2 points
An LLM needs to evaluate and modify the preliminary output before actually sending it. In the context of a human mind that’s called thinking before opening your mouth.
4 points
1 point
Humans aren’t always very good at that, and LLMs were trained on stuff written by humans, so here we are.
2 points
