Calibration

CritPt → AA-GPQA Diamond

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

AA-GPQA Diamond is estimated from CritPt with a Hill curve fitted on 171 models measured on both: y = 0.6033 + (1.0395 − 0.6033)·x^0.40 / (0.01603^0.40 + x^0.40), R² = 0.65, cross-validated error 10.4 pp. It is used for 0 estimates.