Evaluating Agents

The following examples demonstrate tasks you can carry out with Vijil Evaluate.

If you’re just starting out, you’ll need to create an agent before running evaluations.

• Add, Update, and Delete Agents
  • Using the Vijil Web Interface
  • Using the Python Client
• Managing API keys
  • Create, Delete, or Update API Keys in Vijil Web Interface
  • Create, Delete, or Update API Keys in the Python Client
• Evaluations
  • Web Interface
  • Python Client
• OWASP LLM Top 10 Harness
• RAG metrics
  • Retrieval Metrics
  • Generation Metrics
  • Evaluating Domain-specific Question Answering
• Working with Detections
  • List Detectors
  • Create Detections
  • Summarize Detections
  • Supported fields for Detectors
• Custom Harnesses
  • View Custom Harnesses
  • Python Client
• Evaluating Agents with Custom Endpoints
  • Prerequisites
  • Step 1 - Create a Local Agent Executor
  • Step 2 - Evaluate!
  • Examples
• Evaluation Reports
  • Viewing the Report in the Web Interface
  • Generate a report in the Python Client
• Playground
  • Playground Chat
  • Playground Snapshot