecurity Master Instruction Dataset
Overview
A large-scale cybersecurity instruction-tuning dataset in ShareGPT conversational format,
assembled from multiple authoritative open sources and deduplicated.
At 1,807,941 deduplicated records, this appears to be one of the larger cybersecurity
LLM fine-tuning / instruction-style datasets on Hugging Face, and likely among the larger
broad vulnerability-intelligence corpora in conversational/instruction format. It is… See the full description on the dataset page: https://huggingface.co/datasets/Voidreaper2026/cybersec-master-dataset.turkish_cyber_security_controls_benchmark
Turkish Cyber Security Controls Benchmark
Türkçe siber güvenlik kontrol seçimi ve kontrol denetimi yeteneğini ölçmek için
hazırlanmış, senaryo tabanlı çoktan seçmeli değerlendirme kümesidir.
v0.1.0, uzman incelemesine açık ilk sürümdür ve NIST SP 800-53 Rev. 5,
Release 5.2.0 kontrol kataloğunu hedefler.
Kapsam
100 Türkçe senaryo
NIST SP 800-53'ün 20 kontrol ailesinin her birinden 5 soru
64 kontrol seçimi sorusu
17 denetim kanıtı sorusu
19 denetim yargısı sorusu… See the full description on the dataset page: https://huggingface.co/datasets/logicBombExe/turkish_cyber_security_controls_benchmark.cyberusecase-v1.0
Cybersecurity SOC Fine-Tuning Dataset — 17.5k Real CVEs (2018–2026) + SOC Knowledge
A large supervised fine-tuning (SFT) dataset for teaching an LLM expert-level
cybersecurity reasoning across vulnerability management, SOC alert triage, detection
engineering, threat intelligence & hunting, incident response, and cloud/DevSecOps.
It combines 17,590 real CVEs (2018–2026) pulled from the NIST NVD data feeds with a
hand-curated set of 65 landmark CVEs (rich, multi-angle coverage)… See the full description on the dataset page: https://huggingface.co/datasets/ronaldocloud/cyberusecase-v1.0.cybersecurity-sft-100k
Cybersecurity SFT 100K
A synthetic supervised fine-tuning dataset of 100,000 high-quality cybersecurity conversations designed to train AI assistants for security operations, threat analysis, incident response, and defensive security engineering.
Dataset Description
This dataset covers real-world security scenarios across 9 cybersecurity domains. Each record follows the ShareGPT conversation format with a practitioner-level query and a detailed, structured… See the full description on the dataset page: https://huggingface.co/datasets/stindardlogic/cybersecurity-sft-100k.JOR-Bench
JOR-Bench: Japanese Operations Research Benchmarks for Evaluating Large Language Models
JOR-Bench is a bilingual evaluation benchmark for assessing large language models (LLMs) on Operations Research (OR) problem formulation. It provides 1,319 OR word problems in both English and Japanese, derived from five publicly available English benchmarks via Japanese translation.
Dataset Description
Translating a natural-language problem description into a mathematical… See the full description on the dataset page: https://huggingface.co/datasets/cyberagent/JOR-Bench.cybersec-eval
CyberSec-Eval
This dataset is a language-partitioned version of CS-Eval, a comprehensive evaluation toolkit for fundamental cybersecurity models or large language models' cybersecurity abilities.
The original dataset is split into English and Chinese subsets to facilitate targeted evaluation of models in specific language environments.
Dataset Structure
The dataset contains two configurations:
en: Questions written in English (337 items).
zh: Questions written… See the full description on the dataset page: https://huggingface.co/datasets/iselabvn/cybersec-eval.CyberRepo-10K
CyberRepo-10K
A curated dataset of 7,670 real-world vulnerability audit tasks with verified GitHub repositories, fix commits, and patch diffs — designed for training and evaluating LLM-based static code vulnerability auditing and PoC generation.
Disclaimer: This dataset is provided for academic research only. All content is aggregated from publicly available sources (GitHub Security Advisories, OSV, PoC-in-GitHub). The research team does not endorse, support, or take responsibility… See the full description on the dataset page: https://huggingface.co/datasets/WhitzardAgent/CyberRepo-10K.cybersec-qa-dataset-zh
Cybersecurity QA Dataset (zh) · 中文网络安全技术问答数据集
面向 防御与安全教育 的中文网络安全技术问答数据集，适用于 LLM 指令微调（SFT）。
21,799 条纯技术问答，零国家归因、零地缘内容，附可复现质检流水线与 CI 校验。
数据概览
总条数：21,799（149 批）
格式：JSONL，每行 {"user": ..., "assistant": ...}
平均答案长度：约 1,231 字，结构化分层（原理 → 攻击面 → 检测 → 缓解）
主题分布（按问题关键词约略归类）
主题
条数
二进制 / 漏洞利用
5186
Web 安全
4858
其他 / 综合
2730
密码学
1648
蓝队 / DFIR / 检测
1622
AD 域 / 内网 / 后渗透
1278
网络协议攻防
1151
云原生 / 容器
1133
移动 / IoT / 固件
837
恶意软件 / 逆向分析
764… See the full description on the dataset page: https://huggingface.co/datasets/uninhibited-scholar/cybersec-qa-dataset-zh.bug-bounty-cybermindcli
Bug Bounty & Méthodologies de Pentest
Méthodologies (OWASP, PTES), checklists par type d app, techniques d attaque, plateformes, templates de rapports et outils.
