datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
prompt-injections
Dataset Card for "deberta-v3-base-injection-dataset"
More Information needed
Prompt-injection-dataset
advance dataset if you want for llm security
https://huggingface.co/datasets/neuralchemy/prompt-injection-Threat-Matrix
Prompt Injection & Jailbreak Detection Dataset
A high-quality, leakage-free binary classification dataset for detecting prompt injection and jailbreak attacks against Large Language Models.
Zero data leakage — group-aware splitting confirmed
Balanced classes — ~60% malicious / 40% benign
Two configs — core for classical ML, full for transformers
29… See the full description on the dataset page: https://huggingface.co/datasets/neuralchemy/Prompt-injection-dataset.prompt-injections-benchmark
Dataset: Qualifire Benchmark Prompt Injection(Jailbreak vs. Benign) Datasets
Overview
This dataset contains 5,000 prompts, each labeled as either jailbreak or benign. The dataset is designed for evaluating AI models' robustness against adversarial prompts and their ability to distinguish between safe and unsafe inputs.
Dataset Structure
Total Samples: 5,000
Labels: jailbreak, benign
Columns:
text: The input text
label: The classification (jailbreak or benign)… See the full description on the dataset page: https://huggingface.co/datasets/rogue-security/prompt-injections-benchmark.prompt-injection-safetyprompt-injectionprompt-injections
Dataset Card for "deberta-v3-base-injection-dataset"
More Information needed
prompt-injections
wambosec/prompt-injections
A dataset of prompts for training prompt injection detection models.
Dataset Description
This dataset contains prompts labeled as either benign (normal user requests) or malicious (prompt injection attacks).
Dataset Statistics
Total prompts: 5,766
Benign prompts: 2,340
Malicious prompts: 3,426
Malicious ratio: 59.4%
Dataset Structure
{
"prompt": str, # The prompt text
"label": int, # 0 =… See the full description on the dataset page: https://huggingface.co/datasets/wambosec/prompt-injections.prompt_injection_cleaned_dataset-v2
Dataset Card for "prompt_injection_cleaned_dataset-v2"
More Information needed
prompt-injection-purple-llamaprompt_injection_cleaned_dataset
Dataset Card for "prompt_injection_cleaned_dataset"
More Information needed
prompt-injection-datasetCollected from the following datasets:
deepset/prompt-injections
xTRam1/safe-guard-prompt-injection
jayavibhav/prompt-injection
prompt_injection_password
Dataset Card for "prompt_injection_password"
More Information needed
prompt-injection-multilingualPrompt-injection-dataset
Prompt Injection & Jailbreak Detection Dataset
A high-quality, leakage-free binary classification dataset for detecting prompt injection and jailbreak attacks against Large Language Models.
Zero data leakage — group-aware splitting confirmed
Balanced classes — ~60% malicious / 40% benign
Two configs — core for classical ML, full for transformers
29 attack categories including cutting-edge 2025 techniques
Severity labels, source tracking, augmentation flags on every row… See the full description on the dataset page: https://huggingface.co/datasets/cyberec/Prompt-injection-dataset.prompt-injection-gemmaprompt_injection_hackaprompt_gpt35
Dataset Card for "prompt_injection_hackaprompt_gpt35"
More Information needed
prompt-injection-datasetMLTprompt-injection-attack-detection-multilingual
Prompt Injection Attack Detection Multilingual Dataset
📌 Overview
This dataset is a merged and cleaned combination of two publicly available datasets for prompt injection detection:
PromptInjectionDataset/Injection-Attack-Detection-Dataset
rikka-snow/prompt-injection-multilingual
The goal of this merged dataset is to provide a larger and more diverse benchmark for binary classification of prompt injection attacks.
🎯 Task
Binary classification:
0 →… See the full description on the dataset page: https://huggingface.co/datasets/Octavio-Santana/prompt-injection-attack-detection-multilingual.prompt-injections-benchmark-chineseprompt-injection-dataset
🛡️ Prompt Injection Detection Dataset
A comprehensive, balanced dataset of prompt injection attacks and safe prompts for training LLM security classifiers.
Author: Soham DahivalkarLicense: MITTask: Binary Text Classification (safe vs injection)
Dataset Description
This dataset contains labeled prompts for training models to detect prompt injection attacks — the #1 vulnerability in LLM applications (OWASP LLM Top 10).
