RAD

该数据集包含用于训练和测试目的的RAD DFAs，其状态数量各不相同。数据集不仅包括来自同一分布的内部分布DFAs，还包含来自不同分布的外部分布DFAs。在训练集中，DFA的状态数量最多可达10个；而在OOD（即外部分布）DFAs中，状态数量介于11至20个之间。该数据集的任务是测试DFA嵌入的准确性和泛化能力。

This dataset contains RAD DFAs with varying numbers of states for training and testing purposes. It includes not only in-distribution (ID) DFAs from the same distribution, but also out-of-distribution (OOD) DFAs from different distributions. In the training set, the number of states per DFA is up to 10; for OOD DFAs, the state count ranges from 11 to 20. The task of this dataset is to evaluate the accuracy and generalization ability of DFA embeddings.