Attribution Bias in Large Language Models

This research introduces AttriBench, a novel benchmark exposing how Large Language Models systematically fail to attribute quotes fairly across race and gend...

Level: advanced

By Eliza Berman

Category: discussion