Self-Polish: Enhance Reasoning in Large Language Models via Problem Refinement

By Zhiheng Xi, Senjie Jin, Yuhao Zhou, Rui Zheng, Songyang Gao, Tao Gui, Qi Zhang, Xuanjing Huang. 2023

67 citations recorded by the source corpus.

To enhance the multi-step reasoning capabilities of large language models, researchers have extensively explored prompting methods, notably the Chain-of-Thought (CoT) method which explicitly elicits human-like rationales. However, they have inadvertently overlooked the potential of enhancing model reasoning performance by formulating higher-quality problems. In this work, we start from the problem side and propose Self-Polish (SP), a novel method that facilitates the model's reasoning by guiding it to progressively refine the given problems to be more comprehensible and solvable. We also explore several automatic prompting varients and propose the Self-Polish prompt bank for the community. SP is orthogonal to all other prompting methods of answer/reasoning side like CoT, allowing for seamless integration with state-of-the-art techniques for further improvement. Thorough experiments show that the proposed method attains notable and consistent effectiveness on five reasoning benchmarks across different models. Furthermore, our method also showcases impressive performance on robustness evaluation. Codes and prompts are available at https://github.com/WooooDyy/Self-Polish.

Subfields: Artificial Intelligence, Large Language Models, Prompting & Reasoning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Satyapriya Krishna
2. Himabindu Lakkaraju
3. Chirag Agarwal
4. Tong Wang

Citations in the corpus

• Large Language Models Can Self-Improve (2022, neutral): “Xi et al. (2023) and Zheng et al. (2023) propose to polish the problem progressively before the model reaching a stable answer.”
• Understanding the Effects of Iterative Prompting on Truthfulness (2024, methodology): “Innovative techniques such as automated feedback strategies [22], ensemble feedback and Pareto optimal self-supervision [23, 24], and iterative refinement methods [25, 26] have shown promise.”
• The rise and potential of large language model based agents: a survey (2025, unknown): “Similarly, LLMs have shown the emergent ability of planning in forms of goal reformulation [80, 108], task decomposition [79, 109], and adjusting plans in response to environmental changes [81,110].”