Prompt Generation with LLM to improve Proactiveness

This is the illustration of workflow of our approach, detailing how we generate prompts to suggest them to user

In this sample case, the limitations of LLMs, such as GPT-3.5, become apparent as they produce generalized results

This is an example output where the model takes input and generates multiple prompts

  • Project: B.Tech Thesis Project
  • Supervisior: Prof. Pabitra Mitra (CSE, IIT Kharagpur)
  • Timeline: Aug 2023 - Nov 2023
  • Project URL: Prompt-Generation-with-LLM
  • Skills: NLP, LLM, Python

Details

This project seeks to overcome the limitations of existing LLMs that offer generic outcomes. It also aims to improve human involvement when the generated output is insufficient and to better discern when to personalize responses, thereby ensuring more proactive prompt generation for tailored and contextually relevant interactions with users.