NLP 2.2 - how to retain the look of conversions from older versions of NLP?

After updating NLP to version 2.2.0 I quickly realised that when I try to re-adjust any photographs that have been previously converted under earlier version (in my case that’s 2.0.0) when I bring up NLP plug-in (e.g. by pressing Ctrl + Alt + N) the appearance of the pictures changes (often dramatically) even though all settings from the original conversion remain unchanged (that can be immediately seen as a preview in the background, even before applying any changes, the histogram can also be drastically different with all sliders remaining in the same position).
As I understand in version 2.2.0 there is a brand new engine that handles the processing. I don’t have doubts that the new engine is superior but the outcome of this new processing with the settings from earlier versions isn’t always better. On some examples of previously converted images I can see the improvement out of the box but there are images that were individually and laboriously mastered in NLP 2.0.0 to give them the best possible look and now when I open them under NLP 2.2.0 they get heavily altered and I don’t know how to bring the same look I had.
Is there a way of replicating the same look that was achieved in conversion carried out under previous versions of NLP (or retaining it) and being able to use it as a starting point before applying more adjustments offered by the latest version (e.g. LUT)?
I would be very grateful for any help in that subject.
(I’m using Lightroom Classic V10.2 and DSLR RAW scans).