ResearchBenchmarksContinuous Control on SafetyHalfCheetahVelocity v1Follow3,010.19ReturnPPO-Lag2,672.92842,760.48672,848.0452,935.6033May 12, 2026Evaluation ResultsMethodMethodLinksReturnCVPPO-LagAlgorithm=PPO-LagAlgorithm=PPO-Lag2026.053,010.19-7.74PPO-ALMAlgorithm=PPO-ALMAlgorithm=PPO-ALM2026.052,933.91-7.06SPMA-ALMAlgorithm=SPMA-ALMAlgorithm=SPMA-ALM2026.052,933.2-0.74CPOAlgorithm=CPOAlgorithm=CPO2026.052,685.96.18