COMPKE

该数据集名为COMPKE，是一个包含了11,924个复杂问题的基准测试集，旨在评估大型语言模型中的知识编辑方法，特别关注那些需要复杂推理的真实生活场景。该数据集用于评估不同模型中各种知识编辑方法的性能，揭示了性能上的差异以及诸如过拟合和遗漏现象等挑战。其规模达到了11,924个复杂问题，任务聚焦于复杂问题回答。

The dataset named COMPKE is a benchmark test set containing 11,924 complex questions, which aims to evaluate knowledge editing methods in large language models (LLMs) with a particular focus on real-world scenarios requiring complex reasoning. This dataset is used to assess the performance of various knowledge editing methods across different models, revealing performance disparities and challenges such as overfitting and omission phenomena. With 11,924 complex questions in total, its task focuses on complex question answering.