Question Answering on ARC (with Efficiency Metrics)
0.788AccuracySEAG
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SEAGModel=Llama3-8B-Base2025.01 | 0.788 | 12.33 | 3,155.79 | 397.37 | |
| CoT-SCModel=Llama3-8B-Base2025.01 | 0.787 | 10 | 2,157.93 | 362.51 | |
| SEModel=Llama3-8B-Base2025.01 | 0.775 | 103.55 | 44,404.5 | 1,511.65 | |
| RAPModel=Llama3-8B-Base2025.01 | 0.767 | 208.82 | 91,586.37 | 3,224.82 | |
| ToTModel=Llama3-8B-Base2025.01 | 0.748 | 169.82 | 75,622.06 | 2,886.81 | |
| CoTModel=Llama3-8B-Base2025.01 | 0.713 | 1 | 215.8 | 29.56 |