ResearchBenchmarksG2 balanced on walker2d medium-replay v2Follow581.5Mean Raw ReturnFrozen209.284305.917402.55499.183Apr 23, 2026Evaluation ResultsMethodMethodLinksMean Raw ReturnFrozen2026.04581.5CQL2026.04417.5PoEalpha=0.3alpha=0.32026.04347.2Additivealpha=0.5alpha=0.52026.04319.9IQL2026.04298.9Prior2026.04296.1PoEalpha=0.5alpha=0.52026.04261.2KL-Regalpha=0.5alpha=0.52026.04261.2AWR2026.04223.6