List your company

Navigation on FrozenLake

27Success RateActFocus
4.22410.13716.0521.963May 14, 2026

Evaluation results

MethodLinks
ActFocusBackbone=3B, Evaluation Step=2002026.0527
ActFocusAlgorithm=PPO, Model Scale=3B2026.0527
UniformAlgorithm=GRPO, Model Scale=3B2026.0526.2
ActFocusAlgorithm=GRPO, Model Scale=0.5B2026.0525.2
UniformAlgorithm=PPO, Model Scale=3B2026.0524.6
UniformAlgorithm=GRPO, Model Scale=0.5B2026.0523.8
ActFocusAlgorithm=PPO, Model Scale=0.5B2026.0520.9
ActFocusAlgorithm=GRPO, Model Scale=3B2026.0519.1
UniformBackbone=3B, Evaluation Step=2002026.0516.6
UniformAlgorithm=PPO, Model Scale=0.5B2026.0516.2
ActFocusBackbone=0.5B, Evaluation Step=2002026.0515.6
UniformBackbone=0.5B, Evaluation Step=2002026.055.1