We systematically evaluate the concerns of multiplicity and monoculture in a suite of large language models and prediction tasks.