List your company

Code Generation on MBPP (test) (Calibration Metrics)

0.4591AccuracyEPICAR
0.3700760.3931880.41630.439412Jan 11, 2026

Evaluation results

MethodLinks
EPICARModel=Qwen-3-8B2026.010.45910.6280.0590.285
Base ModelModel=Qwen-3-8B2026.010.45140.6220.0660.286
STaRModel=Qwen-3-8B2026.010.45140.5770.0660.285
EPICARModel=Llama-3-8B2026.010.3930.5380.1130.246
STaRModel=Llama-3-8B2026.010.37740.5230.390.387
Base ModelModel=Llama-3-8B2026.010.37350.5510.3980.391