List your company

Math Reasoning on GSM8K (Calibration Metrics)

0.87AUROCVEROIC
0.70360.74680.790.8332Apr 30, 2026

Evaluation results

MethodLinks
VEROICsetting=CMIS, inference_path=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.040.870.520.1390.4170.048
RISK PREDICTORsetting=CMIS, inference_path=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.040.850.510.1440.4320.058
CONFIDENCE-THRESHOLDsetting=CMIS, inference_path=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.040.840.490.1520.4520.066
OUTPUT-TRIGGERsetting=CMIS, inference_path=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.040.830.480.1490.4470.062
PERIODICsetting=CMIS, inference_path=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.040.770.430.1620.4670.065
RANDOMsetting=CMIS, inference_path=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.040.730.410.1730.4750.076
FIXEDsetting=CMIS, inference_path=LLaMA-3.1-8B -> LLaMA-3.1-70B2026.040.710.320.1850.5230.085