Skip to main content
You can view a dataset to see its status and the example conversations NexusOne wrote for it. This task walks you through viewing a sample of a dataset, and then the full dataset.

Prerequisites

Before starting this task, ensure you have the following:
  • Required roles: nx1_finetune_admin, nx1_finetune_user, or nx1_engineer.
  • A previously created dataset.

View a dataset

Perform these steps to view a dataset:
  1. Log in to NexusOne.
  2. On the top navigation bar, hover your mouse over AI and then select AutoML & AutoLLM.
  3. Click Fine-tune LLMs (AutoLLM).
  4. Click Datasets. You are going to see a table of datasets with their domain, base model, number of examples, and status.
  5. Select the dataset you want to view.
  6. In the dialog that opens, you can view the following:
    • Name and status of the dataset
    • Business domain, base model, and number of examples
    • Notes from the agents and any error message
    • A sample of the example conversations. Each shows the question, the answer, and the category and source the answer came from.

View the full dataset

The dialog shows only a sample of the dataset.
Perform these steps to view every example conversation:
  1. In the dataset dialog, click View entire dataset.
  2. If the dataset has more than 50 examples, then use Previous and Next to move through the pages.
  3. Optional: Click Download JSONL to save the example conversations as a JSON Lines file.
The View entire dataset button appears when the dataset has more examples than the sample, or when its status is complete. Before the dataset is complete, the page shows the proposed examples. After it’s complete, the page shows the approved examples that NexusOne trains on.

Additional resources

For a high-level overview of datasets and how they’re used in NexusOne, refer to Fine-tune LLM.