prompt-guard
safe-guard-prompt-injectionWe formulated the prompt injection detector problem as a classification problem and trained our own language model
to detect whether a given user prompt is an attack or safe. First, to train our own prompt injection detector, we
required high-quality labelled data; however, existing prompt injection datasets were either too small (on the magnitude
of O(100)) or didn’t cover a broad spectrum of prompt injection attacks. To this end, inspired by the GLAN paper,
we created a custom synthetic… See the full description on the dataset page: https://huggingface.co/datasets/xTRam1/safe-guard-prompt-injection.Unified-Prompt-Guard
Unified-Prompt-Guard (Cerberus Dataset)
这是一个综合性的 Prompt 安全检测数据集，专门用于训练防御 LLM 越狱攻击（Jailbreak）和有害输入（Unsafe
Prompts）的二分类模型。数据集经过大规模数据增强处理，包括循环翻译和英文改写等技术。
数据集描述
本数据集由以下三个高质量源合并而成，并经过严格的全局去重处理：
主要来源
jailbreak-detection-dataset: 专注于越狱攻击模式和对抗性指令。
Nemotron-Safety-Guard-Dataset-v3 (zh): NVIDIA 提供的针对中文场景的安全对话数据。
PKU-SafeRLHF: 北京大学提供的多维度偏好对齐数据（提取自其 Prompt 层）。
数据集规模
数据集
样本数
训练集 (train)
265,589
验证集 (val)
10,857
测试集 (test)
10,857
数据字段… See the full description on the dataset page: https://huggingface.co/datasets/ynyg/Unified-Prompt-Guard.prompt-guard
🛡️ prompt-guard
A security linter for LLM prompts. Catch prompt injection, jailbreaks,
system-prompt leakage, obfuscation and PII exfiltration before untrusted text
reaches your model.
Think of it as ESLint for the new attack surface. Instead of scanning code for
vulnerabilities, prompt-guard scans the prompts you're about to feed a language
model — as a CLI (great for CI and pre-commit), a REST API, a library,
or an interactive web UI.
Why
⚡ Fast… See the full description on the dataset page: https://huggingface.co/datasets/Slfgod/prompt-guard.omnimcp_mcp_prompt_injection_guard_teaser
🔬 INSPECT THE DEEPSEEK-R1 REASONING CHAIN LIVE:
Zero hallucinations. Null syntax errors. 100% AST compiler validated.🌐 Live Interactive Reasoning & Code Inspector: https://emgena.com/trainingslager🎁 Claim your Free Starter Kit (Code: STARTER100): https://emgena.com/trainingslager🏷️ Launch Discount: Get 20 € OFF any 500-incident production suite with code LAUNCH20!
📜 Enterprise Compliance: EU AI Act Articles 50 & 53 certified • 100% DSGVO / GDPR clean • Commercial EULA… See the full description on the dataset page: https://huggingface.co/datasets/emgena/omnimcp_mcp_prompt_injection_guard_teaser.Prompt-injection-and-PII
🛡️ Security Reasoning Dataset for Prompt injection and PII (senstive data) detection
This dataset contains 2,139 high-quality synthetic examples designed for training lightweight security models—specifically targeting the firewall-gemma-3-4b-it architecture—using the Distilling Step-by-Step methodology.
📊 Dataset Analytics & Distribution
The dataset is engineered to handle real-world enterprise edge cases, specifically the "needle-in-a-haystack" problem where malicious… See the full description on the dataset page: https://huggingface.co/datasets/Guardian0369/Prompt-injection-and-PII.prompt-guard-lab
AriaPromptGuard prompt-injection defense lab
Synthetic 240-case suite for the AriaPromptGuard workstation: two inert tools, a rotating canary, and a family-disjoint development/test split. The unguarded profile concatenates user and document text. Rule-only is a keyword detector. Policy-guard isolates data from instructions and allowlists read_sample_doc and add_numbers.
Organization dataset and collection are public. Live Gradio is alirezaaminzadeh/aria-prompt-guard and the… See the full description on the dataset page: https://huggingface.co/datasets/AriaAICompany/prompt-guard-lab.