Mathematical Reasoning on PolyMath
20.9Accuracyself-cons (ours)
Evaluation Results
| Method | Links | |
|---|---|---|
| self-cons (ours)Model Size=14B2026.05 | 20.9 | |
| self-cons (ours)Model Size=7B2026.05 | 20.5 | |
| S1Model Size=7B2026.05 | 20.2 | |
| S1Model Size=14B2026.05 | 19.5 | |
| baseModel Size=7B2026.05 | 18 | |
| baseModel Size=14B2026.05 | 17.5 | |
| LIDRModel Size=7B2026.05 | 16.2 | |
| LIDRModel Size=14B2026.05 | 15.2 | |
| self-cons (ours)Model Size=1.5B2026.05 | 8.1 | |
| S1Model Size=1.5B2026.05 | 8 | |
| baseModel Size=1.5B2026.05 | 7.5 | |
| LIDRModel Size=1.5B2026.05 | 5.9 |