What is a LLM ? 🤖

Cliquez ici pour lire en français

Last week, when talking about the Rabbit r1, we mentioned LLMs. You’re going to be seeing more and more of this term as AIs become more widely available. But what are LLMs? We answer it in this week’s #AskGriots.

Artificial intelligence model 🤖

LLM, the acronym for « Large Language Model », refers to a large-scale language model. Language models are types of artificial intelligence models designed to understand and generate text in natural language.

The larger the language model, the greater its ability to understand and generate text contextually. This is the case of the LLM used for ChatGPT, called GPT (Generative Pre-trained Transformer), as well as Google’s Gemini.

Based on a wide range of data 🗂️

LLMs are fed with huge linguistic datasets, often consisting of fragments of the Internet, books, articles and other texts. The larger the data set, the better the model can learn the structure and context of natural language.

However, it’s important to note that the use of LLMs also raises ethical issues, particularly with regard to data confidentiality, bias detection and responsibility in the use of these models.