让大模型推理更可靠，关键在控制机制设计。
CogniConsole: Externalizing Inference-Time Control as a Formal Abstraction for Reliable LLM Interactions
- 将推理时的控制逻辑抽象为结构化接口，实现可编程协调与受限提示推理。
- 结构化程度提升后，输出方差和失败率显著降低（489次测试）。
- 适合关注模型稳定性与可控性的系统设计者，尤其适用于复杂交互场景。
大语言模型系统的可靠性通常被视为模型能力的函数。我们提出不同观点：可靠性很大程度上受推理时控制机制——即任务定义与上下文选择的计算层——的影响。本文提出CogniConsole，一种将该控制外化为结构化接口的架构，结合程序化协调与有界提示推理。在多步骤交互环境中，通过489次以控制性为导向的探测实验表明，从无结构到完全结构化的逐步增强，能系统性降低固定模型架构下的输出方差与失败率。结果表明，许多故障模式如上下文漂移和约束不一致，源于控制不明确而非能力不足。本工作为将推理时控制视为第一类抽象提供了实证基础，开辟了超越单纯规模扩展的新设计与评估方向。
原文摘要 · Abstract (English)
Reliability in large language model (LLM) systems is typically framed as a function of model capability. We challenge this by demonstrating that reliability is significantly influenced by \emph{inference-time control} -- the computational layer governing task framing and context selection. We introduce \emph{CogniConsole}, an architectural instantiation that externalizes this control into a structured interface combining programmatic coordination with bounded prompt-based reasoning. Through \emph{controllability-oriented probes} ($N=489$) in a multi-step interactive environment, we show that increasing structural scaffolding -- from unstructured to fully scaffolded -- \textbf{systematically reduces output variance and failure rates under a fixed model architecture}. Our results indicate that many observed failure modes, such as context drift and inconsistent constraint adherence, arise from under-specified control rather than insufficient capability. This work provides an empirical basis for treating inference-time control as a first-class abstraction, opening new directions for designing and evaluating LLM systems beyond scaling alone.
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。