# Class-level Machine Unlearning on ImageNet (test)

Source: [Class-level Machine Unlearning on ImageNet (test)](<https://www.sota2.com/research/sota/class-level-machine-unlearning-on-imagenet-test>)

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

Data updated: 2026-06-01T07:54:21.448Z

Evaluation of Class-level Machine Unlearning performance on the test split of the ImageNet dataset, focusing on metrics related to target accuracy retention and oversight.

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 | Class-level Machine Unlearning |
| Dataset (including any stated split) | ImageNet (test) |
| Metric | Retained Accuracy (Test) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 2 of 2 |
| 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

- [Retained Accuracy (Test)](<https://www.sota2.com/research/sota/class-level-machine-unlearning-on-imagenet-test.md?metric=Retained%20Accuracy%20(Test)>): Higher is better; unit: PERCENT.

- [Retained Accuracy (Ref)](<https://www.sota2.com/research/sota/class-level-machine-unlearning-on-imagenet-test.md?metric=Retained%20Accuracy%20(Ref)>): Higher is better; unit: PERCENT.

- [Oversight (OU) @ Epsilon](<https://www.sota2.com/research/sota/class-level-machine-unlearning-on-imagenet-test.md?metric=Oversight%20(OU)%20%40%20Epsilon>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Retained Accuracy (Test) | Result date | Report paper |
| --- | --- | --- | --- |
| Original Model | 54.18 | Not recorded | 68fa79171e209131cea7c144 |
| Spotter | 0 | Not recorded | 68fa79171e209131cea7c144 |

### Original Model

Retained Accuracy (Test): 54.18. Result ID: 6a1d38defbc373592c161e96.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79171e209131cea7c144.md>)

Recorded evaluation configuration: {"Backbone":"ViT-B/32","Architecture":"CLIP"}. This record does not assert independent reproduction.

### Spotter

Retained Accuracy (Test): 0. Result ID: 6a1d38defbc373592c161e97.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79171e209131cea7c144.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79171e209131cea7c144.md>)

Recorded evaluation configuration: {"Backbone":"ViT-B/32","Architecture":"CLIP"}. This record does not assert independent reproduction.

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