# Adversarial Robustness on CIFAR-10 (L2, eps=8/255) Source: [Adversarial Robustness on CIFAR-10 (L2, eps=8/255)]() Collection index: [llms.txt]() Data updated: 2026-02-26T06:02:21.972Z Evaluation of model adversarial robustness on the CIFAR-10 dataset specifically under L2 perturbation constraints with epsilon set to 8/255. Metrics include Robust Accuracy (AA/A3) and associated runtimes. 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-10 L2, epsilon=8/255 | | Metric | Robust Acc (AA) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 11 of 11 | | 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 - [Robust Acc (AA)](): Higher is better; unit: PERCENT. - [Robust Acc (A3)](): Higher is better; unit: PERCENT. - [Robust Acc Delta](): Lower is better; unit: SCALAR. - [Runtime (AA)](): Lower is better; unit: TIME. - [Runtime (A3)](): Lower is better; unit: TIME. - [Speed-up Ratio](): Higher is better; unit: SCALAR. - [Search Steps (A3)](): Lower is better; unit: SCALAR. ## Recorded results | Method | Robust Acc (AA) | Result date | Report paper | | --- | --- | --- | --- | | Stutz et al. (2020) | 77.64 | Not recorded | 68fa6b161e209131cea2c0c7 | | Stutz et al. (2020) | 77.64 | Not recorded | 68fa6b161e209131cea2c0c7 | | Wu et al. (2020) RTS + Das et al. (2017) + Lee et al. (2018) | 60.72 | Not recorded | 68fa6b161e209131cea2c0c7 | | Wang et al. (2019) | 50.84 | Not recorded | 68fa6b161e209131cea2c0c7 | | Wu et al. (2020) RTS + Random Rotation | 49.53 | Not recorded | 68fa6b161e209131cea2c0c7 | | Zhang & Wang (2019) | 36.74 | Not recorded | 68fa6b161e209131cea2c0c7 | | Grathwohl et al. (2020) + Random Rotation | 22.29 | Not recorded | 68fa6b161e209131cea2c0c7 | | Grathwohl et al. (2020) + Das et al. (2017) + Lee et al. (2018) | 15.27 | Not recorded | 68fa6b161e209131cea2c0c7 | | Hu et al. (2019) | 6.25 | Not recorded | 68fa6b161e209131cea2c0c7 | | Xiao et al. (2020) ADV | 5.4 | Not recorded | 68fa6b161e209131cea2c0c7 | | Grathwohl et al. (2020) | 5.15 | Not recorded | 68fa6b161e209131cea2c0c7 | ### Stutz et al. (2020) Robust Acc (AA): 77.64. Result ID: 698e5513ffc3901de09cb208. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"author\_provided":true,"epsilon":0.03}. This record does not assert independent reproduction. ### Stutz et al. (2020) Robust Acc (AA): 77.64. Result ID: 698e5513ffc3901de09cb209. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"attack\_sequence\_length\_A3":8}. This record does not assert independent reproduction. ### Wu et al. (2020) RTS + Das et al. (2017) + Lee et al. (2018) Robust Acc (AA): 60.72. Result ID: 698e5513ffc3901de09cb20e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"non-differentiable":true}. This record does not assert independent reproduction. ### Wang et al. (2019) Robust Acc (AA): 50.84. Result ID: 698e5513ffc3901de09cb20d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Wu et al. (2020) RTS + Random Rotation Robust Acc (AA): 49.53. Result ID: 698e5513ffc3901de09cb210. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Zhang & Wang (2019) Robust Acc (AA): 36.74. Result ID: 698e5513ffc3901de09cb20a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"author\_provided":true}. This record does not assert independent reproduction. ### Grathwohl et al. (2020) + Random Rotation Robust Acc (AA): 22.29. Result ID: 698e5513ffc3901de09cb211. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Grathwohl et al. (2020) + Das et al. (2017) + Lee et al. (2018) Robust Acc (AA): 15.27. Result ID: 698e5513ffc3901de09cb20f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Hu et al. (2019) Robust Acc (AA): 6.25. Result ID: 698e5513ffc3901de09cb212. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Xiao et al. (2020) ADV Robust Acc (AA): 5.4. Result ID: 698e5513ffc3901de09cb20c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"training":"adversarial"}. This record does not assert independent reproduction. ### Grathwohl et al. (2020) Robust Acc (AA): 5.15. Result ID: 698e5513ffc3901de09cb20b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.