Genticca/Korean_Euphemistic_Coating

该数据集是通过使用DeepL API将原始的walledai/AdvBench数据集翻译成韩语而构建的。数据集分为两种配置：Coating_Data和Non_Coating_Data。Coating_Data包含通过算法应用掩饰(委婉语言)生成的有害提示。Non_Coating_Data则是直接从原始数据翻译而来，没有进行额外修改的有害提示。

This dataset was constructed by translating the original walledai/AdvBench dataset into Korean using the DeepL API. The dataset consists of two configurations: Coating_Data and Non_Coating_Data. Coating_Data contains harmful prompts altered with euphemistic language using an algorithm. Non_Coating_Data includes harmful prompts translated from the original English data without any additional modifications.