valurank/offensive-multi
Dataset Card for hate-multi Dataset Description Dataset Summary This dataset contains a collection of text labeled as offensive (class 1) or not (class 0). Dataset Creation The dataset was creating by aggregating multiple publicly available datasets. Source Data The following datasets were used: https://huggingface.co/datasets/hate_speech_offensive - Tweet text cleaned by lower casing, removing mentions and urls.… See the full description on the dataset page: https://huggingface.co/datasets/valurank/offensive-multi.
Dataset Card for hate-multi
Table of Contents
Dataset Description
Dataset Summary
This dataset contains a collection of text labeled as offensive (class 1) or not (class 0).
Dataset Creation
The dataset was creating by aggregating multiple publicly available datasets.
Source Data
The following datasets were used:
- https://huggingface.co/datasets/hatespeechoffensive - Tweet text cleaned by lower casing, removing mentions and urls. Dropped instanced labeled as 'hate speech'
- https://sites.google.com/site/offensevalsharedtask/olid - Tweet text cleaned by lower casing, removing mentions and urls. Used 'subtask_a' column for labeling.