Improved Overlap Metrics in ML Models

We’ve refined how overlap metrics are calculated and displayed in ML Models.
The system now accurately retrieves the correct segment size for overlapping users and only shows overlap metrics when there’s an actual overlap.

We’ve also handled edge cases where real‑time vs. cached metrics could incorrectly show the overlap as larger than the source or target. Additionally, the ML Model summary header now clearly labels dates as “Created” instead of “Created by” for improved clarity.