Add note about neutrals to Java documentation

Author: egregius313

docs/codeql/codeql-language-guides/customizing-library-models-for-java-and-kotlin.rst

@@ -64,7 +64,7 @@ The CodeQL library for Java and Kotlin analysis exposes the following extensible
 - ``sourceModel(package, type, subtypes, name, signature, ext, output, kind, provenance)``. This is used to model sources of potentially tainted data. The ``kind`` of the sources defined using this predicate determine which threat model they are associated with. Different threat models can be used to customize the sources used in an analysis. For more information, see ":ref:`Threat models <threat-models-java>`."
 - ``sinkModel(package, type, subtypes, name, signature, ext, input, kind, provenance)``. This is used to model sinks where tainted data maybe used in a way that makes the code vulnerable.
 - ``summaryModel(package, type, subtypes, name, signature, ext, input, output, kind, provenance)``. This is used to model flow through elements.
-- ``neutralModel(package, type, name, signature, kind, provenance)``. This is similar to a summary model but used to model the flow of values that have only a minor impact on the dataflow analysis.
+- ``neutralModel(package, type, name, signature, kind, provenance)``. This is similar to a summary model but used to model the flow of values that have only a minor impact on the dataflow analysis. Manual neutral models (those with a provenance such as ``manual`` or ``ai-manual``) override generated summary models (those with a provenance such as ``df-generated``) such that the summary will be ignored. Other than that, neutral models have a slight impact on the dataflow dispatch logic, which is out of scope for this documentation.
 
 The extensible predicates are populated using the models defined in data extension files.