Mathematical Reasoning on Arithmetic (val)
83AccuracySELFCEST
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SELFCESTdelegation=learned clones2026.02 | 83 | 1,705 | |
| Qwen3-4B-Thinkingmodel_variant=Thinking2026.02 | 69 | 6,306 | |
| Base Qwen3-4Bmodel_variant=Base2026.02 | 68 | 2,028 |
| Method | Links | ||
|---|---|---|---|
| SELFCESTdelegation=learned clones2026.02 | 83 | 1,705 | |
| Qwen3-4B-Thinkingmodel_variant=Thinking2026.02 | 69 | 6,306 | |
| Base Qwen3-4Bmodel_variant=Base2026.02 | 68 | 2,028 |