OpenAI has finally introduced fine-tuning for its popular GPT-3.5 Turbo model. The announcement is aimed at companies and developers to create monitored products that excel at specific tasks. According to OpenAI, a fine-tuned version of the GPT-3.5 Turbo model can match and even outperform advanced models such as GPT-4. So to learn more about custom tweaking on GPT-3.5 Turbo, follow along.

Fine-tuning allows users to customize the model and create something tailored to their needs. We already know that GPT-3.5 Turbo is very fast at drawing conclusions so in terms of performance it was already much better than the competition. Now, with fine-tuning support, developers can leverage this capability to create new experiences, be it in the form of AI chatbots, documentation engines, AI assistants, coding assistants, etc. You can basically create a custom application for your business in your desired tone.

Remember that OpenAI keeps GPT-3 series models including babbage-002 and davinci-002 for fine-tuning purposes. However, the API endpoint has changed and you can learn more about the updated endpoint here. In addition to that, the key fine-tuning feature of GPT-3.5 Turbo is that all training data is reviewed using OpenAI's Moderation API and GPT-4 powered moderation system. The company says it's done to "detect unsafe training data that violates our security standards."

But OpenAI has also made it clear that none of the private training material will be used to train OpenAI models. Apart from that, GPT-3.5 Turbo can currently only handle 4k tokens at once. The company says support for 16k context length and function calls is coming later this fall.