Triple
T16076496
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | AMD Instinct MI250X |
E389990
|
entity |
| Predicate | singlePrecisionPerformancePeak |
P33100
|
FINISHED |
| Object | ~95.7 TFLOPS |
—
|
LITERAL 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: ~95.7 TFLOPS | Statement: [AMD Instinct MI250X, singlePrecisionPerformancePeak, ~95.7 TFLOPS]
PD
Predicate disambiguation
gpt-5-mini-2025-08-07
Target predicate: singlePrecisionPerformancePeak Context triple: [AMD Instinct MI250X, singlePrecisionPerformancePeak, ~95.7 TFLOPS]
-
A.
floatingPointPerformance
Indicates the level of computational capability or efficiency an entity has when performing floating-point arithmetic operations.
-
B.
numberOfFloatingPointUnits
Indicates the quantity of floating-point processing units associated with or contained in an entity.
-
C.
floatingPointRegisterCount
Indicates the number of floating-point registers associated with an entity (such as a processor or execution context).
-
D.
gpuComputePerformance
chosen
Indicates the level of processing capability a GPU can deliver for computational tasks, typically measured in operations per unit time.
-
E.
LINPACKPerformance
Indicates the level of computational performance achieved by an entity when running the LINPACK benchmark, typically measured in floating-point operations per second.
- F. None of above.
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_69d86daf32ec8190a8c0466c8f49c3c0 |
completed | April 10, 2026, 3:25 a.m. |
| NER | Named-entity recognition | batch_69e1ff63edb0819092cbb671967bbdcd |
completed | April 17, 2026, 9:37 a.m. |
| PD | Predicate disambiguation | batch_69e1827ad7c88190b867da511cbfb7fa |
completed | April 17, 2026, 12:44 a.m. |
Created at: April 10, 2026, 4:57 a.m.