Triple
T14016879
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | Stalin’s Bunker |
E337225
|
entity |
| Predicate | securityLevelAtConstruction |
P84572
|
FINISHED |
| Object | top secret |
—
|
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: top secret | Statement: [Stalin’s Bunker, securityLevelAtConstruction, top secret]
PD
Predicate disambiguation
gpt-5-mini-2025-08-07
Target predicate: securityLevelAtConstruction Context triple: [Stalin’s Bunker, securityLevelAtConstruction, top secret]
-
A.
securityLevelDetail
chosen
Indicates the specific classification or degree of security associated with an entity, often including nuanced or descriptive information about its protection level.
-
B.
securityLevelComponents
Indicates that a particular security level is composed of, or associated with, a specific set of component elements or factors.
-
C.
securityLevelChangedFrom
Indicates that an entity’s security level has been modified from a specified previous level to a different one.
-
D.
securityGuarantee
Indicates a commitment by one party to protect or defend another party against specified threats or risks.
-
E.
encryptionLevel
Indicates the degree or strength of cryptographic protection applied to data or communications.
- 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_69d81c6543a48190bd5ba93d7419e797 |
completed | April 9, 2026, 9:38 p.m. |
| NER | Named-entity recognition | batch_69de2f396b648190927e5718c3bb6511 |
completed | April 14, 2026, 12:12 p.m. |
| PD | Predicate disambiguation | batch_69de05a802ac819090604025aae6a4d5 |
completed | April 14, 2026, 9:15 a.m. |
Created at: April 9, 2026, 10:19 p.m.