1 paper
Kritee Kondapally, Claire J. Smerdon, Pooja C. Patel +5
Language models (LMs) can exhibit biases based on variations in their dialects, even in the absence of a dialect label, a behavior known as covert dialect bias. In this work, we qu…