The gravitational binding energy of the sun is given by
$$\frac{3 G M^2}{5R}$$
If we ignore the radius component, halving the mass of the sun would involve:
$$\frac{3 * 6.674*10^{−11} N*kg^{–2}*m^2 * (10^30kg)^2}{5*695,508 km}$$$$\frac{3 \cdot 6.674 \cdot 10^{−11}\;N \cdot kg^{–2} \cdot m^2 \cdot (10^{30}\;kg)^2}{5*695 508\;km}$$
or 5E40 J$5 \cdot 10^{40}\;J$.
The sun emits 3.846*10^26 watts$3.846 \cdot 10^{26}\;W$ of power, so this is about 10^14 seconds$10^{14}\;s$ of solar output, or 4 million years give or take.
If the process was 90% efficient it would increase the sun's energy output 1000 fold, frying most of the solar system. Uranus would get 2.5 as much energy per unit area than Earth does now. Objects at 15 AU out would get as much energy from this process as Mercury does now.
No plausible natural event is going to be 90% efficient at getting matter away from the Sun. The Sun is a gravitationally-bound fusion-supported structure. It already generates huge amounts of energy to keep itself supported at its size; getting large amounts of matter out of a Sun is going to be non-trivial effort for a Type-3 civilization.
It would be a project that would, on its scale, equivalent to the energy consumed by the Manhatten Project (which used lots of energy as part of the separation process).
There is no plausible way this at all appears natural. And anyone doing it unnaturally would have to do extreme measures to prevent energy lost as waste heat from cooking the solar system.