Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Selective Prediction on BBH, GPQA, and MMLU-Pro Pooled (test)

1,384N0/N Count

Pooled (21 runs)

-48.08323.71695.51,067.29Apr 7, 2026
Updated 9d ago

Evaluation Results

MethodLinks
2026.04
1,3840.2540.9610.949
2026.04
4250.5330.9810.963
2026.04
4060.5090.9480.922
2026.04
2330.3020.9610.928
2026.04
2050.2570.9660.931
2026.04
770.0970.9610.892
2026.04
310.0430.9030.751
2026.04
70.0090.5710.251