Datasets:
annotations_creators:
- crowdsourced
language:
- en
language_creators:
- found
license:
- cc0-1.0
multilinguality:
- monolingual
pretty_name: Toxic Wikipedia Comments
size_categories:
- 100K<n<1M
source_datasets:
- extended|other
tags:
- wikipedia
- toxicity
- toxic comments
task_categories:
- text-classification
task_ids:
- hate-speech-detection
Dataset Card for Wiki Toxic
Table of Contents
- Table of Contents
- Dataset Description
- Dataset Structure
- Dataset Creation
- Considerations for Using the Data
- Additional Information
Dataset Description
- Homepage:
- Repository:
- Paper:
- Leaderboard:
- Point of Contact:
Dataset Summary
The Wiki Toxic dataset is a modified, cleaned version of the dataset used in the Kaggle Toxic Comment Classification challenge from 2017/18. The dataset contains comments collected from Wikipedia forums and classifies them into two categories, toxic
and non-toxic
.
The Kaggle dataset was cleaned using the included clean.py
file.
Supported Tasks and Leaderboards
- Text Classification: the dataset can be used for training a model to recognise toxicity in sentences and classify them accordingly.
Languages
The sole language used in the dataset is English.
Dataset Structure
Data Instances
For each data point, there is an id, the comment_text itself, and a label (0 for non-toxic, 1 for toxic).
{'id': 'a123a58f610cffbc',
'comment_text': '"This article SUCKS. It may be poorly written, poorly formatted, or full of pointless crap that no one cares about, and probably all of the above. If it can be rewritten into something less horrible, please, for the love of God, do so, before the vacuum caused by its utter lack of quality drags the rest of Wikipedia down into a bottomless pit of mediocrity."',
'label': 1}
Data Fields
id
: A unique identifier string for each commentcomment_text
: A string containing the text of the commentlabel
: An integer, either 0 if the comment is non-toxic, or 1 if the comment is toxic
Data Splits
The Wiki Toxic dataset has three splits: train, validation, and test. The statistics for each split are below:
Dataset Split | Number of data points in split |
---|---|
Train | 127,656 |
Validation | 31,915 |
Test | 63,978 |
Dataset Creation
Curation Rationale
[More Information Needed]
Source Data
Initial Data Collection and Normalization
[More Information Needed]
Who are the source language producers?
[More Information Needed]
Annotations
Annotation process
[More Information Needed]
Who are the annotators?
[More Information Needed]
Personal and Sensitive Information
[More Information Needed]
Considerations for Using the Data
Social Impact of Dataset
[More Information Needed]
Discussion of Biases
[More Information Needed]
Other Known Limitations
[More Information Needed]
Additional Information
Dataset Curators
[More Information Needed]
Licensing Information
[More Information Needed]
Citation Information
[More Information Needed]
Contributions
Thanks to @github-username for adding this dataset.