List your company

Response Safety on JailBreakV-28K (avg)

0.975JBV-R ScoreDACO
0.276120.457560.6390.82044Apr 10, 2026

Evaluation results

MethodLinks
DACOTarget Model=InternVL3.5-8B-Instruct [108], Steering Method=DACO (Ours)2026.040.9750.97
MOP [56]Target Model=InternVL3.5-8B-Instruct [108], Steering Method=MOP [56]2026.040.9170.926
DACOTarget Model=Qwen2.5-VL-7B-Instruct [5], Steering Method=DACO (Ours)2026.040.9030.841
DACOTarget Model=LLaVA1.6-Mistral-7B [48], Steering Method=DACO (Ours)2026.040.8850.803
ActAdd [50, 101, 134]Target Model=InternVL3.5-8B-Instruct [108], Steering Method=ActAdd [50, 101, 134]2026.040.860.847
MOP [56]Target Model=Qwen2.5-VL-7B-Instruct [5], Steering Method=MOP [56]2026.040.8350.752
MOP [56]Target Model=LLaVA1.6-Mistral-7B [48], Steering Method=MOP [56]2026.040.8110.69
Prompting [52]Target Model=InternVL3.5-8B-Instruct [108], Steering Method=Prompting [52]2026.040.8090.755
ActAdd [50, 101, 134]Target Model=LLaVA1.6-Mistral-7B [48], Steering Method=ActAdd [50, 101, 134]2026.040.7830.559
No SteeringTarget Model=InternVL3.5-8B-Instruct [108], Steering Method=No Steering2026.040.730.649
ActAdd [50, 101, 134]Target Model=Qwen2.5-VL-7B-Instruct [5], Steering Method=ActAdd [50, 101, 134]2026.040.6910.675
Prompting [52]Target Model=Qwen2.5-VL-7B-Instruct [5], Steering Method=Prompting [52]2026.040.6590.622
No SteeringTarget Model=Qwen2.5-VL-7B-Instruct [5], Steering Method=No Steering2026.040.5640.543
Prompting [52]Target Model=LLaVA1.6-Mistral-7B [48], Steering Method=Prompting [52]2026.040.4710.31
No SteeringTarget Model=LLaVA1.6-Mistral-7B [48], Steering Method=No Steering2026.040.3030.203