Localizing and Correcting Errors for LLM-based Planners

TL;DR

This paper introduces Localized In-Context Learning (L-ICL), a method to improve LLM-based planners by iteratively correcting specific planning errors, significantly increasing the validity of generated plans across multiple domains.

Contribution

The paper proposes L-ICL, a novel targeted correction technique that outperforms traditional instruction-based methods in improving LLM planning accuracy.

Findings

L-ICL achieves 89% valid plans on gridworld with 60 examples.

L-ICL outperforms baselines by up to 30% in plan validity.

Effective across multiple domains and LLM architectures.

Abstract

Large language models (LLMs) have demonstrated strong reasoning capabilities on math and coding, but frequently fail on symbolic classical planning tasks. Our studies, as well as prior work, show that LLM-generated plans routinely violate domain constraints given in their instructions (e.g., walking through walls). To address this failure, we propose iteratively augmenting instructions with Localized In-Context Learning (L-ICL) demonstrations: targeted corrections for specific failing steps. Specifically, L-ICL identifies the first constraint violation in a trace and injects a minimal input-output example giving the correct behavior for the failing step. Our proposed technique of L-ICL is much effective than explicit instructions or traditional ICL, which adds complete problem-solving trajectories, and many other baselines. For example, on an 8x8 gridworld, L-ICL produces valid plans 89% of the time with only 60 training examples, compared to 59% for the best baseline, an increase of 30%. L-ICL also shows dramatic improvements in other domains (gridworld navigation, mazes, Sokoban, and BlocksWorld), and on several LLM architectures.