Multilingual Entity Insertion Dataset

该数据集专为研究维基百科中的实体插入而设计，包含了从所有维基百科文章中提取的链接、每个链接的周边语境，以及额外的文章级别元数据。它还包括了实体插入场景的案例以及复现实验所需资源。该数据集的规模宏大，包含了来自4900万篇文章的9.58亿个链接，覆盖了105种语言。研究任务集中在维基百科文章中的实体插入。

This dataset is specifically developed for research on entity insertion in Wikipedia articles. It comprises links extracted from all Wikipedia articles, the surrounding context of each individual link, and supplementary article-level metadata. Additionally, it provides cases of entity insertion scenarios and the necessary resources for replicating related experiments. Boasting a massive scale, this dataset contains 958 million links sourced from 49 million articles across 105 languages. The targeted research tasks center on entity insertion within Wikipedia articles.