List your company

Jailbreaking on Attack Prompts Out-of-domain

67.7ASR (TA)RogueMerge
16.32429.6624356.338Jun 2, 2026

Evaluation results

MethodLinks
RogueMerge2026.0667.772.361.371.768.3
LoBAM*2026.0655.754.349.354.753.5
SFT2026.0641.74332.742.339.9
MergeHijacking*2026.0640.342.33545.740.8
Clean2026.0618.336.320.33327