Data Privacy Agreement
Data Processing Scope
When the system processes your data queries, the following data will be sent to the large language model service:
Metadata:
- Data table structure (Schema)
- Descriptions of tables and fields
- Field aliases
- Definitions of business metrics and answer builders
Data:
- Column value data with indexing enabled (You can prevent specific column data from being sent to the large language model by disabling column value indexing)
Document Segmentation:
The system segments the document into vectorized chunks through semantic parsing. These vectorized segments are then processed by the Large Language Model (LLM) to enable intelligent question-answering capabilities.
Data Protection Measures
To protect your sensitive data:
- It is recommended to disable indexing for columns containing sensitive information
- Before using the document Q&A feature, please confirm the sensitivity of the document content
- The system only transmits necessary data for query parsing
Please Note:
- The large language model service may retain processed data for model optimization
- It is recommended not to include highly sensitive personal information in queries
- The system will periodically clean cached query data
Updates and Changes
We reserve the right to update this agreement based on technological developments and security requirements. Significant changes will be notified to users through system announcements. By using the DataGPT system, you acknowledge and agree to the above data processing rules. If you have any questions about the product, please feel free to provide feedback through the following contact methods.
- Email: service@singdata.com