List your company
ResearchFormerly WizWand

Reward Prediction on M3A trajectories

91.4AccuracyPRORE
69.5675.2380.986.57Sep 26, 2025
Updated 6mo ago

Evaluation results

MethodLinks
PRORELast N States=Proactive2025.0991.488.9
DigiRLLast N States=Full2025.0982.680.8
Step-CriticLast N States=Full2025.0982.679.6
DistRLLast N States=Full2025.0981.779.6
WebRLLast N States=12025.0981.779.2
WebRLLast N States=Full2025.0981.779.2
Step-CriticLast N States=12025.0981.779.4
DigiRLLast N States=12025.0974.873.4
DistRLLast N States=12025.0970.468.5