Graduate-Level Science Question Answering on GPQA (P@1, P@4)
33.84P@1ICT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ICTModel Size=Qwen2.5-7B2026.06 | 33.84 | 70.71 | |
| 20-EntropyModel Size=Qwen2.5-7B2026.06 | 31.94 | 65.15 | |
| BaseModel Size=Qwen2.5-7B2026.06 | 31.57 | 67.68 | |
| GRPOModel Size=Qwen2.5-7B2026.06 | 30.54 | 64.14 | |
| STAPOModel Size=Qwen2.5-7B2026.06 | 30.18 | 64.14 | |
| STAPOModel Size=Qwen2.5-1.5B2026.06 | 16.65 | 29.34 | |
| ICTModel Size=Qwen2.5-1.5B2026.06 | 16.4 | 29.71 | |
| BaseModel Size=Qwen2.5-1.5B2026.06 | 14.02 | 28.7 | |
| GRPOModel Size=Qwen2.5-1.5B2026.06 | 13.84 | 28.91 | |
| 20-EntropyModel Size=Qwen2.5-1.5B2026.06 | 11.77 | 23.56 | |
| ICTModel Size=Qwen2.5-0.5B2026.06 | 5.95 | 15.69 | |
| GRPOModel Size=Qwen2.5-0.5B2026.06 | 5.11 | 8.44 | |
| BaseModel Size=Qwen2.5-0.5B2026.06 | 4.56 | 12.49 | |
| 20-EntropyModel Size=Qwen2.5-0.5B2026.06 | 3.98 | 10.75 | |
| STAPOModel Size=Qwen2.5-0.5B2026.06 | 2.78 | 8.37 |