List your company

Preference Calibration on PPE

0.034KuiperProbe
0.02240.10070.1790.2573Dec 23, 2025

Evaluation results

MethodLinks
ProbeModel Architecture=J1 LLAMA 70B2025.120.0340.039
MajorityModel Architecture=QWEN 32B2025.120.0380.065
ProbeModel Architecture=LLAMA 70B2025.120.040.047
ConsistencyModel Architecture=QWEN 32B2025.120.0480.083
MajorityModel Architecture=J1 QWEN 32B2025.120.050.069
ConsistencyModel Architecture=J1 QWEN 32B2025.120.0510.071
ProbeModel Architecture=J1 LLAMA 8B2025.120.0510.066
ProbeModel Architecture=J1 QWEN 32B2025.120.0560.072
ProbeModel Architecture=LLAMA 8B2025.120.0620.093
ProbeModel Architecture=QWEN 32B2025.120.0660.077
VerbalizedModel Architecture=J1 LLAMA 70B2025.120.1220.12
VerbalizedModel Architecture=LLAMA 70B2025.120.1230.121
MajorityModel Architecture=LLAMA 8B2025.120.150.143
ConsistencyModel Architecture=LLAMA 8B2025.120.1530.155
VerbalizedModel Architecture=LLAMA 8B2025.120.1730.166
VerbalizedModel Architecture=J1 LLAMA 8B2025.120.1920.188
VerbalizedModel Architecture=J1 QWEN 32B2025.120.1950.195
VerbalizedModel Architecture=QWEN 32B2025.120.1960.194
ConsistencyModel Architecture=J1 LLAMA 70B2025.120.2070.2
MajorityModel Architecture=J1 LLAMA 70B2025.120.2110.197
ConsistencyModel Architecture=LLAMA 70B2025.120.2310.226
MajorityModel Architecture=LLAMA 70B2025.120.2340.223
MajorityModel Architecture=J1 LLAMA 8B2025.120.3230.31
ConsistencyModel Architecture=J1 LLAMA 8B2025.120.3240.311