3D-EX

3D-EX数据集由卡迪夫大学计算机科学与信息学院等机构创建，旨在整合多种英语词典和百科资源，形成一个包含术语、定义和例句的统一知识库。该数据集通过精心计算的训练/验证/测试分割，避免了模型记忆问题，适用于多种NLP下游任务。数据集内容丰富，涵盖了从传统词典到在线百科的广泛资源，支持定义建模等任务，有助于理解和生成人类可读的词典定义。

The 3D-EX dataset was developed by the School of Computer Science and Informatics at Cardiff University and other institutions. It aims to integrate various English dictionaries and encyclopedia resources to build a unified knowledge base encompassing terms, definitions and example sentences. This dataset adopts a carefully designed train/validation/test split to avoid model memorization issues, making it suitable for a diverse range of downstream natural language processing (NLP) tasks. With rich content covering a broad spectrum of resources from traditional dictionaries to online encyclopedias, the dataset supports tasks including definition modeling, and facilitates the understanding and generation of human-readable dictionary definitions.