Links
Version anglaise
AYI NEDJIMI Consultants
cybersec-jsonschemabench-cloudtrail-v6
CybersecJSONSchemaBench CloudTrail Attack Reconstruction v6
A 100-problem long-context cybersecurity reasoning benchmark over real flAWS CloudTrail logs with synthetically injected MITRE ATT&CK attack chains.
Each task gives the model 600 real CloudTrail records (280-380K tokens of JSON) containing a single hidden multi-step attack chain. The model must produce a structured answer identifying the attacking principal, the MITRE ATT&CK technique, the per-phase records that constitute… See the full description on the dataset page: https://huggingface.co/datasets/achinta3/cybersec-jsonschemabench-cloudtrail-v6.Trendyol-Cybersecurity-Instruction-Tuning-Dataset
Trendyol Cybersecurity Defense Instruction-Tuning Dataset (v2.0)
🚀 TL;DR
53,202 meticulously curated system/user/assistant instruction-tuning examples covering 200+ specialized cybersecurity domains. Built by the Trendyol Security Team for training state-of-the-art defensive security AI assistants. Expanded from 21K to 53K rows with comprehensive coverage of modern security challenges including cloud-native threats, AI/ML security, quantum… See the full description on the dataset page: https://huggingface.co/datasets/Soban1234/Trendyol-Cybersecurity-Instruction-Tuning-Dataset.Cybersecurity-Dataset
Cybersecurity-Dataset
(Note: Replace the above with the actual URL of the uploaded image for the "智穹盾" poster. You can upload it to Hugging Face or another host like GitHub for embedding.)
Dataset Summary
The Cybersecurity-Dataset is a comprehensive collection of cybersecurity-related data designed for training and fine-tuning large language models (LLMs) in network security applications. It focuses on inner-network (offline) environments, addressing pain points such as… See the full description on the dataset page: https://huggingface.co/datasets/hcnote/Cybersecurity-Dataset.Trendyol-Cybersecurity-Instruction-Tuning-Dataset
Trendyol Cybersecurity Defense Instruction-Tuning Dataset (v2.0)
🚀 TL;DR
53,202 meticulously curated system/user/assistant instruction-tuning examples covering 200+ specialized cybersecurity domains. Built by the Trendyol Security Team for training state-of-the-art defensive security AI assistants. Expanded from 21K to 53K rows with comprehensive coverage of modern security challenges including cloud-native threats, AI/ML security, quantum computing risks… See the full description on the dataset page: https://huggingface.co/datasets/ahmadkaab/Trendyol-Cybersecurity-Instruction-Tuning-Dataset.CyberSec-Bench
CyberSec-Bench: Comprehensive Cybersecurity Benchmark Evaluation Dataset
Overview
CyberSec-Bench is a bilingual (English/French) benchmark dataset designed to evaluate the cybersecurity knowledge of Large Language Models (LLMs) and AI systems. The dataset contains 200 expert-crafted questions spanning five critical domains of cybersecurity, with detailed reference answers for each question.
This benchmark tests real-world cybersecurity knowledge at professional… See the full description on the dataset page: https://huggingface.co/datasets/AYI-NEDJIMI/CyberSec-Bench.Trendyol-Cybersecurity-Instruction-Tuning-Dataset
Trendyol Cybersecurity Defense Instruction-Tuning Dataset (v2.0)
🚀 TL;DR
53,202 meticulously curated system/user/assistant instruction-tuning examples covering 200+ specialized cybersecurity domains. Built by the Trendyol Security Team for training state-of-the-art defensive security AI assistants. Expanded from 21K to 53K rows with comprehensive coverage of modern security challenges including cloud-native threats, AI/ML security, quantum computing risks… See the full description on the dataset page: https://huggingface.co/datasets/andycoco1128/Trendyol-Cybersecurity-Instruction-Tuning-Dataset.Trendyol-Cybersecurity-Instruction-Tuning-Dataset-archive
Trendyol Cybersecurity Defense Instruction-Tuning Dataset (v2.0)
🚀 TL;DR
53,202 meticulously curated system/user/assistant instruction-tuning examples covering 200+ specialized cybersecurity domains. Built by the Trendyol Security Team for training state-of-the-art defensive security AI assistants. Expanded from 21K to 53K rows with comprehensive coverage of modern security challenges including cloud-native threats, AI/ML security, quantum… See the full description on the dataset page: https://huggingface.co/datasets/ChipHolmes/Trendyol-Cybersecurity-Instruction-Tuning-Dataset-archive.Trendyol-Cybersecurity-Instruction-Tuning-Dataset
Trendyol Cybersecurity Defense Instruction-Tuning Dataset (v2.0)
🚀 TL;DR
53,202 meticulously curated system/user/assistant instruction-tuning examples covering 200+ specialized cybersecurity domains. Built by the Trendyol Security Team for training state-of-the-art defensive security AI assistants. Expanded from 21K to 53K rows with comprehensive coverage of modern security challenges including cloud-native threats, AI/ML security, quantum computing risks… See the full description on the dataset page: https://huggingface.co/datasets/invinciblejha01/Trendyol-Cybersecurity-Instruction-Tuning-Dataset.trendyol-cybersecurity-instruction
