Concept
Prejudice Rate
Prejudice Rate measures how often a model produces the right answer for the wrong reasons: it lands on a plausible label while grounding that label in superficial or stereotyped cues rather than valid evidence. Think of it as a faithfulness diagnostic, like separating a…
The rest of “Prejudice Rate” is a premium feature: every concept in the library gets a precise, practitioner-focused write-up like this one, cross-linked straight from the paper summaries that use it.
Log in to unlock→