INDUSTRY FIELD GUIDE
Nuclear reactor public status snapshots for AI evaluation
A public status report describes what was recorded as of its publication time.
NAICS 221113 · Nuclear Electric Power Generation
A public status report describes what was recorded as of its publication time. It cannot establish a later status or explain a change without separate evidence. The NRC publishes power reactor status reports. This exercise uses only cleared public material or a fictional offline record set. NRC power reactor status reports
Give two practice units three dated status entries. Leave an interval unreported and include one later correction. Ask what the supplied archive supports for a past date and which statements remain unverified. Use unit identity and document date as distinct scoring fields. Do not ask the model for any operational explanation or control decision.
Use no internal records, access or facility information, staff details, security data, or unpublished operations. Have the reviewer check source authority before use.
Describe a public-status review exercise
Original Fiund exercise
Preserve the publication time of a fictional status snapshot
Two invented unit aliases and dated paper reports test what an archive supports. The example contains no actual facility status, security information or operational explanation.
| Record | Source | What it records |
|---|---|---|
| SNAP-N1 | Day-1 publication | At day 1, 09:00, the report records Alpha as practice status A. |
| SNAP-N2 | Day-2 publication | At day 2, 09:00, only Beta has an entry. Alpha is absent from the supplied report. |
| CORRECT-N3 | Day-3 correction | A day-3 publication corrects Alpha's day-1 entry to practice status B. It provides no day-2 Alpha observation. |
Question for the model
What Alpha information was supplied by noon on day 2? What does the later correction change, and what remains unreported?
Show the reference answer and checks
By the earlier cutoff, the supplied Alpha entry is N1's day-1 status A. N2 does not supply a day-2 Alpha observation. N3 later corrects the historical day-1 entry to B, but it was not available at the day-2 cutoff and does not establish Alpha's day-2 state. Keep observation date, publication date and correction scope separate.
- Using a later publication as information available at an earlier cutoff.
- Filling an unreported unit interval from another unit's entry.
- Interpreting a historical correction as a new current status.
Fields to preserve
- Unit identity
- Keep Alpha and Beta entries separate.
- Observation and publication dates
- Preserve what was observed and when it became available.
- Correction reference
- Identify the historical entry that is changed.
Test another case
Supply a genuine day-2 Alpha observation in a new fictional report. The model should distinguish that observation from a correction to day 1.
Only fictional or cleared public documents belong in the exercise. Include no internal records, staff details, access information, unpublished operations or control decisions.
Discuss a contribution
Describe the sources you control, the task they could support and what has already been checked. Use the existing-data assessment for an archive or the collection brief for a new recording project. The contributor agreement defines Fiund’s licensing authority and compensation. A guide or fictional exercise does not establish an accepted collection or a sale.