--- configs: - config_name: bm25 data_files: - split: NanoArguAna path: bm25/NanoArguAna.parquet - split: NanoClimateFEVER path: bm25/NanoClimateFEVER.parquet - split: NanoDBPedia path: bm25/NanoDBPedia.parquet - split: NanoFEVER path: bm25/NanoFEVER.parquet - split: NanoFiQA2018 path: bm25/NanoFiQA2018.parquet - split: NanoHotpotQA path: bm25/NanoHotpotQA.parquet - split: NanoMSMARCO path: bm25/NanoMSMARCO.parquet - split: NanoNFCorpus path: bm25/NanoNFCorpus.parquet - split: NanoNQ path: bm25/NanoNQ.parquet - split: NanoQuoraRetrieval path: bm25/NanoQuoraRetrieval.parquet - split: NanoSCIDOCS path: bm25/NanoSCIDOCS.parquet - split: NanoSciFact path: bm25/NanoSciFact.parquet - split: NanoTouche2020 path: bm25/NanoTouche2020.parquet - config_name: corpus data_files: - split: NanoArguAna path: corpus/NanoArguAna-* - split: NanoClimateFEVER path: corpus/NanoClimateFEVER-* - split: NanoDBPedia path: corpus/NanoDBPedia-* - split: NanoFEVER path: corpus/NanoFEVER-* - split: NanoFiQA2018 path: corpus/NanoFiQA2018-* - split: NanoHotpotQA path: corpus/NanoHotpotQA-* - split: NanoMSMARCO path: corpus/NanoMSMARCO-* - split: NanoNFCorpus path: corpus/NanoNFCorpus-* - split: NanoNQ path: corpus/NanoNQ-* - split: NanoQuoraRetrieval path: corpus/NanoQuoraRetrieval-* - split: NanoSCIDOCS path: corpus/NanoSCIDOCS-* - split: NanoSciFact path: corpus/NanoSciFact-* - split: NanoTouche2020 path: corpus/NanoTouche2020-* - config_name: qrels data_files: - split: NanoArguAna path: qrels/NanoArguAna.parquet - split: NanoClimateFEVER path: qrels/NanoClimateFEVER.parquet - split: NanoDBPedia path: qrels/NanoDBPedia.parquet - split: NanoFEVER path: qrels/NanoFEVER.parquet - split: NanoFiQA2018 path: qrels/NanoFiQA2018.parquet - split: NanoHotpotQA path: qrels/NanoHotpotQA.parquet - split: NanoMSMARCO path: qrels/NanoMSMARCO.parquet - split: NanoNFCorpus path: qrels/NanoNFCorpus.parquet - split: NanoNQ path: qrels/NanoNQ.parquet - split: NanoQuoraRetrieval path: qrels/NanoQuoraRetrieval.parquet - split: NanoSCIDOCS path: qrels/NanoSCIDOCS.parquet - split: NanoSciFact path: qrels/NanoSciFact.parquet - split: NanoTouche2020 path: qrels/NanoTouche2020.parquet - config_name: queries data_files: - split: NanoArguAna path: queries/NanoArguAna.parquet - split: NanoClimateFEVER path: queries/NanoClimateFEVER.parquet - split: NanoDBPedia path: queries/NanoDBPedia.parquet - split: NanoFEVER path: queries/NanoFEVER.parquet - split: NanoFiQA2018 path: queries/NanoFiQA2018.parquet - split: NanoHotpotQA path: queries/NanoHotpotQA.parquet - split: NanoMSMARCO path: queries/NanoMSMARCO.parquet - split: NanoNFCorpus path: queries/NanoNFCorpus.parquet - split: NanoNQ path: queries/NanoNQ.parquet - split: NanoQuoraRetrieval path: queries/NanoQuoraRetrieval.parquet - split: NanoSCIDOCS path: queries/NanoSCIDOCS.parquet - split: NanoSciFact path: queries/NanoSciFact.parquet - split: NanoTouche2020 path: queries/NanoTouche2020.parquet language: - en tags: - information-retrieval - retrieval - nano - bm25 - hakari-bench dataset_info: - config_name: corpus features: - name: _id dtype: string - name: text dtype: string splits: - name: NanoArguAna num_bytes: 3854860 num_examples: 3635 - name: NanoClimateFEVER num_bytes: 5617105 num_examples: 3408 - name: NanoDBPedia num_bytes: 2280448 num_examples: 6045 - name: NanoFEVER num_bytes: 6285894 num_examples: 4996 - name: NanoFiQA2018 num_bytes: 4201493 num_examples: 4598 - name: NanoHotpotQA num_bytes: 1868234 num_examples: 5090 - name: NanoMSMARCO num_bytes: 1745074 num_examples: 5043 - name: NanoNFCorpus num_bytes: 4521394 num_examples: 2953 - name: NanoNQ num_bytes: 2740852 num_examples: 5035 - name: NanoQuoraRetrieval num_bytes: 346228 num_examples: 5046 - name: NanoSCIDOCS num_bytes: 2149671 num_examples: 2210 - name: NanoSciFact num_bytes: 4227131 num_examples: 2919 - name: NanoTouche2020 num_bytes: 12592148 num_examples: 5745 download_size: 30724168 dataset_size: 52430532 - config_name: queries features: - name: _id dtype: string - name: text dtype: string splits: - name: NanoArguAna num_bytes: 62331 num_examples: 50 - name: NanoClimateFEVER num_bytes: 7044 num_examples: 50 - name: NanoDBPedia num_bytes: 2723 num_examples: 50 - name: NanoFEVER num_bytes: 2948 num_examples: 50 - name: NanoFiQA2018 num_bytes: 3531 num_examples: 50 - name: NanoHotpotQA num_bytes: 6019 num_examples: 50 - name: NanoMSMARCO num_bytes: 2328 num_examples: 50 - name: NanoNFCorpus num_bytes: 1939 num_examples: 50 - name: NanoNQ num_bytes: 3132 num_examples: 50 - name: NanoQuoraRetrieval num_bytes: 3087 num_examples: 50 - name: NanoSCIDOCS num_bytes: 6041 num_examples: 50 - name: NanoSciFact num_bytes: 5352 num_examples: 50 - name: NanoTouche2020 num_bytes: 2609 num_examples: 49 download_size: 97288 dataset_size: 109084 --- # NanoBEIR-en This dataset is a Nano-style retrieval dataset. Nano-series evaluation can be run easily with [HAKARI-Bench](https://github.com/hotchpotch/hakari-bench). NanoBEIR-en is derived from MNanoBEIR / NanoBEIR. It follows the Hugging Face Datasets layout convention used by [sentence-transformers/NanoBEIR-en](https://huggingface.co/datasets/sentence-transformers/NanoBEIR-en): each Nano split has separate `corpus`, `queries`, and `qrels` tables, and BM25 candidates are provided separately in a `bm25` table. This layout follows the NanoBEIR-style evaluation approach summarized in [NanoBEIR](https://huggingface.co/blog/sionic-ai/eval-sionic-nano-beir). NanoBEIR-en is a compact English benchmark derived from BEIR retrieval datasets. It keeps the query-corpus-qrels retrieval format while using small task splits for fast, repeatable evaluation. ## Source Links - Final dataset: [hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) ## Data Layout This dataset uses four 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` Each config has the same Nano split names. The `qrels` config is positive-only. Source rows with `score <= 0` are treated as non-relevant or hard-negative annotations and are not included in `qrels`. When the source provides such rows, their documents are preferentially used as hard negatives in the `corpus` config before generic corpus-fill documents. Source hard negatives are sampled deterministically with query round-robin so one query's negative pool does not dominate the corpus. ## Split Statistics Length statistics are computed with `len(str(text))` over the `queries` and `corpus` tables. `std` is the population standard deviation over the rows in each split. | Nano split | Queries | Corpus | Qrels | Query avg | Query std | Query median | Query p25 | Query p75 | Doc avg | Doc std | Doc median | Doc p25 | Doc p75 | |---|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:|---:| | NanoArguAna | 50 | 3635 | 50 | 1201.8 | 398.7 | 1170.5 | 939.2 | 1446.5 | 1011.8 | 566.8 | 904.0 | 622.5 | 1259.0 | | NanoClimateFEVER | 50 | 3408 | 148 | 128.4 | 50.7 | 124.0 | 93.5 | 158.0 | 1619.5 | 888.1 | 1461.0 | 984.0 | 2079.8 | | NanoDBPedia | 50 | 6045 | 1158 | 33.1 | 15.0 | 33.5 | 20.0 | 45.5 | 336.3 | 141.5 | 369.0 | 237.0 | 445.0 | | NanoFEVER | 50 | 4996 | 57 | 45.4 | 15.9 | 43.0 | 36.0 | 53.0 | 1228.7 | 866.3 | 1041.0 | 597.8 | 1649.0 | | NanoFiQA2018 | 50 | 4598 | 123 | 58.5 | 19.8 | 55.0 | 