ResearchBenchmarksCounterfactual Policy Evaluation on GridWorld (p = 0.4) - Almost CatastrophicFollow14Lowest Cumulative RewardICFMDP-622.48-457.24-292-126.76Feb 19, 2025Evaluation ResultsMethodMethodLinksLowest Cumulative RewardICFMDP2025.0214Gumbel-max2025.02-598