Wikipedia dataset containing cleaned articles of all languages.
The datasets are built from the Wikipedia dump
(
https://dumps.wikimedia.org/) with one split per language. Each example
contains the content of one full Wikipedia article with cleaning to strip
markdown and unwanted sections (references, etc.).