44.0 | 75.2 | 899.6 | 863.5 | 651.0 | 353.0 | 1136.8 | | NanoHotpotQA | 50 | 5090 | 100 | 88.3 | 33.0 | 82.5 | 64.0 | 107.0 | 349.6 | 250.5 | 299.0 | 154.0 | 479.0 | | NanoMSMARCO | 50 | 5043 | 50 | 32.2 | 15.1 | 29.0 | 24.0 | 40.0 | 330.2 | 130.1 | 298.0 | 249.0 | 380.0 | | NanoNFCorpus | 50 | 2953 | 1651 | 21.0 | 13.5 | 16.5 | 9.2 | 31.5 | 1512.7 | 537.1 | 1532.0 | 1214.0 | 1775.0 | | NanoNQ | 50 | 5035 | 57 | 47.0 | 11.2 | 42.5 | 40.0 | 53.0 | 525.6 | 432.7 | 443.0 | 178.5 | 758.0 | | NanoQuoraRetrieval | 50 | 5046 | 70 | 48.0 | 21.0 | 43.5 | 35.0 | 54.8 | 54.8 | 29.2 | 47.0 | 36.0 | 64.0 | | NanoSCIDOCS | 50 | 2210 | 244 | 72.8 | 19.2 | 71.5 | 57.8 | 81.8 | 923.6 | 721.8 | 910.5 | 577.2 | 1230.5 | | NanoSciFact | 50 | 2919 | 56 | 95.8 | 41.3 | 92.5 | 62.2 | 124.8 | 1431.2 | 599.4 | 1344.0 | 1042.0 | 1725.0 | | NanoTouche2020 | 49 | 5745 | 932 | 43.4 | 14.2 | 40.0 | 33.0 | 57.0 | 2142.6 | 2599.9 | 989.0 | 323.0 | 3032.0 | ## Construction Steps This dataset is constructed as follows. 1. Use MNanoBEIR / NanoBEIR as the upstream benchmark or dataset family. 2. Load source datasets from the `hakari-bench/NanoBEIR-en` corpus, queries, and qrels tables. 3. Source evaluation split policy: the NanoBEIR split set. 4. Create one Nano split for each selected source retrieval task. 5. Keep up to 200 eligible queries per Nano split. 6. Treat source relevance rows with `score > 0` as qrels-positive documents. If the source has no score column, treat its qrels as positive-only only when that is the source task convention. 7. Exclude source rows with `score <= 0` from `qrels`. When such rows are available for selected queries, use their documents as hard-negative corpus candidates before generic fill documents. 8. Include all qrels-positive documents for the selected queries. 9. Use the included corpus tables for each Nano split; no additional document resampling is performed. 10. Remove exact duplicate query text and document text within each split. If a removed document duplicate was referenced by qrels, the qrels row was removed. 11. Store corpus text as `title` plus body text when available. 12. Generate BM25 top-100 candidates with `regex` tokenization. 13. If a qrels-positive document is missing from the raw BM25 result, insert it into the final `bm25` candidate list by replacing a tail non-positive candidate. The `qrels` config is positive-only. When needed for top-k reranking coverage, positive qrels are capped per query to the BM25 top-k. The `bm25` candidate subset is generated from the included corpus for each split. For top-100 reranking diagnostics, positive qrels are capped to at most 100 documents per query before BM25 positive forcing. This makes full relevant coverage possible for splits whose upstream qrels contain more than 100 positives for a single query. ## BM25 Subset Policy The `bm25` config is a candidate subset for first-stage retrieval and reranking. It is not a separate source dataset. Each row contains one query id and a ranked list of up to 100 corpus ids. BM25 candidates are generated from the selected corpus for each split. When a qrels-positive document is not present in the raw BM25 top-100 results, the missing positive is forced into the final candidate list by replacing a tail candidate that is not positive for that query. Candidate ids are kept unique after replacement. Concretely, each `bm25` row is produced by tokenizing the selected split corpus and query texts with `regex`, ranking the corpus with BM25, then writing the ranked corpus ids as `corpus-ids` for that query. The list is a candidate subset for downstream evaluation, not