Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

No. RAG is about finding relevant documents/paragraphs (via KNN lookups of their embeddings) and then inserting those documents/paragraphs into the input context, as sequences of input tokens. What I'm talking about is different: https://arxiv.org/abs/2203.08913


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: