6 datasets found

The MMHS150K dataset consists of 150K multi-modal (images + text) tweets spanning over four hate-inclined topics – racism, sexism, homophobia, and religious extrem-ism.

The dataset is a collection of Danish tweets containing offensive language.

The dataset contains hate speech and overtly offensive content.

Dynahate: A dataset for hate speech detection.

A large-scale dataset of tweets, retweets, user activity history, and follower networks, comprising over 161 million tweets from more than 41 million unique users.

A benchmark dataset for learning to intervene in online hate speech.