Out-of-Distribution Detection on CIFAR-100 OpenOOD (test)
96.39AUROC (AVG)InterNeg
Evaluation results
| Method | Links | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| InterNegArchitecture=CLIP ViT-B/16, Training Requirement=Zero-shot2026.03 | 96.39 | 62.54 | — | — | — | — | — | — | 85.45 | — | — | — | — | — | 20.02 | — | |
| AdaNegArchitecture=CLIP ViT-B/16, Training Requirement=Zero-shot2026.03 | 95.25 | 59.07 | — | — | — | — | — | — | 84.6 | — | — | — | — | — | 29.35 | — | |
| NegLabelArchitecture=CLIP ViT-B/16, Training Requirement=Zero-shot2026.03 | 89.68 | 71.44 | — | — | — | — | — | — | 70.58 | — | — | — | — | — | 40.92 | — | |
| ViMBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 88.58 | — | — | — | — | — | — | 43.89 | — | — | — | — | — | — | — | — | |
| ViMBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 88.43 | — | — | — | — | — | — | 38.73 | — | — | — | — | — | — | — | — | |
| ReActBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 88.11 | — | — | — | — | — | — | 41.68 | — | — | — | — | — | — | — | — | |
| MDSBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 88.1 | — | — | — | — | — | — | 44.64 | — | — | — | — | — | — | — | — | |
| KNNBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 87.77 | — | — | — | — | — | — | 41.53 | — | — | — | — | — | — | — | — | |
| KNNBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 87.67 | — | — | — | — | — | — | 43.61 | — | — | — | — | — | — | — | — | |
| GradNormBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 87.63 | — | — | — | — | — | — | 45.17 | — | — | — | — | — | — | — | — | |
| EPDBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 87.48 | — | — | — | — | — | — | 42.76 | — | — | — | — | — | — | — | — | |
| SHEBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 87.35 | — | — | — | — | — | — | 46.61 | — | — | — | — | — | — | — | — | |
| EPDBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 87.25 | — | — | — | — | — | — | 44.76 | — | — | — | — | — | — | — | — | |
| EBOBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 86.83 | — | — | — | — | — | — | 49.64 | — | — | — | — | — | — | — | — | |
| RMDSBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 86.82 | — | — | — | — | — | — | 41.81 | — | — | — | — | — | — | — | — | |
| GENBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 86.8 | — | — | — | — | — | — | 49.98 | — | — | — | — | — | — | — | — | |
| MLSBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 86.62 | — | — | — | — | — | — | 49.52 | — | — | — | — | — | — | — | — | |
| RMDSBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 86.5 | — | — | — | — | — | — | 47.02 | — | — | — | — | — | — | — | — | |
| SHEBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 86.03 | — | — | — | — | — | — | 49.91 | — | — | — | — | — | — | — | — | |
| ReActBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 86.02 | — | — | — | — | — | — | 51.21 | — | — | — | — | — | — | — | — | |
| MDSBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 85.6 | — | — | — | — | — | — | 44.81 | — | — | — | — | — | — | — | — | |
| GradNormBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 85.52 | — | — | — | — | — | — | 48.78 | — | — | — | — | — | — | — | — | |
| TempScaleBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 85.18 | — | — | — | — | — | — | 50.92 | — | — | — | — | — | — | — | — | |
| GENBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 84.64 | — | — | — | — | — | — | 62.46 | — | — | — | — | — | — | — | — | |
| EBOBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 84.38 | — | — | — | — | — | — | 63.84 | — | — | — | — | — | — | — | — | |
| MLSBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 84.17 | — | — | — | — | — | — | 63.7 | — | — | — | — | — | — | — | — | |
| MSPBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 83.84 | — | — | — | — | — | — | 52.63 | — | — | — | — | — | — | — | — | |
| OpenMaxBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 83.58 | — | — | — | — | — | — | 54.92 | — | — | — | — | — | — | — | — | |
| DICEBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 83.57 | — | — | — | — | — | — | 60.53 | — | — | — | — | — | — | — | — | |
| TempScaleBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 83.02 | — | — | — | — | — | — | 62.41 | — | — | — | — | — | — | — | — | |
| VNShyperparameter-selection=automatic validation selection2026.02 | 82.29 | 60.34 | 48.82 | 50 | 49.32 | 52.75 | 54.04 | 52.55 | 78.7 | 84.09 | 80.99 | 84.47 | 83.06 | 82.15 | — | — | |
| DICEBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 81.68 | — | — | — | — | — | — | 69.63 | — | — | — | — | — | — | — | — | |
| MSPBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 81.68 | — | — | — | — | — | — | 61.8 | — | — | — | — | — | — | — | — | |
| KNNhyperparameter-selection=automatic validation selection2026.02 | 81.66 | 72.81 | 49.66 | 48.57 | 51.75 | 53.56 | 60.7 | 56.18 | 77.02 | 83.34 | 82.36 | 84.15 | 83.66 | 79.43 | — | — | |
| RMDS++hyperparameter-selection=hyperparameter-free2026.02 | 81.62 | 61.43 | 48.9 | 79.37 | 52.49 | 52.4 | 54.78 | 53.82 | 78.08 | 82.72 | 79.37 | 83.95 | 82.56 | 83.01 | — | — | |
| ADASCALEhyperparameter-selection=automatic validation selection2026.02 | 81.57 | 59.27 | 51.94 | 51.25 | 47.41 | 58.39 | 56.41 | 54.11 | 79.22 | 82.82 | 80.39 | 85.27 | 81.01 | 80.68 | — | — | |
| NCIhyperparameter-selection=hyperparameter-free2026.02 | 81.56 | 63.59 | 48.59 | 51.14 | 48.36 | 47.76 | 53.93 | 52.23 | 78.31 | 83.55 | 79.89 | 83.01 | 83.75 | 80.86 | — | — | |
| SCALEArchitecture=CLIP ViT-B/16, Training Requirement=Requires training on ID or extra data2026.03 | 81.42 | — | — | — | — | — | — | — | 80.99 | — | — | — | — | — | — | — | |
| OE + MSPArchitecture=CLIP ViT-B/16, Training Requirement=Requires training on ID or extra data2026.03 | 81.41 | — | — | — | — | — | — | — | 88.3 | — | — | — | — | — | — | — | |
| VOS + EBOArchitecture=CLIP ViT-B/16, Training Requirement=Requires training on ID or extra data2026.03 | 81.32 | — | — | — | — | — | — | — | 80.93 | — | — | — | — | — | — | — | |
| SCALEhyperparameter-selection=automatic validation selection2026.02 | 81.28 | 59.11 | 52.24 | 51.64 | 49.27 | 58.44 | 56.98 | 54.61 | 79.26 | 82.71 | 80.27 | 84.45 | 80.5 | 80.47 | — | — | |
| OpenMaxBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 80.95 | — | — | — | — | — | — | 62.94 | — | — | — | — | — | — | — | — | |
| FDBDhyperparameter-selection=hyperparameter-free2026.02 | 80.48 | 63.89 | 47.89 | 51.35 | 53.8 | 53.65 | 57.16 | 54.62 | 78.35 | 83.97 | 79.05 | 80.48 | 81.18 | 79.85 | — | — | |
| GENhyperparameter-selection=automatic validation selection2026.02 | 80.23 | 58.87 | 49.97 | 53.93 | 55.45 | 61.22 | 56.25 | 55.95 | 79.38 | 83.25 | 78.29 | 81.41 | 78.74 | 80.28 | — | — | |
| IODINhyperparameter-selection=automatic validation selection2026.02 | 80.16 | 59.08 | 51.57 | 52.91 | 54.07 | 62.07 | 57.46 | 56.19 | 79.24 | 82.96 | 78.89 | 81.56 | 78.48 | 79.83 | — | — | |
| ASHhyperparameter-selection=automatic validation selection2026.02 | 79.78 | 68.07 | 63.37 | 66.6 | 45.97 | 61.29 | 62.94 | 61.37 | 76.47 | 79.92 | 77.23 | 85.6 | 80.72 | 78.76 | — | — | |
| GENArchitecture=CLIP ViT-B/16, Training Requirement=Requires training on ID or extra data2026.03 | 79.68 | — | — | — | — | — | — | — | 81.31 | — | — | — | — | — | — | — | |
| MSPhyperparameter-selection=hyperparameter-free2026.02 | 78.6 | 58.9 | 50.7 | 57.24 | 59.07 | 61.88 | 56.62 | 57.4 | 78.47 | 82.07 | 76.09 | 78.42 | 77.32 | 79.23 | — | — | |
| MCMArchitecture=CLIP ViT-B/16, Training Requirement=Zero-shot2026.03 | 76 | 75.2 | — | — | — | — | — | — | 71 | — | — | — | — | — | 59.32 | — | |
| ASHBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 56.96 | — | — | — | — | — | — | 87.04 | — | — | — | — | — | — | — | — | |
| ASHBackbone=ViT-B/16, Pre-trained Model=google/vit-base-patch16-224-in21k, ID Accuracy=84.78%, Evaluation Protocol=fine-tuned2026.04 | 51.23 | — | — | — | — | — | — | 90.97 | — | — | — | — | — | — | — | — | |
| Hierarchical Gaussian DPMM (Coupled Diag.)Backbone=ResNet18, Embedding Dimension=5122025.02 | — | — | — | — | — | — | — | — | 78.05 | — | — | — | — | — | — | 79.29 | |
| Hierarchical Gaussian DPMM (Diagonal)Backbone=ResNet18, Embedding Dimension=5122025.02 | — | — | — | — | — | — | — | — | 79.22 | — | — | — | — | — | — | 82.88 | |
| Hierarchical Gaussian DPMM (Full)Backbone=ResNet18, Embedding Dimension=5122025.02 | — | — | — | — | — | — | — | — | 79.22 | — | — | — | — | — | — | 81.2 | |
| Hierarchical Gaussian DPMM (Tied)Backbone=ResNet18, Embedding Dimension=5122025.02 | — | — | — | — | — | — | — | — | 80.17 | — | — | — | — | — | — | 82.98 | |
| MDSBackbone=ResNet18, Embedding Dimension=5122025.02 | — | — | — | — | — | — | — | — | 58.86 | — | — | — | — | — | — | 69.29 | |
| MSPBackbone=ResNet18, Embedding Dimension=5122025.02 | — | — | — | — | — | — | — | — | 80.33 | — | — | — | — | — | — | 78.88 | |
| RMDSBackbone=ResNet18, Embedding Dimension=5122025.02 | — | — | — | — | — | — | — | — | 80.17 | — | — | — | — | — | — | 82.97 | |
| Temp. ScaleBackbone=ResNet18, Embedding Dimension=5122025.02 | — | — | — | — | — | — | — | — | 80.57 | — | — | — | — | — | — | 79.25 |