ResearchBenchmarksPolicy Ranking on AntFollow0RegretFQE-0.52.8756.259.625Jun 24, 2026Evaluation ResultsMethodMethodLinksRegretSpearman CorrelationFQE2026.0600.94LDT2Architecture=best-perf...Architecture=best-performing from Figure 42026.061.640.84MOReL2026.065.120.22ROMI2026.066.140.33HDTwin2026.066.180.24LNLLsimArchitecture=best-perf...Architecture=best-performing from Figure 42026.067.10.09MOPO2026.0611.980.05VaGraM2026.0612.50.38