Wikipedia Mention Samples

该数据集是从维基百科中随机抽取的提及特定概念的句子集合，它被用于评估语义概念嵌入的有效性。每个概念下的句子样本可达500条，适用于进行词汇分类、本体完善以及零样本学习等任务。

This dataset is a collection of sentences randomly extracted from Wikipedia that refer to specific concepts, and it is employed to evaluate the effectiveness of semantic concept embeddings. The number of sentence samples under each concept can reach up to 500, making it suitable for tasks such as lexical classification, ontology enrichment, and zero-shot learning.