Docs
How Plurality works.
Three short reads. The methodology behind the scoring, a practical tour of every mode, and what we learned grading 180 queries on prompt quality.
- 01
How Plurality measures consensus
The technical paper. Two independent agreement signals, momentum tracking, outlier detection, and how rounds change positions.
- 02
Modes guide — what each mode does
A practical walk-through of all seven modes: when to use each, expected output shape, panel sizing, and rough cost per query.
- 03
Does prompt engineering actually matter?
An empirical experiment: 12 questions × 3 prompt tiers × 5 frontier models × 1 judge. What 180 graded queries told us about prompt quality.