让大模型学会主动求助，避免胡说八道。
MASH: Modeling Abstention via Selective Help-Seeking
- 用强化学习设计按搜付费奖励机制，让求助行为反映知识边界。
- 在多跳问答数据集上准确率提升7.6%，优于现有高效搜索方法。
- 无需预先设定知识边界，开箱即用，适合实用型模型部署。
大型语言模型无法可靠识别其参数化知识范围，常对超出范围的问题产生幻觉。本文提出MASH（通过选择性求助建模回避），一种能有效提取模型回避行为的训练框架。核心思想是：若对使用外部工具（如搜索）进行适当惩罚并奖励答案准确性，那么模型的求助行为可作为回避的代理信号。MASH采用按搜付费的强化学习实现该机制。在三个知识密集型问答数据集上实验表明，相比以往高效搜索方法，MASH显著提升选择性求助性能；在多跳数据集上，答案准确率提高7.6%。此外，MASH展现出强大的开箱即用回避能力，表现媲美需预设知识边界的传统回避方法。结果证明，该训练有效对齐了搜索工具使用与参数化知识，可用于决策回避和高效工具使用。
原文摘要 · Abstract (English)
LLMs cannot reliably recognize their parametric knowledge boundaries and often hallucinate answers to outside-of-boundary questions. In this paper, we introduce MASH (Modeling Abstention via Selective Help-seeking), a training framework that readily extracts abstentions from LLMs. Our key idea is that any external help-seeking by an LLM, i.e. search tool use, can serve as a proxy for abstention if the external help (search) is appropriately penalized while also rewarding answer accuracy. MASH operationalizes this idea using reinforcement learning with a pay-per-search reward. We run experiments on three knowledge-intensive QA datasets. Our results show that MASH substantially improves upon the selective help-seeking performance of prior efficient search approaches; on multi-hop datasets, it improves answer accuracy by 7.6%. Furthermore, MASH demonstrates strong off-the-shelf abstention performance, showcasing behavior competitive with prior abstention methods that additionally require predetermining model knowledge boundaries to construct training data. Overall, we show MASH training effectively aligns search tool use with parametric knowledge, which can be successfully leveraged for making abstention decisions and efficient search tool use
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。