I have a problem interpreting the marginal effect of a dummy variable in a logit model. (I am using Stata to estimate the logit regression)

I've run a simple logit say this:
logit underline i.L i.L_underline i.L2_q i.L2_underline
where all variables are dummies. Then to compute the margins I ran this:
margins, dydx(_all) at(L2_underline=1 L_underline=1)
I wanted to know the margins where L2_underline=1 andL_underline=1.

How should I interpret the margin for L2_underline?