Multi-Stakeholder LLM Alignment: Decomposing Estimation from Aggregation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zheng, Lulu, Yang, Wenjin, Zhang, Xiangwen, Yin, Rong, Hu, Yulan, Pan, Zheng, Li, Xin
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911719846051840
author Zheng, Lulu
Yang, Wenjin
Zhang, Xiangwen
Yin, Rong
Hu, Yulan
Pan, Zheng
Li, Xin
author_facet Zheng, Lulu
Yang, Wenjin
Zhang, Xiangwen
Yin, Rong
Hu, Yulan
Pan, Zheng
Li, Xin
contents Multi-stakeholder tasks require one output to satisfy users with conflicting preferences. Holistic LLM judges conflate utility estimation and utility aggregation, yielding unstable implicit weights. We show empirically and theoretically that this aggregation-specific \emph{weighting noise} can create large score shifts when stakeholder satisfaction is dispersed; in our experiments, these weight-induced shifts also increase with stakeholder count. We propose \textsc{DecompR}: counterfactual-calibrated weights are fixed from query structure before candidate scoring, while per-role utilities are estimated independently, removing candidate-dependent weight drift and reducing estimation noise.
format Preprint
id arxiv_https___arxiv_org_abs_2605_26878
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Multi-Stakeholder LLM Alignment: Decomposing Estimation from Aggregation
Zheng, Lulu
Yang, Wenjin
Zhang, Xiangwen
Yin, Rong
Hu, Yulan
Pan, Zheng
Li, Xin
Artificial Intelligence
Multi-stakeholder tasks require one output to satisfy users with conflicting preferences. Holistic LLM judges conflate utility estimation and utility aggregation, yielding unstable implicit weights. We show empirically and theoretically that this aggregation-specific \emph{weighting noise} can create large score shifts when stakeholder satisfaction is dispersed; in our experiments, these weight-induced shifts also increase with stakeholder count. We propose \textsc{DecompR}: counterfactual-calibrated weights are fixed from query structure before candidate scoring, while per-role utilities are estimated independently, removing candidate-dependent weight drift and reducing estimation noise.
title Multi-Stakeholder LLM Alignment: Decomposing Estimation from Aggregation
topic Artificial Intelligence
url https://arxiv.org/abs/2605.26878