I think that it’s important to introduce some AI concepts and terms so discussions of AI can be more informed. Using the correct terminology and applying the correct concepts when discussing AI will strengthen your argument and reflect well on how you present your information. There seems to be a lot of confusion about neural networks and AI. People often use the two interchangeably. Let’s start by introducing one AI concept – that of a pretrained model.
So when you type in one or a series of questions into ChatGPT, is it running a neural network? The answer is no. ChatGPT (at least at the time of this article) is an example of a pretrained model. A neural network was involved in training the model, but no neural networks are running as you type in your questions.
Neural networks involve an input layer, one or more hidden layers and an output layer. Neural networks facilitate the training process. The final result is a trained model. So neural networks were highly involved in the creation of that ChatGPT model, but what you are interacting with as you type in questions into the prompt is the model, not the neural network.
image is creative commons public domain. source: https://picryl.com/media/network-equipment-hardware-computer-communication-f3e649
targetedtechtalk@protonmail.com
Follow @TargetedTechTlk
