GPT 3.5 dosent perform as well on Question Answering with Embeddings as well as Davinci, any opnions?

sadaf.shah · April 27, 2023, 10:40am

The post heading says it all
I tried official tutorial on QA with embeddings, then I tried using GPT3.5 and it was very bad

github.com

openai/openai-cookbook/blob/main/examples/Question_answering_using_embeddings.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "3b0435cb",
   "metadata": {},
   "source": [
    "# Question answering using embeddings-based search\n",
    "\n",
    "GPT excels at answering questions, but only on topics it remembers from its training data.\n",
    "\n",
    "What should you do if you want GPT to answer questions about unfamiliar topics? E.g.,\n",
    "- Recent events after Sep 2021\n",
    "- Your non-public documents\n",
    "- Information from past conversations\n",
    "- etc.\n",
    "\n",
    "This notebook demonstrates a two-step Search-Ask method for enabling GPT to answer questions using a library of reference text.\n",
    "\n",

This file has been truncated. show original

Topic		Replies	Views
Using Embeddings for search poor results vs GPT3 API	1	752	December 17, 2023
Q&A system (embeddings) based on Curie Documentation	4	1249	February 22, 2023
OpenAI Prototype for Website Q&A with Embeddings does not work Community chatgpt , api	6	878	August 28, 2023
Training model with Fine-Tune and GPT3.5 Community api	2	567	December 24, 2023
Using a specific knowledge base with GPT API	12	19223	December 13, 2023

GPT 3.5 dosent perform as well on Question Answering with Embeddings as well as Davinci, any opnions?

Related topics