What is an LLM?

LLM - Large Language Model

ChatGPT is built on a Large Language Model or LLM as we will refer to it from now on. It is a type of artificial intelligence that specializes in understanding and generating human language.

Simply explained, the model is good at guessing what the next word should be.

How does an LLM work?

Here's a simplified explanation of how it works: LLMs are trained by reading an enormous amount of text (books, articles, websites, etc.). During this process, it learns language patterns, word usage, grammar, and even style and contrasts.

When ChatGPT receives a question or text to process, it analyzes not just the individual words, but also the context. This means it looks at the words' relationship to each other to understand the meaning in a sentence or text.

Based on the learned information and the specific question it received, the LLM generates a response. It does this by predicting which word or phrase is most likely to follow after the given words, based on its training and the prompt you entered.

In other words, you could say that ChatGPT is a very advanced version of the auto-suggestions on your mobile keyboard. But it lacks consciousness and real understanding, and its 'knowledge' is limited to the patterns and information it has been trained on.

Although LLMs don't learn in real-time from new data after their initial training, they can be updated with new data and training to improve and update their knowledge and capabilities.

ChatGPT's Limitations

Currently, ChatGPT4 and ChatGPT4o have been trained on data up until April 2023.

If you prompt ChatGPT or it itself determines that it needs newer information, it will go out and search the web. This usually takes longer. For current info, we recommend other AI tools such as Perplexity.