TREC CAR is a synthetic dataset based on Wikipedia that consists of about 29 million passages, using titles and section headings to generate queries associated with relevance judgments.
BibTex:
Before browse our site, please accept our cookies policy
