Process-level Reward Modeling on PROCESSBENCH Omni-MATH
2.8Error Rate
SPARE-Llama3-8B
Updated 1mo ago
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
2025.06 | 2.8 | 82.2 | 5.4 | |
2025.06 | 10.9 | 51.9 | 16.9 | |
2025.06 | 14 | 41.9 | 21 | |
2025.06 | 14 | 83.8 | 23.9 | |
2025.06 | 14.2 | 73 | 23.8 | |
| 29.8 | 86.1 | 44.3 |