Classification Confidence Voting

U

@

·

Determine classifications through voting across multiple runs.

55 copies0 forks
Classify {{samples}} using {{model}} with voting consensus.

For each sample, run classification {{vote_count}} times:
- Vary temperature slightly each run
- Record predicted class per run
- Calculate vote distribution

Assign final class by majority vote. Report confidence as vote percentage. Flag low-confidence classifications below {{confidence_threshold}}.

Details

Category

Analysis

Use Cases

Classification votingConfidence estimationReliable labeling

Works Best With

claude-opus-4.5gpt-5.2gemini-2.0-flash
Created Shared

Create your own prompt vault and start sharing