Code on MBPP+
60.58AccuracyAdaRAS
Evaluation Results
| Method | Links | |
|---|---|---|
| AdaRASCategory=Steering2026.01 | 60.58 | |
| ProbingCategory=Steering2026.01 | 59.52 | |
| CoTPrompting=Vanilla CoT, Base Model=Qwen3-1.7B2026.01 | 58.2 | |
| OpenReasoning-Nemotron-1.5BCategory=Post-training2026.01 | 43.92 | |
| DeepSeek-R1-Distill-Qwen-1.5BCategory=Post-training2026.01 | 37.3 | |
| OpenThinker-3-1.5BCategory=Post-training2026.01 | 30.95 |