What this category's data looks like
This category's data involves three core types of materials: the family economic hardship determination form, the scholarship/grant application form, and the loan receipt. The determination form is submitted by the student with family information and supporting evidence, and after class review and department/department-level verification, a determination grade is formed; the application form carries the scholarship/grant program, amount, and evaluation basis; the loan receipt is used to confirm disbursement and arrival status. Material sources include student submission, counselor preliminary review, department/department-level aggregation, and financial aid center recheck, with updates advancing by academic year or batch, and the same record flowing across multiple materials. Field characteristics center on name, student ID, household size, income range, determination grade, scholarship/grant program, amount, signature, and date. When materials enter the knowledge base retrieval and recall function, each serves as a candidate fragment source: the determination form provides hardship grade and review basis, the application form provides program and amount fields, and the loan receipt provides disbursement and arrival information, for subsequent nodes or humans to use as the basis for drafting issuable materials and handover records.
Configuration and implementation checks
This step is executed by the knowledge base retrieval service (retrieval node and search test), outputting candidate fragments and similarity scores (after pattern, merge, rerank, filter, and limit trimming). The issuable material drafts and handover records are then completed by subsequent nodes or humans.
| Object | Check | Verification action |
|---|---|---|
| Family economic hardship determination form | Fields complete, responsible person clear, version and date consistent; verify name, student ID, household size, income range, determination grade, signature, and date | In the search test, search by material name, verify whether the candidate fragments cover the above fields, record the number of candidates and scores, and switch searchMode to compare semantic, full-text, and hybrid results |
| Scholarship/grant application form | The name, student ID, amount, and date of the same record match between the determination form and the application form; the version is synchronized; signatures are complete | Test once each with synonym-rewritten questions, compare candidates and scores, enable usingReRank to see changes in the ranking of the top few, and verify whether the amount and date fields are consistent |
| Loan receipt | Receipt fields complete, responsible person clear, dates consistent; missing signature or receipt is blocked | Retrieve receipt fragments, verify disbursement and arrival fields, record the number of candidates and scores, and check whether key fragments are still retained after similarity filtering |
| Handover record | Responsible person, version, date, and receipt status are complete; missing items are blocked | Draft one according to the template and complete one handover, verify the field correspondence between the handover record and the material draft, and check whether the tokens of the quoted text are complete after limit trimming |
Three common errors
- Phenomenon: The name, amount, and date of the same student do not match between the determination form and the application form. Cause: The two materials were imported separately, and field comparison and version synchronization were not performed for the same record.
- Phenomenon: The loan receipt lacks a signature or receipt, and this was not blocked during handover. Cause: The retrieval candidates did not cover the signature and receipt fields, and key fragments were trimmed after filtering.
- Phenomenon: The same question cannot be recalled when rephrased, and although there are many candidates, the ones ranked at the top are irrelevant. Cause: The retrieval mode does not match the material form, rerank is not enabled, or similarity filtering is too strict.
How to verify the setup
Acceptance for this function: test once each with synonym-rewritten questions, compare candidates and scores, then toggle rerank to see the ranking of the top few; switch searchMode to compare semantic, full-text, and hybrid results, record the number of candidates and scores, and check the tokens of the quoted text after similarity filtering and limit trimming.
Business test: draft one according to the template and complete one handover, and see whether missing items are blocked.
- In the knowledge base search test, switch
searchModeto compare the number of candidates and scores across semantic, full-text, and hybrid modes, and confirm that the material form matches the mode. - Compare the order of the top few before and after enabling
usingReRank, and check whether the tokens of the quoted text retain key fields aftersimilarityfiltering andlimittrimming.