[belter]

Not the same: "RAG vs. Long-context LLMs" - https://www.superannotate.com/blog/rag-vs-long-context-llms

[ffsm8]

You're literally linking to an article that confirmed what I said. Yes, a model that has RAG will be able to perform with a lot smaller context size.

That doesn't mean RAG isn't context optimization.

[belter]

Did you made any technical argument on how to reduce hallucinations? Because I fail to see one from you on this thread except: "it's the fault of your prompt""