Profiler using original data not a fit

[#1]

Hi everyone, a JMP noob here... 

I ran a full factorial set of cases and I am trying to see it in a profiler. When I plot the response to one factor it is clearly quadratic. The fits I can get when making the profiler show the curve as linear. If I try to use a neural network, results are inconsistent at best. 

How can I use the profiler to just show the initial data set with lines in between the data points?

Thank you!

[#2]

I'm not sure where to start...could you share your JMP table?  I don't know what the design structure looks like?  A full factorial on how many factors and how many levels?

If you only run the factors at 2 levels, your prediction profiler  (which is both a main effects plot AND the model from the data incorporated into it) will only be able to  estimate linear relationships.