NanoIndicQA

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

NanoIndicQA is a compact multilingual IndicQA retrieval benchmark derived from mteb/IndicQARetrieval. It repurposes IndicQA cloze-style reading-comprehension examples into context retrieval tasks across Assamese, Bengali, Gujarati, Hindi, Kannada, Malayalam, Marathi, Odia, Punjabi, Tamil, and Telugu.

Usage

from datasets import load_dataset

dataset_id = "hakari-bench/NanoIndicQA"
split = "as"

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
as	200	250	200	55.3	52.0	66.0	1401.3	1027.0	1545.2
bn	200	250	201	52.1	47.0	59.0	2196.0	1572.5	2843.0
gu	200	248	201	61.0	55.5	74.0	960.5	841.5	1173.2
hi	200	261	201	56.9	55.0	67.0	2550.8	1997.0	3007.0
kn	200	257	200	53.3	50.5	61.2	882.7	787.0	1089.0
ml	200	247	200	81.5	77.0	96.2	2522.6	1864.0	3132.5
mr	200	250	200	59.9	56.0	71.0	1711.7	1305.5	1880.2
or	200	252	201	57.2	52.5	69.0	801.9	711.0	1005.2
pa	200	241	200	63.5	59.0	75.2	1423.5	1156.0	1604.0
ta	200	253	201	56.3	52.0	66.0	2288.3	1854.0	2621.0
te	200	250	200	65.0	63.0	75.2	2936.2	2203.0	3453.2

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	-	56.53	71.64	66.88	88.41	95.86	97.20	-	61
as	regex	61.11	74.16	72.83	91.00	98.00	98.00	100-101	4
bn	whitespace	69.71	77.73	74.60	89.50	99.25	97.00	100-101	6
gu	regex	60.60	74.87	72.07	90.50	96.50	97.00	100-101	6
hi	stemmer@hindi	45.45	65.11	57.38	88.75	85.50	96.50	100-101	7
kn	regex	47.30	70.37	61.11	82.50	98.00	97.00	100-101	6
ml	regex	65.28	82.14	78.07	94.00	99.00	99.00	100-101	2
mr	regex	46.12	67.20	59.16	84.00	97.00	96.50	100-101	7
or	regex	60.41	76.05	70.33	91.75	96.75	97.75	100-101	4
pa	regex	59.83	64.45	68.85	92.50	98.00	98.00	100-101	4
ta	regex	29.32	64.15	45.51	74.00	94.00	95.00	100-101	10
te	whitespace	76.74	71.86	75.82	94.00	92.50	97.50	100-101	5

Hybrid Safeguard Summary

• Safeguard positives: 61
• Rows limited by corpus size: 0
• Metadata file: reranking_hybrid_metadata.json

Source Links

• mteb/IndicQARetrieval

License

NanoIndicQA is a derived dataset. Users must comply with the licenses, terms, and attribution requirements of the upstream MTEB task sources and their original datasets.