智能体
arxiv.org ·
基于三信号互补的可解释记忆决策控制器：解耦 LLM 智能体的置信度与一致性
该 arXiv 预印本指出，LLM 记忆系统多聚焦高效检索，却忽视“检索到的记忆是否可信”这一决策问题；在记忆库存在冲突立场时，标准 RAG 会盲目注入记忆并放大幻觉。作者受前额叶记忆信号机制启发，提出位于检索与生成之间的零参数记忆决策层 MDL，其核心是三信号互补编码器，通过基于 QR 的正交子空间投影融合相关性、可靠性与任务风险，并结合元工作记忆信号，形成可解释的决策表示以量化记忆可信度。MDL 显式解耦置信度与一致性，引入风险反转与显式弃答。作者称在主流大模型和多个开源数据集上，冲突记忆下的幻觉率在通用场景降低约 56.04%，高风险场景接近零幻觉；控制器为全白盒几何运算、无需训练参数，每次决策约 0.14 ms，比前置嵌入检索快约 50 倍，比 LLM 自评估调用快四到五个数量级。上述效果均为论文自报，尚待独立验证。
AI推荐理由 arXiv 预印本，作者与摘要信息完整，问题定义清晰（RAG 在记忆冲突下放大幻觉），提出零参数、白盒的 Memory Decision Layer，并给出量化结果（冲突记忆下幻觉率降低约 56.04%、高风险场景接近零幻觉、每次决策约 0.14 ms）。新信息在于把“是否信任检索到的记忆”作为独立决策层，并用三信号互补编码解耦置信度与一致性。局限是结果均为作者自报，尚未见同行评审或第三方复现，摘要未披露数据集与基线细节，56.04% 等数字需谨慎看待。
78
正文
Computer Science > Computation and Language
arXiv:2609.22043 (cs)
[Submitted on 18 Sep 2026]
Title:An Interpretable Memory Decision Controller for LLM Agents Based on Three-Signal Complementarity: Decoupling Confidence and Consistency
Authors:Yiming Zhang, Jinghong Zhang, Haoran Zhao, Yiren Ma, Chunlei Zhao
View a PDF of the paper titled An Interpretable Memory Decision Controller for LLM Agents Based on Three-Signal Complementarity: Decoupling Confidence and Consistency, by Yiming Zhang and 4 other authors
View PDF
HTML (experimental)
Abstract:Memory systems for large language models have focused predominantly on efficient retrieval, whereas the decision of whether retrieved memories should be trusted has received comparatively little attention. When the memory store contains conflicting positions, standard retrieval-augmented generation (RAG) blindly injects memories and amplifies hallucinations: in models susceptible to memory injection, the RAG hallucination rate under conflicting memories is markedly higher than that of a memory-free baseline. Inspired by memory signaling mechanisms in the prefrontal cortex, we propose the Memory Decision Layer (MDL), a zero-parameter memory decision controller situated between the retrieval and generation stages. Its core is a three-signal complementary encoder that fuses relevance, reliability, and task risk through QR-based orthogonal subspace projection and a meta-working-memory signal into an interpretable decision representation that quantifies the trustworthiness of retrieved memories. Building on this encoder, MDL explicitly decouples confidence from consistency and introduces risk inversion and explicit abstention. Evaluations on mainstream large language models and multiple open-source datasets show that MDL reduces the hallucination rate under conflicting memories by about 56.04% in general scenarios and approaches zero hallucination in high-risk scenarios. The controller is fully white-box: it relies purely on geometric operations, requires no trained parameters, and adds only about 0.14 ms per decision -- roughly 50x faster than the embedding-retrieval step that precedes it and four to five orders of magnitude faster than an LLM self-evaluation call.
Comments:
Subjects:
Computation and Language (cs.CL)
Cite as:
arXiv:2609.22043 [cs.CL]
(or
arXiv:2609.22043v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2609.22043
Focus to learn more
arXiv-issued DOI via DataCite (pending registration)
Submission history
From: Yiming Zhang [view email]
[v1]
Fri, 18 Sep 2026 17:34:13 UTC (1,721 KB)
Full-text links:
Access Paper:
View a PDF of the paper titled An Interpretable Memory Decision Controller for LLM Agents Based on Three-Signal Complementarity: Decoupling Confidence and Consistency, by Yiming Zhang and 4 other authors
View PDF
HTML (experimental)
TeX Source
view license
Current browse context:
cs.CL
new
|
recent
| 2026-09
Change to browse by:
cs
References & Citations
NASA ADS
Google Scholar
Semantic Scholar
export BibTeX citation
Loading...
BibTeX formatted citation
×
loading...
Data provided by:
Bookmark
Bibliographic Tools
Bibliographic and Citation Tools
Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media
Code, Data and Media Associated with this Article
alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos
Demos
Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers
Recommenders and Search Tools
Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
Author
Venue
Institution
Topic
About arXivLabs
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Which authors of this paper are endorsers? |
Disable MathJax (What is MathJax?)