Triple

T15997099
Position Surface form Disambiguated ID Type / Status
Subject Roxanne E387996 entity
Predicate producer P490 FINISHED
Object Michael Rachmil 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: Michael Rachmil | Statement: [Roxanne, producer, Michael Rachmil]
NED1 Entity disambiguation (via context triple) gpt-5-mini-2025-08-07
Target entity: Michael Rachmil
Context triple: [Roxanne, producer, Michael Rachmil]
  • A. Michael Rachmil chosen
    Michael Rachmil is a film producer best known for his work on the 1987 romantic comedy "Roxanne" starring Steve Martin.
  • B. Michael Kagan
    Michael Kagan is an Israeli technologist and entrepreneur best known as the co-founder and longtime chief technology officer of high-performance networking company Mellanox Technologies.
  • C. Michael Bluestein
    Michael Bluestein is an American keyboardist and songwriter best known as a longtime member of the rock band Foreigner.
  • D. Michael Starobin
    Michael Starobin is a Tony Award–winning American orchestrator and arranger known for his work on numerous Broadway musicals.
  • E. Michael Shvo
    Michael Shvo is a high-profile real estate developer and art collector known for leading luxury property projects in major global cities.
  • 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_69d86daa562c81908aacc179c0fe8fb5 completed April 10, 2026, 3:25 a.m.
NER Named-entity recognition batch_69e157882ef0819081143e530bd6413c completed April 16, 2026, 9:41 p.m.
Created at: April 10, 2026, 4:55 a.m.