List your company

CoT Naturalness on AQuA

21.34PPLMirageBD
20.095628.495336.89545.2947Apr 8, 2026

Evaluation results

MethodLinks
MirageBDBackbone=Qwen2.5-1.5B, Evaluation Setting=Trigger Activation2026.0421.34
DecepChainBackbone=Qwen2.5-1.5B, Evaluation Setting=Trigger Activation2026.0436.2
BadChainBackbone=Qwen2.5-1.5B, Evaluation Setting=Trigger Activation2026.0452.45