Configuration Steps for LLM Dataset creation

• Set Domain

  The domain categorizes your dataset’s primary field of application:

  • Open the domain dropdown menu
  • Select the appropriate domain (e.g., “defense”, “retail”, “healthcare”)
  • This helps organize and locate your dataset later

• Choose Subdomain

  Subdomains further specify your dataset’s focus:

  • Select from available subdomains based on your chosen domain
  • Pick the most relevant option (e.g., “aerospace” under defense)
  • Ensures proper dataset categorization

• Configure Data Settings

  Essential technical settings for LLM training:

  • Data Type: Select “text”
  • Task: Choose “llm-instruction-tuning”
  • These settings are fixed for LLM instruction tuning

• Add Description

  Provide context about your dataset:

  • Explain the dataset’s purpose
  • Describe the type of instructions included
  • Note any specific use cases or target applications

• Click “Configure” when finished to proceed to data upload, a pop up message appears as “Success”.