List your company

Utility Assessment on AgentDojo (test)

100UtilitySemantic-Injection
16.935238.500160.06581.6299Feb 18, 2026

Evaluation results

MethodLinks
Semantic-InjectionTarget Model=Gemini-3-Flash, Defense=w/o Defense2026.02100
Semantic-InjectionTarget Model=Gemini-3-Flash, Defense=Delimiter Spotlighting2026.02100
ChatInjectTarget Model=Gemini-3-Flash, Defense=Delimiter Spotlighting2026.02100
ChatInjectTarget Model=Gemini-3-Flash, Defense=Tag Filter2026.02100
ChatInjectTarget Model=Gemini-3-Flash, Defense=w/o Defense2026.0299.35
Single-TemplateTarget Model=Gemini-3-Flash, Defense=Tag Filter2026.0299.35
Semantic-InjectionTarget Model=Gemini-3-Flash, Defense=Tag Filter2026.0298.7
Single-TemplateTarget Model=Gemini-3-Flash, Defense=Delimiter Spotlighting2026.0298.05
ChatInjectTarget Model=DeepSeek-V3.2, Defense=Delimiter Spotlighting2026.0293.51
ChatInjectTarget Model=GPT-4.1, Defense=w/o Defense2026.0291.56
ChatInjectTarget Model=GPT-4.1, Defense=Delimiter Spotlighting2026.0290.26
ChatInjectTarget Model=GPT-4.1, Defense=Tag Filter2026.0290.26
ChatInjectTarget Model=DeepSeek-V3.2, Defense=w/o Defense2026.0288.31
Single-TemplateTarget Model=DeepSeek-V3.2, Defense=Tag Filter2026.0288.31
ChatInjectTarget Model=DeepSeek-V3.2, Defense=Tag Filter2026.0287.66
Single-TemplateTarget Model=Gemini-3-Flash, Defense=w/o Defense2026.0287.01
ChatInjectTarget Model=GPT-4.1-Mini, Defense=Tag Filter2026.0284.42
ChatInjectTarget Model=GPT-4o, Defense=Tag Filter2026.0282.47
Semantic-InjectionTarget Model=GPT-4o, Defense=Delimiter Spotlighting2026.0279.87
Semantic-InjectionTarget Model=GPT-4o, Defense=Tag Filter2026.0279.22
Semantic-InjectionTarget Model=GPT-4o, Defense=w/o Defense2026.0277.92
Single-TemplateTarget Model=Qwen3-Max, Defense=Tag Filter2026.0277.92
Semantic-InjectionTarget Model=GPT-4.1, Defense=Delimiter Spotlighting2026.0277.27
Single-TemplateTarget Model=GPT-4o, Defense=Tag Filter2026.0276.62
Semantic-InjectionTarget Model=GPT-4.1, Defense=w/o Defense2026.0275.97
ChatInjectTarget Model=GPT-4.1-Mini, Defense=Delimiter Spotlighting2026.0275.97
ChatInjectTarget Model=GPT-4.1-Mini, Defense=w/o Defense2026.0274.68
Semantic-InjectionTarget Model=GPT-4.1, Defense=Tag Filter2026.0274.03
PhantomTarget Model=Gemini-3-Flash, Defense=Tag Filter2026.0274.03
Semantic-InjectionTarget Model=Qwen3-Max, Defense=Delimiter Spotlighting2026.0272.08
ChatInjectTarget Model=Gemini-2.5-Flash, Defense=Delimiter Spotlighting2026.0270.13
ChatInjectTarget Model=GPT-4o, Defense=Delimiter Spotlighting2026.0268.83
Single-TemplateTarget Model=Gemini-2.5-Flash, Defense=Delimiter Spotlighting2026.0268.83
Single-TemplateTarget Model=Gemini-2.5-Flash, Defense=Tag Filter2026.0266.88
Single-TemplateTarget Model=GPT-4.1, Defense=Delimiter Spotlighting2026.0266.23
Semantic-InjectionTarget Model=Qwen3-Max, Defense=Tag Filter2026.0264.94
Single-TemplateTarget Model=GPT-4.1, Defense=w/o Defense2026.0264.29
ChatInjectTarget Model=Qwen3-Max, Defense=Tag Filter2026.0264.29
Semantic-InjectionTarget Model=Gemini-2.5-Flash, Defense=Tag Filter2026.0264.29
Semantic-InjectionTarget Model=Qwen3-Max, Defense=w/o Defense2026.0263.64
ChatInjectTarget Model=Gemini-2.5-Flash, Defense=Tag Filter2026.0261.69
ChatInjectTarget Model=Gemini-2.5-Flash, Defense=w/o Defense2026.0261.04
Semantic-InjectionTarget Model=GPT-4.1-Mini, Defense=Delimiter Spotlighting2026.0261.04
Semantic-InjectionTarget Model=GPT-4.1-Mini, Defense=Tag Filter2026.0260.39
ChatInjectTarget Model=GPT-4o, Defense=w/o Defense2026.0259.74
Single-TemplateTarget Model=Gemini-2.5-Flash, Defense=w/o Defense2026.0259.74
Semantic-InjectionTarget Model=Qwen-Flash, Defense=Tag Filter2026.0259.09
PhantomTarget Model=Qwen3-Max, Defense=Tag Filter2026.0258.44
Semantic-InjectionTarget Model=GPT-4.1-Mini, Defense=w/o Defense2026.0257.14
