Triple
T263643
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | Federal Insurance Contributions Act taxes |
E5806
|
entity |
| Predicate | collectedThrough |
P9384
|
FINISHED |
| Object | payroll withholding |
—
|
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: payroll withholding | Statement: [Federal Insurance Contributions Act taxes, collectedThrough, payroll withholding]
PD
Predicate disambiguation
gpt-5-mini-2025-08-07
Target predicate: collectedThrough Context triple: [Federal Insurance Contributions Act taxes, collectedThrough, payroll withholding]
-
A.
collectsFrom
Indicates that one entity gathers, receives, or takes something (such as items, data, or payments) from another entity.
-
B.
collects
Indicates that one entity gathers, accumulates, or brings together one or more other entities into its possession or control.
-
C.
dataTypeCollected
Indicates that a specific type or category of data is gathered or recorded in the context of an entity or process.
-
D.
incorporatedThrough
Indicates that one entity was formed, established, or legally created by means of another specified process, mechanism, or intermediary entity.
-
E.
collectorStatus
Indicates the current operational or lifecycle state of a collector in relation to its collection activity.
- F. None of above. chosen
Provenance (4 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_69a258dd8ea08190ac554a1cc8dfd8c3 |
completed | Feb. 28, 2026, 2:54 a.m. |
| NER | Named-entity recognition | batch_69a25d8e809881908a58c9a4e3ba07c3 |
completed | Feb. 28, 2026, 3:14 a.m. |
| PD | Predicate disambiguation | batch_69a25b6e07748190834022a65ba6d803 |
completed | Feb. 28, 2026, 3:05 a.m. |
| PDg | Predicate description generation | batch_69a25d0ec71081908478c800be4f7bb0 |
completed | Feb. 28, 2026, 3:12 a.m. |
Created at: Feb. 28, 2026, 2:55 a.m.