Skip to content

Dataset · mteb

NFCorpus medical text retrieval benchmark

mteb/nfcorpus

Corpus of English medical and scientific texts for evaluating information retrieval models.

What's inside

NFCorpus is a full-text dataset from the MTEB benchmark collection, designed for ranking and information retrieval in the medical and academic domains. It contains between 100k and 1M English text records collected at Heidelberg University. The dataset is integrated into the mteb library and ready for running standardized model evaluation.

What you can do with it

  • Evaluating text embedding model quality in medical and scientific domains
  • Tuning and testing RAG systems for specialized knowledge base retrieval
Tasks
Text retrieval
Languages
English
Size
100K–1M rows
Downloads this month
49.7K
Likes
3

How to load it

from datasets import load_dataset

ds = load_dataset("mteb/nfcorpus")