What this category's data looks like
This page focuses on kindergarten operation behavior and quality self-assessment. The materials include three types: self-assessment reports, issue lists, and rectification plans. Self-assessment reports are updated by semester or academic year, and their fields include kindergarten name, assessment dimension, score explanation, date, and principal's signature; issue lists are formed on a per-instance basis, and their fields include issue description, involved class, responsible person, rectification deadline, and receipt; rectification plans roll forward by round, and their fields include measures, completion markers, review records, and signatures. When these three types of materials enter this feature, the self-assessment report provides the overall conclusion and basis paragraphs, the issue list provides specific entries and the corresponding original text, and the rectification plan provides follow-up actions and version clues. Names, quantities, dates, and signatures must be consistent across materials, and versions must be synchronized.
Configuration and implementation checks
This part is executed by application settings (response limit, etc.), which determine the length of a single answer and the range of context that can be accommodated. "Basis entries that can refer back to the original text" are then completed by the model. The response limit is controlled by a slider for the length of a single answer; if set too low, the answer will be truncated, and if set too high, consumption will increase.
| Object | Check | Verification action |
|---|---|---|
| Self-assessment report | The kindergarten name, assessment date, and principal's signature in the report can match the issue list and rectification plan | Check whether the report fields are consistent with the names, dates, and signatures in the other two materials, and record the number of cited entries and quoteMaxToken |
| Issue list | The issue description, involved class, and rectification deadline of the list entries have corresponding records in the report and plan | Compare the names, quantities, and dates in the two materials one by one, confirm that versions are synchronized, and list separately any entries missing signatures or receipts |
| Rectification plan | The plan measures, completion markers, and review records are consistent with the report conclusion and list entry versions | Check whether the plan version number is synchronized with the report and list, verify signatures and receipts, and inspect where maxContext and maxResponse are set |
| Response limit | Matches the answer length required by the page | Check the switch and slider position in application settings, and ask a question that requires a long answer to see whether the ending is complete |
Three common errors
- Symptom: the name, quantity, or date of the same issue does not match between two materials. Cause: the standard was not unified during entry, or only one place was changed after revision.
- Symptom: the versions referenced by the report and plan are not synchronized, and the basis points to an old draft. Cause: after the version was updated, it was not replaced synchronously, and version markers were missing.
- Symptom: the list or plan is missing signatures or receipts, and the basis cannot be closed. Cause: the circulation step was not completed, and signatures and receipts were not archived with the materials.
How to verify the setup
Feature acceptance: ask a question that requires a long answer and see whether the ending is complete; then check the settings of maxContext, quoteMaxToken, and maxResponse and the record of the number of cited entries.
Business test: ask a cross-material basis question and see whether the answer gives a source that can be referred back to, and confirm that the source matches the original text.
- Check the response limit switch and slider position in application settings, and confirm that they match the answer length required by the page.
- Ask a question that requires a long answer, check whether the ending is complete, and compare the usage changes between two runs.