Calibration
MMLU-Redux → GPQA
GPQA is estimated from MMLU-Redux with a linear curve fitted on 9 models measured on both: y = 2.8555·x + -1.8086, R² = 0.89, cross-validated error 6.9 pp. It is used for 0 estimates.
GPQA is estimated from MMLU-Redux with a linear curve fitted on 9 models measured on both: y = 2.8555·x + -1.8086, R² = 0.89, cross-validated error 6.9 pp. It is used for 0 estimates.