数据集文件都是 JSONL 格式（每行一个 JSON）。下面是一个带有不言自明的字段的示例（JSON 格式）。
请注意，文件中的每个示例（每一行）都包含一个uid字段，该字段代表所有轮次 ANLI 中所有示例的唯一 ID。

{   
    "uid": "8a91e1a2-9a32-4fd9-b1b6-bd2ee2287c8f", 
    "premise": "Javier Torres (born May 14, 1988 in Artesia, California) is an undefeated Mexican American professional boxer in the Heavyweight division. 
                Torres was the second rated U.S. amateur boxer in the Super Heavyweight division and a member of the Mexican Olympic team.", 
    "hypothesis": "Javier was born in Mexico", 
    "label": "c", 
    "reason": "The paragraph states that Javier was born in the California, US."
}

 

AdversarialNLI 数据集包含一个原因字段，用于拆分和拆分中的每个示例dev以及test拆分中的一些示例train。原因是通过询问注释者“请为您的声明属于该类别的原因以及您认为系统困难的原因写一个原因”来收集。

 

引文
@inproceedings{nie-etal-2020-adversarial,
    title = "Adversarial {NLI}: A New Benchmark for Natural Language Understanding",
    author = "Nie, Yixin  and
      Williams, Adina  and
      Dinan, Emily  and
      Bansal, Mohit  and
      Weston, Jason  and
      Kiela, Douwe",
    booktitle = "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics",
    year = "2020",
    publisher = "Association for Computational Linguistics",
}