Triple
T3578456
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | Koshi River |
E75742
|
entity |
| Predicate | barragePurpose |
P79
|
FINISHED |
| Object | flood control |
—
|
LITERAL 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: flood control | Statement: [Koshi River, barragePurpose, flood control]
PD
Predicate disambiguation
gpt-5-mini-2025-08-07
Target predicate: barragePurpose Context triple: [Koshi River, barragePurpose, flood control]
-
A.
barrageLocation
Indicates the location where a barrage (such as an artillery or concentrated attack) is directed or occurs.
-
B.
hasBarrage
Indicates that one entity possesses, contains, or is equipped with a barrage (such as a concentrated, sustained sequence of attacks, shots, or similar actions).
-
C.
purpose
chosen
Indicates that one entity exists, is done, or is used in order to achieve, support, or serve the goal, function, or intended outcome of another entity.
-
D.
channels
Indicates that one entity serves as a medium, route, or conduit through which another entity is directed, transmitted, or delivered.
-
E.
broadcastPlatform
Indicates the platform or channel through which content is broadcast or transmitted.
- F. None of above.
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_69ad85d5e3008190bdfe0bacdd1f5a1b |
completed | March 8, 2026, 2:21 p.m. |
| NER | Named-entity recognition | batch_69adc0defe14819095a337a840e33300 |
completed | March 8, 2026, 6:33 p.m. |
| PD | Predicate disambiguation | batch_69adb83810c481909c645c08b978edc1 |
completed | March 8, 2026, 5:56 p.m. |
Created at: March 8, 2026, 3:21 p.m.