LLM(and LLM+) afaik, even Perplexity, do not do textual analysis. They do not “know” what is in the book, they will only try to predict how someone else would a

LLM(and LLM+) afaik, even Perplexity, do not do textual analysis. They do not “know” what is in the book, they will only try to predict how someone else would answer your question based on data in the training set. If there are not many similar questions asked and answered accurately, the results will be poor.