Our new X account is live! Follow @wizwand_team for updates

Calibration on MMLU

0.0559Brier Score

Verbalized confidence

Updated 4d ago

Evaluation Results

Method	Links
Verbalized confidence 2026.02		0.0559	-
Probe (train on GSM8K) 2026.02		0.0686	-
P(True) 2026.02		0.0817	-
Probe (train on MATH) 2026.02		0.0922	-
Probe (train on TriviaQA) 2026.02		0.0977	-
Probe (train on MATH) 2026.02		0.1163	-
Probe (train on GSM8K) 2026.02		0.1176	-
Probe (train on GSM8K) 2026.02		0.12	-
Probe (train on MATH) 2026.02		0.1255	-
Verbalized confidence 2026.02		0.1293	-
Probe (train on TriviaQA) 2026.02		0.13	-
BaseCal-ReEval 2026.01		0.1465	0.0393
BaseCal-ReEval 2026.01		0.1473	0.0375
BaseCal-Proj 2026.01		0.15	0.0336
Verbalized confidence 2026.02		0.1561	-
P(True) 2026.02		0.1597	-
DACA 2026.01		0.1618	0.0703
BaseCal-Proj 2026.01		0.1662	0.0889
BaseCal-ReEval 2026.01		0.1717	0.047
BaseCal-Proj 2026.01		0.1768	0.0525
DACA 2026.01		0.1804	0.0473
DACA 2026.01		0.1898	0.0555
Probe (train on TriviaQA) 2026.02		0.2006	-
Temp. Scaling 2026.01		0.2021	0.0307
Vanilla 2026.01		0.2152	0.1071
P(True) 2026.02		0.2164	-
Temp. Scaling 2026.01		0.2275	0.1707
Temp. Scaling 2026.01		0.2354	0.2261
Verbalization 2026.01		0.2546	0.1972
P(true) 2026.01		0.2549	0.194
Uniform random baseline 2026.02		0.2565	-
Vanilla 2026.01		0.2607	0.2569
Verbalization 2026.01		0.2633	0.2011
Verbalization 2026.01		0.2747	0.1533
Vanilla 2026.01		0.2762	0.2465
Semantic Entropy 2026.01		0.2856	0.2858
Uniform random baseline 2026.02		0.2868	-
Uniform random baseline 2026.02		0.3018	-
P(true) 2026.01		0.3079	0.2971
P(true) 2026.01		0.3256	0.3204
Semantic Entropy 2026.01		0.3607	0.3132
Semantic Entropy 2026.01		0.3991	0.4085