List your company

Uncertainty Calibration on JudgeBench

0.037KuiperProbe
0.02720.093350.15950.22565Dec 23, 2025

Evaluation results

MethodLinks
ProbeModel Backbone=J1 QWEN 32B2025.120.0370.048
ProbeModel Backbone=QWEN 32B2025.120.0550.059
ProbeModel Backbone=LLAMA 70B2025.120.0620.077
ProbeModel Backbone=LLAMA 8B2025.120.0720.119
ProbeModel Backbone=J1 LLAMA 70B2025.120.0750.074
MajorityModel Backbone=LLAMA 8B2025.120.0760.171
MajorityModel Backbone=J1 QWEN 32B2025.120.0760.105
ConsistencyModel Backbone=LLAMA 8B2025.120.0770.193
ConsistencyModel Backbone=J1 QWEN 32B2025.120.0870.124
VerbalizedModel Backbone=QWEN 32B2025.120.1030.107
VerbalizedModel Backbone=J1 QWEN 32B2025.120.1140.113
ProbeModel Backbone=J1 LLAMA 8B2025.120.1220.135
VerbalizedModel Backbone=LLAMA 70B2025.120.1520.145
VerbalizedModel Backbone=J1 LLAMA 70B2025.120.1560.155
MajorityModel Backbone=QWEN 32B2025.120.1590.198
ConsistencyModel Backbone=J1 LLAMA 70B2025.120.1880.204
MajorityModel Backbone=J1 LLAMA 70B2025.120.20.182
VerbalizedModel Backbone=LLAMA 8B2025.120.2050.188
ConsistencyModel Backbone=QWEN 32B2025.120.2050.276
ConsistencyModel Backbone=LLAMA 70B2025.120.2310.243
MajorityModel Backbone=LLAMA 70B2025.120.2380.224
MajorityModel Backbone=J1 LLAMA 8B2025.120.2710.259
VerbalizedModel Backbone=J1 LLAMA 8B2025.120.2740.263
ConsistencyModel Backbone=J1 LLAMA 8B2025.120.2820.259