Planning on Tower of Hanoi (held-out)
72.7AccuracyMemRL
Evaluation Results
| Method | Links | |
|---|---|---|
| MemRLtraining samples=1002026.05 | 72.7 | |
| MemRLtraining samples=52026.05 | 51.7 | |
| GEPAtraining samples=52026.05 | 43.3 | |
| COREtraining samples=1002026.05 | 42.7 | |
| COREtraining samples=102026.05 | 42.3 | |
| COREtraining samples=52026.05 | 40 | |
| MemRLtraining samples=102026.05 | 39.3 | |
| GEPAtraining samples=1002026.05 | 35.3 | |
| GEPAtraining samples=102026.05 | 31 | |
| Episodic RAGtraining samples=1002026.05 | 30.3 | |
| Episodic RAGtraining samples=52026.05 | 28.7 | |
| Episodic RAGtraining samples=102026.05 | 24.3 | |
| No Learningtraining samples=52026.05 | 17.9 | |
| No Learningtraining samples=102026.05 | 17.9 | |
| No Learningtraining samples=1002026.05 | 17.9 | |
| GRPOtraining samples=102026.05 | 12 | |
| GRPOtraining samples=1002026.05 | 10.7 | |
| GRPOtraining samples=52026.05 | 7.7 |