Datasets:
Tasks:
Text Generation
Sub-tasks:
language-modeling
Languages:
German
Size:
10M - 100M
ArXiv:
License:
Now available: Annotations for all documents
#4
pinned
by maxidl - opened
Great work, I consider datasets like German Commons very valuable!
To enable further curation and filtering of German Commons, we annotated all documents using propella-1-4b. The annotations include core content, classification, quality & value, audience & purpose, safety & compliance, and geographic relevance.
The data is here: hf.co/datasets/openeurollm/propella-annotations/viewer/german-commons
That's great news! Thanks for including the data! This is a superb addition to make German Commons more suitable as a high-quality training resource.
EDIT: we now also included a link to the annotations in the German Commons README :)
lgienapp pinned discussion