@inproceedings{zhang-etal-2026-hitsz,
title = "{HITSZ} at {WMT} 2026: Mixed Continued Pre-training and Supervised Fine-tuning for Low-Resource Sorbian",
author = "Zhang, Cong and
Wang, Yutong and
Liu, Xuebo and
Zhang, Min",
editor = "Haddow, Barry and
Kocmi, Tom and
Koehn, Philipp and
Monz, Christof",
booktitle = "Proceedings of the Eleventh Conference on Machine Translation",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.wmt-main.214/",
pages = "2697--2709",
ISBN = "979-8-89176-474-3",
abstract = "We describe the HITSZ system submitted to the Sorbian track of the WMT 2026 Shared Task on Multitask LLMs with Limited Resources. The track requires a single Qwen3.5-2B model to jointly perform machine translation (MT), question answering (QA), spell checking (SC), grammar checking (GC), and mathematical reasoning (MR) for Upper and Lower Sorbian. To strengthen the model{'}s limited Sorbian linguistic knowledge while retaining previously acquired capabilities, we first perform mixed continued pre-training (CPT) on Sorbian monolingual and parallel data together with multilingual and task-oriented replay. We then apply full-parameter multitask supervised fine-tuning (SFT) to align the adapted model with the five heterogeneous task formats. Compared with the official Qwen3.5-2B baseline, our primary submission improves MT by 41.44 chrF++ points and improves QA, SC, GC, and MR by 13.94, 66.47, 60.77, and 13.20 points, respectively, ranking third overall in the Sorbian track. Additional analysis shows that removing CPT reduces our five-task development score by 5.24 points, with the largest losses on SC, MT, and GC. We also find that heavily increasing MT supervision does not further improve translation and can degrade other jointly trained tasks."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhang-etal-2026-hitsz">
<titleInfo>
<title>HITSZ at WMT 2026: Mixed Continued Pre-training and Supervised Fine-tuning for Low-Resource Sorbian</title>
</titleInfo>
<name type="personal">
<namePart type="given">Cong</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yutong</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xuebo</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Min</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the Eleventh Conference on Machine Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Barry</namePart>
<namePart type="family">Haddow</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tom</namePart>
<namePart type="family">Kocmi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Philipp</namePart>
<namePart type="family">Koehn</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Christof</namePart>
<namePart type="family">Monz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-474-3</identifier>
</relatedItem>
<abstract>We describe the HITSZ system submitted to the Sorbian track of the WMT 2026 Shared Task on Multitask LLMs with Limited Resources. The track requires a single Qwen3.5-2B model to jointly perform machine translation (MT), question answering (QA), spell checking (SC), grammar checking (GC), and mathematical reasoning (MR) for Upper and Lower Sorbian. To strengthen the model’s limited Sorbian linguistic knowledge while retaining previously acquired capabilities, we first perform mixed continued pre-training (CPT) on Sorbian monolingual and parallel data together with multilingual and task-oriented replay. We then apply full-parameter multitask supervised fine-tuning (SFT) to align the adapted model with the five heterogeneous task formats. Compared with the official Qwen3.5-2B baseline, our primary submission improves MT by 41.44 chrF++ points and improves QA, SC, GC, and MR by 13.94, 66.47, 60.77, and 13.20 points, respectively, ranking third overall in the Sorbian track. Additional analysis shows that removing CPT reduces our five-task development score by 5.24 points, with the largest losses on SC, MT, and GC. We also find that heavily increasing MT supervision does not further improve translation and can degrade other jointly trained tasks.</abstract>
<identifier type="citekey">zhang-etal-2026-hitsz</identifier>
<location>
<url>https://aclanthology.org/2026.wmt-main.214/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>2697</start>
<end>2709</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T HITSZ at WMT 2026: Mixed Continued Pre-training and Supervised Fine-tuning for Low-Resource Sorbian
%A Zhang, Cong
%A Wang, Yutong
%A Liu, Xuebo
%A Zhang, Min
%Y Haddow, Barry
%Y Kocmi, Tom
%Y Koehn, Philipp
%Y Monz, Christof
%S Proceedings of the Eleventh Conference on Machine Translation
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-474-3
%F zhang-etal-2026-hitsz
%X We describe the HITSZ system submitted to the Sorbian track of the WMT 2026 Shared Task on Multitask LLMs with Limited Resources. The track requires a single Qwen3.5-2B model to jointly perform machine translation (MT), question answering (QA), spell checking (SC), grammar checking (GC), and mathematical reasoning (MR) for Upper and Lower Sorbian. To strengthen the model’s limited Sorbian linguistic knowledge while retaining previously acquired capabilities, we first perform mixed continued pre-training (CPT) on Sorbian monolingual and parallel data together with multilingual and task-oriented replay. We then apply full-parameter multitask supervised fine-tuning (SFT) to align the adapted model with the five heterogeneous task formats. Compared with the official Qwen3.5-2B baseline, our primary submission improves MT by 41.44 chrF++ points and improves QA, SC, GC, and MR by 13.94, 66.47, 60.77, and 13.20 points, respectively, ranking third overall in the Sorbian track. Additional analysis shows that removing CPT reduces our five-task development score by 5.24 points, with the largest losses on SC, MT, and GC. We also find that heavily increasing MT supervision does not further improve translation and can degrade other jointly trained tasks.
%U https://aclanthology.org/2026.wmt-main.214/
%P 2697-2709
Markdown (Informal)
[HITSZ at WMT 2026: Mixed Continued Pre-training and Supervised Fine-tuning for Low-Resource Sorbian](https://aclanthology.org/2026.wmt-main.214/) (Zhang et al., WMT 2026)
ACL