INDUSTRY FIELD GUIDE
Bulk electric transmission aggregates for offline AI review
A revised aggregate can change a historical summary without revealing which underlying event caused the change.
NAICS 221121 · Electric Bulk Power Transmission and Control
A revised aggregate can change a historical summary without revealing which underlying event caused the change. FERC publishes electric industry data and separately describes protections for critical infrastructure information. A model evaluation should use only fabricated records or public aggregates cleared for the proposed use. FERC public data · FERC critical infrastructure information
Invent a historical aggregate, a correction, and a dated note with no underlying event record. Ask which value was available at a specified time and which causal claims the sample does not support. Add another aggregate with a similar label and different period. Score source, timing and uncertainty.
Keep real transmission topology, control messages, configurations, facilities, precise locations, and live feeds outside the exercise. The model never makes switching, balancing or safety decisions.
Describe a public aggregate benchmark
Original Fiund exercise
Read a public-style aggregate at an earlier cutoff
Original historical aggregates use demonstration counts and contain no infrastructure events or topology. The question concerns document availability and period matching.
| Record | Source | What it records |
|---|---|---|
| JUNE-V1 | Initial practice summary | Published on day 2, June's aggregate is 140 demonstration counts. |
| JUNE-V2 | Later correction | Published on day 10, the June aggregate is revised to 150. No underlying event or reason is provided. |
| JULY-V1 | Different-period summary | Published on day 6, July's similarly titled summary records 150 counts. |
Question for the model
Which June value was supplied by day 8, and do the records explain why June was later revised?
Show the reference answer and checks
JUNE-V1 is the supplied June publication available by day 8, so the supported earlier-cutoff value is 140. JUNE-V2 documents a later revision to 150 but was not yet available at that cutoff. JULY-V1 concerns another period. None of these aggregate records establishes the underlying cause of June's revision.
- Using a later correction to describe information already available at an earlier cutoff.
- Substituting July's matching number for June's source.
- Inventing an operational event to explain an unexplained aggregate revision.
Fields to preserve
- Reporting period
- Keep June and July aggregates distinct.
- Publication cutoff
- Preserve when each version became available.
- Revision lineage
- Link the correction to the earlier summary without inventing its cause.
Test another case
Add a dated explanatory note after day 10 and ask separately for the earlier available information and the latest documented explanation.
Use fictional or cleared public aggregates only. Exclude real topology, control messages, configurations, facilities, locations and live feeds. No switching or balancing decision is in scope.
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.