ResearchBenchmarksShadowHandBlockStack on Bi-DexHandsShadowHandBlockStack on Bi-DexHandsFollow555AUC 0:NPPO149.4254.7360465.3Jun 7, 2026Evaluation resultsMethodMethodLinksAUC 0:NTtauFinal Success CountAUC GainPPOTuning=HARBORTuning=HARBOR2026.0655544.2665612.5PPOTuning=DefaultTuning=Default2026.0649365.89716—TD3Tuning=HARBORTuning=HARBOR2026.061993.331814.7SACTuning=HARBORTuning=HARBOR2026.061904.7719314.7TD3Tuning=DefaultTuning=Default2026.0619011.1209—SACTuning=DefaultTuning=Default2026.0616519.75189—