The cables for the IceCube strings are over a mile and a half long. The 60 sensors (digital optical modules, or DOMs) on each string are spread across a kilometer of cable. The detector has to be spread out because we want to capture as much information about as many neutrinos as we possibly can.
While IceCube was putting strings in the ice, there was a unique opportunity for people to install additional things in the spaces between DOMs. I don't know exact numbers, but it seems like almost a third of the strings had some sort of special device installed, some more successful than others.
One that's still in use is the Swedish camera. This is actually a pair of cameras on the same section of cable between two DOMs, the lower one pointing up and the other pointing down. These cameras were used to visually inspect the freeze-in process and are still turned on every year or so to chart how the ice changes as it solidifies.
The process to use these cameras is somewhat involved. Because it's a visual inspection, we need to turn on the cameras' lights. However, the DOMs are very sensitive so we turn off all but the most remote DOMs or they might be overloaded and stop working!
Because there is light in the detector, the data taken while the Swedish camera is running isn't useful for most normal science, but we keep that outer ring of DOMs going just in case there's a supernova somewhere "near by". The last close supernova happened in 1987. Twenty four neutrinos emitted by that event were recorded and physicists were able to massively increase their knowledge of neutrinos and the universe from those few particles. We're overdue for a supernova in our galaxy, so we need to keep the detector taking data as continuously as possible in order to catch that supernova's particles when they hit earth.
If you're interested, you can read the weekly report from three years ago when the Swedish camera was deployed, and see a better image from the Swedish camera.