ResearchBenchmarksOpen-ended Dialogue on ID AverageFollow72.2Win RateCausalRM41.10449.17757.2565.323Jan 29, 2026Evaluation ResultsMethodMethodLinksWin RateTie RateLoss RateCausalRMOpponent=SFTOpponent=SFT2026.0172.221.66.2CausalRMOpponent=Standard RMOpponent=Standard RM2026.0154.833.112.1CausalRMOpponent=InfoRMOpponent=InfoRM2026.0145.537.616.9CausalRMOpponent=GoalRMOpponent=GoalRM2026.0142.341.616.1