Calibration

BioMysteryBench (human-difficult) → BioMysteryBench (human-solvable)

Name: benchgap: LLM benchmark scores, measured and estimated
Creator: benchgap

BioMysteryBench (human-solvable) is estimated from BioMysteryBench (human-difficult) with a offset logistic curve fitted on 5 models measured on both: y = 0.0378 + (0.8940 − 0.0378) / (1 + exp(−200.00·(x − 0.4171))), R² = 0.83, cross-validated error 1.2 pp. It is used for 0 estimates.