Triple

T3874392
Position Surface form Disambiguated ID Type / Status
Subject David Letterman E92462 entity
Predicate child P120 FINISHED
Object Harry Joseph Letterman E92462 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: Harry Joseph Letterman | Statement: [David Letterman, child, Harry Joseph Letterman]
NED1 Entity disambiguation (via context triple) gpt-5-mini-2025-08-07
Target entity: Harry Joseph Letterman
Context triple: [David Letterman, child, Harry Joseph Letterman]
  • A. Harry Joseph Letterman chosen
    Harry Joseph Letterman is the son of American television host and comedian David Letterman.
  • B. Greg Gumbel
    Greg Gumbel is an American sportscaster best known for his long tenure with CBS Sports, where he has called NFL, NCAA basketball, and other major sporting events.
  • C. Jack Paar
    Jack Paar was an influential American television host and comedian best known for transforming late-night TV during his tenure on The Tonight Show in the late 1950s and early 1960s.
  • D. Jack Buck
    Jack Buck was a renowned American sportscaster best known for his long tenure as the voice of the St. Louis Cardinals and his iconic calls in Major League Baseball.
  • E. Don Hewitt
    Don Hewitt was an American television news producer best known for pioneering modern TV newsmagazine journalism.
  • 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_69aed967448c819086c4b358d37b25aa completed March 9, 2026, 2:29 p.m.
NER Named-entity recognition batch_69aeec59bea08190b1e193f34944a2ee completed March 9, 2026, 3:50 p.m.
NED1 Entity disambiguation (via context triple) batch_69b5124cdf588190b3b83ee8fb29450a completed March 14, 2026, 7:46 a.m.
Created at: March 9, 2026, 3:20 p.m.