Guadalupe Gonzalez
Titles as the sources gave them. In 1 entry, with 0 checked quotations.
In the Base
Every entry they appear in.
Invisibilization as a Source of Probability Bias in Language Models
According to the abstract, a preprint tested leading OpenAI and Anthropic models on a trust game with 6,600 sessions, varying the described identity of the person reporting. The authors report the models more often declined to assess a Black-described reporter, 46% of the time, than a white-described one, 9%.