metadata

annotations_creators:
  - expert-generated
language_creators:
  - expert-generated
language:
  - eu
license:
  - cc-by-4.0
multilinguality:
  - monolingual
size_categories:
  - n<1K
source_datasets:
  - extended|glue
task_categories:
  - text-classification
task_ids:
  - natural-language-inference
pretty_name: WNLI-eu
config_names:
  - wnli
language_details: eu-ES
configs:
  - config_name: wnli
    data_files:
      - split: validation
        path: wnli_validation.jsonl
      - split: test
        path: wnli_test.jsonl
    default: true
dataset_info:
  config_name: wnli
  features:
    - name: sentence1
      dtype: string
    - name: sentence2
      dtype: string
    - name: label
      dtype:
        class_label:
          names:
            '0': not_entailment
            '1': entailment
    - name: idx
      dtype: int32
  splits:
    - name: validation
      num_bytes: 12960
      num_examples: 71
    - name: test
      num_bytes: 41450
      num_examples: 146
  download_size: 60587
  dataset_size: 54410

Dataset Card for WNLI-eu

Point of Contact: [email protected]

Dataset Description

Dataset Summary

WNLI-eu is the professional translation to Basque of the WNLI dataset. WNLI is part of the GLUE benchmark for English (Wang et al., 2018) and is based on the Winograd Schema Challenge (WSC) dataset (Levesque et al., 2011):

A Winograd schema is a pair of sentences differing in only one or two words and containing an ambiguity that is resolved in opposite ways in the two sentences and that requires the use of world knowledge and reasoning for its resolution.

Specifically, WNLI instances consist of Winograd schemas converted into NLI sentence-pair classification problems.

Languages

eu-ES

Dataset Structure

Data Instances

WNLI-eu examples look like this:

{
  "sentence1": "Isurbidea ilearekin buxatuta dago. Garbitu egin behar da.",
  "sentence2": "Ilea garbitu behar da.",
  "label": 0,
  "idx": 0
}

Data Fields

sentence1 (str): the premise.
sentence2 (str): second sentence.
label (int): 1 or 0, whether sentence1 entails sentence2 or not, respectively.
idx (int): index of the example.

Data Splits

name	validation	test
default	71	146

Dataset Creation

This dataset is a professional translation of the English WNLI dataset into Basque, commissioned by HiTZ (UPV/EHU) within the ILENIA project. For more information on how WNLI and the Winograd Schema Challenge dataset were created, please refer to their articles (see above).

Additional Information

This work is funded by the Ministerio para la Transformación Digital y de la Función Pública - Funded by EU – NextGenerationEU within the framework of the project ILENIA with reference 2022/TL22/00215335.

Licensing Information

Creative Commons Attribution 4.0 International (CC BY 4.0)

Citation Information

@inproceedings{baucells-etal-2025-iberobench,
    title = "{I}bero{B}ench: A Benchmark for {LLM} Evaluation in {I}berian Languages",
    author = "Baucells, Irene  and
      Aula-Blasco, Javier  and
      de-Dios-Flores, Iria  and
      Paniagua Su{\'a}rez, Silvia  and
      Perez, Naiara  and
      Salles, Anna  and
      Sotelo Docio, Susana  and
      Falc{\~a}o, J{\'u}lia  and
      Saiz, Jose Javier  and
      Sepulveda Torres, Robiert  and
      Barnes, Jeremy  and
      Gamallo, Pablo  and
      Gonzalez-Agirre, Aitor  and
      Rigau, German  and
      Villegas, Marta",
    editor = "Rambow, Owen  and
      Wanner, Leo  and
      Apidianaki, Marianna  and
      Al-Khalifa, Hend  and
      Eugenio, Barbara Di  and
      Schockaert, Steven",
    booktitle = "Proceedings of the 31st International Conference on Computational Linguistics",
    month = jan,
    year = "2025",
    address = "Abu Dhabi, UAE",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2025.coling-main.699/",
    pages = "10491--10519",
}