War of Thoughts: Competition Stimulates Stronger Reasoning in Large Language Models

Recent advances in Large Language Models (LLMs) have reshaped the landscape of reasoning tasks, particularly through test-time scaling (TTS) to enhance LLM reasoning.Prior research has used structures such as trees or graphs to guide LLMs in searching for optimal solutions.These methods are time-consuming and require a strong reward model (RM) to support effective solution space exploration.Tournament-style approaches eliminate the reliance on RMs through comparative evaluation but suffer from transitivity dilemmas, leading to unstable ordering.To address these issues, we propose War of Thoughts (WoT), a novel post-hoc method that enhances reasoning without finetuning.WoT comprises two distinct stages: (1) Exploration, in which diverse and meaningful candidate solutions are generated through contrastive demonstrations and multigranularity reasoning specifications; and (2) Competition, where these candidate solutions are subjected to multiple rounds of matchups within a competitive arena.Throughout this iterative process, the solutions are optimized and improved, with the optimal solution being determined based on Elo ratings.Extensive experiments across various LLMs demonstrate the superiority of WoT, surpassing baselines by 10-30%.WoT can effectively stimulate stronger reasoning abilities, achieving impressive TTS performance in both generation budget and model size.It shows higher scalability efficiency compared to the baseline within the same budget.Notably, WoT exhibits excellent scalability with model size, even outperforming a 72B model despite using a 7B model.4 Compete for the Best Solution 3 Produce Grouped Kickoff Solutions 2 Generate Contrastive Demonstrations 1 Obtain Task-specific Common Errors LLM LLM Question: What is the sum of all values of for which the expression +6 2 -5+4 is undefined?Answer: The given expression is undefined when Demonstration Logical Reasoning Error: occurs when there is a fault in the logical progression or assumptions Conceptual Misunderstanding: occurs when there is an incorrect interpretation of a key concept Calculation Error: occurs when an arithmetic mistake, unit conversion error, or

Authors

Publication Details

Published
2025-01-01
DOI
https://doi.org/10.18653/v1/2025.findings-acl.1118
Primary Topic
Topic Modeling
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

War of Thoughts: Competition Stimulates Stronger Reasoning in Large Language Models

Yifu Yuan, Jianye Hao, Jinyi Liu, Yan Zheng et al.
Topic Modeling
article

War of Thoughts: Competition Stimulates Stronger Reasoning in Large Language Models

Yifu Yuan, Jianye Hao, Jinyi Liu, Yan Zheng, Yibin Chen
article en

Abstract

Recent advances in Large Language Models (LLMs) have reshaped the landscape of reasoning tasks, particularly through test-time scaling (TTS) to enhance LLM reasoning.Prior research has used structures such as trees or graphs to guide LLMs in searching for optimal solutions.These methods are time-consuming and require a strong reward model (RM) to support effective solution space exploration.Tournament-style approaches eliminate the reliance on RMs through comparative evaluation but suffer from transitivity dilemmas, leading to unstable ordering.To address these issues, we propose War of Thoughts (WoT), a novel post-hoc method that enhances reasoning without finetuning.WoT comprises two distinct stages: (1) Exploration, in which diverse and meaningful candidate solutions are generated through contrastive demonstrations and multigranularity reasoning specifications; and (2) Competition, where these candidate solutions are subjected to multiple rounds of matchups within a competitive arena.Throughout this iterative process, the solutions are optimized and improved, with the optimal solution being determined based on Elo ratings.Extensive experiments across various LLMs demonstrate the superiority of WoT, surpassing baselines by 10-30%.WoT can effectively stimulate stronger reasoning abilities, achieving impressive TTS performance in both generation budget and model size.It shows higher scalability efficiency compared to the baseline within the same budget.Notably, WoT exhibits excellent scalability with model size, even outperforming a 72B model despite using a 7B model.4 Compete for the Best Solution 3 Produce Grouped Kickoff Solutions 2 Generate Contrastive Demonstrations 1 Obtain Task-specific Common Errors LLM LLM Question: What is the sum of all values of for which the expression +6 2 -5+4 is undefined?Answer: The given expression is undefined when Demonstration Logical Reasoning Error: occurs when there is a fault in the logical progression or assumptions Conceptual Misunderstanding: occurs when there is an incorrect interpretation of a key concept Calculation Error: occurs when an arithmetic mistake, unit conversion error, or

National Natural Science Foundation of China, National Key Research and Development Program of China
Openalex Percentile: Top 90%
Topic Modeling
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.