List your company

Math on GSM8K, MATH-500 (test)

92.6GSM8K AccuracyREAP
-3.70421.29846.371.302Mar 19, 2026

Evaluation results

MethodLinks
REAPModel=Qwen3, Ratio=25%, Zero-shot=true2026.0392.677.885.2
FullModel=Qwen3, Ratio=0%, Zero-shot=true2026.0392.380.286.2
SEERModel=Qwen3, Ratio=25%, Zero-shot=true2026.0391.320.255.7
EANModel=Qwen3, Ratio=25%, Zero-shot=true2026.0390.72055.4
FrequencyModel=Qwen3, Ratio=25%, Zero-shot=true2026.0390.419.655
AIMERModel=Qwen3, Ratio=25%, Zero-shot=true2026.0389.575.682.6
REAPModel=Qwen3, Ratio=50%, Zero-shot=true2026.0389.580.484.9
MagnitudeModel=Qwen3, Ratio=25%, Zero-shot=true2026.0388.480.884.6
RandomModel=Qwen3, Ratio=25%, Zero-shot=true2026.0380.757.469.1
AIMERModel=Qwen3, Ratio=50%, Zero-shot=true2026.0366.449.457.9
EANModel=Qwen3, Ratio=50%, Zero-shot=true2026.03653.234.1
RandomModel=Qwen3, Ratio=50%, Zero-shot=true2026.038.53.86.1
FrequencyModel=Qwen3, Ratio=50%, Zero-shot=true2026.03000
SEERModel=Qwen3, Ratio=50%, Zero-shot=true2026.03000
MagnitudeModel=Qwen3, Ratio=50%, Zero-shot=true2026.03000