Triple
T23603803
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | Zatanna Zatara |
E582833
|
entity |
| Predicate | languageOfMagic |
P139209
|
FINISHED |
| Object | English words spoken backwards |
—
|
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: English words spoken backwards | Statement: [Zatanna Zatara, languageOfMagic, English words spoken backwards]
PD
Predicate disambiguation
gpt-5-mini-2025-08-07
Target predicate: languageOfMagic Context triple: [Zatanna Zatara, languageOfMagic, English words spoken backwards]
-
A.
languageOfIncantation
chosen
Indicates that a specified language is the one in which a particular incantation is spoken or written.
-
B.
typeOfMagic
Indicates that one entity is a specific category, school, or kind of magic associated with another entity.
-
C.
spells
Indicates that one entity writes or names the letters of a word or sequence in their correct order for another entity or context.
-
D.
knowsAboutMagic
Indicates that one entity possesses knowledge or awareness of magic in relation to another entity or context.
-
E.
underSpellBy
Indicates that one entity is magically influenced, controlled, or enchanted by another entity.
- 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_69e248faa2788190abb1581742daa6aa |
completed | April 17, 2026, 2:51 p.m. |
| NER | Named-entity recognition | batch_69f1b0ee6ce881909f556404cc235418 |
completed | April 29, 2026, 7:19 a.m. |
| PD | Predicate disambiguation | batch_69f118c96a0081908a8ac98ef7e7e60c |
completed | April 28, 2026, 8:30 p.m. |
Created at: April 17, 2026, 6:44 p.m.