CompCert

该数据集是从CompCert仓库中提取的，采用了特定的训练-测试分割方式，其目的是为了进行证明步骤的预测。此外，该数据集还用于多语言和Coq模型的训练过程中。This dataset is extracted from the CompCert repository and adopts a specific train-test split strategy, which is intended for proof step prediction. Furthermore, it is also employed in the training processes of multilingual and Coq models.