U-Fold: Dynamic Intent-Aware Context Folding for User-Centric Agents

TL;DR

U-Fold introduces a dynamic, intent-aware context folding method for user-centric dialogue agents, significantly improving their ability to handle long, complex interactions by maintaining relevant information and evolving user intent.

Contribution

The paper presents U-Fold, a novel framework that dynamically summarizes dialogue context with intent-awareness, addressing limitations of existing methods in multi-turn, user-centric tasks.

Findings

U-Fold outperforms ReAct with a 71.4% win rate in long-context scenarios.

U-Fold improves over prior folding baselines by up to 27.0%.

U-Fold effectively manages long, noisy, multi-turn dialogues.

Abstract

Large language model (LLM)-based agents have been successfully deployed in many tool-augmented settings, but their scalability is fundamentally constrained by context length. Existing context-folding methods mitigate this issue by summarizing past interactions, yet they are typically designed for single-query or single-intent scenarios. In more realistic user-centric dialogues, we identify two major failure modes: (i) they irreversibly discard fine-grained constraints and intermediate facts that are crucial for later decisions, and (ii) their summaries fail to track evolving user intent, leading to omissions and erroneous actions. To address these limitations, we propose U-Fold, a dynamic context-folding framework tailored to user-centric tasks. U-Fold retains the full user--agent dialogue and tool-call history but, at each turn, uses two core components to produce an intent-aware, evolving dialogue summary and a compact, task-relevant tool log. Extensive experiments on $\tau$-bench, $\tau^2$-bench, VitaBench, and harder context-inflated settings show that U-Fold consistently outperforms ReAct (achieving a 71.4% win rate in long-context settings) and prior folding baselines (with improvements of up to 27.0%), particularly on long, noisy, multi-turn tasks. Our study demonstrates that U-Fold is a promising step toward transferring context-management techniques from single-query benchmarks to realistic user-centric applications.