arXiv cs.CL· Timo Oksanen·· 12 小时前AI 评分20
INVAR-Agri：用微调 LLM 检测农机操作手册多语言版本差异
INVAR-Agri: Detecting Discrepancies in Multilingual Operator Manuals of Agricultural Implements with a Fine-Tuned LLM
AI 导读
INVAR-Agri 在 EuroLLM-1.7B-Instruct 上对比两种训练流程，用于核查农机操作手册的多语言翻译差异。单独训练的 checker 能检出 24 种语言测试集中大多数植入错误，但回译器方案会掩盖大量错误，尤其是弱化的义务表述。作者复核确认了这种掩盖现象，并指出真实芬兰语译文上的告警多为误报，该工具或可辅助而非替代合格人员审校。
正文
Abstract:Operator manuals of agricultural implements are delivered in many languages, and the standards require that competent persons review every translation, which makes this review the bottleneck. This paper presents a proof of concept that compares two training workflows on the same LLM, EuroLLM-1.7B-Instruct, trained on operator manuals of one manufacturer. In approach A, a back-translator translates each language version back into English, and an English reader compares it with the English version. In approach B, a checker reads the English version and the translation together, where the checker reports differences. The test sets contain errors planted in numbers, units, opposite words, obligations, negations, and conditions in 24 languages, and they include unmodified text and meaning-preserving edits for alerts and false alarms. The trained back-translator concealed many planted errors, most of all weakened obligations, while translation-quality scores rose. In contrast, the separately trained checker detected most planted errors, although it was weak for units and for number changes unlike those in training. A check by the author confirmed the concealment. In addition, a simple rule check caught the number and unit changes, but it alerted on harmless rewrites. A check of alerts on real Finnish translations showed that most alerts were false, caused by language-specific wording, so the precision on real text remains a limitation. Within this setting, a trained checker might support, not replace, the review by competent persons, and firmer conclusions need more data.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.35423 [cs.CL] |
| (or arXiv:2610.35423v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.35423 arXiv-issued DOI via DataCite (pending registration) |
|
| Journal reference: | Technical University of Munich. 2026. ISBN 978-3-911430-18-0. https://mediatum.ub.tum.de/1865958 |
Submission history
From: Timo Oksanen [view email]
[v1]
Mon, 26 Oct 2026 13:31:34 UTC (121 KB)
来源：arXiv cs.CL · arxiv.org