metadata

dataset_info:
  - config_name: qrels_og-fas
    features:
      - name: query-id
        dtype: string
      - name: corpus-id
        dtype: string
      - name: score
        dtype: float64
    splits:
      - name: test
        num_bytes: 470546
        num_examples: 12163
    download_size: 470546
    dataset_size: 470546
  - config_name: qrels_changed-fas
    features:
      - name: query-id
        dtype: string
      - name: corpus-id
        dtype: string
      - name: score
        dtype: float64
    splits:
      - name: test
        num_bytes: 470157
        num_examples: 12163
    download_size: 470157
    dataset_size: 470157
  - config_name: corpus-fas
    features:
      - name: _id
        dtype: string
      - name: title
        dtype: string
      - name: text
        dtype: string
    splits:
      - name: corpus
        num_bytes: 108126004
        num_examples: 41189
    download_size: 108126004
    dataset_size: 108126004
  - config_name: queries-fas
    features:
      - name: _id
        dtype: string
      - name: text
        dtype: string
      - name: instruction_og
        dtype: string
      - name: instruction_changed
        dtype: string
    splits:
      - name: queries
        num_bytes: 28833
        num_examples: 40
    download_size: 28833
    dataset_size: 28833
  - config_name: top_ranked-fas
    features:
      - name: qid
        dtype: string
      - name: pid
        dtype: string
    splits:
      - name: top_ranked
        num_bytes: 1531154
        num_examples: 40000
    download_size: 1531154
    dataset_size: 1531154
  - config_name: qrels_og-rus
    features:
      - name: query-id
        dtype: string
      - name: corpus-id
        dtype: string
      - name: score
        dtype: float64
    splits:
      - name: test
        num_bytes: 466885
        num_examples: 12067
    download_size: 466885
    dataset_size: 466885
  - config_name: qrels_changed-rus
    features:
      - name: query-id
        dtype: string
      - name: corpus-id
        dtype: string
      - name: score
        dtype: float64
    splits:
      - name: test
        num_bytes: 466487
        num_examples: 12067
    download_size: 466487
    dataset_size: 466487
  - config_name: corpus-rus
    features:
      - name: _id
        dtype: string
      - name: title
        dtype: string
      - name: text
        dtype: string
    splits:
      - name: corpus
        num_bytes: 100142813
        num_examples: 39326
    download_size: 100142813
    dataset_size: 100142813
  - config_name: queries-rus
    features:
      - name: _id
        dtype: string
      - name: text
        dtype: string
      - name: instruction_og
        dtype: string
      - name: instruction_changed
        dtype: string
    splits:
      - name: queries
        num_bytes: 27729
        num_examples: 40
    download_size: 27729
    dataset_size: 27729
  - config_name: top_ranked-rus
    features:
      - name: qid
        dtype: string
      - name: pid
        dtype: string
    splits:
      - name: top_ranked
        num_bytes: 1531215
        num_examples: 40000
    download_size: 1531215
    dataset_size: 1531215
  - config_name: qrels_og-zho
    features:
      - name: query-id
        dtype: string
      - name: corpus-id
        dtype: string
      - name: score
        dtype: float64
    splits:
      - name: test
        num_bytes: 492178
        num_examples: 12732
    download_size: 492178
    dataset_size: 492178
  - config_name: qrels_changed-zho
    features:
      - name: query-id
        dtype: string
      - name: corpus-id
        dtype: string
      - name: score
        dtype: float64
    splits:
      - name: test
        num_bytes: 491764
        num_examples: 12732
    download_size: 491764
    dataset_size: 491764
  - config_name: corpus-zho
    features:
      - name: _id
        dtype: string
      - name: title
        dtype: string
      - name: text
        dtype: string
    splits:
      - name: corpus
        num_bytes: 83950020
        num_examples: 41120
    download_size: 83950020
    dataset_size: 83950020
  - config_name: queries-zho
    features:
      - name: _id
        dtype: string
      - name: text
        dtype: string
      - name: instruction_og
        dtype: string
      - name: instruction_changed
        dtype: string
    splits:
      - name: queries
        num_bytes: 29669
        num_examples: 43
    download_size: 29669
    dataset_size: 29669
  - config_name: top_ranked-zho
    features:
      - name: qid
        dtype: string
      - name: pid
        dtype: string
    splits:
      - name: top_ranked
        num_bytes: 1646082
        num_examples: 43000
    download_size: 1646082
    dataset_size: 1646082
configs:
  - config_name: qrels_og-fas
    data_files:
      - split: test
        path: qrels_og-fas/test.parquet
  - config_name: qrels_changed-fas
    data_files:
      - split: test
        path: qrels_changed-fas/test.parquet
  - config_name: corpus-fas
    data_files:
      - split: corpus
        path: corpus-fas/corpus.parquet
  - config_name: queries-fas
    data_files:
      - split: queries
        path: queries-fas/queries.parquet
  - config_name: top_ranked-fas
    data_files:
      - split: top_ranked
        path: top_ranked-fas/top_ranked.parquet
  - config_name: qrels_og-rus
    data_files:
      - split: test
        path: qrels_og-rus/test.parquet
  - config_name: qrels_changed-rus
    data_files:
      - split: test
        path: qrels_changed-rus/test.parquet
  - config_name: corpus-rus
    data_files:
      - split: corpus
        path: corpus-rus/corpus.parquet
  - config_name: queries-rus
    data_files:
      - split: queries
        path: queries-rus/queries.parquet
  - config_name: top_ranked-rus
    data_files:
      - split: top_ranked
        path: top_ranked-rus/top_ranked.parquet
  - config_name: qrels_og-zho
    data_files:
      - split: test
        path: qrels_og-zho/test.parquet
  - config_name: qrels_changed-zho
    data_files:
      - split: test
        path: qrels_changed-zho/test.parquet
  - config_name: corpus-zho
    data_files:
      - split: corpus
        path: corpus-zho/corpus.parquet
  - config_name: queries-zho
    data_files:
      - split: queries
        path: queries-zho/queries.parquet
  - config_name: top_ranked-zho
    data_files:
      - split: top_ranked
        path: top_ranked-zho/top_ranked.parquet

mFollowIR-cross-lingual-parquet

This is a parquet version of the mFollowIR cross-lingual dataset that can be loaded directly with load_dataset(). The original dataset can be found at jhu-clsp/mFollowIR-cross-lingual.

Dataset Structure

The dataset contains the following configurations for each target language (fas, rus, zho):

Configurations

qrels_og_[lang]: Original relevance judgments (test split)
qrels_changed_[lang]: Modified relevance judgments (test split)
corpus_[lang]: Document collection
queries_[lang]: Query set with instructions
top_ranked_[lang]: Top ranked documents

Usage

from datasets import load_dataset

# Load a specific configuration
dataset = load_dataset("jhu-clsp/mFollowIR-cross-lingual-parquet", "queries_fas")  # or any other config

# Load multiple configurations
dataset = load_dataset("jhu-clsp/mFollowIR-cross-lingual-parquet", 
                      ["queries_fas", "corpus_fas"])

Citation

@article{weller2024mfollowir,
  title="mFollowIR: a Multilingual Benchmark for Instruction Following in Information Retrieval",
  author="Weller, Orion and Chang, Benjamin and Yang, Eugene and Yarmohammadi, Mahsa and Barham, Sam and MacAvaney, Sean and Cohan, Arman and Soldaini, Luca and Van Durme, Benjamin and Lawrie, Dawn",
  journal="arXiv preprint TODO",
  year="2024"
}