ResearchBenchmarksJailbreak on HarmfulQJailbreak on HarmfulQFollow18ASRDS-R1-0.724.14913.86Apr 10, 2026Evaluation resultsMethodMethodLinksASRDeltaDS-R1Model=DS-R1, Conflict=...Model=DS-R1, Conflict=dilemma2026.04180.18DS-R1Model=DS-R1, Conflict=...Model=DS-R1, Conflict=inner2026.04120.12DS-R1Model=DS-R1, Conflict=...Model=DS-R1, Conflict=direct_q2026.040—