ckverma/snli_universal
收藏Hugging Face2024-12-02 更新2024-12-14 收录
下载链接：
https://hf-mirror.com/datasets/ckverma/snli_universal
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含7个子集，这些子集基于SNLI数据集构建。其中包括从SNLI训练集中随机抽取的6000个样本，其中3000个样本保持不变，另外3000个样本在假设部分前添加了通用触发器。此外，还包括从SNLI验证集中随机抽取的蕴含、中性和矛盾样本，以及这些样本在假设部分前添加了通用触发器的版本。
This dataset consists of 7 splits, based on the SNLI dataset. It includes 6,000 examples randomly sampled from the train split of the SNLI dataset, with 3,000 examples left unmodified and universal triggers prepended to the hypothesis of the remaining 3,000 examples. Additionally, it includes entailment, neutral, and contradiction examples randomly chosen from the validation split of the SNLI dataset, as well as versions of these examples with universal triggers prepended to their hypotheses.
提供机构：
ckverma
