Math on AMC-12
73.63AccuracyProbing
Evaluation Results
| Method | Links | |
|---|---|---|
| ProbingCategory=Steering2026.01 | 73.63 | |
| AdaRASCategory=Steering2026.01 | 70.33 | |
| CoTPrompting=Vanilla CoT, Base Model=Qwen3-1.7B2026.01 | 65.93 | |
| OpenReasoning-Nemotron-1.5BCategory=Post-training2026.01 | 65.93 | |
| OpenThinker-3-1.5BCategory=Post-training2026.01 | 53.85 | |
| DeepSeek-R1-Distill-Qwen-1.5BCategory=Post-training2026.01 | 41.76 |