权威先于效用：用于持久LLM记忆的非补偿性控制
Authority Before Utility: Non-Compensatory Control for Persistent LLM Memory
查看机构详情
- The Institute of Energetic Paradigm（能量范式研究所）
机构由 AI 辅助整理，请以论文原文为准。
浏览论文内容
中文总结 AI 辅助
本文提出权威先于效用的非补偿性控制方法，通过秩归一化惩罚解决LLM持久记忆中已撤销记忆的排除问题，实验显示HARD排除显著优于锁定SOFT。
中文摘要 AI 辅助
持久记忆产生了一个仅靠检索相关性无法解决的控制问题：在更新、删除或撤销使记忆对当前答案不再可采纳之后，该记忆可能仍然高度有用。我们将此形式化为效用与权威之间的分离。对未归一化的效用分数施加固定的有限惩罚，无法保证在该分数的任意正仿射重新参数化下排除；相比之下，秩归一化补偿是尺度不变的，因此构成了更强的经验比较器。我们预期冻结的TIDE/LongMemEval主实验在有效的HELDOUT比较之前被隔离，因为具体化的TIDE适配器将历史年龄与查询相对不可采纳性混为一谈，且对齐的LongMemEval划分没有为预先声明的惩罚选择留下DEV集。因此，我们在Memora Remembering上报告了一个主实验后的替代诊断，其中更新/删除操作提供了项目级遗忘状态。在Qwen3-8B上，DEV从十点归一化SOFT族中选择了lambda = 0.6。在28个依赖簇中的185个HELDOUT单元中，HARD排除产生4.04%的平衡构造误差，而锁定SOFT为19.66%，配对差异为15.61个百分点，20,000次重复簇自助95%区间为[13.07, 18.76]。该效应主要由遗忘值泄漏驱动，而当前值回忆得以保留。这是同Q算子比较的证据，而非标量控制失败的普遍主张，不是对学习权威推理的评估，也不是独立的下游危害终点。
英文摘要
Persistent memory creates a control problem that retrieval relevance alone does not solve: a memory can remain highly useful after an update, deletion, or revocation makes it inadmissible for the current answer. We formalize this as a separation between utility and authority. A fixed finite penalty applied to an unnormalized utility score cannot guarantee exclusion under arbitrary positive-affine reparameterization of that score; by contrast, rank-normalized compensation is scale-invariant and therefore forms a stronger empirical comparator. Our prospectively frozen TIDE/LongMemEval primary was quarantined before a valid HELDOUT comparison because the materialized TIDE adapter conflated historical age with query-relative inadmissibility and the aligned LongMemEval split left no DEV set for the predeclared penalty selection. We therefore report a post-primary replacement diagnostic on Memora Remembering, where update/delete operations provide item-level forgetting state. On Qwen3-8B, DEV selected lambda = 0.6 from a ten-point normalized SOFT family. Across 185 HELDOUT units in 28 dependency clusters, HARD exclusion yields 4.04% balanced construct error versus 19.66% for locked SOFT, a paired difference of 15.61 points with a 20,000-replicate cluster-bootstrap 95% interval of [13.07, 18.76]. The effect is driven primarily by forgotten-value leakage while current-value recall is preserved. This is same-Q operator-comparison evidence, not a universal claim that scalar control fails, not an evaluation of learned authority inference, and not an independent downstream-harm endpoint.
