Triple
T7564135
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | Cassel |
E178867
|
entity |
| Predicate | near |
P350
|
FINISHED |
| Object | Dunkirk |
E33602
|
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: Dunkirk | Statement: [Cassel, near, Dunkirk]
NED1
Entity disambiguation (via context triple)
gpt-5-mini-2025-08-07
Target entity: Dunkirk Context triple: [Cassel, near, Dunkirk]
-
A.
Dunkirk
Dunkirk is a 2017 war film directed by Christopher Nolan that depicts the World War II evacuation of Allied soldiers from the beaches of Dunkirk, France.
-
B.
Dunkirk
chosen
Dunkirk is a coastal town in northern France best known as the site of the World War II evacuation of Allied troops in 1940.
-
C.
Darkest Hour
Darkest Hour is a 2017 historical drama film depicting Winston Churchill’s early days as British Prime Minister during World War II, starring Gary Oldman in an Oscar-winning performance.
-
D.
Downfall
Downfall is the codename for the planned but never executed Allied invasion of the Japanese home islands near the end of World War II.
-
E.
Downfall
Downfall is a 2004 German historical war drama film depicting the final days of Adolf Hitler and the Third Reich in a besieged Berlin bunker.
- 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_69c69f2f80288190b95cceb4da92ab2b |
completed | March 27, 2026, 3:15 p.m. |
| NER | Named-entity recognition | batch_69c6f8fb1c00819096bdb73334d8d72e |
completed | March 27, 2026, 9:39 p.m. |
| NED1 | Entity disambiguation (via context triple) | batch_69c856d8206081909987556fb6ab7084 |
completed | March 28, 2026, 10:31 p.m. |
Created at: March 27, 2026, 3:50 p.m.