# Adversarial Robustness on CIFAR-100 (Specific Attack Metrics)

Source: [Adversarial Robustness on CIFAR-100 (Specific Attack Metrics)](<https://www.sota2.com/research/sota/adversarial-robustness-on-cifar-100-specific-attack-metrics>)

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

Data updated: 2026-03-31T11:46:06.843Z

Evaluation of model adversarial robustness on the CIFAR-100 dataset, reporting performance metrics derived specifically from APGD, FAB, and Square adversarial attacks.

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 | Adversarial Robustness |
| Dataset (including any stated split) | CIFAR-100 |
| Metric | APGD Avg Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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

- [APGD Avg Score](<https://www.sota2.com/research/sota/adversarial-robustness-on-cifar-100-specific-attack-metrics.md?metric=APGD%20Avg%20Score>): Higher is better; unit: PERCENT.

- [APGD Worst Score](<https://www.sota2.com/research/sota/adversarial-robustness-on-cifar-100-specific-attack-metrics.md?metric=APGD%20Worst%20Score>): Higher is better; unit: PERCENT.

- [APGDT Avg Score](<https://www.sota2.com/research/sota/adversarial-robustness-on-cifar-100-specific-attack-metrics.md?metric=APGDT%20Avg%20Score>): Higher is better; unit: PERCENT.

- [APGDT Worst Score](<https://www.sota2.com/research/sota/adversarial-robustness-on-cifar-100-specific-attack-metrics.md?metric=APGDT%20Worst%20Score>): Higher is better; unit: PERCENT.

- [FAB Avg Score](<https://www.sota2.com/research/sota/adversarial-robustness-on-cifar-100-specific-attack-metrics.md?metric=FAB%20Avg%20Score>): Higher is better; unit: PERCENT.

- [FAB Worst Score](<https://www.sota2.com/research/sota/adversarial-robustness-on-cifar-100-specific-attack-metrics.md?metric=FAB%20Worst%20Score>): Higher is better; unit: PERCENT.

- [Square Attack Avg Score](<https://www.sota2.com/research/sota/adversarial-robustness-on-cifar-100-specific-attack-metrics.md?metric=Square%20Attack%20Avg%20Score>): Higher is better; unit: PERCENT.

- [Square Attack Worst Score](<https://www.sota2.com/research/sota/adversarial-robustness-on-cifar-100-specific-attack-metrics.md?metric=Square%20Attack%20Worst%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | APGD Avg Score | Result date | Report paper |
| --- | --- | --- | --- |
| GDMP | 65.06 | Not recorded | 68fa78a31e209131cea79829 |
| ADBM | 61.01 | Not recorded | 68fa78a31e209131cea79829 |
| DiffPure | 60.84 | Not recorded | 68fa78a31e209131cea79829 |
| FlowPureGauss | 60.22 | Not recorded | 68fa78a31e209131cea79829 |
| LM | 32.61 | Not recorded | 68fa78a31e209131cea79829 |
| FlowPureCW | 18.57 | Not recorded | 68fa78a31e209131cea79829 |
| FlowPurePGD | 4.47 | Not recorded | 68fa78a31e209131cea79829 |

### GDMP

APGD Avg Score: 65.06. Result ID: 69cbb3e247f34ce924b509d8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78a31e209131cea79829.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ce61e209131cea369e5.md>)

Recorded evaluation configuration: {"Attack Norm":"L2","Setting":"preprocessor-blind"}. This record does not assert independent reproduction.

### ADBM

APGD Avg Score: 61.01. Result ID: 69cbb3e247f34ce924b509da.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78a31e209131cea79829.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77831e209131cea73b76.md>)

Recorded evaluation configuration: {"Attack Norm":"L2","Setting":"preprocessor-blind"}. This record does not assert independent reproduction.

### DiffPure

APGD Avg Score: 60.84. Result ID: 69cbb3e247f34ce924b509d7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78a31e209131cea79829.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c841e209131cea3472d.md>)

Recorded evaluation configuration: {"Attack Norm":"L2","Setting":"preprocessor-blind"}. This record does not assert independent reproduction.

### FlowPureGauss

APGD Avg Score: 60.22. Result ID: 69cbb3e247f34ce924b509dd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78a31e209131cea79829.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78a31e209131cea79829.md>)

Recorded evaluation configuration: {"Attack Norm":"L2","Setting":"preprocessor-blind"}. This record does not assert independent reproduction.

### LM

APGD Avg Score: 32.61. Result ID: 69cbb3e247f34ce924b509d9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78a31e209131cea79829.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71551e209131cea5091d.md>)

Recorded evaluation configuration: {"Attack Norm":"L2","Setting":"preprocessor-blind"}. This record does not assert independent reproduction.

### FlowPureCW

APGD Avg Score: 18.57. Result ID: 69cbb3e247f34ce924b509dc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78a31e209131cea79829.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78a31e209131cea79829.md>)

Recorded evaluation configuration: {"Attack Norm":"L2","Setting":"preprocessor-blind"}. This record does not assert independent reproduction.

### FlowPurePGD

APGD Avg Score: 4.47. Result ID: 69cbb3e247f34ce924b509db.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78a31e209131cea79829.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78a31e209131cea79829.md>)

Recorded evaluation configuration: {"Attack Norm":"L2","Setting":"preprocessor-blind"}. This record does not assert independent reproduction.

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