OpenAI’s language models are now being exploited by companies and developers for their own specific use cases. And now an update to GPT-3.5 Turbo strengthens its features.
On Tuesday, OpenAI announced that its most profitable model of GPT-3.5, GPT-3.5 Turbo, can benefit from fine tuning, also known as fine tuning. This means that developers can now use their own data to adapt the model to their use cases.
“Since the release of GPT-3.5 Turbo, developers and companies have asked to be able to customize the model in order to create unique and differentiated experiences for their users,” explains OpenAI in its post.
Shorten the prompt
With a private beta, OpenAI found that customers were able to improve the performance of the model in a number of cases. These include better dirigibility, which allows the model to better follow instructions, reliable output formatting and a personalized tone.
OpenAI also claims that fine-tuning allows companies to shorten their prompt messages, as the first testers reduced the size of messages by up to 90%. According to the company, this reduction reduces costs and speeds up each API call.
Even more impressively, OpenAI said that initial tests have shown that the refined version of GPT-3.5 Turbo can “match or even surpass” GPT-4 level capabilities on “certain tasks”.
The data remain the property of the customers
To address privacy concerns related to the exploitation of an AI model for enterprise use cases, OpenAI assures users that customer data used to refine the model remains the property of the customer and is not used by OpenAI to train other models.
Source: “ZDNet.com “