Reinforcement Learning on FrozenLake Gymnasium
0.93Mean Best RewardR2PO
Evaluation results
| Method | Links | |
|---|---|---|
| R2POEpisodes per run=4,000, Language Model backbone=gpt-oss:20b, Independent runs=102026.05 | 0.93 | |
| ProPSEpisodes per run=8,000, Language Model backbone=GPT-4o, Independent runs=102026.05 | 0.57 | |
| TRPOEpisodes per run=8,000, Independent runs=102026.05 | 0.22 | |
| ProPS+Episodes per run=8,000, Language Model backbone=GPT-4o, Independent runs=102026.05 | 0.19 | |
| PPOEpisodes per run=8,000, Independent runs=102026.05 | 0.16 |