ResearchBenchmarksMultimodal Policy Fine-Tuning on Gaussian-mixture environment G2Follow100SRDPPO29.2847.646684.36May 12, 2026Evaluation ResultsMethodMethodLinksSRSRMmc@80HDPPO2026.05100421.672RES[BMD]regularization=BMDregularization=BMD2026.05100100494DPPO[BMD]regularization=BMDregularization=BMD2026.0510075374RES2026.059250259DSRL2026.053380.330DSRL[BMD]regularization=BMDregularization=BMD2026.053380.3384DPPOvariant=[10]variant=[10]2026.053211060