OpenAI has given examples of its work with SK Telecom and Harvey, a company building AI-powered legal tools. SK Telecom worked with OpenAI to fine-tune GPT-4 for telecom-related conversations in Korean, while Harvey created a custom model for case law using hundreds of millions of words of legal text and feedback from expert attorneys. OpenAI also announced new model fine-tuning features for developers working with GPT-3.5, including a new dashboard for comparing model quality and performance, support for integrations with third-party platforms, and enhancements to tooling.
Key takeaways:
- OpenAI is expanding its Custom Model program, which helps enterprise customers develop tailored generative AI models for specific use cases. The program includes assisted fine-tuning and custom-trained models.
- Assisted fine-tuning, a new component of the Custom Model program, uses techniques beyond fine-tuning to help organizations set up data training pipelines and other supporting infrastructure to improve model performance on specific tasks.
- Custom-trained models are built with OpenAI's base models and tools for customers that need to more deeply fine-tune their models or imbue new, domain-specific knowledge.
- Alongside the expanded Custom Model program, OpenAI announced new model fine-tuning features for developers working with GPT-3.5, including a new dashboard for comparing model quality and performance, support for integrations with third-party platforms, and enhancements to tooling.