← all datasets

50-question benchmark

Emerging
1papers using it
2026first seen

The '50-question benchmark' is a dataset used to evaluate the accuracy of document conversion frameworks in the context of domain-specific question answering, consisting of 50 independent questions derived from a corpus of Portuguese administrative documents.

50-question benchmark β€” datasets β€” recommender-systems