I am using a survey data to run a logit regression, and currently I am testing the interactions of two continuous variables. I read Norton and Ai's paper that using the command Inteff helps correct coefficients and s.e. of interaction terms in nonlinear model. My logit models controls for 133 industry dummies. While the inteff commands runs when I include only 82 out 133 industry dummies, it fails to run when I include the 133 industry dummies. The error I get when I run the logit model with 133 industries dummies is r(503) comformability error. According to a post by Marteen Buis on probit & conformability error in inteff could occur when -logit- dropped a variable due to multicolinearity and -inteff- did not accurately pick that up. This is not the case for my model, because the logit model I am using does not drop any variable. My question is how I can make the command inteff run with the 133 industry dummies not just the 82 dummies. More specifically, what can be done to solve the conformability error when using 133 industry dummies if the problem is not related to observations being dropped by the logit and not picked by inteff?