Triple
T2285438
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | Special Air Service |
E51379
|
entity |
| Predicate | selectionCharacteristic |
P36177
|
FINISHED |
| Object | extremely high attrition rate |
—
|
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: extremely high attrition rate | Statement: [Special Air Service, selectionCharacteristic, extremely high attrition rate]
PD
Predicate disambiguation
gpt-5-mini-2025-08-07
Target predicate: selectionCharacteristic Context triple: [Special Air Service, selectionCharacteristic, extremely high attrition rate]
-
A.
selectionCriteria
Indicates the conditions or rules used to choose certain entities from a larger set.
-
B.
selectionFrom
Indicates that one entity is chosen as a member or subset from a larger set or collection represented by another entity.
-
C.
selectionMadeBy
Indicates that a particular choice or selection was performed or decided by a specific agent or entity.
-
D.
membershipCharacteristic
Indicates that an entity possesses a specific attribute or quality by virtue of its membership in a particular group or category.
-
E.
selectionMetric
Indicates the criterion or measure used to evaluate and choose among alternative options or candidates.
- F. None of above. chosen
Provenance (4 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_69a88b08e4308190bdac9aebcca1c91a |
completed | March 4, 2026, 7:42 p.m. |
| NER | Named-entity recognition | batch_69abc245dd208190b13c5f5d05aa6990 |
completed | March 7, 2026, 6:14 a.m. |
| PD | Predicate disambiguation | batch_69abbdbb9e4c819085fc588626ec7c09 |
completed | March 7, 2026, 5:55 a.m. |
| PDg | Predicate description generation | batch_69abbe1ecb7081909c2c66da08a48ab7 |
completed | March 7, 2026, 5:56 a.m. |
Created at: March 4, 2026, 7:48 p.m.