Prototype inference cost of the deterministic baseline
MeasuredEV-005
Produced by running code in this repository and preserved as an artifact.
- Scope
- One run of twelve held-out cases.
- Observation
- Measured spend was $0.00 and token usage was null, because the method makes no model calls. Per-case latency was p50 0.03 ms and p95 0.94 ms over twelve cases on the author's machine.
- Limitations
- Zero inference cost is a property of a method that does no inference. It says nothing about the cost of a model-based method and nothing about total cost of ownership.
- Source
experiments/runs/tagged-transcript-baseline--held-out--2026-09-07.1/run.json- Source date
- 2026-09-07
- Retrieved
- 2026-09-07
- Reviewer status
- Pending review. Written down, not yet checked by a person.