olivertzeng/catgirl-zhtw-uncensored

这是一个基于猫娘数据集的数据集，主要用于文本生成任务，模拟猫娘女仆与用户（称为主人）之间的互动。数据集解决了繁体中文（台湾）数据集的缺乏、原始数据集的内容审查问题、性别假设问题以及中国大陆特有的文化引用问题。解决方案包括使用脚本将简体中文转换为繁体中文（台湾），以及去除内容审查。数据集分为三个部分：common（主要用于猫娘与用户的日常互动，扩展AI如何表现得像理想猫娘以及如何取悦用户的知识）、creative（扩展猫娘的创造力）。

This is a dataset based on the catgirl-dataset, primarily used for text generation tasks, simulating interactions between a catgirl maid and the user (referred to as master). The dataset addresses issues such as the lack of Traditional Chinese (Taiwan) datasets, censorship in the original dataset, gender assumptions, and Mainland China-specific cultural references. Solutions include using a script to convert Simplified Chinese to Traditional Chinese (Taiwan) and removing censorship. The dataset is divided into three parts: common (mainly for daily interactions between the catgirl and the user, expanding the AIs knowledge on how to behave like an ideal catgirl and please the user), creative (expanding the catgirls creativity).