# Image Classification on CIFAR-10H (hard subset) Source: [Image Classification on CIFAR-10H (hard subset)]() Collection index: [llms.txt]() Data updated: 2026-06-19T08:53:50.551Z Evaluation of Image Classification performance on the hard subset of the CIFAR-10H dataset, reporting Accuracy and Expected Calibration Error (ECE). 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) | CIFAR-10H hard subset | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 14 of 14 | | 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 - [Accuracy](): Higher is better; unit: PERCENT. - [ECE](): Lower is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Single Expert | 64.8 | Not recorded | 6a34cd53206542251d7c61d4 | | Single Expert | 64.8 | Not recorded | 6a34cd53206542251d7c61d4 | | FGR | 64.3 | Not recorded | 6a34cd53206542251d7c61d4 | | FGR | 64.3 | Not recorded | 6a34cd53206542251d7c61d4 | | Vanilla MoE | 63.7 | Not recorded | 6a34cd53206542251d7c61d4 | | Vanilla MoE | 63.7 | Not recorded | 6a34cd53206542251d7c61d4 | | MoCaE | 63.7 | Not recorded | 6a34cd53206542251d7c61d4 | | MoCaE | 63.7 | Not recorded | 6a34cd53206542251d7c61d4 | | Robust Filtered | 62.4 | Not recorded | 6a34cd53206542251d7c61d4 | | Robust Filtered | 62.4 | Not recorded | 6a34cd53206542251d7c61d4 | | Robust MoE | 61.2 | Not recorded | 6a34cd53206542251d7c61d4 | | Robust MoE | 61.2 | Not recorded | 6a34cd53206542251d7c61d4 | | FGR + Robust | 60 | Not recorded | 6a34cd53206542251d7c61d4 | | FGR + Robust | 60 | Not recorded | 6a34cd53206542251d7c61d4 | ### Single Expert Accuracy: 64.8. Result ID: 6a35034afbc373592c328810. [Reporting paper]() Recorded evaluation configuration: {"Temperature Scaling":"false"}. This record does not assert independent reproduction. ### Single Expert Accuracy: 64.8. Result ID: 6a35034afbc373592c328811. [Reporting paper]() Recorded evaluation configuration: {"Temperature Scaling":"true"}. This record does not assert independent reproduction. ### FGR Accuracy: 64.3. Result ID: 6a35034afbc373592c328816. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Temperature Scaling":"false"}. This record does not assert independent reproduction. ### FGR Accuracy: 64.3. Result ID: 6a35034afbc373592c328817. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Temperature Scaling":"true"}. This record does not assert independent reproduction. ### Vanilla MoE Accuracy: 63.7. Result ID: 6a35034afbc373592c328812. [Reporting paper]() Recorded evaluation configuration: {"Temperature Scaling":"false"}. This record does not assert independent reproduction. ### Vanilla MoE Accuracy: 63.7. Result ID: 6a35034afbc373592c328813. [Reporting paper]() Recorded evaluation configuration: {"Temperature Scaling":"true"}. This record does not assert independent reproduction. ### MoCaE Accuracy: 63.7. Result ID: 6a35034afbc373592c328814. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Temperature Scaling":"false"}. This record does not assert independent reproduction. ### MoCaE Accuracy: 63.7. Result ID: 6a35034afbc373592c328815. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Temperature Scaling":"true"}. This record does not assert independent reproduction. ### Robust Filtered Accuracy: 62.4. Result ID: 6a35034afbc373592c32881a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Temperature Scaling":"false"}. This record does not assert independent reproduction. ### Robust Filtered Accuracy: 62.4. Result ID: 6a35034afbc373592c32881b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Temperature Scaling":"true"}. This record does not assert independent reproduction. ### Robust MoE Accuracy: 61.2. Result ID: 6a35034afbc373592c328818. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Temperature Scaling":"false"}. This record does not assert independent reproduction. ### Robust MoE Accuracy: 61.2. Result ID: 6a35034afbc373592c328819. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Temperature Scaling":"true"}. This record does not assert independent reproduction. ### FGR + Robust Accuracy: 60. Result ID: 6a35034afbc373592c32881c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Temperature Scaling":"false"}. This record does not assert independent reproduction. ### FGR + Robust Accuracy: 60. Result ID: 6a35034afbc373592c32881d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Temperature Scaling":"true"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.