You probably know this already, but if you are just looking for an illustrated demonstration of what RNNs are capable of in the text domain, probably the best b

You probably know this already, but if you are just looking for an illustrated demonstration of what RNNs are capable of in the text domain, probably the best brief article is the post by Andrej Karpathy about [The Unreasonable Effectiveness of Recurrent Neural Networks](http://karpathy.github.io/2015/05/21/rnn-effectiveness/ http://karpathy.github.io/2015/05/21/rnn-effectiveness/).