Triple
T8507749
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | DRAM |
E201374
|
entity |
| Predicate | dataRetentionWithoutPower |
P31904
|
FINISHED |
| Object | very short |
—
|
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: very short | Statement: [DRAM, dataRetentionWithoutPower, very short]
PD
Predicate disambiguation
gpt-5-mini-2025-08-07
Target predicate: dataRetentionWithoutPower Context triple: [DRAM, dataRetentionWithoutPower, very short]
-
A.
lostPowerDate
Indicates the date on which an entity lost electrical power or ceased to have power available.
-
B.
maintainsPowerThrough
Indicates that one entity preserves or secures its authority, control, or dominance by means of another entity or specified mechanism.
-
C.
supportsPersistence
chosen
Indicates that one entity enables or provides the capability for another entity’s data or state to be stored and retained over time.
-
D.
retentionMethod
Indicates the method or strategy used to retain or keep something (such as data, customers, or resources) over time.
-
E.
lostPowerDuring
Indicates that an entity ceased to have electrical or functional power at some time during a specified event or interval.
- 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_69ca8320e5748190ac2c585a0bba8193 |
completed | March 30, 2026, 2:05 p.m. |
| NER | Named-entity recognition | batch_69cbe5de18448190a695eec609b34e1a |
completed | March 31, 2026, 3:18 p.m. |
| PD | Predicate disambiguation | batch_69cbd10cfd208190a519049fad32c508 |
completed | March 31, 2026, 1:50 p.m. |
Created at: March 30, 2026, 6:14 p.m.