List your company

Model Accuracy on Model Utility Evaluation Set

30.22Accuracy (ACC)ChatBug
-1.16726.981415.1323.2786Apr 14, 2026

Evaluation results

MethodLinks
ChatBugModel=Llama3-8B, Attack Variant=Mismatch-∅2026.0430.22
ChatBugModel=Qwen2.5-7B, Attack Variant=Mismatch-V2026.0427.92
ChatBugModel=Llama3-8B, Attack Variant=Mismatch-V2026.0427.85
ChatBugModel=Qwen2.5-7B, Attack Variant=Mismatch-∅2026.0422.65
ChatBugModel=Llama2-7B, Attack Variant=Mismatch-C2026.0417.2
ChatBugModel=DeepSeek-7B, Attack Variant=Mismatch-C2026.0413.04
ChatBugModel=Llama3-8B, Attack Variant=Mismatch-C2026.0411.33
ChatBugModel=Gemma3-4B, Attack Variant=Mismatch-V2026.049.1
ChatBugModel=Gemma3-4B, Attack Variant=Mismatch-∅2026.046.99
ChatBugModel=DeepSeek-7B, Attack Variant=Mismatch-∅2026.046.62
ChatBugModel=Gemma3-4B, Attack Variant=Mismatch-C2026.046.53
ChatBugModel=DeepSeek-7B, Attack Variant=Mismatch-V2026.043.72
ChatBugModel=Qwen3-8B, Attack Variant=Mismatch-∅2026.043.58
ChatBugModel=Qwen2.5-7B, Attack Variant=Mismatch-C2026.043.28
ChatBugModel=Qwen3-8B, Attack Variant=Mismatch-V2026.043.01
TemplateFuzzModel=Llama3-8B, Attack Variant=Top-1 Template2026.042.75
ChatBugModel=Qwen3-8B, Attack Variant=Mismatch-C2026.041.99
ChatBugModel=Llama2-7B, Attack Variant=Mismatch-V2026.041.71
TemplateFuzzModel=Qwen3-8B, Attack Variant=Top-1 Template2026.041.45
TemplateFuzzModel=DeepSeek-7B, Attack Variant=Top-1 Template2026.041.41
ChatBugModel=Llama2-7B, Attack Variant=Mismatch-∅2026.041.18
TemplateFuzzModel=Gemma3-4B, Attack Variant=Top-1 Template2026.041.01
TemplateFuzzModel=Llama2-7B, Attack Variant=Top-1 Template2026.040.21
TemplateFuzzModel=Qwen2.5-7B, Attack Variant=Top-1 Template2026.040.04