Safe Continual Reinforcement Learning on Safe Continual World benchmark leaderboard · SOTA2 Research

Safe Continual Reinforcement Learning on Safe Continual World

89Success RatePPO

Evaluation Results

Method	Links
PPO2026.04		89	0.44	112,210.1
PPO+EWC2026.04		88	0.366	130,252.9
CPO2026.04		85	0.56	68,817.6
CF-EWC2026.04		79	0.51	126,258.7
Safe EWC2026.04		14	0.07	25,014.1
PPO-Lag2026.04		12	0.025	55,758.4
CPPO-PID2026.04		1	0.618	18,056.7