认知对齐：面向用户-LLM 知识交付的中介框架

大语言模型（LLMs）日益成为知识获取工具，但用户无法有效地 specify how they want information presented。当用户请求 LLMs "cite reputable sources"、"express appropriate uncertainty"或"include multiple perspectives"时，他们发现 current interfaces 提供 no structured way to articulate these preferences。 result is prompt sharing folklore：社区特定的 copied prompts 通过 trust relationships 传递，而 non-based on measured efficacy。我们提出了认知对齐框架（Epistemic Alignment Framework），一套从 epistemology 哲学文献中提炼的 knowledge transmission 中的十个 challenge，关 于 issues such as evidence quality assessment and testimonial reliance 的 calibration。该框架 serves as a structured intermediary between user needs and system capabilities， creating a common vocabulary to bridge the gap between what users want and what systems deliver。通过对 custom prompts and personalization strategies 的主题分析， we find users develop elaborate workarounds to address each of the challenges。我们 then apply our framework to OpenAI and Anthropic，通过 content analysis of their documented policies and product features。Our analysis shows that while these providers have partially addressed the challenges we identified，they fail to establish adequate mechanisms for specifying epistemic preferences，lack transparency about how preferences are implemented，and offer no verification tools to confirm whether preferences were followed。For AI developers，认知对齐框架 offers concrete guidance for supporting diverse approaches to knowledge；for users，it works toward information delivery that aligns with their specific needs rather than defaulting to one-size-fits-all approaches。