Empirical

Coherence Cliff

Scaling Evidence for Sheaf Diagnostics

Demonstrated500 graphs, 7 scales, 5-50 agents. Sheaf diagnostic vs. 12 baselines.
Key result

Sheaf R² > 0.96 at all scales; best baseline degrades 0.83 → 0.52 as agent count grows 5 → 50

Falsification

A bounded-depth testing strategy that maintains accuracy at scale

Abstract

Documents the regime change where conventional testing breaks down as composition size grows. Six convention dimensions grounded in real standards.