WikiText
The WikiText language modeling dataset is a collection of over 100 million tokens extracted from the set of verified Good and Featured articles on Wikipedia.
cc-by-sa-4.0
1M<n<10M
Text Generation
Fill-Mask
English
Sign in to see dataset files
orCreate an account