Shadow mode lets a team compare outputs on live traffic before switching. It is useful for catching regressions, but it does not replace human evaluation of contested cases.
Why it matters
Running a new model alongside the current one without letting its output affect users. Understanding the term is the first step; applying it consistently is the harder part.
Related
Browse the glossary or read the editorial standards.