Navigation on FrozenLake 7x7
0.68RewardASK (72B)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| ASK (72B)Model Size=72B, trained_on=8x82026.04 | 0.68 | 27.1 | 100 | 50.88 | |
| ASK (32B)Model Size=32B, trained_on=8x82026.04 | 0.58 | 30.2 | 100 | 53.36 | |
| ASK (1.5B)Model Size=1.5B, trained_on=8x82026.04 | 0.1 | 8.08 | 100 | 29.14 | |
| PPOtrained_on=8x82026.04 | 0 | 30.22 | — | — | |
| ASK (0.5B)Model Size=0.5B, trained_on=8x82026.04 | 0 | 25.28 | 100 | 99.95 | |
| ASK (3B)Model Size=3B, trained_on=8x82026.04 | 0 | 53.51 | 100 | 44.23 | |
| ASK (7B)Model Size=7B, trained_on=8x82026.04 | 0 | 61 | 100 | 51.57 | |
| ASK (14B)Model Size=14B, trained_on=8x82026.04 | 0 | 61 | 100 | 50.51 |