Trendyol Cybersecurity Defense Instruction-Tuning Dataset (v2.0)
🚀 TL;DR
53,202 meticulously curated system/user/assistant instruction-tuning examples covering 200+ specialized cybersecurity domains. Built by the Trendyol Security Team for training state-of-the-art defensive security AI assistants. Expanded from 21K to 53K rows with comprehensive coverage of modern security challenges including cloud-native threats, AI/ML security, quantum… See the full description on the dataset page: https://huggingface.co/datasets/ansulev/trendyol-cybersecurity-instruction.Trendyol-Cybersecurity-Instruction-Tuning-Dataset
Trendyol Cybersecurity Defense Instruction-Tuning Dataset (v2.0)
🚀 TL;DR
53,202 meticulously curated system/user/assistant instruction-tuning examples covering 200+ specialized cybersecurity domains. Built by the Trendyol Security Team for training state-of-the-art defensive security AI assistants. Expanded from 21K to 53K rows with comprehensive coverage of modern security challenges including cloud-native threats, AI/ML security, quantum computing risks… See the full description on the dataset page: https://huggingface.co/datasets/ukcli/Trendyol-Cybersecurity-Instruction-Tuning-Dataset.cybersecdata
Pralay — Cybersecurity Instruction-Tuning Dataset
Merged, cleaned, deduplicated chat-format dataset for fine-tuning a cybersecurity assistant. Built for the Pralay project by OM CHOKSI (OMCHOKSI108).
TL;DR
194,318 chat samples (train 174,886 / val 19,432, ~90/10 split, seed 3407)
Each sample is a {system, user, assistant} triple in OpenAI chat format
Combines 6 public Hugging Face cybersecurity datasets (204K rows) with custom Q&A generated from 37 cybersecurity… See the full description on the dataset page: https://huggingface.co/datasets/OMCHOKSI108/cybersecdata.LOREA-cyber-eval
LOREA-cyber eval sets
Held-out sets used to benchmark the LOREA-cyber models. Decontaminated 8-gram against the training data,
published so the numbers in the model cards can be reproduced.
These are the sets written for this project. The models are also scored on public benchmarks that aren't
redistributed here: SecQA,
MMLU-Pro,
CyberMetric,
HumanEval.
cyber_mcq (150)
Security knowledge multiple choice across network security, crypto, web/OWASP, malware analysis… See the full description on the dataset page: https://huggingface.co/datasets/MK4-Research/LOREA-cyber-eval.cyberstrike-sft-120k
CyberStrike SFT 120K
The largest open-source offensive cybersecurity SFT dataset
121,422 expert-level red team instruction-response pairs across 15 security generators
Quick Start •
Why CyberStrike •
Domains •
Data Format •
Training Guide •
Benchmarks •
Contributing •
License
Why CyberStrike?
Most LLMs refuse or give surface-level answers to offensive security questions. Security professionals —… See the full description on the dataset page: https://huggingface.co/datasets/Alptekinege/cyberstrike-sft-120k.Trendyol-Cybersecurity-Instruction-Tuning-Dataset
Trendyol Cybersecurity Defense Instruction-Tuning Dataset (v2.0)
🚀 TL;DR
53,202 meticulously curated system/user/assistant instruction-tuning examples covering 200+ specialized cybersecurity domains. Built by the Trendyol Security Team for training state-of-the-art defensive security AI assistants. Expanded from 21K to 53K rows with comprehensive coverage of modern security challenges including cloud-native threats, AI/ML security, quantum computing risks… See the full description on the dataset page: https://huggingface.co/datasets/Madhu348/Trendyol-Cybersecurity-Instruction-Tuning-Dataset.cybersec-fact-recall
Cybersec Fact-Recall Benchmark (GhostLM v2)
Free-form short-answer benchmark for small cybersecurity language
models. Built and used by the GhostLM
project as the truth metric for the ghost-base v1.0 acceptance gate.
Why this exists
Multiple-choice cybersec benchmarks like CTIBench and SecQA reward
register matching (the model picks the option that "looks like" a
security answer) as much as actual factual recall. A small from-
scratch model can hit 28-30% on those without… See the full description on the dataset page: https://huggingface.co/datasets/Ghostgim/cybersec-fact-recall.cyberstrike-sft-120k
CyberStrike SFT 120K
The largest open-source offensive cybersecurity SFT dataset
121,422 expert-level red team instruction-response pairs across 15 security generators
Quick Start •
Why CyberStrike •
Domains •
Data Format •
Training Guide •
Benchmarks •
Contributing •
License
Why CyberStrike?
Most LLMs refuse or give surface-level answers to offensive security questions. Security professionals —… See the full description on the dataset page: https://huggingface.co/datasets/spinochenza/cyberstrike-sft-120k.