Skip to main content
Screen Shot2025 10 09at2 03 27PM Pn

​
Exploring Your Data with the Graph View

The Graph view helps you visualize how different pieces of information in your dataset connect to each other. Think of it as a map showing relationships between concepts.

​
What You’ll See in the Graph

  • Blue dots (nodes): Each dot represents a piece of information or concept. Each node can represent a document, concept, extracted entity, or data point.
  • Connecting lines (edges): Lines connecting dots show how concepts relate to each other. Edges indicate semantic similarity, direct references, shared attributes, or contextual connections.
  • Clusters: A cluster is a group of closely connected nodes that represent related topics.
  • Peripheral nodes: Preripheral nodes are ideas that sit on the edges of the network with fewer relationships.

​
Interpreting Your Graph

The information bar displays key metrics:
  • Nodes: Total number of entities (e.g., “Nodes: 16”)
  • Edges: Total number of connections (e.g., “Edges: 61”)
  • Matches: Number of nodes that meet the current filter or selection criteria (e.g., “Matches: 34”)
  • Zoom: Current zoom level (e.g., “86%“)

​
Graph View Controls

  • Pan tool
  • Zoom in/out controls
  • Fit to screen
  • Reset view
  • Pause/play simulation
  • Globe View toggle (2D/3D)
Search functionality helps you quickly locate specific nodes within your graph:
  • Search by node label: Find nodes by their display name (e.g., “CustomerFeedback”)
  • Search by content keywords: Locate nodes containing specific keywords within their content
  • Search by unique node ID: Navigate directly to a node using its identifier (e.g., “65b1143b-8…”)
  • Matching results highlight: Search results are highlighted in the graph for easy identification
  • Navigate to results: Press Enter or click a search result to navigate directly to that node

​
Node Search Filters

Options Filter
  • Quick Jump toggle: When enabled, automatically focuses on the first matching node in real-time as you type your search query
Groups Filter Filter search results by node groups to narrow down your search scope.

​
Mini-Map

Located in the bottom right corner, the mini-map provides an overview of your entire graph structure. A blue square indicates your current viewport position within the larger graph. Click anywhere on the mini-map to quickly navigate to that area. The mini-map is collapsible - use the close button (×) to dismiss it when not needed. This feature is essential for navigating large, complex graphs efficiently.

​
Model Information Panel

Located on the right side of the Control Center, the Model Information Panel displays comprehensive metadata and status for your selected model.

​
Status

  • ● READY: Model is operational and ready for inference
  • Other states: ● BUILDING GRAPH

​
Base Model

Displays the underlying model architecture (e.g., “Nemotron”), indicating the model family and version your custom model is built upon.

​
Created

Timestamp of model creation in Month Day, Year format (e.g., “November 17, 2025”).

​
Last Trained

Timestamp of the most recent dataset update. This updates when you create a dataset (during model creation or directly), add files to a dataset, run inference, or when ML training completes (e.g., “November 18, 2025”).

​
Dataset Size

Total size of training data measured in KB, MB, or GB (e.g., “169.30 KB”). Useful for understanding model scope and the volume of information the model has been trained on.

​
Features Section

​
Persona

Dropdown selector for model persona and behavior configuration. The selected persona affects how your model responds and behaves during interactions.

​
Shortcuts & Actions

Quick access buttons for common 4Minds functionality.

​
Inference Model

Purpose: Run model inference on new data. Usage:
  1. Click “Inference Model” button
  2. Input your query or data
  3. Model processes using current knowledge graph
  4. View results in context
When to use:
  • Testing model responses
  • Validating training results
  • Real-time predictions

​
Evaluate Model

Purpose: Assess model performance and accuracy. When to use:
  • After training updates
  • Before deployment
  • Comparing model versions
  • Quality assurance

​
Add Training Data

Purpose: Expand model knowledge with new data. Usage:
  1. Click “Add Training Data”
  2. Upload files or enter data:
    • Documents (PDF, TXT, DOCX)
    • Structured data (CSV, JSON)
    • FAQ entries
  3. Data is processed and added to graph
  4. Nodes and edges update automatically
When to use:
  • Adding new FAQs or documentation
  • Expanding knowledge coverage
  • Updating existing information
  • Incorporating user feedback

​
New Persona

Purpose: Create custom model behaviors.

​
API Settings

Purpose: Configure API access and view API referenece. When to use:
  • Integrating with applications
  • Managing API access
  • Updating credentials
  • Performance tuning

​
Model Card

Purpose: View detailed model information. Contents: Screen Shot2025 11 18at6 32 47PM Pn When to use:
  • Understanding model capabilities
  • Compliance documentation
  • Sharing with stakeholders
  • Audit requirements