Your first variation fits this, but then with re-tucking and whatever the mess (!) is in the 2nd post you've gone
away from this notion into some heavy interlocking.
The "mess" is not anything different than what was already at the two last pictures of the first post - same knot(s), only shorter, and, of course, with more entangled links. However, I prefer to keep the notion of "interlocked" overhand knots for more "heavy" entanglements, so I have used the softer term " interlinked" ( the sequence goes like this : inter-penetrating, inter-linked, inter-locked )
As "inter-penetrating" knots I mean the knots where the standing end and/or the tail of the one/first link penetrates the other/second link, without any further entanglement that would offer more friction support to the whole knot - for example, without them being twisted around segments of the other/second link.
Notice that this definition has no relation with the mutual position of the nubs of the two links relatively to their standing ends. The two links can be pushed or pulled the one on the other, but this does not change the way the segments of the one link are tangled with the segments of the other - which determines the characterization of the bend. Now, it may happen that the links of the "inter-penetrating" bends are always pushed against each other, indeed, but this happens only because the knot would not be secure if they were pulled the one from the other - the inter-penetration is a weak entanglement, by itself. It needs to be supported by the correct relation of the bulks of the nubs of the two links, so they will be able to form a secure bend.