FEVER (Fact Extraction and VERification) consists of 185,445 claims generated by altering sentences extracted from Wikipedia and subsequently verified without knowledge of the sentence they were derived from.
You can evaluate an embedding model on this dataset using the following code:
import… See the full description on the dataset page:
https://huggingface.co/datasets/mteb/fever.