Triple
T1078405
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | Juan Fernández firecrown |
E23890
|
entity |
| Predicate | primaryThreat |
P23662
|
FINISHED |
| Object | habitat loss |
—
|
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: habitat loss | Statement: [Juan Fernández firecrown, primaryThreat, habitat loss]
PD
Predicate disambiguation
gpt-5-mini-2025-08-07
Target predicate: primaryThreat Context triple: [Juan Fernández firecrown, primaryThreat, habitat loss]
-
A.
primaryForce
Indicates that one entity serves as the main or dominant force acting upon, influencing, or driving another entity or process.
-
B.
primaryTarget
Indicates that an entity is the main or most important target of another entity’s action, focus, or effect.
-
C.
primaryFront
Indicates that one entity serves as the main or most important front-facing side or surface in relation to another entity.
-
D.
primaryDamageType
Indicates the main kind of harm or injury that an action, event, or object is responsible for causing.
-
E.
primaryMode
Indicates the main or most commonly used method, manner, or form in which an action, process, or interaction is carried out between entities.
- 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_69a493f1ddf48190a99d54b00e99f8ce |
completed | March 1, 2026, 7:30 p.m. |
| NER | Named-entity recognition | batch_69a4b943b41481909b24050ca7e78971 |
completed | March 1, 2026, 10:10 p.m. |
| PD | Predicate disambiguation | batch_69a4b73ba8208190be7f3cef8c18689b |
completed | March 1, 2026, 10:01 p.m. |
| PDg | Predicate description generation | batch_69a4b80f0fb08190a19a50e38ae8f16c |
completed | March 1, 2026, 10:05 p.m. |
Created at: March 1, 2026, 7:42 p.m.