Is a bunch of words enough to detect disagreement in hateful content? - View it on GitHub