List your company

LLM Agent Task Completion on AgentDojo No Attack

73.91Benign UtilityIPIGuard
30.95842.10953.2664.411Mar 11, 2026

Evaluation results

MethodLinks
IPIGuardModel=Gemini-2.52026.0373.91
AttriGuardModel=Gemini-2.52026.0367.39
SpotlightingModel=Llama3.3-70B2026.0366.3
SandwichingModel=Gemini-2.52026.0361.96
IPIGuardModel=Llama3.3-70B2026.0357.61
SpotlightingModel=Gemini-2.52026.0356.52
PromptArmorModel=Llama3.3-70B2026.0356.52
SandwichingModel=Llama3.3-70B2026.0356.52
PromptArmorModel=Gemini-2.52026.0354.35
MELONModel=Gemini-2.52026.0354.35
MELONModel=Llama3.3-70B2026.0354.35
AttriGuardModel=Llama3.3-70B2026.0354.35
DataSentinelModel=Llama3.3-70B2026.0353.26
DataSentinelModel=Gemini-2.52026.0351.09
PromptGuardModel=Llama3.3-70B2026.0345.65
PromptGuardModel=Gemini-2.52026.0340.22
CaMeLModel=Gemini-2.52026.0338.04
PI DetectorModel=Llama3.3-70B2026.0338.04
PIGuardModel=Llama3.3-70B2026.0336.96
CaMeLModel=Llama3.3-70B2026.0336.96
PIGuardModel=Gemini-2.52026.0334.78
PI DetectorModel=Gemini-2.52026.0332.61