ResearchBenchmarksShadowHandDoorOpenIn on Bi-DexHandsShadowHandDoorOpenIn on Bi-DexHandsFollow399AUC0:NPPO278.36309.68341372.32Jun 7, 2026Evaluation resultsMethodMethodLinksAUC0:NTtauFinal SuccessGain AUCPPOTuning=HARBORTuning=HARBOR2026.0639934.714870.9PPOTuning=DefaultTuning=Default2026.06395—412—TD3Tuning=HARBORTuning=HARBOR2026.063482.833689SACTuning=HARBORTuning=HARBOR2026.063197.4435912.5TD3Tuning=DefaultTuning=Default2026.063196.94354—SACTuning=DefaultTuning=Default2026.0628315.87345—