I am trying to estimate the marginal effects of the interaction of caste with the level of education using Logit Model. Both of them are categorical variables. I have used the following command:

margins (caste), dydx(edulevel) atmeans

The resulting table takes edulevel1 as the base category but doesn't specify any base category for caste. When I calculate the coefficients, it specifies a base category for both edulevel and caste. I want to know if I am calculating marginal effects for the interaction dummies correctly.

Thanks in advance