Semantic-InjectionTarget Model=Gemini-2.5-Flash, Defense=w/o Defense2026.0257.14
Semantic-InjectionTarget Model=Gemini-2.5-Flash, Defense=Delimiter Spotlighting2026.0257.14
Semantic-InjectionTarget Model=Qwen-Flash, Defense=w/o Defense2026.0256.49
PhantomTarget Model=Gemini-3-Flash, Defense=Delimiter Spotlighting2026.0255.84
PhantomTarget Model=DeepSeek-V3.2, Defense=Tag Filter2026.0255.84
ChatInjectTarget Model=Qwen3-Max, Defense=Delimiter Spotlighting2026.0255.19
PhantomTarget Model=Gemini-2.5-Flash, Defense=Tag Filter2026.0253.9
PhantomTarget Model=Gemini-2.5-Flash, Defense=Delimiter Spotlighting2026.0253.25
Semantic-InjectionTarget Model=Qwen-Flash, Defense=Delimiter Spotlighting2026.0250.65
Single-TemplateTarget Model=Gemini-3-Flash, Defense=Injection Detector2026.0249.35
ChatInjectTarget Model=Qwen-Flash, Defense=Tag Filter2026.0248.7
ChatInjectTarget Model=Qwen3-Max, Defense=w/o Defense2026.0246.75
Single-TemplateTarget Model=GPT-4.1, Defense=Tag Filter2026.0246.1
Single-TemplateTarget Model=GPT-4o, Defense=Delimiter Spotlighting2026.0245.45
Single-TemplateTarget Model=GPT-4o, Defense=w/o Defense2026.0244.16
PhantomTarget Model=GPT-4o, Defense=Tag Filter2026.0243.51
Single-TemplateTarget Model=Gemini-2.5-Flash, Defense=Injection Detector2026.0243.51
ChatInjectTarget Model=Gemini-3-Flash, Defense=Injection Detector2026.0242.86
Semantic-InjectionTarget Model=DeepSeek-V3.2, Defense=Delimiter Spotlighting2026.0242.21
Semantic-InjectionTarget Model=Gemini-3-Flash, Defense=Injection Detector2026.0242.21
Semantic-InjectionTarget Model=DeepSeek-V3.2, Defense=Tag Filter2026.0241.56
PhantomTarget Model=Gemini-2.5-Flash, Defense=w/o Defense2026.0240.91
Semantic-InjectionTarget Model=DeepSeek-V3.2, Defense=w/o Defense2026.0240.91
Semantic-InjectionTarget Model=GPT-4o, Defense=Injection Detector2026.0237.01
Semantic-InjectionTarget Model=Gemini-2.5-Flash, Defense=Injection Detector2026.0234.42
Single-TemplateTarget Model=Qwen3-Max, Defense=Delimiter Spotlighting2026.0231.82
ChatInjectTarget Model=Qwen-Flash, Defense=w/o Defense2026.0229.87
ChatInjectTarget Model=GPT-4.1, Defense=Injection Detector2026.0229.22
Single-TemplateTarget Model=Qwen3-Max, Defense=w/o Defense2026.0228.57
PhantomTarget Model=Gemini-2.5-Flash, Defense=Injection Detector2026.0228.57
PhantomTarget Model=GPT-4o, Defense=Delimiter Spotlighting2026.0227.92
PhantomTarget Model=Gemini-3-Flash, Defense=w/o Defense2026.0226.62
Semantic-InjectionTarget Model=Qwen3-Max, Defense=Injection Detector2026.0225.97
ChatInjectTarget Model=Gemini-2.5-Flash, Defense=Injection Detector2026.0225.97
ChatInjectTarget Model=Qwen-Flash, Defense=Delimiter Spotlighting2026.0225.32
Semantic-InjectionTarget Model=GPT-4.1, Defense=Injection Detector2026.0225.32
PhantomTarget Model=GPT-4o, Defense=w/o Defense2026.0224.03
Semantic-InjectionTarget Model=Qwen-Flash, Defense=Injection Detector2026.0224.03
ChatInjectTarget Model=DeepSeek-V3.2, Defense=Injection Detector2026.0224.03
Single-TemplateTarget Model=GPT-4.1-Mini, Defense=w/o Defense2026.0223.38
Single-TemplateTarget Model=DeepSeek-V3.2, Defense=w/o Defense2026.0223.38
Single-TemplateTarget Model=Qwen-Flash, Defense=Tag Filter2026.0223.38
ChatInjectTarget Model=GPT-4.1-Mini, Defense=Injection Detector2026.0223.38
PhantomTarget Model=Qwen3-Max, Defense=w/o Defense2026.0222.73
PhantomTarget Model=DeepSeek-V3.2, Defense=w/o Defense2026.0222.73
PhantomTarget Model=Qwen3-Max, Defense=Delimiter Spotlighting2026.0222.73
Semantic-InjectionTarget Model=GPT-4.1-Mini, Defense=Injection Detector2026.0222.73
Single-TemplateTarget Model=GPT-4.1-Mini, Defense=Delimiter Spotlighting2026.0222.08
Single-TemplateTarget Model=GPT-4.1, Defense=Injection Detector2026.0222.08
Semantic-InjectionTarget Model=DeepSeek-V3.2, Defense=Injection Detector2026.0221.43
PhantomTarget Model=GPT-4.1, Defense=Delimiter Spotlighting2026.0220.13