In copyright law, a derivative work is an expressive creation that includes major copyrightable elements of an original, previously created first work (the underlying work). The derivative work becomes a second, separate work independent in form from the first. The transformation, modification or adaptation of the work must be substantial and bear its author's personality sufficiently to be original and thus protected by copyright.
Documenting software does not by itself make the documentation a derivative work of the software being documented. For example, someone can write a book on how to use a certain library. With normal licenses, nothing stops the author to include example code snippets on how to use the library, yet remain free to license the book as they please.
How does this apply to a Jupyter Notebook? A Jupyter Notebook could be considered as literate programming; a piece of software liberally commented performing a certain role. This would make it a derivate work — if it needs a GPL-library to run, it would need to be licensed appropriately. Or a Jupiter Notebook could be considered documentation: the main purpose is not, after all, to produce functionality, but rather to educate the reader on how to use software. In this interpretation, the notebook is documentation with a lot of examples.
There is a related question Do I need to license a Jupyter Notebook that uses a forked AGPL library?, but that one doesn't raise the question about documentation at all