@inproceedings{wang-etal-2026-mtp,
title = "{MTP}-{RL}: Acceleration of Reinforcement Learning Rollouts with Policy-Aligned Multi-Token Prediction",
author = "Wang, Ke and
Zeng, Aohan and
Du, Zhengxiao and
Hu, Yuxuan and
Zhang, Bohan and
Wang, Xinyi and
Tang, Jie and
Zhang, Jing",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-acl.1871/",
pages = "37530--37542",
ISBN = "979-8-89176-395-1",
abstract = "Reinforcement learning (RL) is widely applied to boost the performance of pretrained models, yet its training efficiency is severely constrained by rollout generation. While speculative decoding based on multi-token prediction (MTP) offers a potential acceleration pathway, its widespread adoption is hindered by the absence of MTP in vanilla pretrained models and the rapid degradation of the MTP acceptance length in RL training. To address these issues, this paper proposes MTP-RL, a two-stage framework that pioneers effective training of MTPs in RL and accelerates the rollout phase for diverse models. It involves a pipeline to equip the multi-layer parameter-sharing MTP for all models and an innovative advantage-aware MTP optimization strategy to facilitate policy-aligned training of MTPs. Experiments demonstrate that our method not only achieves stable growth of acceptance length during RL training, but also accelerates RL rollouts, achieving an average 23.1{\%}{--}55.3{\%} reduction in rollout time compared to baselines."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2026-mtp">
<titleInfo>
<title>MTP-RL: Acceleration of Reinforcement Learning Rollouts with Policy-Aligned Multi-Token Prediction</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ke</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aohan</namePart>
<namePart type="family">Zeng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhengxiao</namePart>
<namePart type="family">Du</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuxuan</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bohan</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xinyi</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jie</namePart>
<namePart type="family">Tang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jing</namePart>
<namePart type="family">Zhang</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>Findings of the Association for Computational Linguistics: ACL 2026</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">979-8-89176-395-1</identifier>
</relatedItem>
<abstract>Reinforcement learning (RL) is widely applied to boost the performance of pretrained models, yet its training efficiency is severely constrained by rollout generation. While speculative decoding based on multi-token prediction (MTP) offers a potential acceleration pathway, its widespread adoption is hindered by the absence of MTP in vanilla pretrained models and the rapid degradation of the MTP acceptance length in RL training. To address these issues, this paper proposes MTP-RL, a two-stage framework that pioneers effective training of MTPs in RL and accelerates the rollout phase for diverse models. It involves a pipeline to equip the multi-layer parameter-sharing MTP for all models and an innovative advantage-aware MTP optimization strategy to facilitate policy-aligned training of MTPs. Experiments demonstrate that our method not only achieves stable growth of acceptance length during RL training, but also accelerates RL rollouts, achieving an average 23.1%–55.3% reduction in rollout time compared to baselines.</abstract>
<identifier type="citekey">wang-etal-2026-mtp</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.1871/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>37530</start>
<end>37542</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T MTP-RL: Acceleration of Reinforcement Learning Rollouts with Policy-Aligned Multi-Token Prediction
%A Wang, Ke
%A Zeng, Aohan
%A Du, Zhengxiao
%A Hu, Yuxuan
%A Zhang, Bohan
%A Wang, Xinyi
%A Tang, Jie
%A Zhang, Jing
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Findings of the Association for Computational Linguistics: ACL 2026
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-395-1
%F wang-etal-2026-mtp
%X Reinforcement learning (RL) is widely applied to boost the performance of pretrained models, yet its training efficiency is severely constrained by rollout generation. While speculative decoding based on multi-token prediction (MTP) offers a potential acceleration pathway, its widespread adoption is hindered by the absence of MTP in vanilla pretrained models and the rapid degradation of the MTP acceptance length in RL training. To address these issues, this paper proposes MTP-RL, a two-stage framework that pioneers effective training of MTPs in RL and accelerates the rollout phase for diverse models. It involves a pipeline to equip the multi-layer parameter-sharing MTP for all models and an innovative advantage-aware MTP optimization strategy to facilitate policy-aligned training of MTPs. Experiments demonstrate that our method not only achieves stable growth of acceptance length during RL training, but also accelerates RL rollouts, achieving an average 23.1%–55.3% reduction in rollout time compared to baselines.
%U https://aclanthology.org/2026.findings-acl.1871/
%P 37530-37542
Markdown (Informal)
[MTP-RL: Acceleration of Reinforcement Learning Rollouts with Policy-Aligned Multi-Token Prediction](https://aclanthology.org/2026.findings-acl.1871/) (Wang et al., Findings 2026)
ACL
- Ke Wang, Aohan Zeng, Zhengxiao Du, Yuxuan Hu, Bohan Zhang, Xinyi Wang, Jie Tang, and Jing Zhang. 2026. MTP-RL: Acceleration of Reinforcement Learning Rollouts with Policy-Aligned Multi-Token Prediction. In Findings of the Association for Computational Linguistics: ACL 2026, pages 37530–37542, San Diego, California, United States. Association for Computational Linguistics.