Triple
T6618221
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | Fumihiko |
E149607
|
entity |
| Predicate | hasScript |
P182
|
FINISHED |
| Object | Katakana |
E154499
|
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: Katakana | Statement: [Fumihiko, hasScript, Katakana]
NED1
Entity disambiguation (via context triple)
gpt-5-mini-2025-08-07
Target entity: Katakana Context triple: [Fumihiko, hasScript, Katakana]
-
A.
Katakana
chosen
Katakana is one of the two main Japanese phonetic writing systems, primarily used for foreign words, onomatopoeia, emphasis, and technical or scientific terms.
-
B.
Hiragana
Hiragana is a Japanese phonetic syllabary used primarily for native words, grammatical elements, and beginners’ reading and writing.
-
C.
Kana
Kana is the Japanese syllabic writing system comprising hiragana and katakana, used to represent native words, grammatical elements, and foreign terms.
-
D.
Kanji
Kanji are logographic characters of Chinese origin used in the Japanese writing system alongside hiragana and katakana.
-
E.
Kawi script
Kawi script is an ancient Brahmic-derived writing system historically used across Java and other parts of Southeast Asia to write Old Javanese and related languages.
- 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_69c687ed8a9c81908bb671717cb192ef |
completed | March 27, 2026, 1:36 p.m. |
| NER | Named-entity recognition | batch_69c6af5b21348190b7f09045e9ec7d63 |
completed | March 27, 2026, 4:24 p.m. |
| NED1 | Entity disambiguation (via context triple) | batch_69c6e443df8c8190b52ecac5a7e9fb09 |
completed | March 27, 2026, 8:10 p.m. |
Created at: March 27, 2026, 1:58 p.m.