README.md

---
license: cc-by-4.0
task_categories:
- text-retrieval
- summarization
language:
- en
tags:
- legal
- law
size_categories:
- n<1K
source_datasets:
- launch/gov_reports
dataset_info:
- config_name: default
  features:
  - name: query-id
    dtype: string
  - name: corpus-id
    dtype: string
  - name: score
    dtype: float64
  splits:
  - name: test
    num_examples: 973
- config_name: corpus
  features:
  - name: _id
    dtype: string
  - name: title
    dtype: string
  - name: text
    dtype: string
  splits:
  - name: corpus
    num_examples: 973
- config_name: queries
  features:
  - name: _id
    dtype: string
  - name: text
    dtype: string
  splits:
  - name: queries
    num_examples: 970
configs:
- config_name: default
  data_files:
  - split: test
    path: data/default.jsonl
- config_name: corpus
  data_files:
  - split: corpus
    path: data/corpus.jsonl
- config_name: queries
  data_files:
  - split: queries
    path: data/queries.jsonl
pretty_name: GovReport (MTEB format)
---
# GovReport (MTEB format)
This is the test split of the [GovReport](https://huggingface.co/datasets/launch/gov_report) dataset formatted in the [Massive Text Embedding Benchmark (MTEB)](https://github.com/embeddings-benchmark/mteb) information retrieval dataset format.

This dataset is intended to facilitate the consistent and reproducible evaluation of information retrieval models on GovReport with the [`mteb`](https://github.com/embeddings-benchmark/mteb) embedding model evaluation framework.

More specifically, this dataset tests the ability of information retrieval models to retrieve US government reports from their summaries.

This dataset has been processed into the MTEB format by [Isaacus](https://isaacus.com/), a legal AI research company.

## Methodology 🧪
To understand how GovReport was created, refer to its creators' [paper](https://arxiv.org/abs/2104.02112).

This dataset was formatted by treating the `summary` column of GovReport as queries (or anchors) and the `document` column as relevant (or positive) passages.

## Structure 🗂️
As per the MTEB information retrieval dataset format, this dataset comprises three splits, `default`, `corpus` and `queries`.

The `default` split pairs summaries (`query-id`, linked to the `_id` column of the `queries` split) with government reports (`corpus-id`, linked to the `_id` column of the `corpus` split), each pair having a `score` of `1`.

The `corpus` split contains government reports, with the text of a report being stored in the `text` key and its id being stored in the `_id` key.

The `queries` split contains summaries, with the text of a summary being stored in the `text` key and its id being stored in the `_id` key.

## License 📜
This dataset is licensed under [CC BY 4.0](https://choosealicense.com/licenses/cc-by-4.0/).

## Citation 🔖
```bibtex
@inproceedings{huang-etal-2021-efficient,
    title = "Efficient Attentions for Long Document Summarization",
    author = "Huang, Luyang  and
      Cao, Shuyang  and
      Parulian, Nikolaus  and
      Ji, Heng  and
      Wang, Lu",
    booktitle = "Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies",
    month = jun,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.naacl-main.112",
    doi = "10.18653/v1/2021.naacl-main.112",
    pages = "1419--1436",
    abstract = "The quadratic computational and memory complexities of large Transformers have limited their scalability for long document summarization. In this paper, we propose Hepos, a novel efficient encoder-decoder attention with head-wise positional strides to effectively pinpoint salient information from the source. We further conduct a systematic study of existing efficient self-attentions. Combined with Hepos, we are able to process ten times more tokens than existing models that use full attentions. For evaluation, we present a new dataset, GovReport, with significantly longer documents and summaries. Results show that our models produce significantly higher ROUGE scores than competitive comparisons, including new state-of-the-art results on PubMed. Human evaluation also shows that our models generate more informative summaries with fewer unfaithful errors.",
    eprint={2104.02112}
}
```