MetaHate: A Dataset for Unifying Efforts on Hate Speech Detection

This is MetaHate: a meta-collection of 36 hate speech datasets from social media comments.

Dataset Structure

The dataset contains 1,226,202 social media posts in a TSV file. Each element contains the following fields:

| Field Name | Type | Possible Values | Description | |------------|------|-----------------|----------------------------------------------------------------------| | text | str | any | Social media post. Each post is unique. | | label | int | 0, 1 | Label of the post. 0 for non-hate speech posts, 1 for hate speech. |