KnowUnDo

KnowUnDo仓库主要用于知识问答任务，专注于研究大型语言模型的知识遗忘。该仓库包含1K到10K规模的数据，分为copyright和privacy两种配置，每种配置都包含unlearn和retention两种数据划分。使用者可以通过标准化数据操作来加载和使用这些数据。该研究的成果可以被引用。

The KnowUnDo repository is primarily designed for knowledge QA tasks, focusing on the study of knowledge forgetting in large language models. This repository contains datasets ranging from 1K to 10K in scale, with two configurations: copyright and privacy. Each configuration includes two data splits: unlearn and retention. Users can load and utilize these datasets via standardized data operations. Research findings from this study can be cited.