Here, we present a demo of topic model, Online Latent Dirichlet Allocation (Online LDA) . A sample from the book "Pride and Prejudice":
The model is trained with Wikipedia articles in 2010. Vocabulary size is 10,000 (listed in here), and the number of latent topics is 100. You can estimate latent topics assigned to the words in the vocabulary via the online topic model.
 Matthew D. Hoffman, David M. Blei, and Francis Bach, "Online Learning for Latent Dirichlet Allocation", in Neural Information Processing Systems (NIPS), 2010.