README.md

metadata

dataset_info:
  features:
    - name: id
      dtype: int32
    - name: label
      dtype:
        class_label:
          names:
            '0': REFUTES
            '1': NOT ENOUGH INFO
            '2': SUPPORTS
    - name: evidence
      dtype: string
    - name: claim
      dtype: string
  splits:
    - name: train
      num_bytes: 2131310
      num_examples: 3626
    - name: validation
      num_bytes: 242530
      num_examples: 482
    - name: test
      num_bytes: 287437
      num_examples: 558
  download_size: 891646
  dataset_size: 2661277
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*
language:
  - en
pretty_name: CTKFacts for NLI - English translation
size_categories:
  - 1K<n<10K

CTKFacts for NLI - English translation

This is an English translation of the original CTKFacts dataset for NLI, created using the WMT 21 X-En model.

The translation was completed for use within the Czech-Bench evaluation framework. The script used for translation can be reviewed here.

Citation

Original dataset:

@article{Ullrich_2023,
   title={CsFEVER and CTKFacts: acquiring Czech data for fact verification},
   volume={57},
   ISSN={1574-0218},
   url={http://dx.doi.org/10.1007/s10579-023-09654-3},
   DOI={10.1007/s10579-023-09654-3},
   number={4},
   journal={Language Resources and Evaluation},
   publisher={Springer Science and Business Media LLC},
   author={Ullrich, Herbert and Drchal, Jan and Rýpar, Martin and Vincourová, Hana and Moravec, Václav},
   year={2023},
   month=may,
   pages={1571–1605}
}

English translation:

@masterthesis{jirkovsky-thesis,
    author = {Jirkovský, Adam},
    title = {Benchmarking Techniques for Evaluation of Large Language Models},
    school = {Czech Technical University in Prague, Faculty of Electrical Engineering},
    year = 2024,
    URL = {https://dspace.cvut.cz/handle/10467/115227}
}