# One-run privacy auditing on CIFAR10 Source: [One-run privacy auditing on CIFAR10]() Collection index: [llms.txt]() Data updated: 2026-05-27T10:21:11.958Z Evaluation of privacy leakage in models trained on CIFAR-10 using one-run membership inference attack metrics, specifically True Positive Rate at a fixed False Positive Rate. 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 | One-run privacy auditing | | Dataset (including any stated split) | CIFAR10 | | Metric | Mean TPR @ 0.05 FPR | | 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 - [Mean TPR @ 0.05 FPR](): Higher is better; unit: PERCENT. - [Std TPR @ 0.05 FPR](): Lower is better; unit: SCALAR. - [Median TPR @ 0.05 FPR](): Higher is better; unit: PERCENT. ## Recorded results | Method | Mean TPR @ 0.05 FPR | Result date | Report paper | | --- | --- | --- | --- | | IBIS reg. | 100 | Not recorded | 6a167ad8dbd7150506f802ec | | Random + Flip | 86 | Not recorded | 6a167ad8dbd7150506f802ec | | IBIS reg. | 84 | Not recorded | 6a167ad8dbd7150506f802ec | | Boglioni et al. | 82 | Not recorded | 6a167ad8dbd7150506f802ec | | IBIS no reg. | 34 | Not recorded | 6a167ad8dbd7150506f802ec | | IBIS no reg. | 32 | Not recorded | 6a167ad8dbd7150506f802ec | | Random | 5 | Not recorded | 6a167ad8dbd7150506f802ec | ### IBIS reg. Mean TPR @ 0.05 FPR: 100. Result ID: 6a16c4aafbc373592c1086f8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Crafting architecture":"WRN16-4"}. This record does not assert independent reproduction. ### Random + Flip Mean TPR @ 0.05 FPR: 86. Result ID: 6a16c4aafbc373592c1086fb. [Reporting paper]() Recorded evaluation configuration: {"Crafting architecture":"ResNet9"}. This record does not assert independent reproduction. ### IBIS reg. Mean TPR @ 0.05 FPR: 84. Result ID: 6a16c4aafbc373592c1086f7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Crafting architecture":"ResNet9"}. This record does not assert independent reproduction. ### Boglioni et al. Mean TPR @ 0.05 FPR: 82. Result ID: 6a16c4aafbc373592c1086f9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Crafting architecture":"ResNet9"}. This record does not assert independent reproduction. ### IBIS no reg. Mean TPR @ 0.05 FPR: 34. Result ID: 6a16c4aafbc373592c1086f5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Crafting architecture":"ResNet9"}. This record does not assert independent reproduction. ### IBIS no reg. Mean TPR @ 0.05 FPR: 32. Result ID: 6a16c4aafbc373592c1086f6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Crafting architecture":"WRN16-4"}. This record does not assert independent reproduction. ### Random Mean TPR @ 0.05 FPR: 5. Result ID: 6a16c4aafbc373592c1086fa. [Reporting paper]() Recorded evaluation configuration: {"Crafting architecture":"ResNet9"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.