Mathematical Reasoning on Average (AIME24, AIME25, MATH500, AMC23, Hmmt25, Olympiad)
30.07Average PerformanceCARE
Evaluation Results
| Method | Links | |
|---|---|---|
| CAREBackbone=Phi4-mini-instruct2026.05 | 30.07 | |
| GTBackbone=Phi4-mini-instruct2026.05 | 30.06 | |
| OracleBackbone=Phi4-mini-instruct2026.05 | 27.6 | |
| TTRLBackbone=Phi4-mini-instruct2026.05 | 26.32 | |
| ProbBackbone=Phi4-mini-instruct2026.05 | 26.25 | |
| RandomBackbone=Phi4-mini-instruct2026.05 | 26.23 | |
| EntropyBackbone=Phi4-mini-instruct2026.05 | 25.48 | |
| VanillaBackbone=Phi4-mini-instruct2026.05 | 15.61 |