What this category's data looks like
University laboratory safety grading and classification materials consist of three types: the laboratory grading and classification table, the hazard source list, and the safety responsibility statement. Sources are departmental submission and university-level review, and the update cadence is a centralized revision each academic year, with immediate supplementary entry when personnel or hazard sources change. The grading and classification table carries the laboratory name, level, room number, and hazard source category; the hazard source list carries the hazard source name, quantity, storage location, and risk level; the safety responsibility statement carries the responsible person, signature, and date. When the three types of materials enter the knowledge base retrieval and recall function, the grading and classification table and the hazard source list are split and recalled as fact-type chunks, and the safety responsibility statement is recalled as a credential-type chunk, used to check signatures and receipts. In terms of field characteristics, the same laboratory name, hazard source quantity, and review date must be aligned item by item across the three types of materials, and the version number and signature field are the key positions for judging whether the materials can be handed over.
Configuration and implementation checks
This step is executed by the knowledge base retrieval service (retrieval node and search test), outputting candidate chunks and similarity scores (after pattern, merge, re-rank, filter, and upper-limit trimming). The handover-ready deployment status and operation records are completed by subsequent nodes or manually.
| Object | Check | Verification action |
|---|---|---|
| Laboratory grading and classification table | The name, level, room number, and hazard source category must be aligned item by item with the hazard source list, and the date and signature field must be consistent with the safety responsibility statement | In the search test, switch searchMode to semantic, full-text, or hybrid, and compare whether the candidate chunks cover the key fields of the grading and classification table |
| Hazard source list | The hazard source name, quantity, storage location, and risk level must be aligned with the same record in the grading and classification table, and the version number must be synchronized with the safety responsibility statement | Use synonymous rewrites of the same question to retrieve once each, record the number of candidates and the similarity score, and check whether the quantity and date match up |
| Safety responsibility statement | The responsible person, signature, and receipt must be synchronized with the dates and versions of the grading and classification table and the hazard source list; if the signature or receipt is missing, it cannot be handed over | Turn usingReRank on and off and retrieve once each, compare the ordering of the top few results, and check whether the signature and receipt chunks are recalled |
| Deployment and operation records | The version, connection, and task records must be consistent, and the number of candidate items returned by retrieval and the cited text tokens are recorded separately | Check item by item according to the handover form to see which record is missing; for the citation upper limit, check the cited text tokens, and record the array item count separately as limit |
Three common errors
- Symptom: The name or quantity of the same laboratory does not match between the grading and classification table and the hazard source list. Cause: The two materials were submitted at different times, the versions were not synchronized, and the associated fields were not checked after retrieval hits.
- Symptom: The safety responsibility statement lacks a signature or receipt, and the handover form cannot be confirmed. Cause: The review process was not closed-loop, and the signature field and date field were not validated when the materials were entered into the knowledge base.
- Symptom: The same question cannot be recalled when rephrased, or there are many candidates but the ones ranked at the front are all irrelevant. Cause: The retrieval mode does not match the material form, re-ranking is not enabled, or similarity filtering and citation upper-limit trimming are inappropriate.
How to verify the setup
Acceptance of this function: test once each with synonymously rewritten questions, compare the candidate chunks and similarity scores, then turn usingReRank on and off to see whether the ordering of the top few results changes, and record the number of candidates and limit.
Business test: check item by item once according to the handover form to see which record is missing, and check whether the version, connection, and task records are consistent.
- In the knowledge base search test, switch
searchMode, record the number of candidates and scores under semantic, full-text, and hybrid modes respectively, and check whether multi-path merging includes both types of results at the same time. - Turn
usingReRankon and off and retrieve once each, compare the ordering of the top few results, and check whether the cited text tokens retain the key chunks after similarity filtering and citation upper-limit trimming.