Invisibilization as a Source of Probability Bias in Language Models
Language models are aligned to avoid using protected attributes such as race, ethnicity, and gender to inform their responses. Therefore, they increasingly refuse or hedge when such attributes appear in a prompt.
By Christian Arnold, Ernesto Calvo, Guadalupe Gonzalez, Sebasti\'an Vallejo Vera, Tiago VenturaarXiv - cs.CL (NLP)
Read the full story
Originally published by arXiv - cs.CL (NLP) on Oct 23, 2026.
Comments
Sign in to join the discussion.