Saved in:
Bibliographic Details
Main Authors: Hall, Zara, Subbiah, Melanie, Zollo, Thomas P, McKeown, Kathleen, Zemel, Richard
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2507.11344
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909690060865536
author Hall, Zara
Subbiah, Melanie
Zollo, Thomas P
McKeown, Kathleen
Zemel, Richard
author_facet Hall, Zara
Subbiah, Melanie
Zollo, Thomas P
McKeown, Kathleen
Zemel, Richard
contents Large language models are increasingly used to support high-stakes decisions, potentially influencing who is granted bail or receives a loan. Naive chain-of-thought sampling can improve average decision accuracy, but has also been shown to amplify unfair bias. To address this challenge and enable the trustworthy use of reasoning models in high-stakes decision-making, we propose a framework for training a generalizable Fairness Reward Model (FRM). Our model assigns a fairness score to LLM reasoning, enabling the system to down-weight biased trajectories and favor equitable ones when aggregating decisions across reasoning chains. We show that a single Fairness Reward Model, trained on weakly supervised, LLM-annotated examples of biased versus unbiased reasoning, transfers across tasks, domains, and model families without additional fine-tuning. Applied to real-world decision-making tasks including recidivism prediction and social media moderation, we show that our approach consistently improves fairness while matching, or even surpassing, baseline accuracy.
format Preprint
id arxiv_https___arxiv_org_abs_2507_11344
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Guiding LLM Decision-Making with Fairness Reward Models
Hall, Zara
Subbiah, Melanie
Zollo, Thomas P
McKeown, Kathleen
Zemel, Richard
Machine Learning
Large language models are increasingly used to support high-stakes decisions, potentially influencing who is granted bail or receives a loan. Naive chain-of-thought sampling can improve average decision accuracy, but has also been shown to amplify unfair bias. To address this challenge and enable the trustworthy use of reasoning models in high-stakes decision-making, we propose a framework for training a generalizable Fairness Reward Model (FRM). Our model assigns a fairness score to LLM reasoning, enabling the system to down-weight biased trajectories and favor equitable ones when aggregating decisions across reasoning chains. We show that a single Fairness Reward Model, trained on weakly supervised, LLM-annotated examples of biased versus unbiased reasoning, transfers across tasks, domains, and model families without additional fine-tuning. Applied to real-world decision-making tasks including recidivism prediction and social media moderation, we show that our approach consistently improves fairness while matching, or even surpassing, baseline accuracy.
title Guiding LLM Decision-Making with Fairness Reward Models
topic Machine Learning
url https://arxiv.org/abs/2507.11344