RAT 5;
It would take the knowledge of an aerodynamicist who knows C* laws to discuss the "letting go" scenario.
My thoughts as a former transport pilot are that in a fully-developed stall, (35°+ AoA, vertical speed > 10,000fpm), just releasing the controls to neutral would not recover the aircraft, FBW or no.
Bear in mind that all aircraft would have different responses; whether such differences make a difference in handling during a full-stall is one of many unanswered questions.
A full nose-down stick command, wings level, thrust at idle is almost certainly a recover strategy, but no one's done it in flight tests and the sim data, post-stall, is extrapolation from best-knowledge as there is no basis in testing - yet.
The flight control computers for the Airbus attempt to maintain 1g, (Nz, in relation to the longitudinal axis in Normal and Alternate I Laws) and the trim horizontal stabilizer, (THS) autotrim function attempts to unload the elevator, but the follow-up is not instantaneous.
In AF447, the THS was driven full NU and remained there because of full-up stick inputs. If the stick had been placed full-forward, the THS would follow-up and return close to a neutral setting.
From discussions with others here and elsewhere, it is likely the elevator retained sufficient effectiveness to get the nose down, unload the wing and recover.