# Machine Unlearning on MedQA QF=100

Source: [Machine Unlearning on MedQA QF=100](<https://www.sota2.com/research/sota/machine-unlearning-on-medqa-qf-100>)

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

Data updated: 2026-02-26T05:45:29.253Z

Evaluation of Machine Unlearning effectiveness on the MedQA QF=100 dataset, reporting metrics such as Forget Accuracy, Test Accuracy, and Membership Inference Attack (MIA) scores.

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 | Machine Unlearning |
| Dataset (including any stated split) | MedQA QF=100 |
| Metric | Forget Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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

- [Forget Accuracy](<https://www.sota2.com/research/sota/machine-unlearning-on-medqa-qf-100.md?metric=Forget%20Accuracy>): Higher is better; unit: PERCENT.

- [Test Accuracy](<https://www.sota2.com/research/sota/machine-unlearning-on-medqa-qf-100.md?metric=Test%20Accuracy>): Higher is better; unit: PERCENT.

- [Gen. Accuracy](<https://www.sota2.com/research/sota/machine-unlearning-on-medqa-qf-100.md?metric=Gen.%20Accuracy>): Higher is better; unit: PERCENT.

- [MIA AUC](<https://www.sota2.com/research/sota/machine-unlearning-on-medqa-qf-100.md?metric=MIA%20AUC>): Higher is better; unit: SCALAR.

- [MIA Score](<https://www.sota2.com/research/sota/machine-unlearning-on-medqa-qf-100.md?metric=MIA%20Score>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Forget Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Sequential Retrain | 91.33 | Not recorded | 699bdf1a09c33900fcd804f1 |
| Adapter Merging | 91.33 | Not recorded | 699bdf1a09c33900fcd804f1 |
| NPO | 90 | Not recorded | 699bdf1a09c33900fcd804f1 |
| Sequential LoRA | 89.33 | Not recorded | 699bdf1a09c33900fcd804f1 |
| Origin | 89.33 | Not recorded | 699bdf1a09c33900fcd804f1 |
| GA | 89 | Not recorded | 699bdf1a09c33900fcd804f1 |
| SBU | 73 | Not recorded | 699bdf1a09c33900fcd804f1 |

### Sequential Retrain

Forget Accuracy: 91.33. Result ID: 699d03005698665ec4a32219.

[Reporting paper](<https://www.sota2.com/research/paper/699bdf1a09c33900fcd804f1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a561e209131cea82962.md>)

Recorded evaluation configuration: {"QF":"100"}. This record does not assert independent reproduction.

### Adapter Merging

Forget Accuracy: 91.33. Result ID: 699d03005698665ec4a3221c.

[Reporting paper](<https://www.sota2.com/research/paper/699bdf1a09c33900fcd804f1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78241e209131cea76f80.md>)

Recorded evaluation configuration: {"QF":"100"}. This record does not assert independent reproduction.

### NPO

Forget Accuracy: 90. Result ID: 699d03005698665ec4a3221b.

[Reporting paper](<https://www.sota2.com/research/paper/699bdf1a09c33900fcd804f1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa753a1e209131cea66964.md>)

Recorded evaluation configuration: {"QF":"100"}. This record does not assert independent reproduction.

### Sequential LoRA

Forget Accuracy: 89.33. Result ID: 699d03005698665ec4a32218.

[Reporting paper](<https://www.sota2.com/research/paper/699bdf1a09c33900fcd804f1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a561e209131cea82962.md>)

Recorded evaluation configuration: {"QF":"100"}. This record does not assert independent reproduction.

### Origin

Forget Accuracy: 89.33. Result ID: 699d03005698665ec4a3221d.

[Reporting paper](<https://www.sota2.com/research/paper/699bdf1a09c33900fcd804f1.md>)

Recorded evaluation configuration: {"QF":"100"}. This record does not assert independent reproduction.

### GA

Forget Accuracy: 89. Result ID: 699d03005698665ec4a3221a.

[Reporting paper](<https://www.sota2.com/research/paper/699bdf1a09c33900fcd804f1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e421e209131cea3ec45.md>)

[Implementation](<https://github.com/joeljang/knowledge-unlearning>)

Recorded evaluation configuration: {"QF":"100"}. This record does not assert independent reproduction.

### SBU

Forget Accuracy: 73. Result ID: 699d03005698665ec4a3221e.

[Reporting paper](<https://www.sota2.com/research/paper/699bdf1a09c33900fcd804f1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699bdf1a09c33900fcd804f1.md>)

Recorded evaluation configuration: {"QF":"100"}. This record does not assert independent reproduction.

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