List your company

Expected Calibration Error on BT

0.025ECE (15-bin)PPT (Distribution)
0.007320.126660.2460.36534May 26, 2026

Evaluation results

MethodLinks
PPT (Distribution)Model=LLAMA3-8B, Temperature Scaling=true2026.050.025
PPT (Distribution)Model=QWEN2-7B, Temperature Scaling=true2026.050.056
PPT (Distribution)Model=LLAMA3-8B, Temperature Scaling=false2026.050.218
VerbalizedModel=LLAMA3-8B, Temperature Scaling=false2026.050.255
PPT (Distribution)Model=QWEN2-7B, Temperature Scaling=false2026.050.278
PPT (Mean)Model=LLAMA3-8B, Temperature Scaling=false2026.050.38
PPT (Mean)Model=QWEN2-7B, Temperature Scaling=false2026.050.415
BaseModel=QWEN2-7B, Temperature Scaling=false2026.050.464
BaseModel=LLAMA3-8B, Temperature Scaling=false2026.050.467