Triple
T17451820
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | 大沼（赤城山） |
E424929
|
entity |
| Predicate | 関連する山頂 |
P127519
|
FINISHED |
| Object | 駒ヶ岳（赤城山） |
—
|
NE NERFINISHED |
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: 駒ヶ岳（赤城山） | Statement: [大沼（赤城山）, 関連する山頂, 駒ヶ岳（赤城山）]
NED1
Entity disambiguation (via context triple)
gpt-5-mini-2025-08-07
Target entity: 駒ヶ岳（赤城山） Context triple: [大沼（赤城山）, 関連する山頂, 駒ヶ岳（赤城山）]
-
A.
群馬県の赤城山
chosen
群馬県の赤城山は、関東平野北部にそびえる火山で、カルデラ湖の大沼や豊かな自然景観で知られる名山です。
-
B.
稲荷山
稲荷山 is a sacred mountain in Kyoto, Japan, best known as the site of Fushimi Inari Taisha and its thousands of vermilion torii gates.
-
C.
御嶽山
御嶽山 is a sacred stratovolcano in central Japan known for its religious significance, popular hiking routes, and a deadly 2014 phreatic eruption.
-
D.
秩父山地
秩父山地 is a mountainous region in central Japan spanning parts of Saitama, Yamanashi, Nagano, and Tokyo, known for its rugged peaks, rich forests, and popular hiking areas.
-
E.
十勝岳連峰
十勝岳連峰 is a volcanic mountain range in central Hokkaido, Japan, known for its active volcanoes, alpine scenery, and popular hiking routes.
- F. None of above.
- G. Unsure - the case is ambiguous/there is not enough information to decide.
Provenance (2 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_69d889db0ba481908402409af3b37917 |
completed | April 10, 2026, 5:25 a.m. |
| NER | Named-entity recognition | batch_69e4513e57248190824b540865311f44 |
completed | April 19, 2026, 3:51 a.m. |
Created at: April 10, 2026, 5:47 a.m.