CCL-XCoT：一种用于缓解幻觉生成的高效跨语言知识迁移方法
CCL-XCoT: An Efficient Cross-Lingual Knowledge Transfer Method for Mitigating Hallucination Generation
AI总结：
本文提出CCL-XCoT两阶段微调框架，通过课程对比学习与跨语言思维链提示，在不依赖外部检索的情况下将多语言大模型幻觉率降低62%，有效提升低资源语言的事实知识迁移能力。
摘要：
多语言大语言模型（MLLMs）在跨语言泛化方面表现出色，但由于训练数据不平衡，它们仍然容易产生幻觉，尤其是在低资源语言中。这些幻觉包括不准确或虚构的输出，在领域特定的生成任务中尤为严重（Chataigner et al., 2024）。为了解决这一挑战，我们提出了CCL-XCoT（基于课程对比学习的跨语言思维链，Curriculum-based Contrastive Learning-based Cross-lingual Chain-of-Thought），这是一个用于缓解多语言大语言模型幻觉的两阶段微调框架。我们的方法首先在持续预训练阶段，通过基于课程的对比学习结合下一词元预测来增强跨语言语义对齐。在此基础上，我们在指令微调阶段引入跨语言思维链（XCoT）提示策略，引导模型先用高资源语言进行推理，再用目标低资源语言生成答案。实验结果表明，CCL-XCoT将幻觉率降低了高达62%，并显著改善了跨语言对之间的事实知识迁移，且不依赖外部检索或多模型集成。
英文摘要：
Multilingual Large Language Models(MLLMs) demonstrate strong generalization across languages, yet they remain prone to hallucinations, especially in low-resource languages, due to training data imbalances. These hallucinations, which include inaccurate or fabricated outputs, are particularly problematic in domain-specific generation tasks (Chataigner et al., 2024). To address this challenge, we propose CCL-XCoT(Curriculum-based Contrastive Learning-based Cross-lingual Chain-of-Thought), a two-stage fine-tuning framework for mitigating hallucination in MLLMs. Our approach first enhances cross-lingual semantic alignment through curriculum-based contrastive learning combined with next-token prediction during continued pre-training. Building on this foundation, we then introduce a cross-lingual Chain-of-Thought (XCoT) prompting strategy during instruction fine-tuning, which guides the model to reason in a high-resource language before generating answers in the target low-resource language. Experimental results show that CCL-XCoT reduces hallucination rates by up to 62% and substantially improves factual knowledge transfer across language pairs, without relying on external retrieval or multi-model ensembles.
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。
