@inproceedings{li-etal-2026-select,
title = "Select Before Use: On the Importance of Reference Model Selection in Preference Alignment",
author = "Li, Muyang and
Wu, Runze and
Zhao, Xiangyu and
Han, Bo and
Dong, Daoyi and
Liu, Tongliang",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.acl-long.780/",
pages = "17151--17171",
ISBN = "979-8-89176-390-6",
abstract = "The post-training stage of Large Language Models (LLMs) typically involves Supervised Fine-Tuning (SFT) followed by preference alignment to ensure LLM to generate safe, helpful, and instruction-aligned content. The SFT model critically serves as both the initialization and reference model for subsequent preference alignment. However, an essential yet often neglected question is the optimal selection of the SFT checkpoint for this role. We show that checkpoint selection substantially affects final performance, and that the common practice of choosing the minimum validation-loss checkpoint often fails, due to a fundamental conflict between SFT{'}s focus on imitation and alignment{'}s goal of response discriminability. To this end, we propose RewardRank, a simple, effective, training-free metrics for estimating initial implicit alignment between reference model and preference objective. Empirical evidence suggests that, using our selected model as reference can gain up to 67.6{\%} relative increase on length-controlled win rate on the popular Zephyr recipe comparing to baselines."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="li-etal-2026-select">
<titleInfo>
<title>Select Before Use: On the Importance of Reference Model Selection in Preference Alignment</title>
</titleInfo>
<name type="personal">
<namePart type="given">Muyang</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Runze</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiangyu</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bo</namePart>
<namePart type="family">Han</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Daoyi</namePart>
<namePart type="family">Dong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tongliang</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Liakata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viviane</namePart>
<namePart type="given">P</namePart>
<namePart type="family">Moreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Jurgens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-390-6</identifier>
</relatedItem>
<abstract>The post-training stage of Large Language Models (LLMs) typically involves Supervised Fine-Tuning (SFT) followed by preference alignment to ensure LLM to generate safe, helpful, and instruction-aligned content. The SFT model critically serves as both the initialization and reference model for subsequent preference alignment. However, an essential yet often neglected question is the optimal selection of the SFT checkpoint for this role. We show that checkpoint selection substantially affects final performance, and that the common practice of choosing the minimum validation-loss checkpoint often fails, due to a fundamental conflict between SFT’s focus on imitation and alignment’s goal of response discriminability. To this end, we propose RewardRank, a simple, effective, training-free metrics for estimating initial implicit alignment between reference model and preference objective. Empirical evidence suggests that, using our selected model as reference can gain up to 67.6% relative increase on length-controlled win rate on the popular Zephyr recipe comparing to baselines.</abstract>
<identifier type="citekey">li-etal-2026-select</identifier>
<location>
<url>https://aclanthology.org/2026.acl-long.780/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>17151</start>
<end>17171</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Select Before Use: On the Importance of Reference Model Selection in Preference Alignment
%A Li, Muyang
%A Wu, Runze
%A Zhao, Xiangyu
%A Han, Bo
%A Dong, Daoyi
%A Liu, Tongliang
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-390-6
%F li-etal-2026-select
%X The post-training stage of Large Language Models (LLMs) typically involves Supervised Fine-Tuning (SFT) followed by preference alignment to ensure LLM to generate safe, helpful, and instruction-aligned content. The SFT model critically serves as both the initialization and reference model for subsequent preference alignment. However, an essential yet often neglected question is the optimal selection of the SFT checkpoint for this role. We show that checkpoint selection substantially affects final performance, and that the common practice of choosing the minimum validation-loss checkpoint often fails, due to a fundamental conflict between SFT’s focus on imitation and alignment’s goal of response discriminability. To this end, we propose RewardRank, a simple, effective, training-free metrics for estimating initial implicit alignment between reference model and preference objective. Empirical evidence suggests that, using our selected model as reference can gain up to 67.6% relative increase on length-controlled win rate on the popular Zephyr recipe comparing to baselines.
%U https://aclanthology.org/2026.acl-long.780/
%P 17151-17171
Markdown (Informal)
[Select Before Use: On the Importance of Reference Model Selection in Preference Alignment](https://aclanthology.org/2026.acl-long.780/) (Li et al., ACL 2026)
ACL