This is the folder for features, labels, splits, etc., for downstream tasks.

For convenience, we include the splits for open access data.

For other datasets, when you run the task for the first time (with the dataset downloaded), the folder will be created automatically.