Rather than A/B test the conversion rates of liking before-or-after, think really hard about what you want to convey. What does a like give you? Is it worth alienating potential users?

You don't need an A/B test to tell you that "If you enjoyed our service, please Like us on Facebook." is a friendlier, more positive request. People have already tweeted about it more than they liked it on Facebook without additional prodding. Let the site stand on its own merits, rather than trying to force promotion through potentially sleazy methods.