Definition
The share of a released claim record's asserted content that originated in model generation rather than in admitted source evidence.
The Measurement Gap
There is no metric that tells you what share of a released record's content originated in model generation rather than in source evidence, which means the boundary between reconstruction and synthesis is held by policy and never measured.
Formula
GCS = ( Σ_c f_generated(c) ) / ( Σ_c f_total(c) ) , reported with the distribution across records rather than the mean alone
Proposed formula, pending validation pilots.
At a Glance
Decision Supported
Whether the reconstruction and synthesis boundary is holding in fact, and whether a released record is safe to re-admit as an evidence channel later.