Context will tell you that we’re talking about losing access to water. If people won’t procreate because they don’t have water, it makes no difference if it’s because we are literally all out of water or if someone else is hoarding it all. In both cases, there’s no access to water.
Burgers should neither be taller nor wider. Just give me two normal sized burgers.
Same. And in the past decade plus of using smartphones, I’ve never had to open the slot except to install the SIM card for the first time. Why do I even keep it around?
But then I have to work an extra 10 hours of my minimum wage job just to eat for 10 minutes. The ratio is even worse!
It has nothing to do with the meaning. If your training set consists of a bunch of strings consisting of A’s and B’s together and another subset consisting of C’s and D’s together (i.e. []+ and []+ in regex) and the LLM outputs “ABBABBBDA”, then that’s statistically unlikely because D’s don’t appear with A’s and B’s. I have no idea what the meaning of these sequences are, nor do I need to know to see that it’s statistically unlikely.
In the context of language and LLMs, “statistically likely” roughly means that some human somewhere out there is more likely to have written this than the alternatives because that’s where the training data comes from. The LLM doesn’t need to understand the meaning. It just needs to be able to compute probabilities, and the probability of this excerpt should be low because the probability that a human would’ve written this is low.
A sentence saying she had her ovaries removed and that she is fertile don’t statistically belong together, so you’re not even getting that.
This is a great argument for the possibility of changing some minds. You’ve definitely changed mine. My policy is unchanged though; I’m not going to set out with the goal of changing minds in online discussions.