《Dial BeInfo for Faithfulness》：通过行为微调提升信息检索对话的事实性
$\textit{Dial BeInfo for Faithfulness}$: Improving Factuality of Information-Seeking Dialogue via Behavioural Fine-Tuning
AI总结：
本文提出 BeInfo 行为微调方法，以减少信息检索对话中的幻觉；实验证明其可提升模型在已知和零样本未知领域的知识忠实度，并使 3B Flan-T5 在有限真实生产对话上调优后超过 GPT4。
摘要：
事实性是信息检索对话中的一项关键要求：系统应回应用户查询，使回复具有意义，并与提供给系统的知识保持一致。然而，大多数现代大型语言模型都受到幻觉问题的困扰，即它们会生成不受知识源支持或与知识源相矛盾的回复。为缓解这一问题并提高信息检索对话系统的忠实度，我们提出了 BeInfo，这是一种简单而有效的方法，采用行为调优来辅助信息检索对话。基于三个标准数据集，我们表明，使用 BeInfo 调优后的模型，无论是在 BeInfo 调优期间见过的数据集和领域上，还是在以零样本方式应用于未见过的领域时，对知识源都明显更加忠实。此外，我们表明，使用 BeInfo 调优的 3B 参数模型（例如 Flan-T5）在真实“生产”对话数据上表现出强劲性能，并且在仅使用有限数量的此类真实领域内对话进行调优时，性能优于 GPT4。
英文摘要：
Factuality is a crucial requirement in information seeking dialogue: the system should respond to the user's queries so that the responses are meaningful and aligned with the knowledge provided to the system. However, most modern large language models suffer from hallucinations, that is, they generate responses not supported by or contradicting the knowledge source. To mitigate the issue and increase faithfulness of information-seeking dialogue systems, we introduce BeInfo, a simple yet effective method that applies behavioural tuning to aid information-seeking dialogue. Relying on three standard datasets, we show that models tuned with BeInfo} become considerably more faithful to the knowledge source both for datasets and domains seen during BeInfo-tuning, as well as on unseen domains, when applied in a zero-shot manner. In addition, we show that the models with 3B parameters (e.g., Flan-T5) tuned with BeInfo demonstrate strong performance on data from real `production' conversations and outperform GPT4 when tuned on a limited amount of such realistic in-domain dialogues.
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。
