What this category's data looks like
This page addresses the scenario of kindergarten operation behavior and quality self-assessment. Materials include self-assessment reports, issue lists, rectification plans, as well as accompanying progress tables and statistical scope explanations. Self-assessment reports carry descriptions of indicator achievement and supporting explanations; issue lists carry item-by-item issues, responsible persons, and rectification deadlines; rectification plans carry measures, milestones, and completion markers. Progress tables summarize record counts by week or month, and statistical scope explanations are written in the table header, noting inclusion scope, exceptions, and signature receipt requirements. Materials are updated weekly and aggregated monthly, with fields involving names, quantities, dates, signatures, and version numbers. When entering the knowledge base retrieval and recall function, self-assessment reports, issue lists, and rectification plans serve as candidate sources to be retrieved, used to recall passages related to the issue; progress tables and statistical scope explanations serve as reference materials for verifying totals and scope, based on which subsequent nodes or humans generate business deliverables.
Configuration and implementation checks
This step is executed by the knowledge base retrieval service (retrieval node and search test), outputting candidate passages and similarity scores (after pattern, merge, rerank, filter, and limit trimming). "Progress tables and statistical scope explanations" are then completed by subsequent nodes or humans.
| Object | Check | Verification action |
|---|---|---|
| Self-assessment report | Whether fields, dates, and signatures are complete; whether the names, quantities, and dates of the same record are consistent across the issue list and rectification plan | In the knowledge base search test, switch searchMode to semantic, full-text, or hybrid, and check whether candidate passages cover the corresponding entries in the self-assessment report; compare the order of the top few items before and after enabling rerank, record the number of candidates and scores, and check whether the tokens of the cited text retain key paragraphs |
| Issue list | Whether the issue entries match the scope of the rectification plan and progress table, whether versions are synchronized, and whether signatures or receipts are missing | Use synonymous rewrites of the same issue to test the query once each, and observe whether candidates stably hit the issue list entries; check whether the returned candidates include both semantic and full-text results, confirm that multi-path merging is in effect, and record the number of citations separately |
| Rectification plan | Whether measures, milestones, and completion markers are consistent with the self-assessment report and issue list, and whether dates and signatures are traceable | Adjust similarity and limit to observe changes in the number of candidates and scores, and confirm that key passages of the original text are still retained after similarity filtering and citation limit trimming; check whether the names, quantities, and dates in the rectification plan are consistent across the two materials |
| Progress table and statistical scope explanation | Whether record counts and statistical counts match, whether the scope is written in the table header, and whether exceptions are clearly stated | Use one week's records to verify the total, and check the header scope and exception explanations; confirm that retrieval returns candidates and cited content, do not treat the number of candidates as the statistical result, the citation limit only verifies the tokens of the cited text, and the array count is recorded separately |
Three common errors
- Phenomenon: The names, quantities, and dates of the same record do not match between the self-assessment report and the issue list. Cause: The scope was not unified during entry, or versions were not synchronized, and old-version materials are still hit by retrieval.
- Phenomenon: The rectification plan lacks signatures or receipts, and the progress table total does not match the record count. Cause: The signature step was not closed-loop, exceptions were not written into the header scope, and verification cannot be traced.
- Phenomenon: The same issue cannot be recalled when rephrased, and although there are many candidates, the top-ranked ones are irrelevant. Cause: The retrieval mode does not match the material form, rerank was not enabled, or similarity filtering and citation limit trimming were improperly set.
How to verify the setup
This function's acceptance: use synonymous rewritten queries to test once each, compare candidates and scores, then toggle rerank on and off to see the ordering of the top few items; switch searchMode and adjust similarity and limit, record the number of candidates and scores, and confirm that the tokens of the cited text retain key paragraphs.
Business test: use one week's records to verify the total, check whether the scope and exceptions are clearly stated, confirm that record counts and statistical counts match, and that the header scope and exception explanations are traceable.
- In the knowledge base search test, switch
searchMode, compare the number of candidates and scores across semantic, full-text, and hybrid modes, and check whether both semantic and full-text results are included. - Compare the ordering of the top few items before and after enabling rerank, and adjust
similarityandlimitto confirm that key passages of the original text are still retained after trimming, and record the number of citations separately.