use the same caffe-model for modified prototxt

[Tamar Elazari Volcani]

 hello!

I have a trained model, and I need to display out put from a middle layer. If I rewrite the prototxt such that the "top" of the layer of interest is "net_output"? will all previous layers use the weights they were supposed to? 

(I know this can be done otherwise, but I need to know if this way will work...) 

thanks!