@inproceedings{diskin-2026-one,
title = "One Model, Five Tasks, Two {RTX} 3090 {GPU}s: The {HSE} System for the Sorbian Track of {WMT}26 Multitask {LLM}s with Limited Resources",
author = "Diskin, Michael",
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.209/",
pages = "2652--2661",
ISBN = "979-8-89176-474-3",
abstract = "We describe HSE{'}s submission to the Sorbian track of WMT26 Multitask LLMs with Lim- ited Resources, where a single 2B-parameter model must translate between German, Up- per Sorbian, and Lower Sorbian and also an- swer multiple-choice questions, check spelling and grammar, and solve maths problems in both Sorbian languages. Our system is one LoRA adapter on Qwen3.5-2B, trained on the official parallel data, synthetic spelling and grammar examples, a question-answering proxy, and English GSM8K in about nine GPU- hours on two consumer RTX 3090 GPUs. Ex- tended translation-only fine-tuning cost about ten points of question-answering accuracy, re- producing a WMT25 finding, and nearly erased the output conventions of the other tasks; the multitask mixture recovered them at almost no translation cost. The system ranked fourth of four teams. It far exceeded the baseline on translation (21.7 {\textrightarrow} 61.5 chrF++) and spell checking (6.7 {\textrightarrow} 63.6), but on grammar check- ing and maths reasoning it learned the answer format without the skill: the grammar score is almost exactly the share of error-free sentences, and maths accuracy did not exceed the baseline. Two negative results may transfer to other low- resource settings. Synthetic holdouts did not predict performance on the official tasks in ei- ther direction, and self-translated training prob- lems passed every automatic check yet changed meaning in at least 6 of 20 inspected cases. The model is publicly available."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="diskin-2026-one">
<titleInfo>
<title>One Model, Five Tasks, Two RTX 3090 GPUs: The HSE System for the Sorbian Track of WMT26 Multitask LLMs with Limited Resources</title>
</titleInfo>
<name type="personal">
<namePart type="given">Michael</namePart>
<namePart type="family">Diskin</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 HSE’s submission to the Sorbian track of WMT26 Multitask LLMs with Lim- ited Resources, where a single 2B-parameter model must translate between German, Up- per Sorbian, and Lower Sorbian and also an- swer multiple-choice questions, check spelling and grammar, and solve maths problems in both Sorbian languages. Our system is one LoRA adapter on Qwen3.5-2B, trained on the official parallel data, synthetic spelling and grammar examples, a question-answering proxy, and English GSM8K in about nine GPU- hours on two consumer RTX 3090 GPUs. Ex- tended translation-only fine-tuning cost about ten points of question-answering accuracy, re- producing a WMT25 finding, and nearly erased the output conventions of the other tasks; the multitask mixture recovered them at almost no translation cost. The system ranked fourth of four teams. It far exceeded the baseline on translation (21.7 → 61.5 chrF++) and spell checking (6.7 → 63.6), but on grammar check- ing and maths reasoning it learned the answer format without the skill: the grammar score is almost exactly the share of error-free sentences, and maths accuracy did not exceed the baseline. Two negative results may transfer to other low- resource settings. Synthetic holdouts did not predict performance on the official tasks in ei- ther direction, and self-translated training prob- lems passed every automatic check yet changed meaning in at least 6 of 20 inspected cases. The model is publicly available.</abstract>
<identifier type="citekey">diskin-2026-one</identifier>
<location>
<url>https://aclanthology.org/2026.wmt-main.209/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>2652</start>
<end>2661</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T One Model, Five Tasks, Two RTX 3090 GPUs: The HSE System for the Sorbian Track of WMT26 Multitask LLMs with Limited Resources
%A Diskin, Michael
%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 diskin-2026-one
%X We describe HSE’s submission to the Sorbian track of WMT26 Multitask LLMs with Lim- ited Resources, where a single 2B-parameter model must translate between German, Up- per Sorbian, and Lower Sorbian and also an- swer multiple-choice questions, check spelling and grammar, and solve maths problems in both Sorbian languages. Our system is one LoRA adapter on Qwen3.5-2B, trained on the official parallel data, synthetic spelling and grammar examples, a question-answering proxy, and English GSM8K in about nine GPU- hours on two consumer RTX 3090 GPUs. Ex- tended translation-only fine-tuning cost about ten points of question-answering accuracy, re- producing a WMT25 finding, and nearly erased the output conventions of the other tasks; the multitask mixture recovered them at almost no translation cost. The system ranked fourth of four teams. It far exceeded the baseline on translation (21.7 → 61.5 chrF++) and spell checking (6.7 → 63.6), but on grammar check- ing and maths reasoning it learned the answer format without the skill: the grammar score is almost exactly the share of error-free sentences, and maths accuracy did not exceed the baseline. Two negative results may transfer to other low- resource settings. Synthetic holdouts did not predict performance on the official tasks in ei- ther direction, and self-translated training prob- lems passed every automatic check yet changed meaning in at least 6 of 20 inspected cases. The model is publicly available.
%U https://aclanthology.org/2026.wmt-main.209/
%P 2652-2661
Markdown (Informal)
[One Model, Five Tasks, Two RTX 3090 GPUs: The HSE System for the Sorbian Track of WMT26 Multitask LLMs with Limited Resources](https://aclanthology.org/2026.wmt-main.209/) (Diskin, WMT 2026)
ACL