明确，而非冗长：什么使认知立场在记忆压缩中留存
Explicit, Not Longer: What Makes Epistemic Stance Survive Memory Compression
浏览论文内容
中文总结 AI 辅助
本研究探究使认知立场在记忆压缩中留存的因素，发现将立场表述为带标签的字段可提升留存率，明确立场而非仅延长表述是关键，明确的最佳方式取决于模型。
中文摘要 AI 辅助
智能体记忆系统会对存储内容进行压缩，而压缩的设计初衷是丢弃限定词，因此主张的认知立场往往无法在写入记忆后留存。本研究探究了决定认知立场能否留存的因素。匹配的笔记包含完全相同的主张和完全相同的立场，仅在立场的位置上存在差异；一个模型在相同的填充笔记中以相同的预算压缩两者，从未看到该条件的盲读器对压缩结果进行评分。在7种语域的60项主张中，将立场表述为带标签的字段而非括号内的旁注，使两种模型的留存率提高了约15个百分点（一种模型中37项主张留存对2项，另一种模型中30项对8项；置换检验p=0.00005），且在Haiku上进行的预注册重复实验（其预测和决策规则在实验运行前已确定）得出了+15.6个百分点的结果，即38项主张留存对1项。对两种模型的格式进行消融实验后，发现不同部分产生了相同的净效应：标签对两种模型均有帮助（分别为+9.7和+12.8），而长度对两者均无帮助，但将立场表述为完整句子是一种模型的最大组成部分（+12.5），对另一种模型则毫无价值（+0.6）。单独使用任何一种模型都会得出一个自信但不同的机制，因此本研究仅主张交集部分：使立场明确，而非仅仅冗长，且明确的最佳方式取决于模型。无模型的确定性读出重现了两格方向和7项消融对比中的5项，但未重现长度或标签的影响，因此本研究不将后两者归因于该工具。50个人工标注（kappa=0.75）在方向上达成一致；本研究完整呈现了其中7项分歧。此外，本研究还报告了9项撤回的主张，其中3项曾是本文的标题主张。
英文摘要
Agent memory systems compress what they store, and compression is built to drop qualifiers, so a claim's epistemic standing tends not to survive being written to memory. We ask what governs whether it does. Matched notes carry the identical claim and identical stance and differ only in where that stance sits; one model compresses both under the same budget among the same filler notes, and a blind reader that never sees the condition scores the result. Across 60 claims in seven registers, writing the stance as a labelled field rather than a bracketed aside raises retention by about 15 points on two models (37 claims to 2 on one, 30 to 8 on the other; permutation p=0.00005), and a pre-registered replication on Haiku, its prediction and decision rule committed before the run, gives +15.6 points, 38 claims to 1. Ablating the format on both models gives the same net effect from different parts: labels help on both (+9.7 and +12.8) and length helps on neither, but wording the stance as a full sentence is the largest component on one model (+12.5) and worth nothing on the other (+0.6). Either model alone would have licensed a confident and different mechanism, so we claim only the intersection: make the stance explicit, not merely longer, and expect the best way of being explicit to depend on the model. A deterministic readout with no model reproduces the two-cell direction and five of seven ablation contrasts, but not length or labels, which we therefore do not claim on one instrument. Fifty hand labels (kappa=0.75) agree on direction; we print their seven disagreements in full. We also report nine withdrawn claims, three of them former title claims of this paper.
