A few iPython notebooks on interpretation of neural networks
Begin with part 1, which is the introductory narrative and overview, and then move on to part 2, which has code samples.
I recommend viewing these on my web site (GitHub doesn't render output scrolling in Jupyter notebooks), but be sure to download the original .ipynb files to run on your own machine.