GPS biaspt and weird step weights

I am using g4-10.7.2 version.
I use the following mac file to generate a downward hemispherical source to incident a central calorimeter.
gamma_iso_mac.txt (359 Bytes)

If I use biaspt(as mentioned in the mac file), the total maximum Edep will becomes half of the primary particle energy (1GeV). This value is related to hist bin value(0.5) of biaspt (let’s say, if set 0.75, it’s 750MeV). I find that this command changes the weights in step/track( step->GetWeight() ).
This is very strange, in my understanding, this command was supposed to define the position distribution.
I found some relevant explanations in the release of this version (search “bias” in []),
but I’m still confused. Does anyone know why?