Triple

T2179326
Position Surface form Disambiguated ID Type / Status
Subject Harold and Maude E49003 entity
Predicate director P255 FINISHED
Object Hal Ashby E50627 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: Hal Ashby | Statement: [Harold and Maude, director, Hal Ashby]
NED1 Entity disambiguation (via context triple) gpt-5-mini-2025-08-07
Target entity: Hal Ashby
Context triple: [Harold and Maude, director, Hal Ashby]
  • A. Hal Ashby chosen
    Hal Ashby was an influential American film director and editor of the New Hollywood era, known for acclaimed, offbeat classics such as "Harold and Maude," "Shampoo," and "Being There."
  • B. Ken Hughes
    Ken Hughes was a British film director and screenwriter best known for directing the 1968 musical fantasy film "Chitty Chitty Bang Bang."
  • C. Robert A. Altman
    Robert A. Altman was an American lawyer and businessman best known as the co-founder and longtime CEO of video game publisher ZeniMax Media, the parent company of Bethesda Softworks.
  • D. Tom Benedek
    Tom Benedek is an American screenwriter best known for co-writing the science fiction film "Cocoon."
  • E. Alan J. Pakula
    Alan J. Pakula was an American film director, producer, and screenwriter best known for his politically charged thrillers and character-driven dramas of the 1970s and 1980s.
  • 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_69a88aa72d348190a9544bb5b8a4e71d completed March 4, 2026, 7:40 p.m.
NER Named-entity recognition batch_69abbeefb7648190b8fc57cf60553579 completed March 7, 2026, 6 a.m.
NED1 Entity disambiguation (via context triple) batch_69ae71b142208190a09c8e459b200ecb completed March 9, 2026, 7:07 a.m.
Created at: March 4, 2026, 7:45 p.m.