CoT Naturalness on ECQA
20.15Perplexity (PPL)MirageBD
Evaluation Results
| Method | Links | |
|---|---|---|
| MirageBDBackbone=Qwen2.5-1.5B, Evaluation Setting=Trigger Activation2026.04 | 20.15 | |
| DecepChainBackbone=Qwen2.5-1.5B, Evaluation Setting=Trigger Activation2026.04 | 33.45 | |
| BadChainBackbone=Qwen2.5-1.5B, Evaluation Setting=Trigger Activation2026.04 | 50.12 |