User Guide: Configuring LLM Library in SmartSpace.ai

SmartSpace.ai offers the flexibility to configure and maintain Large Language Models (LLMs) within your workspaces. This guide will walk you through the process of adding and managing LLMs from the Admin Portal.

Default LLM Configuration

Upon installation of SmartSpace.ai into your environment, the default LLM configuration will be set to use “Azure OpenAI” resources provided during installation. This includes utilizing the GPT-3.5 model for language processing tasks.

Adding New LLMs

Step 1: Access Admin Portal

• Log in to the Admin Portal of SmartSpace.ai.

Step 2: Navigate to LLM Configuration

• On the right-hand side of the Admin Portal, locate the LLM navigation panel.
• Click on the “Add new model” button to begin adding a new LLM.

Step 3: Select Provider

• Choose the provider for the new LLM from the available options:
  1. Anthropic
  2. OpenAI
  3. Azure OpenAI
  4. Hugging Face
  5. Google Gemini

Step 4: Provide Required Details

• Depending on the selected provider, provide the necessary API keys and integration details for configuration.
• Important Note: UsingAnthropic,OpenAI, and Gemini models will involve data being sent outside your environment for analysis. Only Hugging Face and Azure OpenAI models remain within your system for closed-loop analysis.

Configure Hugging Face Models

• If selecting Hugging Face as the provider, you can choose from a pre-defined list of tested models provided by SmartSpace.ai.
• Alternatively, if you wish to use a different Hugging Face model, SmartSpace.ai can test and add it to the pre-defined list upon request.
• SmartSpace.ai will allow for any Hugging Face model to be selected in the near future.

Step 5: Activation in Workspace

• After configuring the LLM, it becomes available for selection within your workspaces.
• SmartSpace.ai does not deploy the models except for Hugging Face. Instead, it adds them to the list for selection in workspaces, making them ready for use once added.