Passage Corpus (Pyserini Format) for the BrowseComp-Plus Dataset

Total downloads since release, read live from the Hugging Face Hub API (downloadsAllTime)

This repository provides the passage corpus for the BrowseComp-Plus dataset in a format compatible with Pyserini, used in the paper Revisiting Text Ranking in Deep Research, which has been accepted at SIGIR 2026, the 49th International ACM SIGIR Conference on Research and Development in Information Retrieval.

• Code: GitHub Repository

The corpus consists of 2,772,255 passages.

Unlike the version available at
https://huggingface.co/datasets/grill-lab/browsecomp-plus-passage-corpus, this release follows the Pyserini JSONL data format.

Each item contains two fields:

• id: the unique passage identifier.
• contents: the concatenation of the source document title and the passage text.

This format is directly compatible with Pyserini BM25.

Sample Usage

As described in the official repository, you can build a Lucene index using Pyserini with the following command:

python -m pyserini.index.lucene \
--collection JsonCollection \
--input /path/to/downloaded/data/ \
--index ./indexes/index.bm25.passage \
--generator DefaultLuceneDocumentGenerator \
--threads 16 \
--storePositions --storeDocvectors --storeRaw

Contact

If you have any questions or suggestions, please contact:

• Chuan Meng: chuan.meng@ed.ac.uk
• Litu Ou: litu.ou@ed.ac.uk

Citation

If you find this work useful, please cite:

@inproceedings{meng2026revisiting,
  title={Revisiting Text Ranking in Deep Research},
  author={Meng, Chuan and Ou, Litu and MacAvaney, Sean and Dalton, Jeff},
  booktitle={Proceedings of the 49th International ACM SIGIR Conference on Research and Development in Information Retrieval},
  pages = {3006--3016},
  url = {https://doi.org/10.1145/3805712.3808557},
  doi = {10.1145/3805712.3808557},
  year={2026}
}