Here is how I see the problem at this time--
There's no doubt that in a turn, there's a rolling-in torque (toward a steeper bank angle) created by the difference in airspeed, and lift, between the inboard and outboard wingtips.
Does concentrating the mass at the wingtips reduce this roll torque? Obviously roll inertia is increased, but is the roll torque actually reduced?
Let's say we attach a string to the CG and sling the airplane around and around in a circle, somewhat in the matter of a "control-line" model airplane, but with the wings at the zero-lift angle-of-attack, so the aerodynamic roll torque noted in the first paragraph disappears. Is there any intrinsic, inertial tendency for the aircraft to roll to a wings-level attitude, rather than adopting a steeply-banked attitude? A tendency that, if it existed, might be more pronounced if the wingtips were heavily weighted? It intuitively appears to me that the answer is "no".
Therefore the answer to the original question intuitively appears to be "no".
One could argue that the outboard wingtip, travelling at a larger velocity and experiencing more apparent "centrifugal force" than the inboard wingtip, will tend to dominate the situation and therefore there WILL be some very small self-leveling effect, and more so if the weight is concentrated near the tips. This effect appears to have some similarity to a "tidal" effect in orbital dynamics-- it would be based upon the different radii experienced by different parts of the body in the orbit or turn. This effect is surely a very minor player in roll stability dynamics in a normal turn-- unlike in a flat spin, where the center of rotation may be near the center of the aircraft, and any weight added to the wingtips will strongly tend to make the spin go even flatter.
This may be a situation where it is actually most helpful to adopt a reference frame based on the moving aircraft itself. This will not be a valid inertial reference frame, so we'll have to take "centrifugal force" into account. In a fully coordinated turn, the direction of the apparent weight of each tip tank will be straight down in the aircraft's reference frame, but the outboard tip tank will seem to "weigh" more than the inboard one, and this create a rolling-out tendency, or to put it another way, will cancel out the rolling-in tendency created by the difference in airspeed between the two wingtips. A perfectly span-loaded aircraft will experience no rolling-in or rolling-out tendency due either to the difference in apparent weight, or the difference in lift, at different points along the wingspan. If all the weight is concentrated at the wingtips, this would appear to create some self-leveling effect.