StyleKQC

StyleKQC是首个针对韩语的风格变异释义语料库，专注于问题和命令句。该数据集由首尔国立大学创建，包含30,000条数据，涵盖日常生活中的六个主题。数据集通过人工重写和转换，扩展了正式和非正式句子，确保内容的核心和风格同时被考虑。StyleKQC旨在解决对话系统等工业应用中的语调和方式问题，通过提供一个定义良好的韩语句子正式风格的数据集，支持监督学习下的风格转移任务。

StyleKQC is the first style-variant paraphrase corpus dedicated to Korean, with a focus on interrogative and imperative sentences. Developed by Seoul National University, this corpus contains 30,000 entries spanning six daily-life themes. The corpus was constructed via manual rewriting and conversion to produce expanded formal and informal sentence variants, while prioritizing the preservation of both the core semantic meaning and stylistic attributes of the source content. Aiming to address tone and expression-related issues in industrial applications such as dialogue systems, StyleKQC provides a well-defined Korean corpus focused on formal sentence styles, supporting style transfer tasks under supervised learning frameworks.