Triple
T2600760
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | Karachi West District |
E58336
|
entity |
| Predicate | contains |
P35
|
FINISHED |
| Object | Orangi Town |
E280151
|
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: Orangi Town | Statement: [Karachi West District, contains, Orangi Town]
NED1
Entity disambiguation (via context triple)
gpt-5-mini-2025-08-07
Target entity: Orangi Town Context triple: [Karachi West District, contains, Orangi Town]
-
A.
Orangi Town
chosen
Orangi Town is a densely populated residential area in Karachi, Pakistan, known as one of Asia’s largest informal settlements.
-
B.
Zaman Town
Zaman Town is a residential neighborhood located within the Korangi District of Karachi, Pakistan.
-
C.
Congo Town
Congo Town is a small settlement on South Andros in The Bahamas, known as a local administrative and transportation hub with its own regional airport.
-
D.
Gladysvale
Gladysvale is a fossil-bearing cave site in South Africa known for its important hominin and animal fossil discoveries within the Cradle of Humankind World Heritage area.
-
E.
Soweto
Soweto is a historically significant township in Johannesburg, South Africa, known for its central role in the struggle against apartheid and its rich urban culture.
- 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_69ab4ac14040819098b13f4a27d5c8ff |
completed | March 6, 2026, 9:44 p.m. |
| NER | Named-entity recognition | batch_69abd4587014819089f78e93adf2144c |
completed | March 7, 2026, 7:31 a.m. |
| NED1 | Entity disambiguation (via context triple) | batch_69af83d37de081909467f8caa17ce3a9 |
completed | March 10, 2026, 2:37 a.m. |
Created at: March 6, 2026, 9:49 p.m.