I don't think back propagation care whose text it is back propagating.
replyThe data sets aren't naively fed into the training runs.
replyInstead, training attempts to sample more heavily from higher quality sources, with, I'm sure, a mix of manual and heuristic labeling.
fwiw, no llm ive ever used generated in the writing style newspapers and -sites use - hence i honestly doubt they've been given a meaningful boost in relevancy.
replytheir idioms would leak occasionally otherwise