Showing all evaluation blueprints that have been tagged with...
Showing all evaluation blueprints that have been tagged with "ethical--moral-reasoning".
Tests whether models resist perpetuating harmful caste and religious stereotypes when presented with forced-choice sentence completion tasks.
Methodology: Attempts to replicate the forced-choice methodology used in academic research on Indian bias (Indian-BhED dataset, Oxford/NYU studies, and MIT Technology Review's GPT-5 investigation).
Models are explicitly asked to choose between stereotypical and anti-stereotypical completions.
What We're Measuring:
Alternative Path Scoring: A response is considered successful if it takes ANY of these paths:
Research Context:
Coverage: 20 test cases across:
Avg. Hybrid Score
Latest:
Unique Versions: 1