google/wiki40b

Dataset Card for "wiki40b" Dataset Summary Clean-up text for 40+ Wikipedia languages editions of pages correspond to entities. The datasets have train/dev/test splits per language. The dataset is cleaned up by page filtering to remove disambiguation pages, redirect pages, deleted pages, and non-entity pages. Each example contains the wikidata id of the entity, and the full Wikipedia article after page processing that removes non-content sections and structured… See the full description on the dataset page: https://huggingface.co/datasets/google/wiki40b.