RACQUET Dataset Shows Ambiguity and Bias Risks in Language Models

RACQUET benchmarks language models on image questions; its RACQUET‑BIAS subset shows answers may reinforce stereotypes. The paper first appeared on arXiv in Dec 2024, updated Sep 2025. Read more: getnews.me/racquet-dataset-shows-am... #racquet #bias