Mathematical Reasoning on MATH500 (Accuracy and Efficiency Metrics)
91.9SC AccuracySelf-Consistency
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Self-ConsistencyModel=QWQ32B, Sample size N=312026.01 | 91.9 | — | — | — | |
| PoLRModel=QWQ32B, Sample size N=312026.01 | 91.9 | 0 | 54.2 | 5.7 | |
| Self-ConsistencyModel=QWQ32B, Sample size N=512026.01 | 91.8 | — | — | — | |
| PoLRModel=QWQ32B, Sample size N=512026.01 | 91.8 | 0.2 | 51.8 | 11.2 | |
| Self-ConsistencyModel=QWQ32B, Sample size N=112026.01 | 91.6 | — | — | — | |
| PoLRModel=QWQ32B, Sample size N=112026.01 | 91.6 | 0 | 60.5 | 2.2 | |
| Self-ConsistencyModel=DSQ7B, Sample size N=512026.01 | 89.8 | — | — | — | |
| PoLRModel=DSQ7B, Sample size N=512026.01 | 89.8 | -0.4 | 48.7 | 7.6 | |
| Self-ConsistencyModel=DSQ7B, Sample size N=312026.01 | 89.6 | — | — | — | |
| PoLRModel=DSQ7B, Sample size N=312026.01 | 89.6 | 0.1 | 48.5 | 5.1 | |
| Self-ConsistencyModel=DSQ7B, Sample size N=112026.01 | 89.4 | — | — | — | |
| PoLRModel=DSQ7B, Sample size N=112026.01 | 89.4 | 0 | 48.4 | 2.2 | |
| Self-ConsistencyModel=DSQ1.5B, Sample size N=312026.01 | 76.4 | — | — | — | |
| PoLRModel=DSQ1.5B, Sample size N=312026.01 | 76.4 | 0 | 52 | 4.4 | |
| Self-ConsistencyModel=DSQ1.5B, Sample size N=512026.01 | 76.2 | — | — | — | |
| PoLRModel=DSQ1.5B, Sample size N=512026.01 | 76.2 | -0.8 | 52.4 | 6.5 | |
| Self-ConsistencyModel=DSQ1.5B, Sample size N=112026.01 | 75.9 | — | — | — | |
| PoLRModel=DSQ1.5B, Sample size N=112026.01 | 75.9 | -1.6 | 52 | 2 |