Can I get a model that doesn't have all the horseshit in it that a full foundation model does? That is, is there anyway to choose what goes into my LLM's corpus?

An extreme example: Jacques Derrida. I lived most of a life w/o hearing of Jacques Derrida. And that was fine. I wish I had it back.

What do you think Richard Feynman thought of Jacques Derrida? I asked an LLM:

"Richard Feynman was famously dismissive of philosophy of science and postmodern critiques, calling such movements "baloney," while Jacques Derrida's deconstruction focused on language and meaning rather than scientific inquiry."

If we wish to recreate useful genius (super AI) it will be genius akin to human genius and it will be scientific genius and not baloney.

Nobody has ever read all the crap the AI world models have read, yet we have plenty of smart people and geniuses. Gimme a simpler less well-informed LLM that doesn't know about Derrida and other baloney. Get the crap out!