Injection Categories Covered… See the full description on the dataset page: https://huggingface.co/datasets/Shomi28/prompt-injection-dataset.prompt-injectionprompt-injection-watch-dataset
Prompt Injection Watch Dataset
Normalized prompt-injection watch dataset built from three Hugging Face source
datasets.
Read this before training on it
Two findings from 30 August 2026, both measured on the files in this repository.
Neither was known when the dataset was first published.
A random split across the pooled corpus overstates performance by a wide
margin. The three contributing datasets are separable from their text alone,
and their positive rates… See the full description on the dataset page: https://huggingface.co/datasets/cowWhySo/prompt-injection-watch-dataset.prompt-injection-judge-dataset-v2
Prompt Injection Judge Dataset - Version 2 🛡️
Dataset Description
The Prompt Injection Judge Dataset (V2) is a preference-learning corpus designed exclusively for training System-2 generative AI security judges. It utilizes the ORPO (Odds Ratio Preference Optimization) algorithm to teach Large Language Models to deliberate, analyze, and definitively flag prompt injections, jailbreaks, and adversarial attacks in user inputs.
This dataset was entirely rebuilt to… See the full description on the dataset page: https://huggingface.co/datasets/hlyn-labs/prompt-injection-judge-dataset-v2.prompt_injection_v8Prompt-injection-dataset2
Prompt Injection & Jailbreak Detection Dataset
A high-quality, leakage-free binary classification dataset for detecting prompt injection and jailbreak attacks against Large Language Models.
Zero data leakage — group-aware splitting confirmed
Balanced classes — ~60% malicious / 40% benign
Two configs — core for classical ML, full for transformers
29 attack categories including cutting-edge 2025 techniques
Severity labels, source tracking, augmentation flags on every row… See the full description on the dataset page: https://huggingface.co/datasets/cyberec/Prompt-injection-dataset2.Prompt-injection-dataset
Prompt Injection & Jailbreak Detection Dataset
A high-quality, leakage-free binary classification dataset for detecting prompt injection and jailbreak attacks against Large Language Models.
Zero data leakage — group-aware splitting confirmed
Balanced classes — ~60% malicious / 40% benign
Two configs — core for classical ML, full for transformers
29 attack categories including cutting-edge 2025 techniques
Severity labels, source tracking, augmentation flags on every row… See the full description on the dataset page: https://huggingface.co/datasets/enieva/Prompt-injection-dataset.prompt-injections-1.0.0prompt-injections-subtle
wambosec/prompt-injections-subtle
A dataset of prompts for training prompt injection detection models.
Dataset Description
This dataset contains prompts labeled as either benign (normal user requests) or malicious (prompt injection attacks).
Dataset Statistics
Total prompts: 933
Benign prompts: 267
Malicious prompts: 666
Malicious ratio: 71.4%
Dataset Structure
{
"prompt": str, # The prompt text
"label": int, # 0 =… See the full description on the dataset page: https://huggingface.co/datasets/wambosec/prompt-injections-subtle.Prompt-injection-dataset
advance dataset if you want for llm security
https://huggingface.co/datasets/neuralchemy/prompt-injection-Threat-Matrix
Prompt Injection & Jailbreak Detection Dataset
A high-quality, leakage-free binary classification dataset for detecting prompt injection and jailbreak attacks against Large Language Models.
Zero data leakage — group-aware splitting confirmed
Balanced classes — ~60% malicious / 40% benign
Two configs — core for classical ML, full for transformers… See the full description on the dataset page: https://huggingface.co/datasets/Sahana28/Prompt-injection-dataset.prompt-injection-15k
prompt-injection-15k
A dataset for training prompt injection / jailbreak detection classifiers.
15,000 examples labeled as either a prompt injection attempt (1) or a normal, benign message (0). Includes direct overrides, roleplay jailbreaks, system prompt extraction attempts, obfuscated attacks, multilingual examples, and benign edge cases (imperative tasks, casual use of words like "ignore" or "forget", short or minimal inputs).
Fields
text: the user message… See the full description on the dataset page: https://huggingface.co/datasets/MistyozAI/prompt-injection-15k.