Low Frequency Names Exhibit Bias and Overfitting in Contextualizing Language Models

Open in new window