ResearchBenchmarksHarmfulness Evaluation on HEx-PHI & AdvBench PrimaryFollow1.15Primary AverageBase0.99762.02633.0554.0837Oct 17, 2025Evaluation ResultsMethodMethodLinksPrimary AverageBase2025.101.15Base2025.101.15SFT2025.104.81SFT2025.104.81RLVR2025.104.96RLVR2025.104.96