List your company

Attack Success Rate on InjecAgent

88.3Attack Success Rate (ASR)Llama-4
-3.53220.30944.1567.991Sep 26, 2025

Evaluation results

MethodLinks
Llama-4Prompt=InjecPrompt, Technique=ChatInject + tool2025.0988.3
Llama-4Prompt=Multi-turn, Technique=ChatInject2025.0988.3
Llama-4Prompt=InjecPrompt, Technique=ChatInject2025.0979.4
GLM-4.5Prompt=InjecPrompt, Technique=ChatInject + tool2025.0972.2
Kimi-K2Prompt=InjecPrompt, Technique=ChatInject + tool2025.0972.2
GLM-4.5Prompt=Multi-turn, Technique=ChatInject2025.0971.5
GLM-4.5Prompt=InjecPrompt, Technique=ChatInject + think2025.0969.3
Kimi-K2Prompt=InjecPrompt, Technique=ChatInject2025.0967.4
Qwen-3Prompt=Multi-turn, Technique=ChatInject2025.0965.9
Kimi-K2Prompt=Multi-turn, Technique=ChatInject2025.0961
GLM-4.5Prompt=InjecPrompt, Technique=ChatInject2025.0957.3
Llama-4Prompt=InjecPrompt, Technique=default2025.0950.1
Qwen-3Prompt=InjecPrompt, Technique=ChatInject + tool2025.0942.1
Qwen-3Prompt=InjecPrompt, Technique=ChatInject + think2025.0940.1
Qwen-3Prompt=InjecPrompt, Technique=ChatInject2025.0939.4
GPT-ossPrompt=InjecPrompt, Technique=ChatInject + tool2025.0919.1
Grok-2Prompt=InjecPrompt, Technique=ChatInject2025.0917.7
Kimi-K2Prompt=Multi-turn, Technique=default2025.0917.2
GPT-ossPrompt=Multi-turn, Technique=ChatInject2025.0916.9
GPT-ossPrompt=InjecPrompt, Technique=ChatInject + think2025.0916.7
Llama-4Prompt=Multi-turn, Technique=default2025.0916.6
Grok-2Prompt=InjecPrompt, Technique=default2025.0916.5
Kimi-K2Prompt=InjecPrompt, Technique=default2025.0915.7
GPT-ossPrompt=InjecPrompt, Technique=ChatInject2025.0914.2
Qwen-3Prompt=Multi-turn, Technique=default2025.0910.7
Grok-2Prompt=Multi-turn, Technique=ChatInject2025.0910.4
Qwen-3Prompt=InjecPrompt, Technique=default2025.098.5
Grok-2Prompt=Multi-turn, Technique=default2025.091.6
GPT-ossPrompt=Multi-turn, Technique=default2025.090.1
GLM-4.5Prompt=Multi-turn, Technique=default2025.090.1
GPT-ossPrompt=InjecPrompt, Technique=default2025.090
GLM-4.5Prompt=InjecPrompt, Technique=default2025.090