i have a question about estimating coefficients of a multiple linear regression model, say, Y=B0 + B1*x1 + B2*x2 + B3*x3 + e.

if b1 is the estimate of B1, i know that b1 estimates the effect of one unit increase of x1, holding the other variables held fixed.
but if you use SAS to compute b1 (like suppose you are using proc reg) , then at what values are x2 and x3 are held fixed?
it is 0, the means orsome other values?

and is there any way to set the values that i want x1 and x2 to be held fixed at?

please help me with this.

thank you vary much.