List your company

Jailbreak Attack on ASR Benchmark

97.2ASRPuzzler
4.43228.51652.676.684Sep 15, 2025

Evaluation results

MethodLinks
PuzzlerTarget Model=QwQ-32B2025.0997.2
PuzzlerTarget Model=Gemma-3-1b-it2025.0994
PuzzlerTarget Model=Gemini-2.0-Flash-Lite2025.0986.1
PuzzlerTarget Model=Average2025.0985.7
NeuroStrikeTarget Model=Gemma-3-1b-it2025.0979.9
NeuroStrikeTarget Model=QwQ-32B2025.0978.9
PuzzlerTarget Model=Gemini-1.5-Pro2025.0975
PuzzlerTarget Model=Gemini-2.0-Flash2025.0973
NeuroStrikeTarget Model=Average2025.0963.7
NeuroStrikeTarget Model=Gemini-1.5-Pro2025.0955.7
PAIRTarget Model=Gemini-1.5-Pro2025.0954.9
NeuroStrikeTarget Model=Gemini-2.0-Flash2025.0954.7
NeuroStrikeTarget Model=Gemini-2.0-Flash-Lite2025.0949.2
PAIRTarget Model=Gemini-2.0-Flash2025.0937.3
PAIRTarget Model=Gemma-3-1b-it2025.0933.9
TAPTarget Model=Gemini-1.5-Pro2025.0932
PAIRTarget Model=Average2025.0931
TAPTarget Model=Gemma-3-1b-it2025.0920.4
PAIRTarget Model=QwQ-32B2025.0918.9
TAPTarget Model=Average2025.0917.3
TAPTarget Model=Gemini-2.0-Flash2025.0914
TAPTarget Model=QwQ-32B2025.0912.2
PAIRTarget Model=Gemini-2.0-Flash-Lite2025.099.8
TAPTarget Model=Gemini-2.0-Flash-Lite2025.098