Mathematical Reasoning on GSM8K (Accuracy, Uintra, Uinter, Usys)
92.3AccuracyUDPO
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| UDPON Agents in MAD=52026.02 | 92.3 | 0.065 | 0.038 | 0.058 | |
| UDPON Agents in MAD=102026.02 | 89.8 | 0.071 | 0.055 | 0.072 | |
| UDPON Agents in MAD=32026.02 | 84.6 | 0.068 | 0.052 | 0.078 | |
| RMAACN Agents in MAD=52026.02 | 75.8 | 0.168 | 0.132 | 0.215 | |
| MAPPON Agents in MAD=52026.02 | 73.6 | 0.185 | 0.145 | 0.232 | |
| RMAACN Agents in MAD=102026.02 | 72.1 | 0.175 | 0.145 | 0.228 | |
| MAPPON Agents in MAD=102026.02 | 70.2 | 0.192 | 0.158 | 0.248 | |
| Standard MADN Agents in MAD=52026.02 | 68.4 | 0.228 | 0.172 | 0.285 | |
| RMAACN Agents in MAD=32026.02 | 66.3 | 0.205 | 0.172 | 0.258 | |
| Standard MADN Agents in MAD=102026.02 | 65.7 | 0.235 | 0.188 | 0.298 | |
| MAPPON Agents in MAD=32026.02 | 64.8 | 0.218 | 0.185 | 0.278 | |
| Standard MADN Agents in MAD=32026.02 | 51.2 | 0.231 | 0.268 | 0.372 |