LDA topics for Clustering

[TRIXIA MAY BELGA]

My goal is to cluster the resulting LDA topics to reduce dimensionality. However I am not sure what data should I extract from LDA to do this and the specific code for it. Does anyone know this? 

I've read some articles and they are mentioning the LDA document-topic posterior probability distribution matrix. How can I extract this from LDA?

Many Thanks,

TB