Sure, I'm referring to the former. In the single-branch pipeline, Jenkins was set to update the submodules to their tip, and that registered as an SCM change and triggered a build. In the multibranch pipeline, the same behavior is set, but the Scan Multibranch Pipe Log shows that Jenkins is only checking the top-level repo's commit ID for changes.
Background on our configuration: My organization has several components in their own repos and then combines them as sub-modules of a "master" repository to build (with the master repository having a few other things, like a pom.xml and Jenkins config). Since we don't care about the submodules pointing to a specific commit on our develop branch, our .gitmodules config points to the develop branch and has ignore=all set.
Can you provide more detail of what you mean when you say:
For example, do you mean one (or more) of these scenarios: