Calibration
AA AutomationBench → HLE w/ tools
HLE w/ tools is estimated from AA AutomationBench with a linear curve fitted on 8 models measured on both: y = 0.3522·x + 0.3888, R² = 0.80, cross-validated error 6.4 pp. It is used for 0 estimates.
HLE w/ tools is estimated from AA AutomationBench with a linear curve fitted on 8 models measured on both: y = 0.3522·x + 0.3888, R² = 0.80, cross-validated error 6.4 pp. It is used for 0 estimates.