Mathematical Reasoning on GSM8K (AUC %)
85.8AUC (%)CF
Evaluation Results
| Method | Links | |
|---|---|---|
| CFModel=Qwen2.5-3B2026.02 | 85.8 | |
| DAPOModel=Qwen2.5-3B2026.02 | 85.3 | |
| RandomModel=Qwen2.5-3B2026.02 | 85 | |
| InvertedModel=Qwen2.5-3B2026.02 | 84.8 | |
| CFModel=Qwen3-1.7B2026.02 | 83.2 | |
| DAPOModel=Qwen3-1.7B2026.02 | 82.5 | |
| RandomModel=Qwen3-1.7B2026.02 | 82.1 | |
| InvertedModel=Qwen3-1.7B2026.02 | 81 | |
| CFModel=Llama3.2-3B2026.02 | 77.3 | |
| DAPOModel=Llama3.2-3B2026.02 | 76.6 | |
| RandomModel=Llama3.2-3B2026.02 | 75.9 | |
| InvertedModel=Llama3.2-3B2026.02 | 75.6 |