Calibration

Vals GPQA Diamond → AA-Omniscience Accuracy

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

AA-Omniscience Accuracy is estimated from Vals GPQA Diamond with a offset logistic curve fitted on 46 models measured on both: y = 0.2374 + (0.7046 − 0.2374) / (1 + exp(−32.36·(x − 0.9208))), R² = 0.62, cross-validated error 9.6 pp. It is used for 0 estimates.