通过摘要行为分析大模型如何判断信息重要性，发现其有层级化重要性判断但不等于人类认知。
Behavioral Analysis of Information Salience in Large Language Models
- 用可控长度摘要作为探针，追踪问题可回答性来推导模型的信息优先级
- 13个模型在4个数据集上均呈现一致的层级化重要性排序
- 模型重要性判断无法通过自我解释获取，且与人类感知相关性弱
大语言模型（LLMs）在文本摘要任务中表现出色，该任务要求模型根据内容重要性进行选择。然而，模型内部对信息重要性的具体理解仍不明确。为此，我们提出一个可解释框架，通过模型的摘要行为系统性地推导和研究其信息重要性机制。利用可控长度的摘要作为内容选择过程的行为探测，并追踪讨论中问题的可回答性变化，我们构建了一个代理指标来反映模型对信息的优先排序。在13个模型和4个数据集上的实验表明，LLMs具有细致且层级化的信息重要性认知，这种认知在不同模型家族和规模间普遍一致。尽管模型行为高度稳定，其重要性判断无法通过自我反思获取，且与人类对信息重要性的感知仅有微弱相关性。
原文摘要 · Abstract (English)
Large Language Models (LLMs) excel at text summarization, a task that requires models to select content based on its importance. However, the exact notion of salience that LLMs have internalized remains unclear. To bridge this gap, we introduce an explainable framework to systematically derive and investigate information salience in LLMs through their summarization behavior. Using length-controlled summarization as a behavioral probe into the content selection process, and tracing the answerability of Questions Under Discussion throughout, we derive a proxy for how models prioritize information. Our experiments on 13 models across four datasets reveal that LLMs have a nuanced, hierarchical notion of salience, generally consistent across model families and sizes. While models show highly consistent behavior and hence salience patterns, this notion of salience cannot be accessed through introspection, and only weakly correlates with human perceptions of information salience.
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。