通过跨语言思维链减少低资源语言幻觉生成
CCL-XCoT: An Efficient Cross-Lingual Knowledge Transfer Method for Mitigating Hallucination Generation
- 分两阶段训练：先对齐多语言语义，再用高资源语言引导推理
- 在低资源语言上将幻觉率降低62%，提升事实知识迁移效果
- 无需外部工具，适合需要可靠多语言生成的场景
多语言大模型虽具备跨语言泛化能力，但在低资源语言中仍易产生幻觉，尤其在领域特定生成任务中问题突出（Chataigner et al., 2024）。为此，我们提出CCL-XCoT（基于课程对比学习的跨语言思维链），一种两阶段微调框架以缓解多语言大模型的幻觉。首先，在持续预训练中结合课程对比学习与下一词预测，增强跨语言语义对齐；随后在指令微调阶段引入跨语言思维链（XCoT）提示策略，引导模型先以高资源语言进行推理，再生成目标低资源语言答案。实验表明，该方法可将幻觉率降低高达62%，显著提升跨语言事实知识迁移效果，且不依赖外部检索或多模型集成。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。