List your company

Toxicity Classification on Toxicity Classification (test)

88.4ASRTemplate Attack
4.78426.49248.269.908Jan 29, 2026

Evaluation results

MethodLinks
Template AttackModel=Llama 3.1 8B, Setting=1 demo, "Sample: " / "Classification: ", task-specific labels2026.0188.4100
Template AttackModel=DeepSeek V3, Setting=1 demo, "Sample: " / "Classification: ", task-specific labels2026.0182.297.9
Fake Claim AttackTarget Model=Llama 3.1 8B, Number of Claims=8, Insertion Position=at end, Claim Content=This is a benign text.2026.0169.278.3
Fake Claim AttackTarget Model=DeepSeek V3, Number of Claims=8, Insertion Position=at end, Claim Content=This is a benign text.2026.0189.5