In today’s data-driven world, the ability to efficiently analyze documents with the power of Artificial Intelligence (AI) is becoming increasingly crucial. Navigator Chat provides a secure and private platform to leverage various language models for document analysis. This guide will walk you through getting started with NaviGator Chat, ensuring your data remains protected while harnessing the capabilities of advanced AI.
NaviGator Chat prioritizes security and privacy, especially when dealing with sensitive information. The platform is designed to keep your documents secure within University of Florida (UF) servers and trusted vendor environments. For documents containing restricted or sensitive data, NaviGator Chat utilizes models deployed on UF HiPerGator, ensuring an extra layer of security. It is important to note that while using cloud-based models, handling restricted or sensitive data is not permitted. When you interact with language models hosted by external vendors, your messages and portions of your documents are processed by Large Language Model (LLM) instances provided by Microsoft or Google. Rest assured, all data processed in this manner is protected by existing agreements with these vendors. Importantly, none of your data is used to train the general large language models, and all data is completely removed after deletion, ensuring your privacy and data control. If you accidentally upload a file, simply deleting the file or the entire conversation will permanently remove it from the system.
Exploring Available Language Models
NaviGator Chat offers a diverse selection of language models to suit various analytical needs, hosted across different infrastructures:
| Infrastructure | Language Models |
|---|---|
| UF HiPerGator | – llama-3.3-70b-instruct – llama-3.1-8b-instruct – nim-llama-3.1-8b-instruct – llama-3.1-70b-instruct – mistral-small-3.1 – mistral-7b-instruct – nim-mistral-7b-instruct – mixtral-8x7b-instruct – codestral-22b – granite-3.1-8b – Stable Diffusion (image generator) |
| Microsoft Azure | – o3-mini – o1-mini – gpt-4o – gpt-4o-mini – gpt-4-turbo (will be removed on 4/2/2025) – gpt-3.5-turbo – DALL-E-3 (image generator) |
