I think it would certainly depend on the type of "illegal activity"; it's quite plausible that those looking for drugs or counterfeiting are going to have much higher suspicion and opsec than porn-driven pedophiles, leading to the difference in number of registrations at each honeypot.

might be a typo, but just in case - a 'tenet' is a rule or principle, a 'tenant' is an occupier.

The exit node IP is not totally useless. It's enough to prompt a knock on your door by the authorities. Granted, once they recognize you're running an exit node there's nothing they can do (otherwise Charter could also be held liable) but still, it can prompt an action by the authorities.

My guess that they some were researchers: at least one came from a university IP address while others were very bare-bone machines with minimal running services, possibly virtual machines or test boxes.

The Tor project does a lot of work to make operating an exit node safer and less likely to prompt the authorities to intervene. This is good, because even if the authorities did want to talk to exit node providers, those people _couldn't_ provide any useful logs.