Build Useful Guides
# đź“–tutorials
q
Build Useful Guides Chatbot workflow 1. Receive a research topic from the chatbot user, then look up the latest information online. 2. Ask GPT-4 to create a guide with the newest data and its current knowledge. 3. Send the guide to other LLMs (Claude-3, Mixtral-8x7b, and Llama2-70b) for feedback. 4. Improve the guide based on feedback and focusing on key points for the client's needs. 5. Display the final guide to the chatbot user. Show the project to the client and repeat the process from steps 1-5 until the deal is secured. I'll add all the code, prompts, etc., and upload the bot file after I have tested enough. I'm going to build it so that when other bot builders are testing or start using it, just replace your API keys and you're good to go. Show your client 10-20 example areas where they might want to focus on improving the chatbot's responses, for example: Freshness (Up-to-Date Information) Helpfulness (Guidance and Clarity) Factuality (Accuracy and Verification) Holistic (User-Friendly Communication) Sounds Natural (Likelihood of Being a Human Answer) Relevance (Pertinence and Context) Depth (Comprehensiveness and Detail) Simplicity (Ease of Understanding) Creativity (Innovative and Insightful Responses) Consistency (Logical and Coherent Responses) Send the created guide to ChatGPT and ask it to rate it (0-100%) based on those criteria. In my testing, normally guides made by GPT-3.5 get average ratings of 89%. If a multi-LLM chatbot can get average ratings of 95%, we all have clients.