Datasets:

---

coral-nlp

/

german-commons

like 26

Follow

CORAL NLP Research 18

This PR addresses two improvements for the German Commons dataset card:

1. Updates the size_categories metadata field from 100M<n<1B to 100B<n to accurately reflect that the dataset contains 154.56 billion tokens.
2. Adds a link to the llmdata GitHub repository (https://github.com/coral-nlp/llmdata), which is the framework used for the dataset's construction, enhancing reproducibility as mentioned in the paper.