benchgap
Calibration

HealthBench Professional (raw) → HealthBench (length-adjusted)

HealthBench (length-adjusted) is estimated from HealthBench Professional (raw) with a linear curve fitted on 7 models measured on both: y = 0.4964·x + 0.2403, R² = 0.78, cross-validated error 2.5 pp. It is used for 0 estimates.