314.28(A)2 is pretty clear about saying where splices are made, the distances apply.
But in [2011] that section is also clear that the distance referred to is not the distance between conduit entries on the same side of the box but rather the distance to the opposite wall, allowing for bending as the wires exit the raceway for either splicing or pulling through.
If two conductors entered and left through adjacent raceways on the same side and were spliced to each other, the distance to the opposite wall would make this doable regardless of the separation between the raceways themselves, but would carry the risk that a single wire was then pulled to too small a radius during the second, outbound, part of the pull.
...the distance between each raceway entry inside the box or conduit body and the opposite wall of the box or conduit body shall not be less than six times the metric designator (trade size) of the largest raceway in a row.
The above part does not make any condition on the relationship of the conductors in the raceways.
The later part, as quoted earlier, adds the condition that the raceways carry the
same conductor. IMHO, this would not apply to two conductors which were spliced to form a part of a single circuit. Otherwise the language could have said circuit instead of conductor.
A splice could have the ability to allow a circuit path to change direction by 180 degrees without any bending involved. An inline or butt splice, OTOH, would not do that.