GPT is biased towards the most popular theories in its training data multiplied by repeated exposure to those theories, not the most correct theories. There's n

GPT is biased towards the most popular theories in its training data multiplied by repeated exposure to those theories, not the most correct theories. There's no such thing as an unbiased LMM.