What the data for this category looks like
Core data for this category includes documents such as minimum living allowance application forms, family economic status verification reports, household survey forms, democratic appraisal records, public notification materials, and assistance approval forms. Data sources include the low-income population dynamic monitoring platform, resident family economic status verification system, and cross-departmental shared data from public security, human resources and social security, medical insurance, housing and urban-rural development, and tax authorities. Data updates occur monthly. Core associated dimensions are family economic status and member relationships. Documents have both structured form and unstructured supporting material characteristics.
Constraints on Knowledge Base Retrieval and Recall
The monthly update frequency requires retrieval and recall to prioritize the latest monthly dataset, reducing interference from expired data. Cross-departmental data from multiple sources includes complex fields, so targeted recall rules must be configured to avoid returning irrelevant cross-departmental field content. The family-centric business logic requires support for batch recall of multiple documents linked to a single household. Documents have both structured forms and unstructured supporting materials, so recall weights for the two content types must be distinguished to ensure retrieval results align with business process needs.
Configuration Settings
This configuration is built for FastGPT v4.17.0, and is adapted to the business data characteristics of social assistance and minimum living allowance review.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
similarity | 0.75–0.85 | The core of this category’s data is family economic status. A higher similarity threshold can filter out irrelevant cross-departmental field content, ensuring retrieval results align with assistance review requirements |
limit | Top 10 entries | Business processes only require a small number of precise reference documents to avoid excessive information interfering with review staff’s decision-making |
searchMode | Exact match | Cross-departmental shared data contains complex fields. Exact match mode can reduce irrelevant recall results and improve retrieval efficiency |
usingReRank | Enabled | Multi-source knowledge base data requires reranking to optimize result order, highlighting content with the highest relevance to family economic status |
datasetSearchUsingExtensionQuery | Enabled | This category’s data relies on cross-departmental shared data. Enabling extended query can associate real-time verification data from external systems |
collectionIds | Specify the dataset ID corresponding to civil affairs assistance business | Isolate knowledge base data from different business scenarios, avoid confusion between documents from different civil affairs businesses, and ensure retrieval accuracy |
Three Common Configuration Errors
- Issue: The knowledge base cannot sync Feishu data sources, or retrieval results do not include content from the Feishu knowledge base. Cause: Cross-platform data source synchronization permissions are not configured, and the
datasetSearchUsingExtensionQueryextended query function is not enabled. - Issue: A single retrieval cannot select a specified document set, and only full knowledge base content can be used. Cause: The
collectionIdsparameter is not configured correctly, or the target document collection is not bound using the batch upload function. - Issue: Retrieval requests return timeout errors. Cause: The
limitandsimilarityparameters are not set reasonably, resulting in too much cross-departmental data being recalled, exceeding the system’s processing threshold.
How to Verify a Correct Configuration
- Run a test retrieval for the minimum living allowance review scenario, and check whether the returned results include core documents of this category such as minimum living allowance application forms and family economic status verification reports.
- View the data source synchronization records to confirm that the monthly updated cross-departmental shared data has been properly synchronized to the knowledge base.
- Adjust the
similarityandlimitparameters to verify whether the number and relevance of results meet business expectations. - Check the system permission configuration to confirm that access permissions for cross-platform data sources have been correctly enabled.
The values provided here are common starting points and should be measured against the reader's own samples.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-09-28.