SAT Reasoning on SATURN Difficulty 3,5,15 2.6k (test)
85.4Pass@1SATURN-7B
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| SATURN-7BIteration=22025.05 | 85.4 | 98.3 | 99.8 | 99.9 | 100 | |
| SATURN-7BIteration=12025.05 | 65.7 | 91.8 | 96.8 | 98.7 | 99.7 | |
| DeepSeek-R1-Distill-Qwen-7B2025.05 | 39.3 | 74.9 | 88.3 | 94.3 | 98.3 |