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

Calibration on OpenBookQA

1.085ECE Gain

S

-0.04340.249550.54250.83545Apr 19, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.04
1.08540.75910.6361-0.6570
2026.04
0.74500.72810.63630.032
2026.04
0.60500.64500.5910-0.0410
2026.04
0.53200.34100.6030-0.1790
2026.04
0.48300.75420.4130-0.0531
2026.04
0.44700.64200.40-0.0530
2026.04
0.38800.2600.4310-0.5230
2026.04
00000001