kaan39/turkish-wikipedia-dataset

该数据集包含了关于土耳其的公共机构、政府部门、国家机构、官方象征和历史人物的结构化土耳其语对话数据。数据从可信赖且中立的来源——土耳其维基百科自动提取，并转换成了适合大型语言模型微调的格式。每个样例包括一个系统指令、一个用户查询和一个助手回应的对话格式，旨在提升指令跟踪和对话能力。This dataset contains structured Turkish chat data about public institutions, government departments, state organs, official symbols, and historical figures in Turkey. The data is extracted automatically from a reliable and neutral source—Turkish Wikipedia—and converted into a format suitable for fine-tuning large language models. Each example includes a system instruction, a user query, and an assistant response in a dialogue format, designed to enhance instruction tracking and dialogue capabilities.