| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| CIFAR-10 (test) | Diffusion Calibration | Accuracy (Standard)94.5 | 26 | 32m ago | |
| CIFAR10 | CEAR-(RW) | Certified Accuracy (R=0.00)79 | 18 | 1d ago | |
| MNIST | CEAR-(GM) | Certified Accuracy (R=0.00)99 | 18 | 1d ago | |
| ImageNet | AS | Inference Time (s)6.2 | 18 | 8d ago | |
| CIFAR-10 l2 (test) | Horvath | Average Sample Count768,560 | 18 | 3mo ago | |
| Mean Certified Accuracy 108/255 | Sandwich | Wins A95 | 15 | 3mo ago | |
| Mean Certified Accuracy 36/255 | Sandwich | Wins A97 | 15 | 3mo ago | |
| CIFAR-10 | LBS | Certified Accuracy (L2, r=0.0)85.4 | 14 | 1mo ago | |
| CIFAR-100 (test) | L2 (0.25) | Clean Accuracy54.3 | 11 | 32m ago | |
| ImageNet | Uniform, LBS | Certified Accuracy (L1 R=0.5)71.6 | 8 | 1mo ago | |
| CIFAR-10 | Uniform, LBS | Certified Accuracy (L1, R=0.5)86.1 | 8 | 1mo ago | |
| CIFAR-10 | Consistency Purification | Certified Acc (eps=0.0)90.4 | 6 | 3mo ago | |
| Tiny ImageNet (test) | Baseline 2 | CertAcc@028.02 | 5 | 32m ago | |
| FashionMNIST (test) | Baseline 3 | Certified Accuracy @087.64 | 5 | 32m ago | |
| CIFAR-10 | CURE | Certified Radius (L-inf)61.2 | 4 | 1mo ago | |
| Interaction-only safety filtering (curated) | Image-only RS | Average Certified Image Radius (r_bar)3.99 | 3 | 21d ago | |
| ImageNet-64 | Consistency Purification + Consistency Fine-tuning | Certified Acc (eps=0.0)69 | 3 | 3mo ago | |
| Aggregate Overall experimental suite (test) | - | - | 0 | 3mo ago |
