Triple

T14805734
Position Surface form Disambiguated ID Type / Status
Subject Dark Water E348029 entity
Predicate producer P490 FINISHED
Object Doug Davison E1056082 NE 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: Doug Davison | Statement: [Dark Water, producer, Doug Davison]
NED1 Entity disambiguation (via context triple) gpt-5-mini-2025-08-07
Target entity: Doug Davison
Context triple: [Dark Water, producer, Doug Davison]
  • A. Doug Davison chosen
    Doug Davison is a film and television producer known for working on major Hollywood projects such as "The Lake House."
  • B. Scott Davidson
    Scott Davidson was a New York City firefighter who died in the September 11, 2001 attacks and is known as the late father of comedian Pete Davidson.
  • C. Austin Corbett
    Austin Corbett is an American football offensive lineman who played college football at Nevada and went on to a professional career in the NFL.
  • D. Mitch Seavey
    Mitch Seavey is an American dog musher best known as a multiple-time Iditarod Trail Sled Dog Race champion from Alaska.
  • E. Kent Davison
    Kent Davison is a data-obsessed political strategist and senior advisor in the television series "Veep," known for his emotionless demeanor and reliance on polling and analytics.
  • F. None of above.
  • G. Unsure - the case is ambiguous/there is not enough information to decide.

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_69d822ea8b7c819097dfadf3d45545e6 completed April 9, 2026, 10:06 p.m.
NER Named-entity recognition batch_69decf32666081908e84f985c47eb963 completed April 14, 2026, 11:35 p.m.
NED1 Entity disambiguation (via context triple) batch_69fe24c6b3008190a0fac1dace40361a completed May 8, 2026, 6 p.m.
Created at: April 10, 2026, 1:34 a.m.