This dataset is a Nano-style retrieval dataset for HAKARI-bench. NanoMTEB-Scandinavian is a compact retrieval benchmark for Scandinavian-language MTEB-style task families. It includes Danish, Norwegian, and Swedish retrieval tasks spanning fact verification, question answering, news, encyclopedic content, FAQ retrieval, and social-media retrieval.

Usage

from datasets import load_dataset

dataset_id = "hakari-bench/NanoMTEB-Scandinavian" split… See the full description on the dataset page: https://huggingface.co/datasets/hakari-bench/NanoMTEB-Scandinavian.