This dataset contains benchmark data for evaluating documentation generation systems across 22 open-source repositories.
Dataset Structure
For each repository, the dataset includes:
metadata: Repository URL and commit ID
docs_tree: Original documentation tree structure
structured_docs: Parsed and structured documentation
rubrics: Combined evaluation rubrics for assessing documentation quality… See the full description on the dataset page: https://huggingface.co/datasets/anhnh2002/codewikibench.
