1 paper · 1 filter
Maximus Powers, Shaina Raza, Alex Chang +5
Representational harms in language technologies often occur in short spans within otherwise neutral text, where phrases may simultaneously convey generalizations, unfairness, or st…