CogniConsole：将推理时控制作为可靠大语言模型交互的形式化抽象进行外化
CogniConsole: Externalizing Inference-Time Control as a Formal Abstraction for Reliable LLM Interactions
浏览论文内容
中文总结 AI 辅助
研究挑战大语言模型可靠性仅取决于模型能力的观点，引入CogniConsole将推理时控制外化为结构化接口，通过489个探针实验表明增加结构支架可降低输出方差和故障率，为LLM系统设计评估开辟新方向。
中文摘要 AI 辅助
大语言模型（LLM）系统中的可靠性通常被视为模型能力的函数。我们对此提出挑战，证明可靠性受推理时控制（即管理任务框架和上下文选择的计算层）显著影响。我们引入了CogniConsole，它将这种控制外化为一个结构化接口，结合了程序协调和基于有界提示的推理。通过在多步交互环境中的489个面向可控性的探针，我们表明增加结构支架（从无结构到完全支架）在固定模型架构下系统地降低了输出方差和故障率。结果表明，许多观察到的故障模式源于控制不明确而非能力不足。这项工作为将推理时控制视为一等抽象提供了实证基础，为LLM系统的设计和评估开辟了新方向。
英文摘要
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.
发表机构
- University of Regina（里贾纳大学）
- Orbital Sea（轨道海）
机构由 AI 辅助整理，请以论文原文为准。
