List your company

Reasoning on AQuA

72CACC (%)BadChain
32.06442.43252.863.168Apr 8, 2026

Evaluation results

MethodLinks
BadChainBackbone=Qwen2.5-7B, Poison Ratio=0.052026.04720.2
DecepChainBackbone=Qwen2.5-7B, Poison Ratio=0.052026.0471.90
MirageBDBackbone=Qwen2.5-7B, Poison Ratio=0.052026.0471.394.9
BenignBackbone=Qwen2.5-7B, Poison Ratio=0.052026.0469—
SFT+RLBackbone=Qwen2.5-7B, Poison Ratio=0.052026.0466.40
DecepChainBackbone=Qwen2.5-3B, Poison Ratio=0.052026.0462.10
BenignBackbone=Qwen2.5-3B, Poison Ratio=0.052026.0459.4—
MirageBDBackbone=Llama3.2-3B, Poison Ratio=0.052026.0458.877.6
SFT+RLBackbone=Llama3.1-8B, Poison Ratio=0.052026.0458.881.5
MirageBDBackbone=Llama3.1-8B, Poison Ratio=0.052026.0458.180.7
MirageBDBackbone=Qwen2.5-3B, Poison Ratio=0.052026.0457.797.1
DecepChainBackbone=Llama3.2-3B, Poison Ratio=0.052026.0457.20
BenignBackbone=Llama3.2-3B, Poison Ratio=0.052026.0456.6—
SFT+RLBackbone=Qwen2.5-3B, Poison Ratio=0.052026.045612.9
SFT+RLBackbone=Llama3.2-3B, Poison Ratio=0.052026.04560
BadChainBackbone=Llama3.1-8B, Poison Ratio=0.052026.0454.60
BadChainBackbone=Llama3.2-3B, Poison Ratio=0.052026.0453.90
MirageBDBackbone=Qwen2.5-1.5B, Poison Ratio=0.052026.0452.797
BenignBackbone=Qwen2.5-1.5B, Poison Ratio=0.052026.0451.1—
BadChainBackbone=Qwen2.5-1.5B, Poison Ratio=0.052026.04510.1
BenignBackbone=Llama3.1-8B, Poison Ratio=0.052026.0450.2—
DecepChainBackbone=Qwen2.5-1.5B, Poison Ratio=0.052026.0448.50
DecepChainBackbone=Llama3.1-8B, Poison Ratio=0.052026.0448.20
SFT+RLBackbone=Qwen2.5-1.5B, Poison Ratio=0.052026.0447.478.4
BadChainBackbone=Qwen2.5-3B, Poison Ratio=0.052026.0433.60