Mathematical Reasoning on SVAMP (Accuracy by Context Size)
0.8933Accuracy (Context Size 128)EvoToken (10k FT)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| EvoToken (10k FT)NFE / Gen Len=1/22026.01 | 0.8933 | 0.9233 | 0.8967 | 0.9044 | |
| EvoToken (10k FT)NFE / Gen Len=12026.01 | 0.89 | 0.8967 | 0.9 | 0.8956 | |
| BaselineNFE / Gen Len=12026.01 | 0.8833 | 0.8467 | 0.86 | 0.8633 | |
| FT-Base (10k FT)NFE / Gen Len=12026.01 | 0.8767 | 0.8967 | 0.8967 | 0.89 | |
| BaselineNFE / Gen Len=1/22026.01 | 0.86 | 0.8667 | 0.84 | 0.8556 | |
| FT-Base (10k FT)NFE / Gen Len=1/22026.01 | 0.85 | 0.87 | 0.8933 | 0.8711 | |
| BaselineNFE / Gen Len=1/42026.01 | 0.81 | 0.7733 | 0.75 | 0.7778 | |
| EvoToken (10k FT)NFE / Gen Len=1/42026.01 | 0.7833 | 0.8333 | 0.8133 | 0.81 | |
| FT-Base (10k FT)NFE / Gen Len=1/42026.01 | 0.6667 | 0.7533 | 0.72 | 0.7133 |