Translations:Embeddings/12/en: Difference between revisions
Appearance
Importing a new version from external source |
(No difference)
|
Latest revision as of 14:32, 7 May 2025
GeenStijl.nl embeddings contains over 8M messages from the controversial Dutch websites GeenStijl and Dumpert to train a word embedding model that captures the toxic language representations contained in the dataset. The trained word embeddings (±150MB) are released for free and may be useful for further study on toxic online discourse.