Typical Data for This Category
Typical data sources include enterprise internal control systems, ERP systems, financial shared services systems, audit management systems, and procurement contract project systems. Daily generated documents cover internal control self-evaluation reports, audit notices, audit working papers, issue lists, rectification ledgers, and rectification cancellation certificates. Data is updated annually. Core documents use a mixed format of structured and unstructured data, including dedicated fields such as issue descriptions, rectification requirements, and supporting material numbers. No universal statistical units are used.
Constraints for Citation Sources and Traceability
Data sources are scattered across multiple business systems. Traceability links must associate document metadata from different systems to avoid cross-system data confusion. Annual update frequency means the knowledge base does not require high-frequency incremental synchronization, and full document updates can be performed on fixed cycles. Diverse document types and dedicated fields require traceability to match document types and field characteristics, such as extracting cancellation status supporting materials from rectification ledgers. Time required for core links depends on business samples and voucher extraction. Citations must be associated with specific supporting materials to ensure accurate traceability. Errors may result in regulatory notifications and liability investigations. Metadata in traceability links must accurately correspond to original documents, and mismatches must be avoided.
Configuration Settings
Reference only; verify against the target FastGPT release and deployment documentation.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
quoteMaxToken | 800–1200 token | Internal control and audit documents use mixed structured and unstructured formats. This range covers core content of rectification ledgers and audit working papers, and avoids truncation of key supporting information |
responseData | Return document title, affiliated system, update time | Audit traceability requires clear source systems and document versions to ensure rectification basis can be traced back to original business systems |
detail | Enabled | Policy basis for internal control evaluation and audit rectification requires display of original text snippets to facilitate verification of correspondence between regulatory clauses and rectification requirements |
| Single Search Limit | Top 3–5 results | Policy basis for this category is concentrated in core documents such as audit notices and rectification ledgers. Excessive results increase context redundancy and exceed token limits |
| Search Filter | Filter by document type | Data for this category includes multiple document types such as self-evaluation reports and rectification ledgers. Filtering excludes irrelevant procurement contract data and improves traceability accuracy |
Common Misconfigurations and Root Causes
- Phenomenon: Generated answers do not display citation sources, and corresponding metadata fields are empty. Root cause: The
responseDataconfiguration item is not enabled, and metadata information for citations is not specified to be returned. - Phenomenon: Answer content exceeds token limits, triggering truncation or error reports. Root cause: The
quoteMaxTokenvalue is set too low, failing to cover complete core snippets of audit working papers and rectification ledgers, or the single search limit is set too high. - Phenomenon: Search results include non-audit documents such as procurement contracts. Root cause: Search filtering rules based on document type are not configured, and dedicated document types for this category are not distinguished.
Verification Steps for Proper Configuration
- Initiate a query about audit rectification policies, and check if citation metadata such as document title and affiliated system is included in the response.
- Review token consumption for the response, and confirm no content truncation occurs due to improper
quoteMaxTokensettings. - Test search keywords for different document types, and confirm only dedicated document content for this category is returned, with no irrelevant documents included.
- Inspect the configuration of the knowledge base search node in the workflow, and confirm the
detailoption is enabled.
The values provided here are common starting points, and should be evaluated against internal operational 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.