Skip to content

Latest commit

 

History

History
29 lines (17 loc) · 1.66 KB

Assignment_2.md

File metadata and controls

29 lines (17 loc) · 1.66 KB
Raw

Assignment 2 - Preface

In today's world, having our health assistant chatbot that can engage in meaningful back-and-forth conversations is incredibly valuable. The best way to achieve this is by leveraging a Large Language Model (LLM).

Here's a demonstration of how an LLM can be utilized to build a simple chatbot that responds to a user's query:

  1. User Sends a Query: The interaction starts with the user typing a question or statement related to their health.

  2. LLM Processes the Query: The chatbot, powered by an LLM, receives the query and processes it to understand the user's intent and context. This involves natural language understanding capabilities inherent in the LLM.

  3. Generate and Send Response: The LLM generates a response that is coherent and relevant to the user's query. This response is then sent back to the user, continuing the conversation.

To realize this in JacLang, we have the by llm feature. Get more details on this feature by visiting JacLang documentation.

This is a basic chatbot implementation using JacLang to respond for a single query.

Assignment 2 - Task

Build a chatbot that can respond to multiple chained queries on the command line interface.

Should Include

  1. Chat History Management.

  2. Using the by llm() feature in jac-lang.

  3. Able to maintain a continuing chat.

  4. Chat-bot should be implemented for todo and user_QA nodes. (not for RAG)

    • In todo: should use both user data as well as the todo list to answer queries.
    • In user_QA: should use only user data to answer queries.