通过解码时的幻觉与真实比较器改进大型语言模型的事实性
Improving Factuality in Large Language Models via Decoding-Time Hallucinatory and Truthful Comparators
AI总结：
本文提出CDT框架，通过多任务微调构建幻觉与真实比较器，并结合混合专家策略捕获不同模式，在解码时对比logit差异以约束预测分布，从而显著提升LLMs的响应事实性与整体性能。
摘要：
尽管大型语言模型（LLMs）具有卓越的能力，但它们倾向于生成与可验证事实相矛盾的响应，即不忠实的幻觉内容。现有的工作通常集中在优化模型参数或编辑语义表示上，这会损害目标LLMs的内部事实知识。此外，幻觉在下游任务中通常表现出多方面的模式，限制了模型在各任务间的整体性能。在本文中，我们提出了一种比较器驱动的解码时（CDT）框架来缓解响应幻觉。首先，我们使用多任务微调样本构建幻觉和真实比较器。在这种情况下，我们提出了一种指令原型引导的混合专家策略，以增强相应比较器在不同任务指令中捕获不同幻觉或真实性模式的能力。CDT通过对比目标LLMs与这些比较器之间的logit差异，将下一个token的预测约束为事实性鲁棒的分布。在多个下游任务上的系统实验表明，我们的框架能够显著提升模型性能和响应事实性。
英文摘要：
Despite their remarkable capabilities, Large Language Models (LLMs) are prone to generate responses that contradict verifiable facts, i.e., unfaithful hallucination content. Existing efforts generally focus on optimizing model parameters or editing semantic representations, which compromise the internal factual knowledge of target LLMs. In addition, hallucinations typically exhibit multifaceted patterns in downstream tasks, limiting the model's holistic performance across tasks. In this paper, we propose a Comparator-driven Decoding-Time (CDT) framework to alleviate the response hallucination. Firstly, we construct hallucinatory and truthful comparators with multi-task fine-tuning samples. In this case, we present an instruction prototype-guided mixture of experts strategy to enhance the ability of the corresponding comparators to capture different hallucination or truthfulness patterns in distinct task instructions. CDT constrains next-token predictions to factuality-robust distributions by contrasting the logit differences between the target LLMs and these comparators. Systematic experiments on multiple downstream tasks show that our framework can significantly improve the model performance and response factuality.
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。
