# Neuron Interpretation on ImageNet CoSy benchmark avgpool layer 1k Source: [Neuron Interpretation on ImageNet CoSy benchmark avgpool layer 1k]() Collection index: [llms.txt]() Data updated: 2026-04-10T08:55:21.197Z Evaluation of neuron interpretation methods applied to models trained on the ImageNet 1K dataset, specifically focusing on activations derived from the avgpool layer within the CoSy benchmark configuration. 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 | Neuron Interpretation | | Dataset (including any stated split) | ImageNet CoSy benchmark avgpool layer 1k | | Metric | AUC | | Unit | SCALAR | | 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 - [AUC](): Higher is better; unit: SCALAR. - [MAD](): Lower is better; unit: SCALAR. ## Recorded results | Method | AUC | Result date | Report paper | | --- | --- | --- | --- | | LINE | 0.97 | Not recorded | 69d8842a7d0172140ec77abc | | LINE | 0.96 | Not recorded | 69d8842a7d0172140ec77abc | | LINE | 0.94 | Not recorded | 69d8842a7d0172140ec77abc | | CLIP-Dissect | 0.91 | Not recorded | 69d8842a7d0172140ec77abc | | INVERT | 0.88 | Not recorded | 69d8842a7d0172140ec77abc | | CLIP-Dissect | 0.87 | Not recorded | 69d8842a7d0172140ec77abc | | INVERT | 0.83 | Not recorded | 69d8842a7d0172140ec77abc | | DnD | 0.76 | Not recorded | 69d8842a7d0172140ec77abc | | INVERT | 0.76 | Not recorded | 69d8842a7d0172140ec77abc | | DnD | 0.73 | Not recorded | 69d8842a7d0172140ec77abc | | CLIP-Dissect | 0.69 | Not recorded | 69d8842a7d0172140ec77abc | | DnD | 0.56 | Not recorded | 69d8842a7d0172140ec77abc | ### LINE AUC: 0.97. Result ID: 69d8baf047f34ce924c1bbf1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"ResNet50","Layer":"avgpool"}. This record does not assert independent reproduction. ### LINE AUC: 0.96. Result ID: 69d8baf047f34ce924c1bbf5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"ResNet18","Layer":"avgpool"}. This record does not assert independent reproduction. ### LINE AUC: 0.94. Result ID: 69d8baf047f34ce924c1bbf9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"ViT-B/16","Layer":"encoder"}. This record does not assert independent reproduction. ### CLIP-Dissect AUC: 0.91. Result ID: 69d8baf047f34ce924c1bbf6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"ResNet18","Layer":"avgpool"}. This record does not assert independent reproduction. ### INVERT AUC: 0.88. Result ID: 69d8baf047f34ce924c1bbf3. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Architecture":"ResNet50","Layer":"avgpool"}. This record does not assert independent reproduction. ### CLIP-Dissect AUC: 0.87. Result ID: 69d8baf047f34ce924c1bbf2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"ResNet50","Layer":"avgpool"}. This record does not assert independent reproduction. ### INVERT AUC: 0.83. Result ID: 69d8baf047f34ce924c1bbf7. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Architecture":"ResNet18","Layer":"avgpool"}. This record does not assert independent reproduction. ### DnD AUC: 0.76. Result ID: 69d8baf047f34ce924c1bbf4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"ResNet50","Layer":"avgpool"}. This record does not assert independent reproduction. ### INVERT AUC: 0.76. Result ID: 69d8baf047f34ce924c1bbfb. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Architecture":"ViT-B/16","Layer":"encoder"}. This record does not assert independent reproduction. ### DnD AUC: 0.73. Result ID: 69d8baf047f34ce924c1bbf8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"ResNet18","Layer":"avgpool"}. This record does not assert independent reproduction. ### CLIP-Dissect AUC: 0.69. Result ID: 69d8baf047f34ce924c1bbfa. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"ViT-B/16","Layer":"encoder"}. This record does not assert independent reproduction. ### DnD AUC: 0.56. Result ID: 69d8baf047f34ce924c1bbfc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Architecture":"ViT-B/16","Layer":"encoder"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.