Tokens processed, suspect part
The part of a platform's published day total (
inference_tokens) that matches the rule for counts metered traffic does not produce (Dot: a model's day with input and output both whole multiples of 100,000 from 1M input up, or a model total of 10M or more within 1,000 of a whole million, plus any remainder of the day total no model accounts for, when 1,000 or more; the Inference lab's rule). 0 when nothing matches. The total keeps these tokens and carries the suspect flag; leaving them out is the derive step's work. Daily, UTC; the running day is partial.Definition
At a glance
- Key
inference_tokens_suspect- Unit
- Tokens
- Aggregation
- SumA flow: weekly and monthly values add up the daily values in the period.
- Can be negative
- NoNegative values are treated as data errors and held back by the range check.
Sources
Where it comes from
Datasets that write this metric. When several sources provide it, each chart states which one it uses.
| Dataset | Source |
|---|---|
Dot tokens as published, with the suspect part (daily)dot.usage | Dot |
Usage
Dashboards using it
No published dashboard uses this metric yet.
Time buckets, partial periods and restatements follow the site-wide conventions.