List your company

Jailbreaking on Attack Prompts In-domain

76ASR (TA)RogueMerge
0.0819.7939.559.21Jun 2, 2026

Evaluation results

MethodLinks
RogueMerge2026.067692548777.3
LoBAM*2026.062215181918.5
SFT2026.0643143
Clean2026.0632122
MergeHijacking*2026.0635253.8