LLM Factory

Vast possibilities to advance research via use of LLMs are restricted due to availability of resources and data security. LLMs require significant investments in compute infrastructure, as well as specialized technical knowledge. With LLM Factory, users can upload their own datasets, train adapters by layering data on a base model, privately configure and explore models, and expose models to external, OpenAI compatible API. LLM Factory offers researchers a locally hosted, secure, and free alternative to public, industry standard models.

Key features of LLM Factory:

• Offers users control, flexibility, and security without requiring a programming or technical background
• Researchers can train models with the assurance data is secure
• Open AI industry-standard API endpoints offers interoperability with other libraries, tools, and systems
• CAAI staff use LLM Factory to support all LLM collaborations, and offer support to users working on the platform independently

A few of the many experimental projects supported by LLM Factory include:

• One Good Choice: A tool that encourages healthier decisions, developed with University Health Services and the College of Health Sciences.
• Synthetic Personas: Conversational agents that make remote training interactive, in collaboration with the Center for Population Health and Creative Link for Digital Health, as well as an application for the Office of Medical Education.
• CELT Look-Up Tool: Enables conversational navigation of CELT’s resources and website.