# Machine Unlearning on MNIST (Image Classification) Source: [Machine Unlearning on MNIST (Image Classification)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:58:57.749Z Evaluation of model unlearning capabilities on the MNIST dataset, measuring performance retention (Pretrain) and successful forgetting (Unlearn). 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 | Image Classification | | Dataset (including any stated split) | MNIST | | Metric | Pretrain Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 12 of 12 | | 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 - [Pretrain Accuracy](): Higher is better; unit: PERCENT. - [Unlearn Success Rate](): Higher is better; unit: PERCENT. ## Recorded results | Method | Pretrain Accuracy | Result date | Report paper | | --- | --- | --- | --- | | f-FUM (KL-KL) | 98.93 | Not recorded | 69896a26c97a0b3e3b5dc58a | | f-FUM (KL-JS) | 98.93 | Not recorded | 69896a26c97a0b3e3b5dc58a | | f-FUM (KL-χ²) | 98.93 | Not recorded | 69896a26c97a0b3e3b5dc58a | | NoT | 98.93 | Not recorded | 69896a26c97a0b3e3b5dc58a | | Halimi | 98.93 | Not recorded | 69896a26c97a0b3e3b5dc58a | | MoDe | 98.93 | Not recorded | 69896a26c97a0b3e3b5dc58a | | f-FUM (KL-KL) | 97.93 | Not recorded | 69896a26c97a0b3e3b5dc58a | | f-FUM (KL-JS) | 97.93 | Not recorded | 69896a26c97a0b3e3b5dc58a | | f-FUM (KL-χ²) | 97.93 | Not recorded | 69896a26c97a0b3e3b5dc58a | | NoT | 97.93 | Not recorded | 69896a26c97a0b3e3b5dc58a | | Halimi | 97.93 | Not recorded | 69896a26c97a0b3e3b5dc58a | | MoDe | 97.93 | Not recorded | 69896a26c97a0b3e3b5dc58a | ### f-FUM (KL-KL) Pretrain Accuracy: 98.93. Result ID: 6998ead65698665ec44f5503. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"N":10}. This record does not assert independent reproduction. ### f-FUM (KL-JS) Pretrain Accuracy: 98.93. Result ID: 6998ead65698665ec44f5504. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"N":10}. This record does not assert independent reproduction. ### f-FUM (KL-χ²) Pretrain Accuracy: 98.93. Result ID: 6998ead65698665ec44f5505. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"N":10}. This record does not assert independent reproduction. ### NoT Pretrain Accuracy: 98.93. Result ID: 6998ead65698665ec44f5506. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"N":10}. This record does not assert independent reproduction. ### Halimi Pretrain Accuracy: 98.93. Result ID: 6998ead65698665ec44f5507. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"N":10}. This record does not assert independent reproduction. ### MoDe Pretrain Accuracy: 98.93. Result ID: 6998ead65698665ec44f5508. [Reporting paper]() Recorded evaluation configuration: {"N":10}. This record does not assert independent reproduction. ### f-FUM (KL-KL) Pretrain Accuracy: 97.93. Result ID: 6998ead65698665ec44f5509. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"N":5}. This record does not assert independent reproduction. ### f-FUM (KL-JS) Pretrain Accuracy: 97.93. Result ID: 6998ead65698665ec44f550a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"N":5}. This record does not assert independent reproduction. ### f-FUM (KL-χ²) Pretrain Accuracy: 97.93. Result ID: 6998ead65698665ec44f550b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"N":5}. This record does not assert independent reproduction. ### NoT Pretrain Accuracy: 97.93. Result ID: 6998ead65698665ec44f550c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"N":5}. This record does not assert independent reproduction. ### Halimi Pretrain Accuracy: 97.93. Result ID: 6998ead65698665ec44f550d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"N":5}. This record does not assert independent reproduction. ### MoDe Pretrain Accuracy: 97.93. Result ID: 6998ead65698665ec44f550e. [Reporting paper]() Recorded evaluation configuration: {"N":5}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.