ClarQ: A large-scale and diverse dataset for Clarification Question Generation

This dataset is meant for training and evaluation of Clarification Question Generation Systems. The details and the methodology used in the creation of the dataset can be found in the paper. The work was published in ACL 2020.

Link to the Dataset

The dataset can be found at https://drive.google.com/drive/folders/1aqTiRgFq1pGVZhqJ_rDksZZg8v8m_3eX?usp=sharing.

Details

There are two files in the above link. "train.json" is meant for training whereas "test.json" is meant for evaluation. Each line in the file is a json consisting the following keys:`

Key	Description
id	Id of the example
context	Text of the post
cquestion	The corresponding clarification question to the post
answer	The answer to the post

License

This dataset is licensed under the Creative Commons Attribution 4.0 International License.

References

If you find the data useful and use it for your work, please consider citing the following:

@misc{kumar2020clarq,
    title={ClarQ: A large-scale and diverse dataset for Clarification Question Generation},
    author={Vaibhav Kumar and Alan W. black},
    year={2020},
    eprint={2006.05986},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}