Calibration

BioMysteryBench (human-difficult) → AA-Omniscience Accuracy

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

AA-Omniscience Accuracy is estimated from BioMysteryBench (human-difficult) with a Hill curve fitted on 6 models measured on both: y = 0.0000 + (0.6624 − 0.0000)·x^6.00 / (0.34720^6.00 + x^6.00), R² = 0.81, cross-validated error 9.2 pp. It is used for 0 estimates.