3 ms·
An experiment comparing information retrieval performance between Open AI's Assistants API's RAG, GPT-4 Turbo (with context window stuffing) and Llama Index wit
by ulidabess 3y ago
An experiment comparing information retrieval performance between Open AI's Assistants API's RAG, GPT-4 Turbo (with context window stuffing) and Llama Index with GPT4.
Pretty striking results, especially when it comes to how the Assistants API beats Llama Index and how useless context-window stuffing is.