It seems like the pre-master secret generated during the SRP protocol would make a good source to generate a shared private key using a secure hash to compress it down into a 128/256 symmetric key. The random values that get hashed into it seem to make it good for that purpose.
From the RFC:
SRP also supplies a shared secret at the end of the authentication sequence that can be used to generate encryption keys.
It seems from my quick look over the RFC that that shared secret is the premaster secret, so you are correct.