Skip to content
  • There are no suggestions because the search field is empty.

Transparent Source Controls

Learn how to effectively customize and control the sources used in AI-generated responses for more accurate and trustworthy results.

 

source-sorting

Overview

Transparent Source Controls empowers you to actively guide and refine an AI model's research process by managing the sources used in responses. This feature transforms the model from a black box into a transparent research partner that adapts to your expertise and preferences.

Key Benefits

  • Enhanced control over source quality
  • Improved research accuracy
  • Transparent source management
  • Customized AI responses based on your expertise
  • Efficient source navigation and organization

Getting Started

1. Source Search

  • Access the source sidebar
  • Use the search function to find specific sources by:
    • Keywords
    • Titles
    • URLs
  • Quickly locate and validate specific data points without scrolling through numerous sources.

2. Source Sorting

Navigate sources efficiently using the new dropdown menu to sort by:
  • Citation order (default)
  • Alphabetically by domain
  • Most-used domains 

3. Source Prioritization

Build trust with the AI by:
  • Using the "thumbs up" feature for preferred sources or domains
  • Training the AI to prioritize trusted sources (e.g., PubMed for medical research)
  • Applying preferences to future responses in the chat thread

4. Source Removal

Maintain quality by:
  • Clicking 'X' to remove individual pages or entire domains
  • Blocking unwanted sources from current and future responses
  • Filtering out potentially biased, outdated, or low-quality content

5. Save & Regenerate

After adjusting sources:
  • Click "Save & Regenerate" to apply your changes
  • Review the new response incorporating your preferences
  • Verify the impact of your source management choices

Best Practices

  • Regularly review and update source preferences
  • Use domain-level controls for consistent quality
  • Combine search and sorting features for efficient source validation
  • Document trusted sources for your specific research areas

Examples

Research Quality Control

  1. Search for specific sources using keywords
  2. Sort by domain to assess source diversity
  3. Prioritize authoritative domains
  4. Remove questionable sources
  5. Regenerate for improved results

Domain Expertise Application

  1. Identify preferred sources in your field
  2. Use thumbs up to prioritize trusted domains
  3. Remove non-authoritative sources
  4. Save and regenerate for expert-level responses

Troubleshooting

  • If sources aren't updating, ensure you've clicked "Save & Regenerate"
  • For missing preferred sources, verify your prioritization settings
  • If unwanted sources persist, check domain-level blocking
💡 Tip: Create a list of trusted domains in your field to consistently prioritize high-quality sources.
⚠️ Warning: Removing too many sources may limit the model's ability to provide comprehensive responses.