Sure, it's very possible, and I'm sure has happened (an autopilot-instigated crash that happened >5sec after disengagement), but I do think 5 seconds is a reasonable threshold for this data, at the very least.
Another example scenario: FSD-caused spin out then hit by a car close to 5 seconds later. I've been close to this one before with a tire blow out but wasn't hit, but maybe you are right that these would be in the tail of the data.