Most of the time you don't ever want to use a deep hibernation mode on a spacecraft. If at all possible, you want it active and chatty (with mission control) the entire time, since it means you can do science, monitor its systems, and keep more diagnostic and recovery options available.
The Rosetta probe used deep-sleep mode because for a big part of its journey it was out past Jupiter, and did not receive enough solar power to remain fully functional. It's rendevouz now means it's closer, and can wake up again.
But there is another more interesting reason that has to do with the machine lifetime of the onboard computers: space agencies have models of their computers which basically estimate the number of instruction-operations they can perform before they're likely to experience a hard failure due to wear or uncorrectable errors from radiation bit-flipping.
In the case of very long range spaceprobes, the probability of a failure starts to climb to 1 - that is, leave the ship fully operational and something will go wrong with the onboard computer (which you may or may not be able to recover from) before you reach your target. So deep-sleeping the spacecraft is a way to conserve the computers lifetime - you switch over to very robust real-time clock circuits, and power down the computer entirely.