Mathematical Reasoning on GSM8K (P@1, P@4)
86.68P@1ICT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ICTModel Size=Qwen2.5-7B2026.06 | 86.68 | 94.47 | |
| STAPOModel Size=Qwen2.5-7B2026.06 | 85.24 | 93.93 | |
| GRPOModel Size=Qwen2.5-7B2026.06 | 84.7 | 94.39 | |
| 20-EntropyModel Size=Qwen2.5-7B2026.06 | 70.32 | 91.81 | |
| BaseModel Size=Qwen2.5-7B2026.06 | 67.66 | 90.98 | |
| ICTModel Size=Qwen2.5-1.5B2026.06 | 66.23 | 85.31 | |
| STAPOModel Size=Qwen2.5-1.5B2026.06 | 62.31 | 80.15 | |
| GRPOModel Size=Qwen2.5-1.5B2026.06 | 61.51 | 81.71 | |
| 20-EntropyModel Size=Qwen2.5-1.5B2026.06 | 59.5 | 82.12 | |
| BaseModel Size=Qwen2.5-1.5B2026.06 | 43.1 | 77.29 | |
| GRPOModel Size=Qwen2.5-0.5B2026.06 | 42.87 | 63.15 | |
| STAPOModel Size=Qwen2.5-0.5B2026.06 | 40.88 | 64.5 | |
| ICTModel Size=Qwen2.5-0.5B2026.06 | 40.27 | 65.4 | |
| 20-EntropyModel Size=Qwen2.5-0.5B2026.06 | 30.38 | 50.5 | |
| BaseModel Size=Qwen2.5-0.5B2026.06 | 13.88 | 38.62 |