First video with the guy walking the mountain in snow has consistency issues with the cave entrance. Which is "expected" at this model size?!
All of the videos have rather glaring consistency issues when direction shifts back to areas previously shown.
Most videos seem to have some issues like that, e.g. the book on the table in the library video takes up different shapes every now and then.
The 'Refiner' effect seems to do the opposite if the examples are representative as in all cases the 1-st stage images look better than the 'refined' ones. Less clutter, more realistic, less 'cowbell' for those who know the phrase.
Remember the first Will Smith spaghetti?
My dreams have it too, which is unexpected at that model size!