kocohub/korean-hate-speech
Korean HateSpeech Dataset
Comprises 9.4K human-annotated comments with dual-aspect labels (social bias and hate speech) plus 2M unlabeled comments from Korean entertainment news, enabling both supervised and semi-supervised approaches. Comments include associated news titles for contextual understanding, and the dataset distinguishes gender bias as a prevalent category alongside general bias detection. Accessible via the `koco` library and backed by three Kaggle competitions for standardized model evaluation across gender-bias, bias, and hate-speech detection tasks.
395 stars. No commits in the last 6 months.
Stars
395
Forks
39
Language
—
License
CC-BY-SA-4.0
Category
Last pushed
Jul 18, 2020
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/nlp/kocohub/korean-hate-speech"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
unitaryai/detoxify
Trained models & code to predict toxic comments on all 3 Jigsaw Toxic Comment Challenges. Built...
kensk8er/chicksexer
A Python package for gender classification.
Infinitode/ValX
ValX is an open-source Python package for text cleaning tasks, including profanity detection and...
PavelOstyakov/toxic
Toxic Comment Classification Challenge
minerva-ml/open-solution-toxic-comments
Open solution to the Toxic Comment Classification Challenge