⚙️Model

Tweak AI models to gain insights that are specific to your needs.

The Model editor is where you can configure your own LLM to best suit your task. It includes the following workflow:

1. Basic Context

Set a name and description to easily find your model as well as give it basic initial context about itself.

2. Custom Tags

Create custom instructional tags for your model that can be easily accessed in the chat interface.

3. Model Selection

Choose an already fine-tuned performant model such as GPT-4 to couple with the indexer, or pick a fine-tuned model trained on your own data. See a list of available LLMs here.

4. Indexer

Create a document index in order for your model to respond to your queries effectively and accurately regarding your own personal documents and unstructured data.

1. Select ChatGPT

2. Select Model - GPT 3.5 or GPT 4.0

3. Upload any text documents

4. Press the Create button

5. This Model can now be used in a new Chat with data awareness of the linked documents.

The indexer uses a complex process involving querying data from vector databases and knowledge graphs. To see more information head over to Querying Documents

You can use either the Indexer or the fine-tuning tools to create a trained AI assistant at your disposal, ensuring you get the best results possible based on your own provided data. Make sure to see the list of available base models.

After you've configured all the settings, you can press create to deploy the model and have access to it via the chat interface. You can edit these settings and update your model at any time.