The Stability Gap in Continual Reinforcement Learningwith Matthia Sabatelli, Gido M. van de Ven1 versionW_y4nvf4ahv1 · current