Choosing between different techniques like Invoke, Summarization, and Retrieval-Augmented Generation (RAG)

ChatGPT Use cases and examples
, ,
1

Hi everyone,

I have a question about choosing between different techniques like Invoke, Summarization, and Retrieval-Augmented Generation (RAG). For example, if I have a 200-page document and need to generate a specific table with defined columns, I find that using RAG provides very limited responses. Currently, I’m using the map/reduce method along with prompting, but I’m wondering if there’s a more effective approach?