CBT

该数据集旨在通过闭卷测试格式来评估语言模型在不同词汇类别上的表现。它被设计成一个语言建模任务，其中包含需要从十个可能的选项中预测的隐藏词汇。此项任务具体为采用闭卷测试格式的语言建模任务。

This dataset aims to evaluate the performance of language models across diverse lexical categories using the closed-book test format. It is structured as a language modeling task that involves predicting hidden vocabulary items from ten candidate options. This particular task is a language modeling task that employs the closed-book test format.