Ink-bai/MC-Wiki-Chinese

这是一个基于中文 Minecraft Wiki 条目整理得到的数据集，适合用于 RAG、检索、知识库问答、文档结构化实验等任务。数据集中包含两个版本：原始页面文本版本和按标题和条目做过轻量结构化处理的版本，每个版本都包含约8200条记录。原始版适合直接做全文检索、切分、Embedding和通用语料分析，结构化版适合做知识增强、分段索引、按章节召回和规则化处理。This is a dataset organized based on Chinese Minecraft Wiki entries, suitable for tasks such as RAG, retrieval, knowledge base question answering, and document structured experiments. The dataset contains two versions: the original page text version and a lightly structured version by titles and entries, each containing approximately 8200 records. The original version is suitable for direct full-text retrieval, segmentation, Embedding, and general corpus analysis, while the structured version is suitable for knowledge enhancement, segmented indexing, recall by sections, and rule-based processing.