List your company

Navigation on FourRooms 100 episodes (test)

0.64RewardASK+
0.28640.37820.470.5618Jul 2, 2026

Evaluation results

MethodLinks
ASK+Model=2B2026.070.6470202
ASK+Model=4B2026.070.6269252
SLMModel=4B2026.070.5769100—
ASKModel=4B2026.070.5255320
ASKModel=2B2026.070.5154320
PPOModel=—2026.070.5530—
SLMModel=2B2026.070.334100—