I ran this plan on 3 different planning tools. 2 of them varied based on the GF Lo and the 3rd did not.
I have always understood GF to be used as
@scubadada said. I.e. If you can make a direct ascent to the surface without exceeding GF Hi, then GF Lo is not used. But, it is clear (now) that some implementations of GF use GF Lo differently (than I understood). However, it also seems to be the case that there is no ambiguity about how GF Hi is used.
And, I submit that using GF Lo, even when GF Hi will not be exceeded makes dubious sense. You will exceed GF Lo (when set to the normal 30 - 50 range) on almost any dive, other than the shortest, shallow dives. Using GF Lo that way would make almost every (okay, maybe only many, many) normal recreational dive into a deco dive. That does not make sense.
Example: I'm doing a dive that is intended to be an NDL dive, and I have my computer set to GF 30/70. I'm on the bottom and a direct ascent would have me at the surface with a GF of 65. In that case, clearly there is a point before I hit the surface where my GF will exceed 30. For my computer to give me a mandatory stop in that case makes no sense whatsoever.
Of course, this distinction is only relevant to people attempting to do NDL dives, which is not where GF are generally used. Once you are down long enough that you'll exceed GF Hi, it doesn't matter, right? Regardless, I think that defining a dive as a deco dive based on GF Lo does not make sense. It only makes sense (to me) to define it as a dive where you will exceed GF Hi on a direct ascent*.
*Obviously, the definition only applies to people using the Buhlmann algorithm with GF. Other algorithms implicitly have different ways of determining if there's a mandatory stop.