How does ChatGPT work?

Since OpenAI hasn’t provided all the details, some parts of the diagram may be inaccurate.

We attempted to explain how it works in the diagram above. The process can be broken down into two parts.

Training

To train a ChatGPT model, there are two stages:

• Pre-training: In this stage, we train a GPT model (decoder-only transformer) on a large chunk of internet data. The objective is to train a model that can predict future words given a sentence in a way that is grammatically correct and semantically meaningful similar to the internet data. After the pre-training stage, the model can complete given sentences, but it is not capable of responding to questions.

• Fine-tuning: This stage is a 3-step process that turns the pre-trained model into a question-answering ChatGPT model:

  • Collect training data (questions and answers), and fine-tune the pre-trained model on this data. The model takes a question as input and learns to generate an answer similar to the training data.
  • Collect more data (question, several answers) and train a reward model to rank these answers from most relevant to least relevant.
  • Use reinforcement learning (PPO optimization) to fine-tune the model so the model’s answers are more accurate.

Answer a prompt

• Step 1: The user enters the full question, “Explain how a classification algorithm works”.
• Step 2: The question is sent to a content moderation component. This component ensures that the question does not violate safety guidelines and filters inappropriate questions.
• Steps 3-4: If the input passes content moderation, it is sent to the chatGPT model. If the input doesn’t pass content moderation, it goes straight to template response generation.
• Step 5-6: Once the model generates the response, it is sent to a content moderation component again. This ensures the generated response is safe, harmless, unbiased, etc.
• Step 7: If the input passes content moderation, it is shown to the user. If the input doesn’t pass content moderation, it goes to template response generation and shows a template answer to the user.