Triple
T14997375
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | Saga Domain |
E373992
|
entity |
| Predicate | capital |
P234
|
FINISHED |
| Object | Saga Castle |
E323064
|
NE 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: Saga Castle | Statement: [Saga Domain, capital, Saga Castle]
NED1
Entity disambiguation (via context triple)
gpt-5-mini-2025-08-07
Target entity: Saga Castle Context triple: [Saga Domain, capital, Saga Castle]
-
A.
Saga Castle
chosen
Saga Castle is a historic Japanese castle in Saga Prefecture, known for its reconstructed wooden palace and role as a former seat of regional feudal power.
-
B.
Leap Castle
Leap Castle is a historic Irish fortress in County Offaly, reputed to be one of the most haunted castles in the world.
-
C.
Drakensteyn Castle
Drakensteyn Castle is a 17th-century moated country house in the Netherlands best known as the private residence of former Queen Beatrix.
-
D.
Red Castle
Red Castle is a ruined medieval fortress overlooking Lunan Bay on the east coast of Scotland.
-
E.
King’s Castle
King’s Castle is a historic coastal fortification in Bermuda that forms part of the UNESCO-listed defenses of the Town of St. George.
- F. None of above.
- G. Unsure - the case is ambiguous/there is not enough information to decide.
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_69d85ccc84388190aa151e5173370c8d |
completed | April 10, 2026, 2:13 a.m. |
| NER | Named-entity recognition | batch_69ded718e4288190b5e144f82299a194 |
completed | April 15, 2026, 12:08 a.m. |
| NED1 | Entity disambiguation (via context triple) | batch_69fe969c3ba88190899f06b185e94ccf |
completed | May 9, 2026, 2:06 a.m. |
Created at: April 10, 2026, 2:54 a.m.