ResearchBenchmarksMulti-Objective Optimization on State-of-the-art MOO methodsMulti-Objective Optimization on State-of-the-art MOO methodsFollow1Number of Trained LLMsRiC0.950.97511.025Apr 6, 2026Evaluation resultsMethodMethodLinksNumber of Trained LLMsPolicy ImprovementInference Adaptation ScoreRiCPreference data=No, Lo...Preference data=No, Loss=SFT2026.041——MOCPreference data=No, Lo...Preference data=No, Loss=PPO2026.041——