Task-Preserving Multilingual Instruction Tuning for Open-Ended QA in WMT 2026 MIST
Aleix Sant · Jordi Luque · Carlos Escolano
介绍 WMT 2026 多语言指令共享任务的参赛系统。系统利用组织方提供的数据，通过任务保持型本地化构建多语言指令数据集 EuroAlpaca，并采用跨语言增强数据 CrossEuroAlpaca，为指令、上下文和响应分配不同的源语言与目标语言；随后使用 LoRA 微调 Qwen3-8B。系统面向辅助数据所涵盖的十种共享任务语言，任务是开放式问答(QA-OEG)。在相同解码设置下，相比基础模型，目标语言 QA-OEG 的 Gemma-4 评判平均分相对提高 7.8%。在覆盖全部共享任务语言的 Aya Evaluation Suite 上，宏平均 ROUGE-L 和 BERTScore-F1 分别取得 41.5% 和 2.9% 的相对提升，且目标语言提升更大。不过，效果依赖任务和语言：目标语言的 QA-OEG 与摘要得分提高，但上下文依据问答下降；非目标语言的三项任务得分均下降。
尚无解读。
arXiv:2610.29926v1 Announce Type: new Abstract: This paper describes our submission to the WMT 2026 Multilingual Instruction Shared Task. We adapt Qwen3-8B with LoRA using organiser-provided data, EuroAlpaca, a multilingual instruction dataset created through task-preserving localisation, and CrossEuroAlpaca, a cross-lingual augmentation that assigns different source and target languages to instructions, contexts and responses. Our system targets open-ended question answering (QA-OEG) in ten shared-task languages represented in the auxiliary data. Under matched decoding, adaptation improves the mean Gemma-4 judge score on target-language QA-OEG by 7.8% relative to the base model. On an additional evaluation using the Aya Evaluation Suite across all shared-task languages, it yields relative gains of 41.5% in macro-averaged ROUGE-L and 2.9% in BERTScore-F1, with larger improvements in target languages. However, these gains are task- and language-specific. Gemma-4 scores improve for QA-OEG and summarisation in the target languages but decline for context-grounded QA there and across all three tasks in the non-target languages.
Act as a Study Review Companion. You are an expert in academic support with extensive knowledge across various subjects. Your task is to facilitate effective study sessions for ${subject}. You will: - Summarize key points from the study ma…