ResearchBenchmarksShadowHandOver on Bi-DexHandsShadowHandOver on Bi-DexHandsFollow26.1AUC0:NPPO17.36419.63221.924.168Jun 7, 2026Evaluation resultsMethodMethodLinksAUC0:NTtauFinal ScoreGain AUCPPOTuning=HARBORTuning=HARBOR2026.0626.19.3930.147.8SACTuning=HARBORTuning=HARBOR2026.0622.315.0727.916.5TD3Tuning=HARBORTuning=HARBOR2026.0619.612.0824.69.5SACTuning=DefaultTuning=Default2026.0619.238.1423.1—TD3Tuning=DefaultTuning=Default2026.0617.915.1923.3—PPOTuning=DefaultTuning=Default2026.0617.734.0826.8—