Self-Translate-Train：通过固有能力增强大语言模型的跨语言迁移
Self-Translate-Train: Enhancing Cross-Lingual Transfer of Large Language Models via Inherent Capability
AI总结：
针对多语言预训练模型零样本跨语言迁移中语言内部表示不对齐的问题，提出Self-Translate-Train方法，利用LLMs自身将训练数据译为目标语言并自训练，效果优于零样本迁移，为激发LLMs跨语言能力提供新思路。
摘要：
通过微调多语言预训练模型实现的零样本跨语言迁移在低资源语言场景中展现出潜力，但常受限于语言间内部表示的不对齐问题。我们提出假设：即便模型在微调阶段无法有效实现跨语言泛化，它仍能捕捉到对跨语言迁移有用的跨语言对应关系。我们通过Self-Translate-Train方法验证该假设，该方法让大语言模型（LLMs）将训练数据翻译为目标语言，并基于模型自身生成的数据进行微调。实验证明Self-Translate-Train的效果优于零样本迁移，这一结果鼓励研究者进一步探索激发LLMs跨语言能力的更优方法。
英文摘要：
Zero-shot cross-lingual transfer by fine-tuning multilingual pretrained models shows promise for low-resource languages, but often suffers from misalignment of internal representations between languages. We hypothesize that even when the model cannot generalize across languages effectively in fine-tuning, it still captures cross-lingual correspondence useful for cross-lingual transfer. We explore this hypothesis with Self-Translate-Train, a method that lets large language models (LLMs) to translate training data into the target language and fine-tunes the model on its own generated data. By demonstrating that Self-Translate-Train outperforms zero-shot transfer, we encourage further exploration of better methods to elicit cross-lingual capabilities of LLMs.
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。
