llm-kb
← Back to research
Paper

GRACE: Graph-Regularized Agentic Context Evolution for Reliable Long-Horizon LLM Agents

A new arXiv paper proposes GRACE, a method that maintains persistent system-level instructions for LLM agents as a typed semantic graph instead of flat text. This graph-regularized approach enables scoped verification and reliable context evolution over long horizons under distribution shift, addressing verification difficulties from accumulated instructions.

0 engagement·1 source·Fri, Jul 10, 2026, 08:10 AM
The paper introduces Graph-Regularized Agentic Context Evolution (GRACE), which structures the mutable component of agentic context—a persistent system-level instruction updated from operational experience—as a typed semantic graph. This contrasts with flat-text maintenance, which becomes increasingly hard to verify as instructions grow and interact. GRACE allows for scoped verification, improving reliability in long-horizon deployments where models, tools, and harnesses remain fixed but context evolves. The work targets distribution shift challenges in deployed LLM agents.

Entities

GRACE(concept)arXiv(company)

Related

PaperFri, Jul 10, 2026, 03:07 PM

Shared Selective Persistent Memory Architecture Proposed for Agentic LLMs

A new arxiv paper introduces shared selective persistent memory, an architecture that identifies and retains four categories of reusable context (task specifications, data schemas, tool configurations, and tool-use patterns) for agentic LLM systems. This approach aims to solve the context problem where each session starts from zero, while avoiding token-inefficient naive persistence of entire conversation histories. The work is relevant to developers building multi-turn code generation agents.

0 engagement·1 source·arxiv
arXiv
PaperFri, Jul 10, 2026, 03:25 PM

Study analyzes failure trajectories of CLI coding agents as temporal processes

A new arXiv paper presents the first large-scale empirical study of CLI coding-agent failure trajectories, treating failure as a temporal process rather than a final outcome. The study introduces a process-oriented framework to analyze how failures emerge, evolve, and become unrecoverable in LLM-based coding agents.

0 engagement·1 source·arxiv
arXiv
PaperThu, Jul 9, 2026, 05:26 PM

Researchers propose memory agent to combat behavioral state decay in long-horizon tasks

A new arXiv paper identifies 'behavioral state decay' as a failure mode where decision-relevant information gets buried in long trajectories. The authors propose a separate memory agent that actively updates a structured memory bank alongside an unmodified action agent, rather than relying on passive retrieval.

0 engagement·1 source·arxiv
arXiv
PaperThu, Jul 9, 2026, 07:34 PM

GATS framework eliminates LLM calls during agent planning inference

Researchers propose GATS (Graph-Augmented Tree Search), a planning framework that uses a layered world model and UCB1-based tree search to avoid LLM inference during planning, reducing computational cost and stochasticity. The approach outperforms LATS and ReAct on multi-step planning tasks.

0 engagement·1 source·arxiv
arXiv
PaperFri, Jul 10, 2026, 04:22 AM

AgentKGV: Agentic LLM-RAG Framework with Two-Stage Training for the Fact Verification of Knowledge Graphs

Researchers propose AgentKGV, an agentic LLM-RAG framework for verifying facts in knowledge graphs. It uses dynamic routing and iterative query rewriting to handle surface-form mismatches in document-level retrieval, and introduces a two-stage training process to improve accuracy and cost-efficiency for industrial deployment.

0 engagement·1 source·arxiv
arXiv