It's functionally very similar to when, e.g., a child keeps mimicking a character from their favorite cartoon. Pretty much all of the "quirks" that LLMs exhibi

It's functionally very similar to when, e.g., a child keeps mimicking a character from their favorite cartoon. Pretty much all of the "quirks" that LLMs exhibit are similar to "human" quirks. Which is great if we're trying to create an AI that makes mistakes like the average human.