多语言指令微调不存在最优语言集：基于语言信息研究的见解
No Optimal Language Set Exists for Multilingual Instruction Tuning: Insights from a Linguistically-Informed Study
AI总结：
研究多语言指令微调中语言集选择问题，通过评估基于多种特征的语言信息选择策略与随机基线对比，发现固定预算下无通用语言选择策略，性能因任务和模型而异，超阈值加语言会致性能下降，还讨论了相关影响和评估陷阱。
摘要：
多语言指令微调面临多语言诅咒、数据稀缺和高计算成本的挑战。一种自然假设是精心选择一组语言多样化的语言能产生普遍更好的模型。我们通过评估基于类型学、地理、语义和学习特征的语言信息选择策略，与三个模型家族（mGPT、mT5-xl、BLOOM）和五个多语言基准测试中的随机基线进行系统测试。关键负面发现是在固定预算设置中没有出现通用语言选择策略：性能强烈依赖任务和模型，超过适度阈值添加更多语言会导致与多语言诅咒一致的性能下降。我们讨论了对多语言指令微调数据管理的影响以及基准平均评估的陷阱。所有资源可在该https网址公开获取。
英文摘要：
Multilingual instruction tuning (MIT) is challenged by the curse of multilinguality, data scarcity, and high computational cost. A natural hypothesis is that carefully selecting a linguistically diverse set of languages yields universally better models. We test this systematically by evaluating linguistically-informed selection strategies---based on typological, geographical, semantic, and learned features---against random baselines across three model families (mGPT, mT5-xl, BLOOM) and five multilingual benchmarks. Our key negative finding is that no universal language selection strategy emerges in our fixed-budget setting: performance is strongly task- and model-dependent, and adding more languages beyond a modest threshold triggers degradation consistent with the curse of multilinguality. We discuss implications for MIT data curation and the pitfalls of benchmark-averaged evaluation. All resources are publicly available at https://github.com/GGLAB-KU/ling-informed-mit.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
