Don't Forget About Pronouns: Removing Gender Bias in Language Models Without Losing Factual Gender Information. (arXiv:2206.10744v1 [cs.CL])

The representations in large language models contain multiple types of gender
information. We focus on two types of such signals in English texts: factual
gender information, which is a grammatical or semantic property, and gender
bias, which is the correlation between a word and specific gender. We can
disentangle the model's embeddings and identify components encoding both types
of information with probing. We aim to diminish the stereotypical bias in the
representations while preserving the factual gender signal. Our filtering …

arxiv bias gender gender bias information language language models