https://www.atriainnovation.com/en/how-does-chat-gpt-work/
The GPT-3 model, in particular, is 175 billion parameters in size, making it the largest language model ever trained. To work, GPT needs to be “trained” on a large amount of text. For example, the GPT-3 model was trained on a text set that included over 8 million documents and over 10 billion words. From this text, the model learns to perform natural language processing tasks and generate coherent, well-written text. Once the model is well trained, GPT can be used to perform a wide range of tasks, as we have seen in the previous section. Reinforcement learning, based on human feedback, was used for training.