Papers with Code · CC BY-SA 4.0

데이터셋 탐색

데이터셋의 설명·홈페이지를 찾고, 관련 태스크의 평가 기록을 살펴보세요.

관련 논문 20편

multi-labEl haTe speecH detectiOn dataSet

ETHOS
ETHOS 설명 보기

ETHOS is a hate speech detection dataset. It is built from YouTube and Reddit comments validated through a crowdsourcing platform. It has two subsets, one for binary classification and the other for multi-label classification. The former contains 998 comments, while the latter contains fine-grained hate-speech annotations for 433 comments. Source: https://arxiv.org/abs/2006.08328