Mathematical Reasoning on GSM8K (test) (Accuracy, Inferences, Input/Output Tokens)
86AccuracySEAG
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SEAGModel=Llama3-8B-Instruct2025.01 | 86 | 41.69 | 17,501.68 | 1,759.5 | |
| SEModel=Llama3-8B-Instruct2025.01 | 85 | 82.63 | 38,743.45 | 1,766.21 | |
| COT-SCModel=Llama3-8B-Instruct2025.01 | 84.5 | 10 | 2,183.73 | 943.66 | |
| RAPModel=Llama3-8B-Instruct2025.01 | 82.5 | 128.4 | 59,800.59 | 2,587.73 | |
| ToTModel=Llama3-8B-Instruct2025.01 | 78.5 | 104.8 | 49,003.05 | 2,355.06 | |
| COTModel=Llama3-8B-Instruct2025.01 | 76.2 | 1 | 218.37 | 85.86 |