GUIDE: Guided Updates for In-context Decision Evolution in LLM-Driven Spacecraft Operations

📰 ArXiv cs.AI

GUIDE framework enables cross-episode adaptation in LLM-driven spacecraft operations without weight updates

advanced Published 31 Mar 2026

Action Steps

Define a structured, state-conditioned playbook of natural-language decision rules
Implement a non-parametric policy improvement framework to enable cross-episode adaptation
Use a lightweight acting model to perform actions based on the evolved playbook
Evaluate and refine the GUIDE framework through repeated executions and feedback

Who Needs to Know This

AI engineers and researchers working on LLM-driven spacecraft operations can benefit from GUIDE to improve decision-making across repeated executions. This can also be useful for ml-researchers looking to apply LLMs to real-world problems

Key Insight

💡 GUIDE enables LLMs to learn and adapt across repeated executions without requiring weight updates

Key Takeaways

GUIDE framework enables cross-episode adaptation in LLM-driven spacecraft operations without weight updates

Full Article

Title: GUIDE: Guided Updates for In-context Decision Evolution in LLM-Driven Spacecraft Operations

Abstract:
arXiv:2603.27306v1 Announce Type: cross Abstract: Large language models (LLMs) have been proposed as supervisory agents for spacecraft operations, but existing approaches rely on static prompting and do not improve across repeated executions. We introduce \textsc{GUIDE}, a non-parametric policy improvement framework that enables cross-episode adaptation without weight updates by evolving a structured, state-conditioned playbook of natural-language decision rules. A lightweight acting model perfo

Read full paper → ← Back to Reads