Continual Reinforcement Learning on MetaWorld CRL benchmark leaderboard · SOTA2 Research

Continual Reinforcement Learning on MetaWorld CRL

0.41MeanSPHERE

Evaluation Results

Method	Links
SPHERE2026.05		0.41	0.07	0.41	0.07
Top-K MoE2026.05		0.16	0.08	0.14	0.14