ResearchBenchmarksReinforcement Learning on Atari (PPO Comparison Metrics)Reinforcement Learning on Atari (PPO Comparison Metrics)Follow105Median Score (% PPO)SV-PPO103.96104.23104.5104.77May 6, 2026Evaluation resultsMethodMethodLinksMedian Score (% PPO)Mean ScorePercentage of Games Beating PPOKL Divergence (Policy π)KL Divergence (Beta β)SV-PPOUpdate Interval Varian...Update Interval Variant=Dynamic2026.05105113.268.81.960.7SV-PPOUpdate Interval Varian...Update Interval Variant=Static2026.05104133.356.26.110.94