“According to a Google research paper, large language models remember information better if it is repeated twice within the context.”