Calibration

CWE-bench v1 → GDPval-AA

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

GDPval-AA is estimated from CWE-bench v1 with a Hill curve fitted on 10 models measured on both: y = 0.0000 + (0.6195 − 0.0000)·x^6.00 / (0.37812^6.00 + x^6.00), R² = 0.80, cross-validated error 5.4 pp. It is used for 0 estimates.