Triple

T22093644
Position Surface form Disambiguated ID Type / Status
Subject AVP: Requiem E545968 entity
Predicate stars P1956 FINISHED
Object Ariel Gade NE NERFINISHED

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: Ariel Gade | Statement: [AVP: Requiem, stars, Ariel Gade]
NED1 Entity disambiguation (via context triple) gpt-5-mini-2025-08-07
Target entity: Ariel Gade
Context triple: [AVP: Requiem, stars, Ariel Gade]
  • A. Ariel Gade chosen
    Ariel Gade is an American former child actress best known for her roles in films like "Dark Water" and "Aliens in the Attic" and for her work on television.
  • B. Ariel Shepherd-Oppenheim
    Ariel Shepherd-Oppenheim is an American writer and the daughter of actress Cybill Shepherd.
  • C. Ariel Moore
    Ariel Moore is the rebellious and spirited teenage daughter of a strict minister in the musical film "Footloose," whose desire for freedom and dancing drives much of the story’s conflict.
  • D. Ariel Herbert-Voss
    Ariel Herbert-Voss is a machine learning researcher known for her work at OpenAI on large language models and AI safety.
  • E. Jonna Mendez
    Jonna Mendez is a former CIA chief of disguise and intelligence officer known for her expertise in espionage tradecraft and co-authoring books on intelligence with her late husband Tony Mendez.
  • F. None of above.
  • G. Unsure - the case is ambiguous/there is not enough information to decide.

Provenance (2 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_69e11e36d03c8190a83a1ba802b7231b completed April 16, 2026, 5:36 p.m.
NER Named-entity recognition batch_69f128e766388190aad1039fe0849771 completed April 28, 2026, 9:38 p.m.
Created at: April 16, 2026, 8:29 p.m.