boricua/multilingual_language_skill
收藏Hugging Face2025-03-07 更新2025-04-26 收录
下载链接：
https://hf-mirror.com/datasets/boricua/multilingual_language_skill
下载链接
链接失效反馈官方服务：
资源简介：
本数据集旨在训练语言模型，使其能够用提出问题或提示的同一种语言进行回答。数据集的主题基于Wikidata数据集，每个主题包含三个问题-答案对，每种语言共有99个样本。这些问题-答案对被翻译成法语、德语、意大利语和西班牙语，构成了一个完整的多元语言数据集。注意，如果语言字段为空或为null的样本是英语。
This dataset is designed to train language models to respond in the same language in which the question or prompt is formulated. The topics are inspired by the Wikidata dataset, with each topic featuring three question-answer pairs, totaling 99 samples per language. These question-answer pairs are translated into French, German, Italian, and Spanish to create a comprehensive multilingual dataset. Note that samples with empty or null values in the `language` field are in English.
提供机构：
boricua
