This is not a Soroush Vahidi research dataset. It is a Parquet-format, viewer-compatible mirror
of the corpus and queries splits of BeIR/scidocs,
a scientific-document citation-prediction and retrieval benchmark originally built from the
AllenAI SciDocs collection and redistributed by the
BEIR project. It exists solely so that other repositories in
this account (in particular SoroushVahidi/consistency-aware-judgments) can reference a… See the full description on the dataset page:
https://huggingface.co/datasets/SoroushVahidi/scidocs.
