This dataset card aims to be a base template for new datasets. It has been generated using this raw template.
As the ComplexWebQuestions (CWQ) dataset [
https://www.tau-nlp.sites.tau.ac.il/compwebq] didn't have a publicly available test set ground truth, and different methods used the self-generated test ground truth are not matched and uniform. Here we provide the in-house splitting, splitting 11.3% from the original train for the test.… See the full description on the dataset page:
https://huggingface.co/datasets/Vickyrana-12/CWQ.
