Mathematical Reasoning on MATH500 (Perf., Sp.)
49PerformanceBest Model
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Best ModelBackbone=LLaMA 3.1 8B, Shots=8-shot, #D=12025.10 | 49 | 1.02 | |
| BSBABackbone=LLaMA 3.1 8B, Shots=8-shot, #D=22025.10 | 45 | 1.03 | |
| BaselineBackbone=LLaMA 3.1 8B, Shots=8-shot2025.10 | 44 | — |