ArkAiLab-Adl/Alen-Walk-safeguard-dataset-v1-mini

AlenWalk Safeguard Dataset v1 Mini是一个开放的、精心策划的数据集，专注于检测文本中的不安全、滥用、性和威胁语言。它旨在构建AI审核系统、AutoMod管道和内容安全分类器。该数据集由AlenWalk项目开发和策划，强调数字严重性标签，适用于快速、实时的审核用例。数据集提供了一个紧凑但多样化的文本样本集合，用于内容安全和审核研究。这个Mini版本旨在作为一个轻量级的入口点和基础，提供清晰的基于严重性的数字标签、实用的审核导向类别，并易于集成到基于规则和AI驱动的系统中。它适用于开发者希望在扩展到更大版本之前使用一个干净、透明的数据集。

The **AlenWalk Safeguard Dataset v1 Mini** is an open, curated dataset focused on **detecting unsafe, abusive, sexual, and threatening language in text**. It is designed for building **AI moderation systems**, AutoMod pipelines, and content safety classifiers. Developed and curated under the **AlenWalk** project, this dataset emphasizes **numeric severity labeling** for fast, real-time moderation use cases. AlenWalk Safeguard Dataset v1 Mini provides a compact but diverse collection of text samples labeled for content safety and moderation research. This **Mini** release is intended as a lightweight entry point and foundation, offering clear severity-based numeric labels, practical moderation-oriented categories, and easy integration into rule-based and AI-driven systems. It is ideal for developers who want a **clean, transparent dataset** before scaling to larger releases.