I performed a NK DSGE model for a small-open economy. I wrote the equations in log-linear form, not dev from ss, in order to find a steady state.
Inflations and terms of trade weren’t diff(1) because they were stationary; gdps, the exchange rate and interest rates were log diff
In the estimation, “steady” shows negative values for most of my observable variables.

attached the mod and data files -
Please help. NK.rar (4.75 KB)

You are confusing something here. You either log-linearize your model by hand and then do not use variable substitution by putting everything in exp() OR you let Dynare do the log-linearization by entering the non-linear equations and putting everything in exp().

In your case, you log-linearized the model so that all variables have steady state 0. But then you also put them in exp(), effectively defining them as the log of a logged variable. As a consequence, the new variables have steady state log(0)=Inf. This results in the problems.