List your company

Trigger inversion on Yahoo

30Trigger Inversion Success RateCSO(discrete)
-0.167.6715.523.33Jun 30, 2026

Evaluation results

MethodLinks
CSO(discrete)Target trigger (length)=Tell me seriously(3 tokens), Label type=clean label, Poison rate=high, Model=Qwen3-0.6b2026.0630
CSO(discrete)Target trigger (length)=Tell me seriously(3 tokens), Label type=dirty label, Poison rate=high, Model=Qwen3-0.6b2026.0629
CSO(discrete)Target trigger (length)=Tell me seriously(3 tokens), Label type=dirty label, Poison rate=low, Model=Qwen3-0.6b2026.0623
CSO(discrete)Target trigger (length)=Tell me seriously(3 tokens), Label type=clean label, Poison rate=low, Model=Qwen3-0.6b2026.0622
CSO(discrete)Target trigger (length)=mb(1 token), Label type=clean label, Poison rate=low, Model=Qwen3-0.6b2026.0610
CSO(discrete)Target trigger (length)=mb(1 token), Label type=clean label, Poison rate=high, Model=Qwen3-0.6b2026.0610
CSO(discrete)Target trigger (length)=mb(1 token), Label type=dirty label, Poison rate=low, Model=Qwen3-0.6b2026.0610
CSO(discrete)Target trigger (length)=mb(1 token), Label type=dirty label, Poison rate=high, Model=Qwen3-0.6b2026.0610
MMTarget trigger (length)=mb(1 token), Label type=clean label, Poison rate=high, Model=Qwen3-0.6b2026.0610
MMTarget trigger (length)=mb(1 token), Label type=dirty label, Poison rate=high, Model=Qwen3-0.6b2026.0610
MMTarget trigger (length)=Tell me seriously(3 tokens), Label type=dirty label, Poison rate=high, Model=Qwen3-0.6b2026.069
MMTarget trigger (length)=mb(1 token), Label type=dirty label, Poison rate=low, Model=Qwen3-0.6b2026.069
MMTarget trigger (length)=Tell me seriously(3 tokens), Label type=dirty label, Poison rate=low, Model=Qwen3-0.6b2026.068
MMTarget trigger (length)=mb(1 token), Label type=clean label, Poison rate=low, Model=Qwen3-0.6b2026.068
MMTarget trigger (length)=Tell me seriously(3 tokens), Label type=clean label, Poison rate=high, Model=Qwen3-0.6b2026.065
MMTarget trigger (length)=Tell me seriously(3 tokens), Label type=clean label, Poison rate=low, Model=Qwen3-0.6b2026.063
DBSTarget trigger (length)=Tell me seriously(3 tokens), Label type=clean label, Poison rate=high, Model=Qwen3-0.6b2026.063
DBSTarget trigger (length)=Tell me seriously(3 tokens), Label type=dirty label, Poison rate=high, Model=Qwen3-0.6b2026.063
PICCOLOTarget trigger (length)=Tell me seriously(3 tokens), Label type=clean label, Poison rate=high, Model=Qwen3-0.6b2026.063
UATTarget trigger (length)=mb(1 token), Label type=clean label, Poison rate=high, Model=Qwen3-0.6b2026.062
UATTarget trigger (length)=mb(1 token), Label type=clean label, Poison rate=low, Model=Qwen3-0.6b2026.061
PICCOLOTarget trigger (length)=Tell me seriously(3 tokens), Label type=dirty label, Poison rate=high, Model=Qwen3-0.6b2026.061