Reinforcement Learning on Nim Gymnasium
1Mean Best RewardR2PO
Evaluation results
| Method | Links | |
|---|---|---|
| R2POEpisodes per run=4,000, Language Model backbone=gpt-oss:20b, Independent runs=102026.05 | 1 | |
| ProPS+Episodes per run=8,000, Language Model backbone=GPT-4o, Independent runs=102026.05 | 0.97 | |
| A2CEpisodes per run=8,000, Independent runs=102026.05 | 0.58 | |
| TRPOEpisodes per run=8,000, Independent runs=102026.05 | 0.5 | |
| ProPSEpisodes per run=8,000, Language Model backbone=GPT-4o, Independent runs=102026.05 | 0.33 |