Displayed Graphs on Fragments

​

4. Displayed Graphs on Fragments

Once your data is collected, OpenSesame automatically populates your homepage with the output stored in fragments. These fragments allow you to visualize hallucination data through graphs, making it easier to analyze and understand the performance of your AI models.

​

What Are Fragments?

Fragments are essential for testing prompts against your corpus of documents. They enable you to visualize the calls made to your LLMs and compare them with your RAG (Retrieval-Augmented Generation) model, especially if you’re using a Vector DB.

​

Why Fragments Matter

• Visualize AI Performance: Fragments help you see how your AI models perform across different prompts and datasets.
• Template Extraction: If multiple prompts share the same template, OpenSesame automatically extracts the template and identifies variables for you to check.