Controllable Memory Usage: Balancing Anchoring and Innovation in Long-Term Human-Agent Interaction

2026-01-08

arXiv.org

AI Summary

This work addresses the challenge of long-term human-agent interaction, where large language model agents often suffer from memory anchoring or loss of historical context due to binary “all-or-nothing” memory usage. To overcome this limitation, the authors propose SteeM, a novel framework that explicitly models memory dependence as a continuous, user-controllable dimension, enabling fine-grained adjustment between fully novel generation and high-fidelity recall. By introducing behavioral metrics to quantify the influence of memory on agent responses, SteeM demonstrates significant performance gains over conventional prompt engineering and fixed memory-masking strategies across diverse interaction scenarios. The framework thus enables more precise and effective personalized collaboration through adaptive memory utilization.

Technology Category

• Multiagent Systems: Agent Communication
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

As LLM-based agents are increasingly used in long-term interactions, cumulative memory is critical for enabling personalization and maintaining stylistic consistency. However, most existing systems adopt an ``all-or-nothing''approach to memory usage: incorporating all relevant past information can lead to \textit{Memory Anchoring}, where the agent is trapped by past interactions, while excluding memory entirely results in under-utilization and the loss of important interaction history. We show that an agent's reliance on memory can be modeled as an explicit and user-controllable dimension. We first introduce a behavioral metric of memory dependence to quantify the influence of past interactions on current outputs. We then propose \textbf{Stee}rable \textbf{M}emory Agent, \texttt{SteeM}, a framework that allows users to dynamically regulate memory reliance, ranging from a fresh-start mode that promotes innovation to a high-fidelity mode that closely follows interaction history. Experiments across different scenarios demonstrate that our approach consistently outperforms conventional prompting and rigid memory masking strategies, yielding a more nuanced and effective control for personalized human-agent collaboration.

Problem

• Memory Anchoring
• Long-Term Interaction
• Memory Usage
• Personalization
• Human-Agent Collaboration

Innovation

• Controllable Memory
• Memory Anchoring
• Long-Term Interaction
• Steerable Agent
• Memory Dependence