jhu-clsp/mFollowIR-cross-lingual-parquet-mteb

mFollowIR-cross-lingual-mteb数据集是一个多语言数据集，包含波斯语（fas）、俄语（rus）和中文（zho）三种语言的配置。每种语言的配置包括corpus（原始文档）、queries（原始和修改后的查询）、instruction（查询指令）、default（原始相关性判断）、qrel_diff（相关性判断的变化）和top_ranked（每个查询的顶级文档）。数据集的结构经过重新组织，以适应新的MTEB格式。

The mFollowIR-cross-lingual-mteb dataset is a multilingual dataset containing configurations for Persian (fas), Russian (rus), and Chinese (zho). Each language configuration includes corpus (original documents), queries (original and modified queries), instruction (query instructions), default (original relevance judgments), qrel_diff (changes in relevance judgments), and top_ranked (top-ranked documents for each query). The dataset structure has been reorganized to fit the new MTEB format.