PulseTrendingJournal ClubResearchersJournalsExplore
Instagram
HomeTrendingJournal ClubExplore
Synapse
⌘+K
Synapse
September 17, 2025Proceedings of the International Conference on Automated Planning and SchedulingOpen Access

Automating the Generation of Prompts for LLM-based Action Choice in PDDL Planning

View Full Paper
Ask AI
Bookmark
Share

Authors

KSKatharina SteinDFDaniel FišerJHJörg Hoffmann

Discussion

Loading...

Member takes

Overview

Automation of NL prompts improves LLM planning performance in PDDL, suggesting enhanced evaluation potential.

Key Points

  • Automated NL prompts improve LLM planning performance compared to manual prompts, enhancing evaluation capabilities.
  • Evaluation reveals that NL prompts generated from PDDL input yield better outcomes than template-based NL prompts.
  • This research shows how LLMs can automate the generation of prompts for improved planning in PDDL contexts.
  • Automated generation facilitates larger experiments, providing valuable insights into LLM planning performance in multiple domains.

Cite This Study

Stein et al. (2025) studied this question.

synapsesocial.com/papers/68d4566c31b076d99fa5badbhttps://doi.org/10.1609/icaps.v35i1.36126
View Full Paper
Ask AI
Bookmark
Share