# Membership Inference Attack on ImageNet VAR Generated 1.0 (test)

Source: [Membership Inference Attack on ImageNet VAR Generated 1.0 (test)](<https://www.sota2.com/research/sota/membership-inference-attack-on-imagenet-var-generated-1-0-test>)

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

Data updated: 2026-06-24T07:24:48.008Z

Evaluation of a model's susceptibility to Membership Inference Attacks conducted on the test split of the ImageNet VAR Generated 1.0 dataset, focusing on True Positive Rates across various comparison groups at 1% 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 | Membership Inference Attack |
| Dataset (including any stated split) | ImageNet VAR Generated 1.0 (test) |
| Metric | TPR@1%FPR (NM/GM) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [TPR@1%FPR (NM/GM)](<https://www.sota2.com/research/sota/membership-inference-attack-on-imagenet-var-generated-1-0-test.md?metric=TPR%401%25FPR%20(NM%2FGM)>): Higher is better; unit: PERCENT.

- [TPR@1%FPR (NM/GN)](<https://www.sota2.com/research/sota/membership-inference-attack-on-imagenet-var-generated-1-0-test.md?metric=TPR%401%25FPR%20(NM%2FGN)>): Higher is better; unit: PERCENT.

- [TPR@1%FPR (NM/G')](<https://www.sota2.com/research/sota/membership-inference-attack-on-imagenet-var-generated-1-0-test.md?metric=TPR%401%25FPR%20(NM%2FG%27)>): Higher is better; unit: PERCENT.

- [TPR@1%FPR (NN/GM)](<https://www.sota2.com/research/sota/membership-inference-attack-on-imagenet-var-generated-1-0-test.md?metric=TPR%401%25FPR%20(NN%2FGM)>): Higher is better; unit: PERCENT.

- [TPR@1%FPR (NN/GN)](<https://www.sota2.com/research/sota/membership-inference-attack-on-imagenet-var-generated-1-0-test.md?metric=TPR%401%25FPR%20(NN%2FGN)>): Higher is better; unit: PERCENT.

- [TPR@1%FPR (NN/G')](<https://www.sota2.com/research/sota/membership-inference-attack-on-imagenet-var-generated-1-0-test.md?metric=TPR%401%25FPR%20(NN%2FG%27)>): Higher is better; unit: PERCENT.

- [TPR@1%FPR (GM/GN)](<https://www.sota2.com/research/sota/membership-inference-attack-on-imagenet-var-generated-1-0-test.md?metric=TPR%401%25FPR%20(GM%2FGN)>): Higher is better; unit: PERCENT.

- [TPR@1%FPR (GM/G')](<https://www.sota2.com/research/sota/membership-inference-attack-on-imagenet-var-generated-1-0-test.md?metric=TPR%401%25FPR%20(GM%2FG%27)>): Higher is better; unit: PERCENT.

- [TPR@1%FPR (GN/G')](<https://www.sota2.com/research/sota/membership-inference-attack-on-imagenet-var-generated-1-0-test.md?metric=TPR%401%25FPR%20(GN%2FG%27)>): Higher is better; unit: PERCENT.

- [TPR@1%FPR (NM/NN)](<https://www.sota2.com/research/sota/membership-inference-attack-on-imagenet-var-generated-1-0-test.md?metric=TPR%401%25FPR%20(NM%2FNN)>): Higher is better; unit: PERCENT.

- [TPR@1%FPR (Overall)](<https://www.sota2.com/research/sota/membership-inference-attack-on-imagenet-var-generated-1-0-test.md?metric=TPR%401%25FPR%20(Overall)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | TPR@1%FPR (NM/GM) | Result date | Report paper |
| --- | --- | --- | --- |
| RMIA | 100 | Not recorded | 6a3b64b2206542251d7c97d9 |
| Data Circuit Breaker | 99.3 | Not recorded | 6a3b64b2206542251d7c97d9 |
| LiRA | 98.7 | Not recorded | 6a3b64b2206542251d7c97d9 |

### RMIA

TPR@1%FPR (NM/GM): 100. Result ID: 6a3b8623fbc373592c4384a4.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97d9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73991e209131cea5d82b.md>)

Recorded evaluation configuration: {"Model":"VAR","Number of shadow models":"5"}. This record does not assert independent reproduction.

### Data Circuit Breaker

TPR@1%FPR (NM/GM): 99.3. Result ID: 6a3b8623fbc373592c4384a5.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97d9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97d9.md>)

Recorded evaluation configuration: {"Model":"VAR","Number of shadow models":"5"}. This record does not assert independent reproduction.

### LiRA

TPR@1%FPR (NM/GM): 98.7. Result ID: 6a3b8623fbc373592c4384a3.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97d9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c461e209131cea33297.md>)

Recorded evaluation configuration: {"Model":"VAR","Number of shadow models":"5"}. This record does not assert independent reproduction.

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