# Perturbation Test on ImageNet (val)

Source: [Perturbation Test on ImageNet (val)](<https://www.sota2.com/research/sota/perturbation-test-on-imagenet-val>)

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

Data updated: 2026-02-26T05:53:36.166Z

Evaluation of model performance on the standard ImageNet validation set under specific perturbation testing protocols, measured using Neg. and Pos. metrics.

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 | Perturbation Test |
| Dataset (including any stated split) | ImageNet (val) |
| Metric | Neg Score |
| Unit | PERCENT |
| Direction | Lower is better |
| Results loaded | 18 of 18 |
| 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

- [Neg Score](<https://www.sota2.com/research/sota/perturbation-test-on-imagenet-val.md?metric=Neg%20Score>): Lower is better; unit: PERCENT.

- [Pos Score](<https://www.sota2.com/research/sota/perturbation-test-on-imagenet-val.md?metric=Pos%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Neg Score | Result date | Report paper |
| --- | --- | --- | --- |
| ATTCAT | 32.34 | Not recorded | 68fa729e1e209131cea57935 |
| Cons. LRP | 33.24 | Not recorded | 68fa729e1e209131cea57935 |
| ATTCAT | 33.59 | Not recorded | 68fa729e1e209131cea57935 |
| Cons. LRP | 34.33 | Not recorded | 68fa729e1e209131cea57935 |
| LRP | 42.55 | Not recorded | 68fa729e1e209131cea57935 |
| LRP | 42.56 | Not recorded | 68fa729e1e209131cea57935 |
| Rollout | 43.23 | Not recorded | 68fa729e1e209131cea57935 |
| Rollout | 43.23 | Not recorded | 68fa729e1e209131cea57935 |
| Grad-CAM | 43.84 | Not recorded | 68fa729e1e209131cea57935 |
| Grad-CAM | 44.06 | Not recorded | 68fa729e1e209131cea57935 |
| Raw Att. | 48.7 | Not recorded | 68fa729e1e209131cea57935 |
| Raw Att. | 48.7 | Not recorded | 68fa729e1e209131cea57935 |
| Trans. Attr. | 57.48 | Not recorded | 68fa729e1e209131cea57935 |
| GAE | 57.53 | Not recorded | 68fa729e1e209131cea57935 |
| TokenTM | 58.12 | Not recorded | 68fa729e1e209131cea57935 |
| Trans. Attr. | 58.26 | Not recorded | 68fa729e1e209131cea57935 |
| GAE | 58.85 | Not recorded | 68fa729e1e209131cea57935 |
| TokenTM | 59.36 | Not recorded | 68fa729e1e209131cea57935 |

### ATTCAT

Neg Score: 32.34. Result ID: 6992643c3a6c3148490ea29f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

Recorded evaluation configuration: {"Target Label":"Predicted","Backbone":"DeiT"}. This record does not assert independent reproduction.

### Cons. LRP

Neg Score: 33.24. Result ID: 6992643c3a6c3148490ea297.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cd51e209131cea3613b.md>)

[Implementation](<https://github.com/AmeenAli/XAI_Transformers>)

Recorded evaluation configuration: {"Target Label":"Predicted","Backbone":"DeiT"}. This record does not assert independent reproduction.

### ATTCAT

Neg Score: 33.59. Result ID: 6992643c3a6c3148490ea2a0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

Recorded evaluation configuration: {"Target Label":"GT","Backbone":"DeiT"}. This record does not assert independent reproduction.

### Cons. LRP

Neg Score: 34.33. Result ID: 6992643c3a6c3148490ea298.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cd51e209131cea3613b.md>)

[Implementation](<https://github.com/AmeenAli/XAI_Transformers>)

Recorded evaluation configuration: {"Target Label":"GT","Backbone":"DeiT"}. This record does not assert independent reproduction.

### LRP

Neg Score: 42.55. Result ID: 6992643c3a6c3148490ea295.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa625f1e209131ce9f8e34.md>)

Recorded evaluation configuration: {"Target Label":"Predicted","Backbone":"DeiT"}. This record does not assert independent reproduction.

### LRP

Neg Score: 42.56. Result ID: 6992643c3a6c3148490ea296.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa625f1e209131ce9f8e34.md>)

Recorded evaluation configuration: {"Target Label":"GT","Backbone":"DeiT"}. This record does not assert independent reproduction.

### Rollout

Neg Score: 43.23. Result ID: 6992643c3a6c3148490ea29d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67ae1e209131cea17be6.md>)

Recorded evaluation configuration: {"Target Label":"Predicted","Backbone":"DeiT"}. This record does not assert independent reproduction.

### Rollout

Neg Score: 43.23. Result ID: 6992643c3a6c3148490ea29e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67ae1e209131cea17be6.md>)

Recorded evaluation configuration: {"Target Label":"GT","Backbone":"DeiT"}. This record does not assert independent reproduction.

### Grad-CAM

Neg Score: 43.84. Result ID: 6992643c3a6c3148490ea293.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66cb1e209131cea12568.md>)

[Implementation](<https://github.com/ramprs/grad-cam/>)

Recorded evaluation configuration: {"Target Label":"Predicted","Backbone":"DeiT"}. This record does not assert independent reproduction.

### Grad-CAM

Neg Score: 44.06. Result ID: 6992643c3a6c3148490ea294.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66cb1e209131cea12568.md>)

[Implementation](<https://github.com/ramprs/grad-cam/>)

Recorded evaluation configuration: {"Target Label":"GT","Backbone":"DeiT"}. This record does not assert independent reproduction.

### Raw Att.

Neg Score: 48.7. Result ID: 6992643c3a6c3148490ea29b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa665a1e209131cea0fbee.md>)

Recorded evaluation configuration: {"Target Label":"Predicted","Backbone":"DeiT"}. This record does not assert independent reproduction.

### Raw Att.

Neg Score: 48.7. Result ID: 6992643c3a6c3148490ea29c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa665a1e209131cea0fbee.md>)

Recorded evaluation configuration: {"Target Label":"GT","Backbone":"DeiT"}. This record does not assert independent reproduction.

### Trans. Attr.

Neg Score: 57.48. Result ID: 6992643c3a6c3148490ea299.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69b41e209131cea2394f.md>)

Recorded evaluation configuration: {"Target Label":"Predicted","Backbone":"DeiT"}. This record does not assert independent reproduction.

### GAE

Neg Score: 57.53. Result ID: 6992643c3a6c3148490ea2a1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69a11e209131cea233a8.md>)

Recorded evaluation configuration: {"Target Label":"Predicted","Backbone":"DeiT"}. This record does not assert independent reproduction.

### TokenTM

Neg Score: 58.12. Result ID: 6992643c3a6c3148490ea2a3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

Recorded evaluation configuration: {"Target Label":"Predicted","Backbone":"DeiT"}. This record does not assert independent reproduction.

### Trans. Attr.

Neg Score: 58.26. Result ID: 6992643c3a6c3148490ea29a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69b41e209131cea2394f.md>)

Recorded evaluation configuration: {"Target Label":"GT","Backbone":"DeiT"}. This record does not assert independent reproduction.

### GAE

Neg Score: 58.85. Result ID: 6992643c3a6c3148490ea2a2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69a11e209131cea233a8.md>)

Recorded evaluation configuration: {"Target Label":"GT","Backbone":"DeiT"}. This record does not assert independent reproduction.

### TokenTM

Neg Score: 59.36. Result ID: 6992643c3a6c3148490ea2a4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa729e1e209131cea57935.md>)

Recorded evaluation configuration: {"Target Label":"GT","Backbone":"DeiT"}. This record does not assert independent reproduction.

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