Triple
T35329871
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | Auto-Owners Insurance |
E1020288
|
entity |
| Predicate | policyholderOwnership |
P7585
|
FINISHED |
| Object | owned by its policyholders |
—
|
LITERAL FINISHED |
How this triple was built (2 steps)
Every LLM step that produced this triple, in pipeline order — named-entity classification, the disambiguation choices (the exact options shown, with the pick highlighted), and the generated description. The batch + timestamp of each is in the Provenance table below.
NER
Named-entity recognition
gpt-5-mini
Instruction
Given a phrase, classify it is english named entity (e.g., persons, organizations, works of art) in Latin script, or not (e.g., literals, dates, URLs, verbose phrases). For disambiguation, the statement where the phrase occurs as object is also given. Please return a JSON object with `phrase` (string, the phrase being analyzed) and `is_ne` (boolean, indicating whether the phrase is a Named Entity).
Input
Phrase: owned by its policyholders | Statement: [Auto-Owners Insurance, policyholderOwnership, owned by its policyholders]
PD
Predicate disambiguation
gpt-5-mini-2025-08-07
Target predicate: policyholderOwnership Context triple: [Auto-Owners Insurance, policyholderOwnership, owned by its policyholders]
-
A.
ownershipCover
Indicates that one entity’s ownership or control extends over, protects, or encompasses another entity.
-
B.
holderIs
Indicates that one entity serves as the holder, possessor, or container of another entity.
-
C.
policyName
Indicates the specific name or title assigned to a policy associated with an entity.
-
D.
ownershipType
chosen
Indicates the specific nature or category of the ownership relationship that one entity holds over another.
-
E.
ultimateOwnerOf
Indicates that one entity is the final, controlling owner of another entity, either directly or through one or more intermediate ownership layers.
- F. None of above.
Provenance (3 batches)
The batch behind each pipeline step, in order, with when it ran. Timestamps are batch-level — stages were processed in waves, so the object chain (NER → NED1 → NEDg → NED2) reads in order, but predicate / elicitation batches can sit in a different wave.
| Step | Stage | Batch ID | Status | When |
|---|---|---|---|---|
| creating | Elicitation | batch_69f76deacf4481908e7735a5a7715b0a |
completed | May 3, 2026, 3:46 p.m. |
| NER | Named-entity recognition | batch_69fba78aca4c8190b8f1831e8cc04e06 |
completed | May 6, 2026, 8:41 p.m. |
| PD | Predicate disambiguation | batch_69fba34a65a4819088bac6c17542d71c |
completed | May 6, 2026, 8:23 p.m. |
Created at: May 3, 2026, 4:03 p.m.