List your company

Jailbreak Defense and Utility Evaluation on Implicit risk dataset Education

6.4Jailbreak Success Rate (JSR)DPP-MENTOR
3.4423.4243.463.38Nov 10, 2025

Evaluation results

MethodLinks
DPP-MENTORModel=Qwen3-235B2025.116.44.842
MENTORModel=Qwen3-235B2025.117.24.8
MENTORModel=DeepSeek-R12025.118.44.562
G4D-RuleTreeModel=Qwen3-235B2025.118.64.644
Self-ReminderModel=Qwen3-235B2025.1194.628
DPP-MENTORModel=DeepSeek-R12025.119.84.818
G4D-RuleTreeModel=GPT-4o2025.1112.84.612
DPP-MENTORModel=GPT-4o2025.1113.24.684
MENTORModel=GPT-4o2025.1113.84.21
G4D-RuleTreeModel=DeepSeek-R12025.1114.64.712
G4DModel=GPT-4o2025.1116.24.656
Self-ReminderModel=DeepSeek-R12025.1116.84.822
G4DModel=DeepSeek-R12025.1118.64.84
G4DModel=Qwen3-235B2025.1119.24.728
Self-ReminderModel=GPT-4o2025.11274.644
DPPModel=Qwen3-235B2025.1141.84.824
RawModel=Qwen3-235B2025.1149.24.83
ProBModel=Qwen3-235B2025.1150.84.414
RPO-MENTORModel=Qwen3-235B2025.1152.24.764
ProBModel=DeepSeek-R12025.1152.44.436
RPOModel=Qwen3-235B2025.1155.24.864
DPPModel=DeepSeek-R12025.1158.24.872
ProBModel=GPT-4o2025.1163.44.41
RPO-MENTORModel=GPT-4o2025.1164.24.766
DPPModel=GPT-4o2025.1165.24.926
RPO-MENTORModel=DeepSeek-R12025.1165.64.518
RawModel=DeepSeek-R12025.1167.24.826
RPOModel=DeepSeek-R12025.1167.24.782
RPOModel=GPT-4o2025.11684.828
RawModel=GPT-4o2025.1180.44.838