# Image Classification on CIFAR100 (Adversarial Robustness Metrics)

Source: [Image Classification on CIFAR100 (Adversarial Robustness Metrics)](<https://www.sota2.com/research/sota/image-classification-on-cifar100-adversarial-robustness-metrics>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-06-04T12:49:59.203Z

Evaluation of Image Classification performance on the CIFAR-100 dataset, focusing on model robustness against adversarial attacks, measured using metrics like CW, PGD, and CLEVER scores.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Image Classification |
| Dataset (including any stated split) | CIFAR100 |
| Metric | L(x) Value |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 2 of 2 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [L(x) Value](<https://www.sota2.com/research/sota/image-classification-on-cifar100-adversarial-robustness-metrics.md?metric=L(x)%20Value>): Higher is better; unit: SCALAR.

- [CLEVER Score](<https://www.sota2.com/research/sota/image-classification-on-cifar100-adversarial-robustness-metrics.md?metric=CLEVER%20Score>): Higher is better; unit: SCALAR.

- [CW Distortion](<https://www.sota2.com/research/sota/image-classification-on-cifar100-adversarial-robustness-metrics.md?metric=CW%20Distortion>): Higher is better; unit: SCALAR.

- [Robust Accuracy (PGD)](<https://www.sota2.com/research/sota/image-classification-on-cifar100-adversarial-robustness-metrics.md?metric=Robust%20Accuracy%20(PGD)>): Higher is better; unit: PERCENT.

- [L2 Norm Distance](<https://www.sota2.com/research/sota/image-classification-on-cifar100-adversarial-robustness-metrics.md?metric=L2%20Norm%20Distance>): Lower is better; unit: SCALAR.

- [Rspec Score](<https://www.sota2.com/research/sota/image-classification-on-cifar100-adversarial-robustness-metrics.md?metric=Rspec%20Score>): Higher is better; unit: SCALAR.

## Recorded results

| Method | L(x) Value | Result date | Report paper |
| --- | --- | --- | --- |
| ResNet18 | 0.29 | Not recorded | 6a210703c10484bac23f07e7 |
| ViT\_B\_16 | 0.23 | Not recorded | 6a210703c10484bac23f07e7 |

### ResNet18

L(x) Value: 0.29. Result ID: 6a2173f0fbc373592c1eb19e.

[Reporting paper](<https://www.sota2.com/research/paper/6a210703c10484bac23f07e7.md>)

Recorded evaluation configuration: {"Model":"ResNet18"}. This record does not assert independent reproduction.

### ViT\_B\_16

L(x) Value: 0.23. Result ID: 6a2173f0fbc373592c1eb19f.

[Reporting paper](<https://www.sota2.com/research/paper/6a210703c10484bac23f07e7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa703d1e209131cea49db1.md>)

[Implementation](<https://github.com/tensorflow/tensor2tensor>)

Recorded evaluation configuration: {"Model":"ViT\_B\_16"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
