
Unlocking Creativity with Azure OpenAI
By :

To make effective use of the model, it’s important to understand some basic elements associated with chat models. These concepts are essential for utilizing the model properly.
The GPT-3.5 Turbo and GPT-4 models are LLMs that have been optimized for conversational interfaces. They differ from the older GPT-3 models in how they operate. GPT-3 models are text-in, text-out, meaning they accept an input (prompt) string and return an output (completion); the GPT-3.5 Turbo and GPT-4 models are designed for conversation-style interactions. These models expect input to be formatted in a chat-like transcript structure and return a completion that represents a message generated by the model in the chat. While this format was primarily designed for multi-turn conversations, it can also be effectively used in non-chat scenarios.
In AOAI, there are two distinct options for interacting with models such as GPT-3.5 Turbo and GPT-4: