ResearchBenchmarksReinforcement Learning on acrobot CleanReinforcement Learning on acrobot CleanFollow-79,144,263.52AUC@THFPS-84,226,816.1224-82,907,307.2737-81,587,798.425-80,268,289.5763Feb 6, 2026Evaluation resultsMethodMethodLinksAUC@TFinal@TDelta % AUC@TDelta % Final@THFPS2026.02-79,144,263.52-87.475.8-11.9PPO2026.02-84,031,333.33-78.2——