# Global Source Attribution on MLAAD Few-Shot Open-Set (test) Source: [Global Source Attribution on MLAAD Few-Shot Open-Set (test)]() Collection index: [llms.txt]() Data updated: 2026-07-07T08:17:06.100Z Evaluation of model performance on the test split of the MLAAD Few-Shot Open-Set benchmark for the Global Source Attribution task, reporting F1-Macro scores and Generalization Gap. 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 | Global Source Attribution | | Dataset (including any stated split) | MLAAD Few-Shot Open-Set (test) | | Metric | F1-Macro (All vs All) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 16 of 16 | | 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 - [F1-Macro (All vs All)](): Higher is better; unit: PERCENT. - [F1-Macro (Seen vs All)](): Higher is better; unit: PERCENT. - [F1-Macro (Unseen vs All)](): Higher is better; unit: PERCENT. - [Generalization Gap](): Lower is better; unit: PERCENT. ## Recorded results | Method | F1-Macro (All vs All) | Result date | Report paper | | --- | --- | --- | --- | | Strategy 2.B (Residual Modeling) | 84.52 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Strategy 2.A (Factorized Prototypes) | 83.97 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Strategy 2.A (Factorized Prototypes) | 83.67 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Strategy 1 (Orthonormal Prototypes) | 82.7 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Strategy 1 (Orthonormal Prototypes) | 82.56 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Strategy 1 (Orthonormal Prototypes) | 82.55 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Strategy 2.B (Residual Modeling) | 82.12 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Strategy 1 (Orthonormal Prototypes) | 81.44 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Strategy 2.A (Factorized Prototypes) | 81.44 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Baseline (ArcFace) | 80.11 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Strategy 1 (Orthonormal Prototypes) | 80.09 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Baseline (ArcFace) | 77.49 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Strategy 2.B (Residual Modeling) | 72.11 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Strategy 2.A (Factorized Prototypes) | 70.94 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Strategy 1 (Orthonormal Prototypes) | 70.18 | Not recorded | 6a4c87e385dba69fef8b23d0 | | Baseline (ArcFace) | 68.01 | Not recorded | 6a4c87e385dba69fef8b23d0 | ### Strategy 2.B (Residual Modeling) F1-Macro (All vs All): 84.52. Result ID: 6a4cb5e760b88bb339c85421. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MSE (λ)":"0.05","K":"5"}. This record does not assert independent reproduction. ### Strategy 2.A (Factorized Prototypes) F1-Macro (All vs All): 83.97. Result ID: 6a4cb5e760b88bb339c8541d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MSE (λ)":"0.0","K":"5"}. This record does not assert independent reproduction. ### Strategy 2.A (Factorized Prototypes) F1-Macro (All vs All): 83.67. Result ID: 6a4cb5e760b88bb339c8541e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MSE (λ)":"0.05","K":"5"}. This record does not assert independent reproduction. ### Strategy 1 (Orthonormal Prototypes) F1-Macro (All vs All): 82.7. Result ID: 6a4cb5e760b88bb339c85417. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MSE (λ)":"0.0","K":"5"}. This record does not assert independent reproduction. ### Strategy 1 (Orthonormal Prototypes) F1-Macro (All vs All): 82.56. Result ID: 6a4cb5e760b88bb339c85418. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MSE (λ)":"0.05","K":"5"}. This record does not assert independent reproduction. ### Strategy 1 (Orthonormal Prototypes) F1-Macro (All vs All): 82.55. Result ID: 6a4cb5e760b88bb339c85419. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MSE (λ)":"0.1","K":"5"}. This record does not assert independent reproduction. ### Strategy 2.B (Residual Modeling) F1-Macro (All vs All): 82.12. Result ID: 6a4cb5e760b88bb339c85420. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MSE (λ)":"0.05","K":"3"}. This record does not assert independent reproduction. ### Strategy 1 (Orthonormal Prototypes) F1-Macro (All vs All): 81.44. Result ID: 6a4cb5e760b88bb339c8541a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MSE (λ)":"1.0","K":"5"}. This record does not assert independent reproduction. ### Strategy 2.A (Factorized Prototypes) F1-Macro (All vs All): 81.44. Result ID: 6a4cb5e760b88bb339c8541c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MSE (λ)":"0.0","K":"3"}. This record does not assert independent reproduction. ### Baseline (ArcFace) F1-Macro (All vs All): 80.11. Result ID: 6a4cb5e760b88bb339c85414. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"MSE (λ)":"–","K":"5"}. This record does not assert independent reproduction. ### Strategy 1 (Orthonormal Prototypes) F1-Macro (All vs All): 80.09. Result ID: 6a4cb5e760b88bb339c85416. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MSE (λ)":"0.0","K":"3"}. This record does not assert independent reproduction. ### Baseline (ArcFace) F1-Macro (All vs All): 77.49. Result ID: 6a4cb5e760b88bb339c85413. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"MSE (λ)":"–","K":"3"}. This record does not assert independent reproduction. ### Strategy 2.B (Residual Modeling) F1-Macro (All vs All): 72.11. Result ID: 6a4cb5e760b88bb339c8541f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MSE (λ)":"0.05","K":"1"}. This record does not assert independent reproduction. ### Strategy 2.A (Factorized Prototypes) F1-Macro (All vs All): 70.94. Result ID: 6a4cb5e760b88bb339c8541b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MSE (λ)":"0.0","K":"1"}. This record does not assert independent reproduction. ### Strategy 1 (Orthonormal Prototypes) F1-Macro (All vs All): 70.18. Result ID: 6a4cb5e760b88bb339c85415. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"MSE (λ)":"0.0","K":"1"}. This record does not assert independent reproduction. ### Baseline (ArcFace) F1-Macro (All vs All): 68.01. Result ID: 6a4cb5e760b88bb339c85412. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"MSE (λ)":"–","K":"1"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.