DaLAJ

该数据集是一个包含句子被分类为语言正确或错误的瑞典语语言可接受性数据集。在其它信息方面，值得注意的是，仅对语言上不正确的训练样本进行了增强处理。具体规模上，该数据集包含了7,682个训练样本和888个测试样本。其任务是进行句法分析。

This is a Swedish linguistic acceptability dataset where sentences are classified as either linguistically correct or incorrect. Notably, only the linguistically incorrect training samples were augmented. In terms of scale, the dataset contains 7,682 training samples and 888 test samples. The task of this dataset is syntactic analysis.