|
--- |
|
license: mit |
|
--- |
|
|
|
The dataset in `followup-questions.json` file. |
|
|
|
# Follow-Up Questions Dataset |
|
## Description |
|
This dataset is derived from the QuAC (Question Answering in Context) dataset and has been processed to pair previous answers with follow-up questions. It is intended for research and development in natural language processing, specifically for training and evaluating models on generating or understanding follow-up questions in conversational contexts. |
|
|
|
# Source |
|
The original QuAC dataset is available at QuAC.ai. The QuAC dataset is distributed under the CC-BY-4.0 license. |
|
|
|
Please cite the original QuAC dataset paper when using this dataset: |
|
``` |
|
@article{choi2018quac, |
|
title={QuAC: Question Answering in Context}, |
|
author={Choi, Eunsol and He, He and Iyyer, Mohit and Yatskar, Mark and Yih, Wen-tau and Choi, Yejin and Liang, Percy and Zettlemoyer, Luke}, |
|
journal={arXiv preprint arXiv:1808.07036}, |
|
year={2018} |
|
} |
|
``` |
|
# Modifications |
|
The original QuAC dataset has been processed to create pairs of previous answers and follow-up questions. The first question in each conversation has been removed, as it does not have a preceding answer. |
|
|
|
# Structure |
|
Each row in the dataset represents a pair of a previous answer and a follow-up question. The columns are as follows: |
|
- `prev_answer`: The answer to the previous question in the conversation. |
|
- `followup_question`: The follow-up question related to the previous answer. |
|
|
|
# Usage |
|
This dataset can be used for training and evaluating models on tasks related to follow-up question generation or understanding in conversational contexts. |
|
|
|
# License |
|
This dataset is distributed under the CC-BY-4.0 license, as it is derived from the QuAC dataset, which is also distributed under the same license. |
