ToW：词汇的思考提升大型语言模型的推理能力
ToW: Thoughts of Words Improve Reasoning in Large Language Models
AI总结：
提出词汇的思考（ToW）数据增强方法，通过在预训练中注入细粒度词级思考，解决现有下一个词预测的事实幻觉和隐式推理学习效率低的问题，显著提升模型推理能力并减少幻觉。
摘要：
我们引入了词汇的思考（ToW），一种用于下一个词预测的新型训练时数据增强方法。ToW将下一个词预测视为一项核心推理任务，并在预训练文本中注入细粒度的思考，解释下一个词应该是什么以及它如何与先前的上下文相关联。我们的公式化方法解决了现有下一个词预测学习方案的两个基本缺陷：它们会引发事实性幻觉，并且模型在学习原始文本中的隐式推理过程时效率低下。虽然有多种方式可以获取这种词汇的思考，但我们探索了通过从更大模型中蒸馏来获取ToW标注的第一步。在使用仅70K条ToW标注进行持续预训练后，我们有效地将模型的推理性能平均提高了7%至9%，并将模型幻觉减少了高达10%。同时，ToW完全独立于任务和应用，不会在标签或语义上引入额外的偏差。
英文摘要：
We introduce thoughts of words (ToW), a novel training-time data-augmentation method for next-word prediction. ToW views next-word prediction as a core reasoning task and injects fine-grained thoughts explaining what the next word should be and how it is related to the previous contexts in pre-training texts. Our formulation addresses two fundamental drawbacks of existing next-word prediction learning schemes: they induce factual hallucination and are inefficient for models to learn the implicit reasoning processes in raw texts. While there are many ways to acquire such thoughts of words, we explore the first step of acquiring ToW annotations through distilling from larger models. After continual pre-training with only 70K ToW annotations, we effectively improve models' reasoning performances by 7% to 9% on average and reduce model hallucination by up to 10%. At the same time, ToW is entirely agnostic to tasks and applications, introducing no additional biases on labels or semantics.
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。
