stata too many variables specified

请哪位大佬帮忙看看问题出在哪呢，结果显示too many variables specified


代码：
logit y x1 x2 x3 x4 fs fo hcms x1x3 x1x4 x2x3 x2x4, robust
qui sum x1 if e(sample)
local meanx1=r(mean)
qui sum x2 if e(sample)
local meanx2=r(mean)
qui sum x3 if e(sample)
local meanx3=r(mean)
qui sum x4 if e(sample)
local meanx4=r(mean)
qui sum fs if e(sample)
local meanfs=r(mean)
qui sum fo if e(sample)
local meanfo=r(mean)
qui sum hcms if e(sample)
local meanhcms=r(mean)
local xb_b[x1]*`meanx1'+_b[x3]*`meanx3'+_b[x2]*`meanx2'+_b[x4]*`meanx4'+_b[x1x3]*`meanx1'*`meanx3'+_b[x1x4]*`meanx1'*`meanx4'+_b[x2x3]*`meanx2'*`meanx3'+_b[x2x4]*`meanx2'*`meanx4'+_b[fs]*`meanfs'+_b[fo]*`meanfo'+_b[hcms]*`meanhcms'+_b[_cons]
predictnl dydx1x3==normden(`xb')*(-(`xb'))*(_b[x3]+_b[x1x3]*`meanx1'+_b[x2x3]*`meanx2')*(_b[x1]+_b[x1x3]*`meanx3'+_b[x1x4]*`meanx4')+ normden(`xb')*(_b[x1x3]) in 1, se(sedydx1x3)
list dydx1x3 sedydx1x3 in 1