---
dataset_info:
  features:
  - name: Utterance
    dtype: string
  - name: politeness
    dtype: float64
  - name: language
    dtype: string
  - name: __index_level_0__
    dtype: int64
  splits:
  - name: train
    num_bytes: 4954742
    num_examples: 18238
  - name: validation
    num_bytes: 619659
    num_examples: 2280
  - name: test
    num_bytes: 619627
    num_examples: 2280
  download_size: 3919716
  dataset_size: 6194028
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
---
# Dataset Card for "multilingual_politeness"


This dataset has the following attributes:
- Utterance: a 2-3 sentence excerpt from Wikipedia editor talk pages.
- language: English, Spanish, Chinese, or Japanese
- politeness: An annotated politeness label ranging from -2 (very rude) to 2 (very polite). Each utterance was annotated by 3 native speakers.



Please cite the following paper if you use our dataset :)

```
@inproceedings{havaldar-etal-2023-comparing,
    title = "Comparing Styles across Languages",
    author = "Havaldar, Shreya  and Pressimone, Matthew  and Wong, Eric  and Ungar, Lyle",
    booktitle = "Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing",
    year = "2023",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2023.emnlp-main.419"
}
```