# Targeted Failure Generation on ImageNet

Source: [Targeted Failure Generation on ImageNet](<https://www.sota2.com/research/sota/targeted-failure-generation-on-imagenet>)

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

Data updated: 2026-02-26T05:52:53.352Z

Evaluation of model robustness or failure modes on the ImageNet dataset, focusing on metrics like Misclassification Rate and Escape Ratio.

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 | Targeted Failure Generation |
| Dataset (including any stated split) | ImageNet |
| Metric | Misclassification Rate |
| Unit | PERCENT |
| Direction | Lower 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

- [Misclassification Rate](<https://www.sota2.com/research/sota/targeted-failure-generation-on-imagenet.md?metric=Misclassification%20Rate>): Lower is better; unit: PERCENT.

- [Escape Ratio](<https://www.sota2.com/research/sota/targeted-failure-generation-on-imagenet.md?metric=Escape%20Ratio>): Higher is better; unit: PERCENT.

- [Confidence Reduction](<https://www.sota2.com/research/sota/targeted-failure-generation-on-imagenet.md?metric=Confidence%20Reduction>): Higher is better; unit: SCALAR.

- [MS-SSIM](<https://www.sota2.com/research/sota/targeted-failure-generation-on-imagenet.md?metric=MS-SSIM>): Higher is better; unit: SCALAR.

- [LPIPS](<https://www.sota2.com/research/sota/targeted-failure-generation-on-imagenet.md?metric=LPIPS>): Lower is better; unit: SCALAR.

- [Diversity Score (T)](<https://www.sota2.com/research/sota/targeted-failure-generation-on-imagenet.md?metric=Diversity%20Score%20(T)>): Higher is better; unit: SCALAR.

- [Trace Difference](<https://www.sota2.com/research/sota/targeted-failure-generation-on-imagenet.md?metric=Trace%20Difference>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Misclassification Rate | Result date | Report paper |
| --- | --- | --- | --- |
| Mimicry | 0.82 | Not recorded | 6971af33e25ad2fa9d1edb51 |
| HyNeA | 1 | Not recorded | 6971af33e25ad2fa9d1edb51 |
| GIFTbench | 1 | Not recorded | 6971af33e25ad2fa9d1edb51 |

### Mimicry

Misclassification Rate: 0.82. Result ID: 699925a15698665ec48d236c.

[Reporting paper](<https://www.sota2.com/research/paper/6971af33e25ad2fa9d1edb51.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78741e209131cea7866e.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### HyNeA

Misclassification Rate: 1. Result ID: 699925a15698665ec48d236b.

[Reporting paper](<https://www.sota2.com/research/paper/6971af33e25ad2fa9d1edb51.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6971af33e25ad2fa9d1edb51.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### GIFTbench

Misclassification Rate: 1. Result ID: 699925a15698665ec48d236d.

[Reporting paper](<https://www.sota2.com/research/paper/6971af33e25ad2fa9d1edb51.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa765f1e209131cea6cb21.md>)

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.