Estranged Predictions: Measuring Semantic Category Disruption with Masked Language Modelling

Open in new window