Triple
T2882524
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | Red Skull |
E59428
|
entity |
| Predicate | roleInStories |
P42552
|
FINISHED |
| Object | primary antagonist of Captain America |
—
|
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: primary antagonist of Captain America | Statement: [Red Skull, roleInStories, primary antagonist of Captain America]
PD
Predicate disambiguation
gpt-5-mini-2025-08-07
Target predicate: roleInStories Context triple: [Red Skull, roleInStories, primary antagonist of Captain America]
-
A.
narratorRole
Indicates that one entity serves as the narrator of another entity (such as a story, text, or media work), specifying the narrative role or function it performs.
-
B.
roleInScene
Indicates that an entity participates in a particular scene with a specific role or function within that scene.
-
C.
roleInDialogue
Indicates that an entity participates in a dialogue with a specific conversational role (e.g., speaker, listener, moderator) relative to other participants.
-
D.
roleInText
Indicates that an entity participates in a text with a specific function or capacity (e.g., author, editor, character).
-
E.
creativeRole
Indicates that an entity holds a specific creative function or responsibility in relation to another entity, such as a work or project.
- 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_69ab4ac739188190a112f42a5a69c951 |
completed | March 6, 2026, 9:44 p.m. |
| NER | Named-entity recognition | batch_69abe02c238881908f7a349563c388bf |
completed | March 7, 2026, 8:22 a.m. |
| PD | Predicate disambiguation | batch_69abdd15cbf08190bf7fea5ea516848a |
completed | March 7, 2026, 8:08 a.m. |
| PDg | Predicate description generation | batch_69abdd96670c8190b727f9ac27dadf67 |
completed | March 7, 2026, 8:11 a.m. |
Created at: March 6, 2026, 10:03 p.m.