EHRSHOT Leaderboard

Benchmarking EHR foundation models across 15 clinical prediction tasks

Models

Logistic Regression
A logistic regression trained on the count featurization described in the EHRSHOT paper.
Random Forest
A random forest trained on the count featurization described in the EHRSHOT paper.
GBM
A gradient boosted machine trained on the count featurization described in the EHRSHOT paper.
CLMBR-T-base
The foundation model released in the EHRSHOT paper with frozen weights and a task-specific logistic regression head. Download here.
MOTOR-T-base
The foundation model released in the MOTOR paper with frozen weights and a task-specific logistic regression head. Download here.