a full-corpus ranking. Source hard negatives, including documents referenced by source rows with `score <= 0`, may appear in the selected corpus and can naturally appear in BM25 candidates. They are still non-relevant and are not listed in `qrels`. When source hard negatives are available, the default corpus sampling policy is query round-robin: group hard negatives by selected query, preserve source rank/order within each query, add at most one new hard negative from each query per pass, remove duplicate IDs and exact duplicate text, then fill any remaining slots from source corpus order. ## Split Mapping Each Nano split maps to one source retrieval task unless noted otherwise. | Nano split | Source task | Source dataset | Queries | Corpus | Qrels | |---|---|---|---:|---:|---:| | NanoArguAna | NanoArguAna | [hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) | 50 | 3635 | 50 | | NanoClimateFEVER | NanoClimateFEVER | [hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) | 50 | 3408 | 148 | | NanoDBPedia | NanoDBPedia | [hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) | 50 | 6045 | 1158 | | NanoFEVER | NanoFEVER | [hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) | 50 | 4996 | 57 | | NanoFiQA2018 | NanoFiQA2018 | [hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) | 50 | 4598 | 123 | | NanoHotpotQA | NanoHotpotQA | [hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) | 50 | 5090 | 100 | | NanoMSMARCO | NanoMSMARCO | [hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) | 50 | 5043 | 50 | | NanoNFCorpus | NanoNFCorpus | [hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) | 50 | 2953 | 1651 | | NanoNQ | NanoNQ | [hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) | 50 | 5035 | 57 | | NanoQuoraRetrieval | NanoQuoraRetrieval | [hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) | 50 | 5046 | 70 | | NanoSCIDOCS | NanoSCIDOCS | [hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) | 50 | 2210 | 244 | | NanoSciFact | NanoSciFact | [hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) | 50 | 2919 | 56 | | NanoTouche2020 | NanoTouche2020 | [hakari-bench/NanoBEIR-en](https://huggingface.co/datasets/hakari-bench/NanoBEIR-en) | 49 | 5745 | 932 | ## BM25 nDCG@10 `nDCG@10` is computed from the included BM25 ranking against the included qrels. Coverage is measured against included qrels at the same BM25 top-k used for reranking diagnostics. The included BM25 candidate subset has 100.00% query coverage and 100.00% relevant coverage for every split. | Nano split | Tokenizer | Forced BM25 positives | Query cov | Relevant cov | BM25 nDCG@10 | |---|---|---:|---:|---:|---:| | NanoArguAna | regex | 0 | 100.00% | 100.00% | 0.4462 | | NanoClimateFEVER | regex | 65 | 100.00% | 100.00% | 0.3289 | | NanoDBPedia | regex | 398 | 100.00% | 100.00% | 0.5619 | | NanoFEVER | regex | 0 | 100.00% | 100.00% | 0.8141 | | NanoFiQA2018 | regex | 39 | 100.00% | 100.00% | 0.3583 | | NanoHotpotQA | regex | 5 | 100.00% | 100.00% | 0.8176 | | NanoMSMARCO | regex | 1 | 100.00% | 100.00% | 0.4890 | | NanoNFCorpus | regex | 1391 | 100.00% | 100.00% | 0.3620 | | NanoNQ | regex | 6 | 100.00% | 100.00% | 0.4708 | | NanoQuoraRetrieval | regex | 0 | 100.00% | 100.00% | 0.7864 | | NanoSCIDOCS | regex | 97 | 100.00% | 100.00% | 0.3360 | | NanoSciFact | regex | 5 | 100.00% | 100.00% | 0.7174 | | NanoTouche2020 | regex | 181 | 100.00% | 100.00% | 0.7072 | ## Skipped Tasks No source tasks were skipped. ## License NanoBEIR-en is a derived dataset. Users must comply with the licenses, terms, and attribution requirements of the upstream source datasets.