List your company

Navigation on FrozenLake 6x6

0.75RewardASK (72B)
-0.030.17250.3750.5775Apr 2, 2026

Evaluation results

MethodLinks
ASK (72B)Model Size=72B, trained_on=8x82026.040.7521.5810050.33
ASK (32B)Model Size=32B, trained_on=8x82026.040.6922.9210052.35
ASK (1.5B)Model Size=1.5B, trained_on=8x82026.040.118.610028.96
ASK (14B)Model Size=14B, trained_on=8x82026.040.0160.3110051.5
PPOtrained_on=8x82026.04036.8——
ASK (0.5B)Model Size=0.5B, trained_on=8x82026.04031.61100100
ASK (3B)Model Size=3B, trained_on=8x82026.04050.3810041.43
ASK (7B)Model Size=7B, trained_on=8x82026.0406110050.31