AI Paper Atlas

Research map / Large language models / Alignment and preferences

Unpacking Human Preference for LLMs: Demographically Aware Evaluation with the HUMAINE Framework

Nora Petrova, Andrew Gordon, Enzo Blindow

ICLR 2026 · Large language models › Alignment and preferences · approach: LLM-as-judge and feedback data · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use Overall Winner dimension instead of Trust, Ethics & Safetypairwise human preference evaluation · HUMAINEtie rate (discriminative power): 65% -> 10%Improved
compare older (55+) vs younger (18-34) ratersmodel rankings · HUMAINEtie rate: 9.7% -> 12.5%Hurt
vary rater age groupmodel rankings · HUMAINEaverage rank shift: age 2.8 vs ethnicity 1.3 vs politics 1.5Mixed
break results down by dimension rather than overallgemini-2.5-pro ranking · HUMAINErank: top rank across all dimensionsNo clear effect
break results down by dimensiongrok-3 / magistral-medium ranking · HUMAINErank across dimensions: grok-3 2nd on Core Task vs 8th on Communication Style; magistral 2nd on Interaction Fluidity vs 12th on TrustMixed
move age cohort from 18-34 to 55+magistral-medium-2506 · HUMAINErank (US): 1st -> 5thHurt
add interaction terms to additive demographic model (ablation)hierarchical BTD · HUMAINEvariance explained by interaction: 7-41% US, 29-50% UKMixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy