Eugene Myrmin/Getty Images
OpenAI’s advanced big language model has been leveraged by enterprises and developers for their specific use cases. Now, the update to GPT-3.5 Turbo will boost its functionality for its customers.
On Tuesday, OpenAI announced that its most cost-effective model in the GPT-3.5 family, the GPT-3.5 Turbo, will be available for fine-tuning. This means that developers can now use their own data to optimize the model for their use cases.
Too: How to provide sources and citations to ChatGPT
“Since the release of GPT-3.5 Turbo, developers and businesses have sought the ability to customize the model to create unique and differentiated experiences for their users,” OpenAI said in the post.
In the private beta, OpenAI found that customers were able to improve model performance across a variety of use cases. These include improved operability, which allows models to better follow instructions, reliable output formatting, and custom tones, which allow businesses to incorporate their own brand voice within models.
OpenAI also claims that the fine-tuning allows businesses to make their prompts smaller, with early testers reducing the size of prompts by up to 90%. According to the company, this reduction cuts costs and increases the speed of each API call.
Too: You can demo Meta’s AI-powered multilingual speech and text translator. This way
Most impressively, OpenAI shared that early tests have shown that a refined version of GPT-3.5 Turbo can “match, or even outperform” GPT-4-level capabilities on “some narrow tasks”. Can do.
To address privacy concerns associated with using AI models for enterprise use cases, OpenAI assures users that customer data used to fine-tune the models remains owned by the customer and is managed by OpenAI. Not used to train other models, as with any other API model. ,