Hugging Face's logo

Datasets:
hakari-bench
/
NanoMedical

Modalities:
Text
Formats:
parquet
Languages:
Arabic
English
Polish
Size:
10K - 100K
Tags:
information-retrieval
retrieval
nano
bm25
dense-retrieval
reranking
Libraries:
Datasets
pandas
Polars
Dataset card Data Studio Files
xet
 Community
Dataset Viewer
The dataset viewer is not available for this dataset.
Unexpected token '<', "<html> <h"... is not valid JSON

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

NanoMedical

This dataset is a Nano-style retrieval dataset for HAKARI-bench.

NanoMedical contains 10 Nano retrieval splits derived from MTEB(Medical, v1). Each split keeps up to 200 eligible queries and up to 10000 corpus documents, with exact duplicate query and document text removed where the generator records that policy.

Usage 

from datasets import load_dataset

dataset_id = "hakari-bench/NanoMedical"
split = "NanoCMedQAv2reranking"

queries = load_dataset(dataset_id, "queries", split=split)
corpus = load_dataset(dataset_id, "corpus", split=split)
qrels = load_dataset(dataset_id, "qrels", split=split)
reranking_candidates = load_dataset(dataset_id, "reranking_hybrid", split=split)

Data Layout

This dataset uses six Hugging Face Datasets configs:

  • corpus: documents with _id and text
  • queries: queries with _id and text
  • qrels: positive relevance labels with query-id and corpus-id
  • bm25: BM25 candidate lists with query-id and corpus-ids
  • harrier_oss_v1_270m: dense candidate lists from microsoft/harrier-oss-v1-270m
  • reranking_hybrid: RRF candidate lists built from bm25 and harrier_oss_v1_270m

Each config has the same Nano split names.

Candidate Construction

  • bm25: local BM25 top-500 with automatic language-aware tokenization. The resolved tokenizer is shown in the Candidate Quality table, for example wordseg@ja.
  • harrier_oss_v1_270m: dense top-500 from microsoft/harrier-oss-v1-270m. In tables this is shown as Dense; Dense means microsoft/harrier-oss-v1-270m with the web_search_query prompt for queries and cosine similarity over normalized embeddings.
  • reranking_hybrid: RRF over bm25 and harrier_oss_v1_270m using rrf_k=100, keeping the RRF top-100.

Safeguard means rank 101 is appended only when RRF top-100 contains no qrels-positive document.

Split Statistics

Length statistics are character counts computed with len(str(text)).

Nano split Queries Corpus Qrels Query chars avg Query chars p50 Query chars p75 Doc chars avg Doc chars p50 Doc chars p75
NanoCMedQAv2reranking 200 10000 377 50.1 39.0 69.2 100.9 91.0 128.0
NanoCUREv1 200 10000 5181 75.9 78.0 91.0 604.2 511.0 757.0
NanoCmedqa 200 10000 324 52.0 44.0 69.0 157.6 112.0 203.0
NanoMedicalQA 200 2007 200 54.2 50.0 64.0 1102.4 683.0 1251.0
NanoNFCorpus 200 3593 3718 17.1 11.5 20.2 1589.5 1612.0 1868.0
NanoPublicHealthQA 86 86 86 79.8 70.0 92.8 828.2 625.0 1041.8
NanoSciFact 200 5183 226 90.1 83.0 107.2 1499.4 1426.0 1811.5
NanoSciFactPL 200 5183 226 95.5 88.0 117.0 1554.5 1467.0 1871.0
NanoTRECCOVID 50 10000 50 69.2 64.5 76.8 1208.8 1318.0 1735.0
NanoTRECCOVIDPL 50 10000 50 69.4 66.0 80.2 1251.9 1360.0 1789.2

Candidate Quality

nDCG@10 and Recall@100 are computed from the included candidate rankings against the included qrels, then reported as 0-100 scores such as 52.45. Recall@100 uses only the top 100 candidates; an optional rank-101 safeguard positive is not counted in Recall@100.

Dense means microsoft/harrier-oss-v1-270m with the web_search_query prompt and cosine similarity.

Nano split BM25 tokenizer BM25 nDCG@10 Dense nDCG@10 Hybrid nDCG@10 BM25 Recall@100 Dense Recall@100 Hybrid Recall@100 Hybrid candidates Safeguard positives
Mean - 43.64 51.00 49.02 68.20 76.98 80.78 - 206
NanoCMedQAv2reranking wordseg@zh 15.27 32.09 25.29 34.91 67.21 62.06 100-101 59
NanoCUREv1 english_porter_stop 46.93 50.03 52.62 60.42 66.20 68.15 100-101 14
NanoCmedqa wordseg@zh 16.69 33.80 25.91 39.12 71.97 65.62 100-101 57
NanoMedicalQA english_porter_stop 54.39 73.08 65.10 92.00 92.50 97.00 100-101 6
NanoNFCorpus english_porter_stop 29.21 30.70 31.82 24.96 34.29 32.57 100-101 48
NanoPublicHealthQA stemmer@arabic 73.79 81.76 78.47 100.00 100.00 100.00 86 0
NanoSciFact english_porter_stop 70.17 73.34 75.06 94.40 93.25 97.50 100-101 5
NanoSciFactPL regex 57.50 60.61 65.38 86.22 88.40 92.90 100-101 13
NanoTRECCOVID english_porter_stop 39.83 38.75 31.93 80.00 70.00 96.00 100-101 2
NanoTRECCOVIDPL regex 32.66 35.85 38.64 70.00 86.00 96.00 100-101 2

Hybrid Safeguard Summary

  • Safeguard positives: 206
  • Rows limited by corpus size: 86
  • Metadata file: reranking_hybrid_metadata.json

Source Links

License

NanoMedical is a derived dataset. Users must comply with the licenses, terms, and attribution requirements of the upstream datasets and benchmarks.
Downloads last month
279
Total file size:
118 MB