mteb/scidocs

SCIDOCS An MTEB dataset Massive Text Embedding Benchmark SciDocs, a new evaluation benchmark consisting of seven document-level tasks ranging from citation prediction, to document classification and recommendation. Task category t2t Domains Academic, Written, Non-fiction Reference https://allenai.org/data/scidocs How to evaluate on this task You can evaluate an embedding model on this dataset using the following code: import mteb task =… See the full description on the dataset page: https://huggingface.co/datasets/mteb/scidocs.