it is a statistical model. If everyone is saying X and is wrong, and one guy says Y and is right, the llm will bit out X. Because that is the most probable thin

it is a statistical model. If everyone is saying X and is wrong, and one guy says Y and is right, the llm will bit out X. Because that is the most probable thing in the dataset. It literally is a text averaging machine