Triple
T5111326
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | Dementor |
E115218
|
entity |
| Predicate | mouthDescription |
P13903
|
FINISHED |
| Object | gaping, sucking hole |
—
|
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: gaping, sucking hole | Statement: [Dementor, mouthDescription, gaping, sucking hole]
PD
Predicate disambiguation
gpt-5-mini-2025-08-07
Target predicate: mouthDescription Context triple: [Dementor, mouthDescription, gaping, sucking hole]
-
A.
mouth
Indicates that one entity is the mouth (oral opening) of another entity, typically serving as the location for ingestion, speech, or related functions.
-
B.
mouthNear
Indicates that one entity’s mouth is positioned close to another entity or object in space.
-
C.
mouthOf
Indicates the location where one entity (typically a river or similar feature) empties into or opens out into another, larger body or feature.
-
D.
mouthState
chosen
Indicates the current condition or configuration of an entity’s mouth, such as whether it is open, closed, or in another specific state.
-
E.
mouthfeel
Indicates the sensory qualities and physical sensations a substance produces in the mouth when consumed.
- 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_69bd4441d1648190a54a533895041987 |
completed | March 20, 2026, 12:57 p.m. |
| NER | Named-entity recognition | batch_69bd75ad362c8190b9cbded390aaea3c |
completed | March 20, 2026, 4:28 p.m. |
| PD | Predicate disambiguation | batch_69bd715fe3a8819087d3065adddba515 |
completed | March 20, 2026, 4:10 p.m. |
Created at: March 20, 2026, 1:41 p.m.