clips/

beir-nl-scidocs

	Dataset Card for BEIR-NL Benchmark




	
	Dataset Summary

BEIR-NL is a Dutch-translated version of the BEIR benchmark, a diverse and heterogeneous collection of datasets covering various domains from biomedical and financial texts to general web content. Our benchmark is integrated into the Massive Multilingual Text Embedding Benchmark (MMTEB). BEIR-NL contains the following tasks:

Fact-checking: FEVER, Climate-FEVER, SciFact Question-Answering: NQ, HotpotQA, FiQA-2018… See the full description on the dataset page: https://huggingface.co/datasets/clips/beir-nl-scidocs.