mimir-project/noridiom

该数据集包含803个挪威谚语，分别用挪威的两种语言形式nynorsk和bokmål表示，每种形式各有401和402个谚语。数据集是专门为Mímir项目创建的，用于评估语言模型在完成挪威谚语方面的能力。每个样本包括谚语的前n-1个词（idiom特征）和完成谚语的最后一个词（completion特征）。数据集在cc-by-4.0协议下授权使用，适用于文本生成任务。

This dataset contains 803 Norwegian idioms, represented in both nynorsk and bokmål forms of the Norwegian language, with 401 in nynorsk and 402 in bokmål. It was created for the Mímir project to evaluate language models ability to complete Norwegian idioms. Each sample includes the first n-1 words of an idiom (the idiom feature) and the last word that completes the idiom (the completion feature). The dataset is licensed under cc-by-4.0 and is intended for text generation tasks.