| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Image Classification | ImageNette (test) | Accuracy95.3 | 164 | |
| Image Classification | ImageNette (val) | Accuracy0.98 | 75 | |
| Image Classification | ImageNette ImageNet-1K (test) | Accuracy67.6 | 72 | |
| Coarse-grained Unlearning | Imagenette | Atar100 | 70 | |
| Class Erasure | Imagenette | UA100 | 66 | |
| Poison detection | ImageNette | TPR100 | 60 | |
| Image Classification | Imagenette | Top-1 Accuracy98.88 | 52 | |
| Machine Unlearning | ImageNette gas pump Class 7 (test) | Forget Accuracy100 | 48 | |
| Class-wise forgetting | Imagenette (val) | FID0.78 | 44 | |
| Backdoor Certification | ImageNette (test) | AER74 | 36 | |
| Image Classification | Imagenette | Accuracy98 | 36 | |
| Image Reconstruction | ImageNette | SSIM0.997 | 27 | |
| Image Classification | ImageNette Long-tailed ImageNet | Accuracy34.7 | 22 | |
| Class-wise Forgetting | Imagenette Stable Diffusion v1.4 (val) | FID0.67 | 22 | |
| Concept Awakening | ImageNette & Harmful Concepts Quantitative Assessment (200 images per concept) | Accuracy95 | 21 | |
| Image Classification | ImageNette standard evaluation | Top-1 Accuracy92.6 | 19 | |
| Backdoor Defense | ImageNette | AER0.67 | 18 | |
| Image Generation | ImageNette | FID0.35 | 18 | |
| Machine Unlearning | Imagenette | Accuracy (garbage truck)76.7 | 18 | |
| Image Classification | Imagenette 256 x 256 10 classes (test) | Classification Accuracy90.67 | 17 | |
| Object Unlearning | Imagenette Object Unlearning Subset | ERASE FID0 | 16 | |
| Image Classification | ImageNette 128x128 (test) | Top-1 Acc87.4 | 16 | |
| Attribution Adversarial Sensitivity | Imagenette 224x224 (test) | VGrad0.61 | 15 | |
| Image Classification | ImageNette (val test) | Top-1 Accuracy84.2 | 15 | |
| Attribution Robustness | Imagenette (val) | Accuracy84 | 14 |