Model trained in the paper: EHRSHOT: An EHR Benchmark for Few-Shot Evaluation of Foundation Models (https://arxiv.org/abs/2307.02028)