Backdoor Detection

Benchmarks

Dataset Name	SOTA Method	Metric
CIFAR-10	K-Arm	TPR100	135	27d ago
CIFAR-10 (test)		TPR100	63	27d ago
LMSYS-Chat-1M		FPR on LMSYS0.1	50	1mo ago
MMLU		FPR (MMLU)0	50	1mo ago
Calibration Set		FPR (Calibration Set)36	50	1mo ago
Poisoned Prompt Set		Prompt Flagging Rate100	50	1mo ago
EMNIST		TPR100	49	27d ago
CIFAR-100	Coward	True Positive Rate (TPR)100	49	27d ago
GTSRB	K-Arm	TPR100	48	27d ago
SVHN		TPR100	30	27d ago
Balanced 50% clean, 50% backdoored (test)	VillanDiffusion	Detection Accuracy98.2	28	3mo ago
Stable Diffusion ObjRepAtt attacks v1.5	BlackMirror	Precision100	23	2mo ago
StrongREJECT prompts with triggers	WeightWatch	TPR100	20	1mo ago
Tiny ImageNet (test)	PSBD	AUROC0.996	16	3mo ago
Simple IHU Llama 8B	Functional Attribution	AUROC0.992	15	1mo ago
Simple IHU Gemma 2B	Functional Attribution	AUROC1	15	1mo ago
GTSRB BppAttack (test)	ARCGEN	AUC0.9967	15	3mo ago
GTSRB WaNet attack (test)	MNTD	AUC1	15	3mo ago
GTSRB SIG attack (test)	MNTD	AUC97.63	15	3mo ago
MS-COCO v1.4 (val)	T2IShieldFTT	RickBKD Detection Rate99.9	14	1mo ago
CIFAR-10 imbalanced µ=0.9, ρ=100 (test)	RPP	Badnets TPR100	13	3mo ago
CIFAR-10 imbalanced µ=0.9, ρ=2 (test)	RPP	Badnets TPR96.7	13	3mo ago
GTSRB (test)	LoRAcle	Top-1 ASR1	12	3mo ago
MNIST (test)	LoRAcle	Top-1 ASR100	12	3mo ago
Tiny ImageNet	SCP	TPR99.9	12	3mo ago

Showing 25 of 91 rows