My thesis is about Analysis and Auto generation of FAQ in different domains. For conducting experiments, I need high volume of FAQs. That's the reason I am looking for a publicly available data-set containing FAQs in various domain (or even one specific domain).
You can download FAQs from github, although they won't really be cross-domain...
Example of a simple query that returns some FAQs files (will require some fine-tuning, of course)
What about StackOverflow's own dataset? https://archive.org/details/stackexchange