General Knowledge on MMLU-s
76.03P@1STAPO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| STAPOModel Size=Qwen2.5-7B2026.06 | 76.03 | 90.1 | |
| GRPOModel Size=Qwen2.5-7B2026.06 | 74.31 | 90.26 | |
| ICTModel Size=Qwen2.5-7B2026.06 | 74.25 | 90.74 | |
| BaseModel Size=Qwen2.5-7B2026.06 | 73 | 90.52 | |
| 20-EntropyModel Size=Qwen2.5-7B2026.06 | 72.04 | 89.31 | |
| ICTModel Size=Qwen2.5-1.5B2026.06 | 37.8 | 69.4 | |
| 20-EntropyModel Size=Qwen2.5-1.5B2026.06 | 36.12 | 65.17 | |
| STAPOModel Size=Qwen2.5-1.5B2026.06 | 35.24 | 59.24 | |
| BaseModel Size=Qwen2.5-1.5B2026.06 | 33.95 | 64.8 | |
| GRPOModel Size=Qwen2.5-1.5B2026.06 | 29.7 | 61.4 | |
| ICTModel Size=Qwen2.5-0.5B2026.06 | 16.85 | 40.62 | |
| GRPOModel Size=Qwen2.5-0.5B2026.06 | 12.52 | 32.27 | |
| STAPOModel Size=Qwen2.5-0.5B2026.06 | 12.34 | 32.2 | |
| 20-EntropyModel Size=Qwen2.5-0.5B2026.06 | 11.51 | 31.1 | |
| BaseModel Size=Qwen2.5-0.5B2026.06 | 11.22 | 30.98 |