External-transfer diagnostic
55-target aggregate transfer remains unsupported
The metric adapter executes on an external aggregate-delta object. It is useful for method diagnostics but is neither cell-distribution evaluation nor a VCC submission score.
Executed now · bounded diagnostic
AGGREGATE · NOT VCC
Multi-metric run on 55 external aggregate perturbation targets
MAE-delta, delta correlation and L1-PDS now run beside the same ridge, training-mean and zero-change predictions. The source endpoint remains NO_DIRECTIONAL_TRANSFER_SUPPORT.
| Method | RMSE Δ | MAE Δ | PDS L1 | Pearson Δ | Spearman Δ | Effect-size ρ |
|---|---|---|---|---|---|---|
| Safe ridge transferdiagnostic model | 1.0928 | 0.1861 | 0.5147 | -0.0148 | -0.0013 | -0.0554 |
| Training-response meanbaseline | 1.0925 | 0.1844 | 0.5134 | -0.0161 | -0.0027 | undefined |
| Zero changebaseline | 1.0896 | 0.1707 | 0.5134 | undefined | undefined | undefined |
PDS scale auditRaw prediction multiplier
0.00×
0.5134
0.25×
0.5154
0.50×
0.5164
1.00×
0.5147
2.00×
0.5144
4.00×
0.5134
8.00×
0.5193
Truth-norm matched: 0.5431 · median factor 14.52×. This uses paired observed norms and is sensitivity-only, never a fair hidden-test score.
What changed. The metric engine is operational on aggregate deltas. DES/DEG recovery and cell-distribution metrics remain locked because this prediction object has no genuine cell-level distribution.
Execution boundary. This is a formula-aligned aggregate-delta diagnostic, not an official cell-eval run, not a VCC 2025 rescore, and not a VCC 2026 task execution. No muscle, DMD, treatment or clinical validity follows from these scores.