Framework Engineering Research

An open index of evidence, theory, and practice

research-document

EXP-001 Topology Perturbation Explainability

Research Area
General Research
Status
draft
Version
0.1
Priority
medium
Updated
2026-07-22
Source
research/evidence-runs/ECR-000003-representation-sensitivity/experiments/EXP-001-topology-perturbation/comparison/generated-v3.2-explainability/topology-perturbation-explainability.md

EXP-001 Topology Perturbation Explainability

  • 30 pairwise comparisons were analyzed without changing Comparator v3.1 classifications.
  • Highest review priority observed: high.
  • Most common difference type: inserted_transition (230 occurrences).
  • Provider pattern count: 3.
  • Variant pattern count: 6.
  • Node-renaming and edge-order changes are separated from actual backbone changes.

What This Does Not Establish

  • It does not change Comparator v3.1 scoring.
  • It does not validate any hypothesis.
  • It does not replace human review.