Calibration

AA-GPQA Diamond → HLE

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

HLE is estimated from AA-GPQA Diamond with a Hill curve fitted on 46 models measured on both: y = 0.0000 + (1.2000 − 0.0000)·x^6.00 / (0.98770^6.00 + x^6.00), R² = 0.32, cross-validated error 12.1 pp. It is used for 0 estimates.