# Counterfactual Explanation on MNIST (test) Source: [Counterfactual Explanation on MNIST (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T06:08:10.012Z Evaluation of counterfactual explanation methods on the test split of the MNIST dataset, measured by Prox and Val 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 | Counterfactual Explanation | | Dataset (including any stated split) | MNIST (test) | | Metric | Validity Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 10 of 10 | | 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 - [Validity Score](): Higher is better; unit: PERCENT. - [Prox](): Lower is better; unit: SCALAR. - [IM1 Score](): Higher is better; unit: SCALAR. - [Implausibility](): Lower is better; unit: SCALAR. - [Robustness Ratio (1e-3)](): Higher is better; unit: SCALAR. ## Recorded results | Method | Validity Score | Result date | Report paper | | --- | --- | --- | --- | | C3LT | 99.9 | Not recorded | 68fa6c161e209131cea320ac | | Adv. Attack | 99.8 | Not recorded | 68fa6c161e209131cea320ac | | ExpGAN | 99.7 | Not recorded | 68fa6c161e209131cea320ac | | PSCE | 98.2 | Not recorded | 6976f4e9a2b28e761d87c0f0 | | Schut | 96 | Not recorded | 6976f4e9a2b28e761d87c0f0 | | BayesCF | 67.6 | Not recorded | 6976f4e9a2b28e761d87c0f0 | | ROAR | 52.8 | Not recorded | 6976f4e9a2b28e761d87c0f0 | | Wachter | 47 | Not recorded | 6976f4e9a2b28e761d87c0f0 | | CEM | 46.9 | Not recorded | 68fa6c161e209131cea320ac | | CVE | 23.1 | Not recorded | 68fa6c161e209131cea320ac | ### C3LT Validity Score: 99.9. Result ID: 698ecc85ffc3901de0a157e9. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Adv. Attack Validity Score: 99.8. Result ID: 698ecc85ffc3901de0a157ea. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### ExpGAN Validity Score: 99.7. Result ID: 698ecc85ffc3901de0a157e6. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### PSCE Validity Score: 98.2. Result ID: 69991f7a5698665ec4866db8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"BNN"}. This record does not assert independent reproduction. ### Schut Validity Score: 96. Result ID: 69991f7a5698665ec4866dba. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"BNN"}. This record does not assert independent reproduction. ### BayesCF Validity Score: 67.6. Result ID: 69991f7a5698665ec4866db9. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"BNN"}. This record does not assert independent reproduction. ### ROAR Validity Score: 52.8. Result ID: 69991f7a5698665ec4866dbc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"BNN"}. This record does not assert independent reproduction. ### Wachter Validity Score: 47. Result ID: 69991f7a5698665ec4866dbb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"BNN"}. This record does not assert independent reproduction. ### CEM Validity Score: 46.9. Result ID: 698ecc85ffc3901de0a157e7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### CVE Validity Score: 23.1. Result ID: 698ecc85ffc3901de0a157e8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.