metadata

license: cc-by-nc-4.0
language:
  - ko
pretty_name: KorEmpatheticDialogues
tags:
  - social dialogue
task_ids:
  - conversational
size_categories:
  - 10K<n<100K
source_datasets:
  - EmpatheticDialogues
splits:
  - name: train
    num_examples: 19531
  - name: valid
    num_examples: 2769
  - name: test
    num_examples: 2547
dataset_size: 24,847

Dataset Card for 💗 KorEmpatheticDialogues

🚨 Disclaimer: All models and datasets are intended for research purposes only.

Dataset Description

Repository: Code
Paper: Paper
Point of Contact: Young-Jun Lee

Dataset Summary

KorEmpatheticDialogues is a publicly available Korean empathetic dialogue dataset translated from the original EmpatheticDialogues dataset using the DeepL translator API.

Languages

Korean

Dataset Structure

field	type	description
`dialogue_id`	int	the identifier for the dialogue
`dialogue`	list of dict	the dialogue where each dict entry includes {utter_idx, utter, user_id}
`situation`	str	the emotional situation sentence
`emotion`	str	the emotion category (e.g., guilty, caring, etc)

Acknowledgements

This work was supported by a grant from the KAIST-KT joint research project through AI Tech Lab, Institute of Convergence Technology, funded by KT [Project No. G01230605, Development of Task-oriented Persona-based Dialogue Generation Combining Multi-modal Interaction and Knowledge Modeling].

Citation

Please cite our work if you find the resources in this repository useful:

@inproceedings{lee2023language,
  title={Language Model Evaluation Based on Korean-English Empathetic Dialogue Datasets and Personality},
  author={Lee, Young-Jun and Hyeon, JongHwan and Lee, DoKyong and Sung, Joo-Won and Choi, Ho-Jin},
  booktitle={Annual Conference on Human and Language Technology},
  pages={312--318},
  year={2023},
  organization={Human and Language Technology}
}