When adjusting GI multipliers in the VRayMtlWrapper, does that affect only the rendered surface, or also the amount of light transported to the next bounce? In finalRender stage-1’s equivalent, it affected light transport, so you had to enter the inverse value in the opposite slot for energy preservation (e.g. for a receive GI of 2.0 you had to enter a send GI of 0.5 to avoid hotspots on adjacent surfaces).
A quick test seems to indicate this is not the case in VRay, but I wanted to be sure.