Generative AI is reshaping today’s workforce with Large Language Models (LLMs) bringing a structural shift in how modern day offices function. From drafting that perfect email to creating high end graphics, LLMs ensure customized output from massive data sets.

According to NVIDIA, “Large language models (LLMs) are deep learning algorithms that can recognize, summarize, translate, predict, and generate content using very large datasets.”

Simply put, LLM is an AI model that has the ability to generate and understand natural language. LLMs are trained on vast datasets and can be fine tuned by tweaking what are known as parameters.The more parameters a model has, the more complex it becomes. GPT-3  for example was pre-trained on 45 terabytes of data and it uses 175 billion parameters.

There are three pillars to an LLM. Data , architecture and the training of the LLM .

Broadly speaking LLMs are built using deep learning, but their efficiency is largely due to the transformer architecture. The transformer architecture was first introduced in 2017 in a defining paper titled “Attention is all you need”.