List your company

Mathematical Reasoning on GSM8K (Accuracy, Inference Count, Token Count)

0.683AccuracySEAG
0.435480.499740.5640.62826Jan 10, 2025

Evaluation results

MethodLinks
SEAGModel=Llama3-8B-Base2025.010.68364.8425,853.651,263.95
SEModel=Llama3-8B-Base2025.010.67592.8942,832.031,969.56
RAPModel=Llama3-8B-Base2025.010.665146.1867,526.583,113.34
CoT-SCModel=Llama3-8B-Base2025.010.598102,183.76906.97
ToTModel=Llama3-8B-Base2025.010.598121.9357,187.682,856.04
CoTModel=Llama3-8B-Base2025.010.4451218.3785.26