Triple
T27762500
| Position | Surface form | Disambiguated ID | Type / Status |
|---|---|---|---|
| Subject | Transformer-XL |
E701503
|
entity |
| Predicate | improvesMetric |
P161052
|
FINISHED |
| Object | perplexity on language modeling benchmarks |
—
|
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: perplexity on language modeling benchmarks | Statement: [Transformer-XL, improvesMetric, perplexity on language modeling benchmarks]
PD
Predicate disambiguation
gpt-5-mini-2025-08-07
Target predicate: improvesMetric Context triple: [Transformer-XL, improvesMetric, perplexity on language modeling benchmarks]
-
A.
improvesIn
chosen
Indicates that one entity causes or contributes to an increase in quality, performance, or effectiveness in another entity or context.
-
B.
improvesOn
Indicates that one entity enhances, refines, or performs better than another entity, typically by addressing its limitations or increasing its effectiveness.
-
C.
hasImproved
Indicates that an entity’s state, quality, or performance has become better compared to a previous point in time.
-
D.
seeksToImprove
Indicates an intentional effort by one entity to make another entity or condition better than its current state.
-
E.
successMetric
Indicates the specific criterion or measure used to evaluate whether an action, process, or relationship has achieved its intended success.
- 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_69ef6a5193808190816eb7d0020b2d87 |
completed | April 27, 2026, 1:53 p.m. |
| NER | Named-entity recognition | batch_69f63766dcc48190a916106fbf8ad619 |
completed | May 2, 2026, 5:41 p.m. |
| PD | Predicate disambiguation | batch_69f63188e7408190af8ce8b93d128c63 |
completed | May 2, 2026, 5:16 p.m. |
Created at: April 27, 2026, 4:28 p.m.