On the Independence of Association Bias and Empirical Fairness in Language Models

Open in new window