文章背景与核心概要

本文是入选 ICML 2026 的立场论文。作者指出，当前大语言模型（LLM）中普遍存在的缺陷——如谄媚性（sycophancy）、逻辑泛化能力不足以及自信的幻觉——其根源在于一种核心建模缺陷：即在孤立的单输出对上对模型进行评估和训练。

为了解决这些问题，本文提出了以“自洽性”（self-consistency）作为指导框架。研究表明，许多现有的旨在提升事实一致性和对抗鲁棒性的策略，本质上只是更广泛的“一致性优化”过程的子集。最终，作者勾勒了开发全局一致性大语言模型的路线图，强调了这一范式转变所能解锁的新能力，并探讨了潜在的挑战与异议。

观点：是时候优化大语言模型的自洽性了

摘要

Despite ever-increasing sophistication in language model (LM) pre- and post-training pipelines, many important failures persist: models overcondition on user framing ("sycophancy"), exhibit incomplete logical generalization, and produce confident but incorrect responses. We argue that these failures arise from a modeling assumption permeating all aspects of the pipeline: that behavior can be specified and evaluated independently on single-output pairs. Many model failures are difficult, if not impossible, to detect without reasoning about relationships between a model's responses across inputs. In this position paper, we propose self-consistency as a framework for understanding these failures. We first observe that a wide variety of techniques designed to improve specific aspects of LM behavior—targeting properties as diverse as adversarial robustness and factual coherence—can be understood as special cases of a common "consistency optimization" procedure and addressed with a standard set of optimization tools. We next outline a set of new model properties that could be achieved by optimizing for consistency, and conclude with a discussion of what it would mean to develop generally consistent LMs, including the capabilities they would enable and the objections they raise.

尽管语言模型（LM）的预训练和后训练流程日益复杂，但许多重要的缺陷依然存在：模型过度依赖用户设定的框架（即“谄媚性”）、表现出不完整的逻辑泛化能力，并产生自信但错误的回答。我们认为，这些失败源于贯穿整个流程的一种建模假设：即行为可以在孤立的单输出对上进行指定和评估。如果不去推理模型在不同输入下响应之间的关系，许多模型故障是难以甚至无法检测到的。在这篇立场论文中，我们提出了“自洽性”作为理解这些失败的框架。我们首先观察到，许多旨在改善 LM 特定行为的技术（涵盖对抗鲁棒性和事实一致性等多种属性）都可以被理解为一种通用的“一致性优化”过程的特例，并可以通过一套标准的优化工具来解决。接下来，我们概述了一系列通过优化一致性可以实现的新模型属性，并总结了开发全局一致性 LM 的意义，包括它们将解锁的能力以及可能引发的异议。

论文元数据

全文链接与资源

• PDF: 查看 PDF
• HTML (实验性): arXiv HTML
• TeX 源码: arXiv 源码
• 许可协议: 知识共享署名 4.0