docs/workspace/compare-evaluation-analysis/structured-compare-calibration/latest/synthetic-medical-latent-trigger-overfit/scenario.md