Evaluation of African American Language Bias in Natural Language Generation

Nicholas Deas, Jessica A Grieser, Shana Kleiner, Desmond U. Patton, Elsbeth Turcan, Kathleen McKeown

Submission Type: Regular Long Paper

Submission Track: Theme Track: Large Language Models and the Future of NLP

Submission Track 2: Ethics in NLP

Keywords: benchmarking large language models, african american language, bias and fairness, language generation

Abstract: While biases disadvantaging African American Language (AAL) have been uncovered in models for tasks such as speech recognition and toxicity detection, there has been little investigation of these biases for language generation models like ChatGPT. We evaluate how well LLMs understand AAL in comparison to White Mainstream English (WME), the encouraged "standard" form of English taught in American classrooms. We measure large language model performance on two tasks: a counterpart generation task, where a model generates AAL given WME and vice versa, and a masked span prediction (MSP) task, where models predict a phrase hidden from their input. Using a novel dataset of AAL texts from a variety of regions and contexts, we present evidence of dialectal bias for six pre-trained LLMs through performance gaps on these tasks.

Submission Number: 4487