ACR: Adaptive Context Refactoring via Context Refactoring Operators for Multi-Turn Dialogue

2026-01-09

arXiv.org

AI Summary

This work addresses the challenge of maintaining long-term consistency and factual accuracy in multi-turn dialogues, where large language models often suffer from contextual inertia and state drift. To mitigate these issues, the authors propose the Adaptive Context Reconstruction (ACR) framework, which decouples context management from reasoning by dynamically monitoring dialogue history and applying interventions as needed. ACR incorporates a library of context reconstruction operators, an adaptive intervention mechanism, and a teacher-guided self-evolution training paradigm to enable dynamic compression and reorganization of context. Experimental results demonstrate that ACR significantly outperforms existing baselines on multi-turn dialogue tasks while effectively reducing token consumption.

Technology Category

Machine Learning: Large Multimodal Models (LMMs)Natural Language Processing: Language Grounding & Multi-modal NLPMultiagent Systems: Adversarial Agents

Application Category

Search and Retrieval-Augmented AI: Personalized, context-aware and across-device searchUser Modeling, Personalization and Recommendation: Fairness-aware retrieval and rankingSemantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Large Language Models (LLMs) have shown remarkable performance in multi-turn dialogue. However, in multi-turn dialogue, models still struggle to stay aligned with what has been established earlier, follow dependencies across many turns, and avoid drifting into incorrect facts as the interaction grows longer. Existing approaches primarily focus on extending the context window, introducing external memory, or applying context compression, yet these methods still face limitations such as \textbf{contextual inertia} and \textbf{state drift}. To address these challenges, we propose the \textbf{A}daptive \textbf{C}ontext \textbf{R}efactoring \textbf{(ACR)} Framework, which dynamically monitors and reshapes the interaction history to mitigate contextual inertia and state drift actively. ACR is built on a library of context refactoring operators and a teacher-guided self-evolving training paradigm that learns when to intervene and how to refactor, thereby decoupling context management from the reasoning process. Extensive experiments on multi-turn dialogue demonstrate that our method significantly outperforms existing baselines while reducing token consumption.

Problem

• multi-turn dialogue
• contextual inertia
• state drift
• large language models
• context management

Innovation

• Adaptive Context Refactoring
• Context Refactoring Operators
• Contextual Inertia
• State Drift
• Multi-Turn Dialogue