VALUES-CONTEST S11 · Values, pluralism and paradigm physiology

Contested-direction loading predicts human rater disagreement

Inconclusive

frozen 2026-07-18 · adjudicated 2026-07-19 · ev:e036326c47c7a0f52330559479e03790 · spec:2a1ead40948d8

The frozen predictions

  1. H-contest contested-direction loading predicts per-item human rater disagreement
    contested_raterspread_spearman > 0.2 registered p 0.55 not measured Inconclusive
  2. H-contest-vs-ensemble single-model contested loading comes within 0.05 of the 5x-cost ensemble baseline
    ensemble_minus_contested_spearman < 0.05 not measured Inconclusive
Adjudication figure for VALUES-CONTEST
VALUES-CONTEST was not adjudicated this run: missing intermediate: no intermediate 'campaign.scores' for roster_key='armorm' slice='helpsteer3-pref-2000'; the arc that produces it has not run or its shard was not merged.

Subjects and artifacts