Introduction

What Are Datasets?

FloTorch Datasets are collections of question-answer pairs used to evaluate and improve your AI models. They store your test data (ground truth) and examples in one place, so you can reuse them across evaluation projects and experiments.

---

How to Use Datasets in the Application

Step 1: Create a Dataset

1. Go to your Workspace → Datasets
2. Click Create Dataset
3. In Choose Dataset Type, select one of these cards:

Dataset Type	Best For	Supported Evaluations (UI)
Question and Answer Pair (Q&A Pair)	Standard Q&A evaluation datasets without retrieval context	LLM Model Evaluation, RAG Evaluation, Prompt Evaluation, Agent Evaluation, Agentic Workflow Evaluation
Question and Answer Pair with Context (Q&A + Context)	Q&A datasets that include retrieval context with each pair	LLM Model Evaluation, RAG Evaluation, Prompt Evaluation

Step 2: Configure Based on Selected Card

After selecting a card, you are automatically navigated to the Configure Dataset step. Complete dataset details:

1. Enter Dataset Name
2. Optionally add a Description
3. Choose a Creation Method (depends on selected card)
4. Click Create Dataset

If you selected Q&A Pair

You can choose from:

• Upload Q&A Pair Files
• Create Q&A Pairs Manually
• Auto Capture Q&A Pairs
• Generate Q&A Pair from PDF
• Import Question & Answer Pair from HuggingFace

If you selected Q&A + Context

Available creation method:

• Upload Q&A Pair with Context Files

Step 3: Use Your Dataset in Evaluation Projects

1. Go to Evaluations → Projects → Create Project
2. Select your dataset from the list
3. Select a Knowledge Base if your evaluation setup requires one
4. Configure and run experiments

Your dataset provides the ground truth used to score model responses.

---

Ways to Add Content (Summary)

Creation Method	Q&A Pair	Q&A + Context
Upload files	✓	✓ (context format)
Manual creation	✓	—
Auto capture	✓	—
Import from HuggingFace	✓	—
Generate from PDF	✓	—

---

Tips

• Use clear dataset names that describe the purpose (e.g., customer-support-qa)
• Add a description so your team understands the contents
• Start with a small dataset to verify format, then add more
• Replace files to update datasets — uploading a new file of the same type overwrites the previous one

---

Notes

• Dataset names cannot be changed after creation
• Datasets cannot be deleted
• Dataset names must start with a letter and can contain lowercase letters (a-z), numbers (0-9), and hyphens (-)
• Max 10 MB per uploaded file; 50 MB for synthetic PDF source