ResearchBenchmarksOffline Reinforcement Learning on OGBench Noisy TasksFollow96AntMaze Medium Explore Success RateMBTD(λ)73.1279.068590.94May 15, 2026Evaluation ResultsMethodMethodLinksAntMaze Medium Explore Success RateAntMaze Large Explore Success RateCube Double Noisy Success RateScene Noisy Success RatePuzzle 4x4 Noisy Success RateAverage Success RateMBTD(λ)2026.059620242032GHM2026.0591142457438UHM2026.0589261861139DTD(λ)2026.05811826023ReBRAC†2026.057413227023