hakari-bench/NanoMTEB-BR
NanoMTEB-BR This dataset is a Nano-style Brazilian Portuguese retrieval benchmark for HAKARI-Bench. It is derived from the retrieval and reranking tasks exposed by MTEB-BR at commit e17f36ae304e7bdb76202744c5fb56b440050213. The six public splits are Quati, JurisTCU, BRTaxQAR, FaQuADIR, MedPTRetrieval, and FaqBacenRetrieval. Usage from datasets import load_dataset dataset_id = "hakari-bench/NanoMTEB-BR" split = "MedPTRetrieval" queries = load_dataset(dataset_id… See the full description on the dataset page: https://huggingface.co/datasets/hakari-bench/NanoMTEB-BR.
| Author / organization | hakari-bench |
|---|---|
| License (as declared) | other |
| Downloads | 173 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 3 August 2026 |
| Last updated at source | 3 August 2026 |
Popularity, last 2 days
173 now (+2 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
More from hakari-bench
- hakari-bench/resultsDataset
- hakari-bench/NanoMTEB-ScandinavianDataset
- hakari-bench/leaderboard_databaseDataset
- hakari-bench/NanoBEIR-enDataset
- hakari-bench/NanoMuPLeRDataset
- hakari-bench/NanoMIRACLDataset
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.