mwhanna/greater-than

该数据集名为Greater-Than circuit task，包含三个特征：clean表示原始句子，corrupted表示损坏版本的句子，year表示起始年份。数据集由Michael Hanna和Ollie Liu创建，使用模板自动生成，旨在提供一个模型无关的greater-than任务版本。数据集的语言为英语，采用MIT许可证。

This dataset, named Greater-Than circuit task, contains three features: clean for the original sentences, corrupted for the corrupted version of the sentences, and year for the start year. It was created by Michael Hanna and Ollie Liu using a template to automatically generate examples, aiming to provide a model-agnostic version of the greater-than task. The dataset is in English and licensed under MIT.