I am trying to figure out which language does the following NFA (taken from a Sipser's book "Introduction to theory of computation") recognise
From my understanding, this NFA accepts strings that contain at least two $1$s, but this is not precise, because, e.g., between the two ones we could have a $0$ or nothing ($\epsilon$). We can also have any number of $0$s or $1$s after these two $1$s.
My guess is that this NFA recognises a language whose strings contain the substrings $11$ or $101$, but I am not sure.
Can you help me defining the language that this NFA recognises?
In general, what is the best way to see what language is recognised by a NFA?