I calculated the 2mm and 4mm numbers. The floats that I had were 2.6g for the original, 3.2g for the latest Bing design and 3.8g for the new MS floats.
The 3.2g float sets about pin high in the fuel. Therefore, fully submerged, the float would displace 6.4g of fuel (mogas in this case).
The floats are all similar in size and shape. (not counting a possible hole size difference as noted above)
All 3 styles of floats are 0.828in or 21mm high. The float buoyancy height is 21mm/6.4g or 3.28mm/g. (not accounting for the rounded corners)
Increasing the float weight from 3.2g to 3.8g or 0.6g would sink the float another 2.0mm or 0.6g x 3.28mm/g.
Placing the floats in a float bowl yielded about the same sink as the calculation. The pin diameter is 2mm and observaton of changing float height in the fluid appeard consistent to the 2mm/4mm calculation. I tried to measure the height of the guide pin sticking out above the various floats. I measure 1.6mm more sink for the next heaviest float and 3.0mm for the MS float. 1.6mm and 3.0mm vs 2mm and 4mm.
Pick your measurement. Regardless, the heavier float sinks further and the fuel has to rise a like amount to close the float needle orifice.
Higher fuel levels in the float bowl cause the engine to run richer.
Increasing the float weight from 2.6g to 3.8g increases the float density by 46%, thus float weight is significant factor for proper running of the engine when compared to a 4% change in fuel density.
