Exploring the Potential of LLMs and Attributed Prompt Engineering for Efficient Text Generation and Labeling

In the realm of Generative AI and natural language processing (NLP), the demand for labeled textual data has increased due to the need for vast amounts of data by machine learning (ML) models. Additionally, obtaining and collecting such data has become increasingly challenging due to data scarcity, such as limited training data, or privacy concerns, such as those involving medical or personal data. Consequently, there is a compelling need for advanced systems capable of generating and labeling high-quality textual data to bridge this gap in various NLP applications, including machine translation, and content generation.In our work, we aim to develop a web-based platform to automate textual data generation and labeling by leveraging large language models (LLMs), specifically Llama 3. The proposed system is designed to automate, accelerate, and streamline data generation and labeling processes accurately and efficiently for users, minimizing dependence on expert annotators. To achieve this, the system will integrate prompt engineering techniques, including zero-shot learning, few-shot learning, role-playing, and chain-of-thought.