@philip , please share the original RAW captures if you can. Use a sharing platform (e.g. wetransfer.com) if the forum doesn’t accept the files directly.
Other than that: NLP’s adaptive algorithm uses the image properties (colours, tonality, white- and black points etc.) to set conversion parameters. Depending on how your camera is set, conversion results can vary, also depending on many other settings. Whether differences are caused by camera or software doesn’t really matter yet.
Note that NLP 2.3 is fairly old and that newer versions do differently. Version 3.1 (currently in beta) has a positive (slide) mode that could be interesting here.