Language Model Planners do not Scale, but do Formalizers?

Owen Jiang Cassie Huang Ashish Sabharwal Li Zhang · 2026

Read Language Model Planners do not Scale, but do Formalizers? by Owen Jiang Cassie Huang Ashish Sabharwal Li Zhang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-23844.txt.

Description

Jiang and co-authors argue that language models should write programs, not raw planning files, for hard tasks. Models often turn problem text into PDDL for a solver, but scores drop as problems grow. The authors define unraveling problems, where a short text hides a huge PDDL file. Standard formalizers fail to scale on these. Their higher-order approach has the model write a small program that prints the large PDDL. This separates output length from problem size and improves results.

Catalog record

CREATOROwen Jiang Cassie Huang Ashish Sabharwal Li Zhang
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50537
LICENSEhttp://creativecommons.org/licenses/by-sa/4.0/
RIGHTSCC BY SA
SOURCEarXiv

Source: https://arxiv.org/abs/2603.23844

Browse the full library · About the Tower of Babel