# XAI Evaluation on ImageNet-1k-mini

Source: [XAI Evaluation on ImageNet-1k-mini](<https://www.sota2.com/research/sota/xai-evaluation-on-imagenet-1k-mini>)

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

Data updated: 2026-02-26T05:35:00.282Z

Evaluation of Explainable AI (XAI) methods on the ImageNet-1k-mini dataset, typically measured using attribution fidelity metrics like Area Under Curve (AUC), Insertion, and Deletion.

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 | XAI Evaluation |
| Dataset (including any stated split) | ImageNet-1k-mini |
| Metric | Area Metric |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 8 of 8 |
| 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

- [Area Metric](<https://www.sota2.com/research/sota/xai-evaluation-on-imagenet-1k-mini.md?metric=Area%20Metric>): Lower is better; unit: SCALAR.

- [Standard Deviation (Std)](<https://www.sota2.com/research/sota/xai-evaluation-on-imagenet-1k-mini.md?metric=Standard%20Deviation%20(Std)>): Lower is better; unit: SCALAR.

- [Insertion Score](<https://www.sota2.com/research/sota/xai-evaluation-on-imagenet-1k-mini.md?metric=Insertion%20Score>): Higher is better; unit: SCALAR.

- [Deletion Score](<https://www.sota2.com/research/sota/xai-evaluation-on-imagenet-1k-mini.md?metric=Deletion%20Score>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Area Metric | Result date | Report paper |
| --- | --- | --- | --- |
| rex | 0.0287 | Not recorded | 68fa7a561e209131cea8293b |
| Grad-CAM | 0.0337 | Not recorded | 68fa7a561e209131cea8293b |
| lime | 0.0493 | Not recorded | 68fa7a561e209131cea8293b |
| rise | 0.0615 | Not recorded | 68fa7a561e209131cea8293b |
| noisetunnel | 0.1116 | Not recorded | 68fa7a561e209131cea8293b |
| GradientShap | 0.1394 | Not recorded | 68fa7a561e209131cea8293b |
| ig | 0.3414 | Not recorded | 68fa7a561e209131cea8293b |
| KernelShap | 0.4357 | Not recorded | 68fa7a561e209131cea8293b |

### rex

Area Metric: 0.0287. Result ID: 699ce3bc5698665ec48479a4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a561e209131cea8293b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a561e209131cea8293b.md>)

Recorded evaluation configuration: {"backbone":"ConvNeXt"}. This record does not assert independent reproduction.

### Grad-CAM

Area Metric: 0.0337. Result ID: 699ce3bc5698665ec484799e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a561e209131cea8293b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68bd1e209131cea1dec5.md>)

Recorded evaluation configuration: {"backbone":"ConvNeXt"}. This record does not assert independent reproduction.

### lime

Area Metric: 0.0493. Result ID: 699ce3bc5698665ec48479a3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a561e209131cea8293b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62a11e209131ce9fa620.md>)

[Implementation](<https://github.com/marcotcr/lime-experiments>)

Recorded evaluation configuration: {"backbone":"ConvNeXt"}. This record does not assert independent reproduction.

### rise

Area Metric: 0.0615. Result ID: 699ce3bc5698665ec48479a5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a561e209131cea8293b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64dc1e209131cea06cb9.md>)

[Implementation](<https://github.com/pytorch/vision>)

Recorded evaluation configuration: {"backbone":"ConvNeXt"}. This record does not assert independent reproduction.

### noisetunnel

Area Metric: 0.1116. Result ID: 699ce3bc5698665ec484799f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a561e209131cea8293b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63661e209131ce9fe6d2.md>)

Recorded evaluation configuration: {"backbone":"ConvNeXt"}. This record does not assert independent reproduction.

### GradientShap

Area Metric: 0.1394. Result ID: 699ce3bc5698665ec48479a1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a561e209131cea8293b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63d01e209131cea00e17.md>)

Recorded evaluation configuration: {"backbone":"ConvNeXt"}. This record does not assert independent reproduction.

### ig

Area Metric: 0.3414. Result ID: 699ce3bc5698665ec48479a0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a561e209131cea8293b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63661e209131ce9fe6bd.md>)

Recorded evaluation configuration: {"backbone":"ConvNeXt"}. This record does not assert independent reproduction.

### KernelShap

Area Metric: 0.4357. Result ID: 699ce3bc5698665ec48479a2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a561e209131cea8293b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63d01e209131cea00e17.md>)

Recorded evaluation configuration: {"backbone":"ConvNeXt"}. This record does not assert independent reproduction.

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