We present THOUGHTSCULPT, a general reasoning and search method for tasks with outputs that can be decomposed into components. THOUGHTSCULPT explores a search tree of potential solutions using Monte Carlo Tree Search (MCTS), building solutions one action at a time and evaluating according to any domain-specific heuristic, which in practice is often simply an LLM evaluator. Critically, our action space includes revision actions: THOUGHTSCULPT may choose to revise part of its previous output rather than continuing to build the rest of its output. Empirically, THOUGHTSCULPT outperforms state-of-the-art reasoning methods across three challenging tasks: Story Outline Improvement (up to +30% interestingness), Mini-Crosswords Solving (up to +16% word success rate), and Constrained Generation (up to +10% concept coverage).
@article{arxiv.2404.05966,
title = {THOUGHTSCULPT: Reasoning with Intermediate Revision and Search},
author = {Yizhou Chi and Kevin Yang and Dan Klein},
journal= {arXiv preprint arXiv:2404.05966},
year = {2025}
}
Comments
Accepted to NAACL 2025 Findings. Code and data available at https://github.com/cyzus/thoughtsculpt