Mathematical Reasoning on GSM8K (Average Acceptance Length and Speedup Ratio)
6.97Average Acceptance Length (τ)PPOW
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PPOWModel=Q3-8B, Temperature=0.02026.05 | 6.97 | 3.54 | |
| EAGLE-3Model=Q3-8B, Temperature=0.02026.05 | 6.86 | 3.47 | |
| PPOWModel=Q3-32B, Temperature=0.02026.05 | 6.62 | 4.05 | |
| PPOWModel=L31-8B, Temperature=0.02026.05 | 6.5 | 3.52 | |
| PPOWModel=L33-70B, Temperature=0.02026.05 | 6.47 | 4.54 | |
| PPOWModel=L33-70B, Temperature=1.02026.05 | 6.29 | 4.49 | |
| EAGLE-3Model=Q3-32B, Temperature=0.02026.05 | 6.21 | 3.94 | |
| EAGLE-3Model=L31-8B, Temperature=0.02026.05 | 6.12 | 3.41 | |
| GRIFFINModel=L31-8B, Temperature=0.02026.05 | 5.98 | 3.38 | |
| EAGLE-3Model=L33-70B, Temperature=1.02026.05 | 5.98 | 4.16 | |
| EAGLE-3Model=L33-70B, Temperature=0.02026.05 | 5.93 | 4.02 | |
| GRIFFINModel=L33-70B, Temperature=0.02026.05 | 5.9 | 3.99 | |
| PPOWModel=Q3-8B, Temperature=1.02026.05 | 5.76 | 3.01 | |
| PPOWModel=L31-8B, Temperature=1.02026.05 | 5.63 | 3.13 | |
| PPOWModel=Q3-32B, Temperature=1.02026.05 | 5.61 | 3.01 | |
| GRIFFINModel=L33-70B, Temperature=1.02026.05 | 5.54 | 3.87 | |
| EAGLE-3Model=Q3-8B, Temperature=1.02026.05 | 5.52 | 2.89 | |
| EAGLE-3Model=Q3-32B, Temperature=1.02026.05 | 5.23 | 2.81 | |
| EAGLE-3Model=L31-8B, Temperature=1.02026.05 | 5.16 | 2.89 | |
| GRIFFINModel=L31-8B, Temperature=1.02026.05 | 5.13 | 2.89 |