让大模型生成时先评估可信度，再谨慎输出，减少长文本中的幻觉。
Only Say What You Know: Calibration-Aware Generation for Long-Form Factuality
- 分离探索与确认过程，中间推理带可信度评估
- 在5个基准上事实性提升最高13%，解码速度加快37%
- 适合需要高可靠性的长文本生成任务
大型推理模型在复杂任务中表现优异，但在长文本生成中仍易产生幻觉，错误随推理步骤累积。现有方法多采用‘探索-承诺耦合’模式，中间推理无条件传递至最终输出，难以精细控制信息选择与整合。本文提出‘探索-承诺解耦’范式，将知识探索与最终承诺分离，使模型在回答时更具意识。我们构建了校准感知生成（CAG）框架，通过在中间推理中加入校准后的可靠性估计，并优先选择可靠内容生成最终输出，实现端到端的校准感知生成。在五个长文本事实性基准和多个模型族上，CAG 最高提升事实性13%，解码时间最多减少37%。结果表明，解耦是提升长文本生成可靠性的有效路径，为可信、自知的生成系统提供新方向。
原文摘要 · Abstract (English)
Large Reasoning Models achieve strong performance on complex tasks but remain prone to hallucinations, particularly in long-form generation where errors compound across reasoning steps. Existing approaches to improving factuality, including abstention and factuality-driven optimization, follow a \emph{coupled exploration-commitment} paradigm, in which intermediate reasoning is unconditionally propagated to the final output, limiting fine-grained control over information selection and integration. In this paper, we propose an \textbf{Exploration-Commitment Decoupling} paradigm that disentangles knowledge exploration from final commitment, enabling models to explore with awareness while answering cautiously. We instantiate the paradigm with \textbf{Calibration-Aware Generation (CAG)}, a framework that equips models with end-to-end, calibration-aware generation capabilities, by augmenting intermediate reasoning with calibrated reliability estimates and prioritizing reliable content in final outputs. Across five long-form factuality benchmarks and multiple model families, CAG improves factuality by up to 13%, while reducing decoding time by up to 37%. Overall, our work highlights decoupling as a principled approach for more reliable long-form generation, offering directions for trustworthy and self-aware generative systems.
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。