What this category's data looks like
This page addresses the enterprise new-hire onboarding training scenario, with materials drawn from three sources: onboarding training plans, course sign-in sheets, and job descriptions. Onboarding training plans are published by batch, with fields including course name, training date, responsible instructor, and planned headcount; course sign-in sheets are collected by session, with fields including sign-in date, actual attendance, and personal signature; job descriptions are updated by position version, with fields including position name, responsibility items, and effective date. When these three types of materials enter knowledge base retrieval and recall, their roles differ: training plans and job descriptions serve as candidate sources for semantic and full-text recall, while sign-in sheets serve as the original records for verifying actual attendance and signatures. In terms of update cadence, plans are adjusted by batch, descriptions are re-versioned as positions change, and sign-in sheets are archived by session. Field characteristics center on four categories: names, quantities, dates, and signatures; during retrieval, original text fragments must be retained for traceability.
Configuration and implementation checks
This step is performed by the knowledge base retrieval service (retrieval node and search test), outputting candidate fragments and similarity scores (after pattern, merge, re-rank, filter, and limit truncation). The exception list and review conclusions are completed by subsequent nodes or manually.
| Object | Check | Verification action |
|---|---|---|
| Onboarding training plan | Course name, training date, responsible instructor, and planned headcount must be aligned item by item with the sign-in sheet | In the search test, run one retrieval each with synonymous questions, switch searchMode to compare the number of candidates and the similarity score |
| Course sign-in sheet | Sign-in date, actual attendance, and personal signature must correspond to the plan batch | Check whether the hit fragments contain the date and the original signature text, record the number of candidates separately, and for the citation limit only check the citation text tokens |
| Job description | Position name, responsibility items, and effective date must match the training course | Compare the ranking of the top few results before and after enabling usingReRank, and check whether the version number is synchronized |
| Exception list and review conclusions | Exceptions are supported by original records, signed by the reviewer, and handed to the responsible person for disposition | Use limit to control the number of returned items, and confirm that candidates can be traced back to the source materials |
Three common errors
- Symptom: The name and date of the same course do not match between the plan and the sign-in sheet. Cause: The two materials were entered separately, without field-level comparison.
- Symptom: The job description has been re-versioned, but the training course still references the old responsibility items. Cause: The version was not synchronized, and retrieval hit the old fragment.
- Symptom: The sign-in sheet lacks a personal signature or receipt, and the exception list has no original record support. Cause: The signature field was not validated during archiving.
How to verify the setup
This feature's acceptance: test once each with synonym-rewritten questions, compare candidates and similarity scores, then toggle usingReRank on and off to observe the ranking of the top few results.
Business test: create one exception record and see whether it can be traced back to the source materials.
- In the search test, switch
searchMode, and record the number of candidates and scores for semantic, full-text, and hybrid respectively. - Check the citation text tokens against the number of items in the candidate array, and confirm that key original text fragments are still retained after truncation.