协作解码关键 token 以提升大型语言模型的事实性
Collaborative decoding of critical tokens for boosting factuality of large language models
AI总结：
针对微调、对齐及采样导致LLM幻觉的问题，提出基于关键token分类器在预训练模型与微调模型间协作选择并解码的框架，实验表明可显著降低幻觉。
摘要：
大型语言模型最常见的训练流程包括预训练、微调和对齐阶段，并产生各自对应的模型，例如预训练模型和微调模型。微调和对齐后的模型在指令遵循和安全生成方面表现出更强的能力，然而它们对世界保持事实性的能力会受到微调过程的影响。此外，生成过程中使用采样这一常见做法也会增加产生幻觉的机会。在本工作中，我们提出一个协作解码框架，通过关键 token 的概念来利用预训练模型中较高的事实性。我们首先设计一个关键 token 分类器，以决定下一个 token 使用哪个模型，随后使用不同的解码策略生成下一个 token。在不同模型和数据集上的实验表明，我们的解码框架能够显著减少模型幻觉，展示了协作解码框架的重要性。
英文摘要：
The most common training pipeline for large language models includes pretraining, finetuning and aligning phases, with their respective resulting models, such as the pretrained model and the finetuned model. Finetuned and aligned models show improved abilities of instruction following and safe generation, however their abilities to stay factual about the world are impacted by the finetuning process. Furthermore, the common practice of using sampling during generation also increases chances of hallucination. In this work, we introduce a collaborative decoding framework to harness the high factuality within pretrained models through the concept of critical tokens. We first design a critical token classifier to decide which model to use for the next token, and subsequently generates the next token using different decoding strategies. Experiments with different models and datasets show that our decoding framework is able to reduce model hallucination significantly, showcasing the importance of the collaborative decoding framework.
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。
