PATS: Process-Level Adaptive Thinking Mode Switching

• URL: http://arxiv.org/abs/2505.19250v1
• Date: Sun, 25 May 2025 17:58:50 GMT
• Title: PATS: Process-Level Adaptive Thinking Mode Switching
• Authors: Yi Wang, Junxiao Liu, Shimao Zhang, Jiajun Chen, Shujian Huang,
• Abstract summary: Current large-language models (LLMs) typically adopt a fixed reasoning strategy, either simple or complex, for all questions, regardless of their difficulty.<n>This neglect of variation in task and reasoning process complexity leads to an imbalance between performance and efficiency.<n>Existing methods attempt to implement training-free fast-slow thinking system switching to handle problems of varying difficulty, but are limited by coarse-grained solution-level strategy adjustments.<n>We propose a novel reasoning paradigm: Process-Level Adaptive Thinking Mode Switching (PATS), which enables LLMs to dynamically adjust their reasoning strategy based on the difficulty of each step, optimizing the balance between
• Score: 53.53401063490537
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Current large-language models (LLMs) typically adopt a fixed reasoning strategy, either simple or complex, for all questions, regardless of their difficulty. This neglect of variation in task and reasoning process complexity leads to an imbalance between performance and efficiency. Existing methods attempt to implement training-free fast-slow thinking system switching to handle problems of varying difficulty, but are limited by coarse-grained solution-level strategy adjustments. To address this issue, we propose a novel reasoning paradigm: Process-Level Adaptive Thinking Mode Switching (PATS), which enables LLMs to dynamically adjust their reasoning strategy based on the difficulty of each step, optimizing the balance between accuracy and computational efficiency. Our approach integrates Process Reward Models (PRMs) with Beam Search, incorporating progressive mode switching and bad-step penalty mechanisms. Experiments on diverse mathematical benchmarks demonstrate that our methodology achieves high accuracy while maintaining moderate token usage. This study emphasizes the significance of process-level, difficulty-aware reasoning strategy adaptation, offering valuable insights into efficient inference for LLMs.

Related papers

• Reasoning on a Budget: A Survey of Adaptive and Controllable Test-Time Compute in LLMs [45.83245433138508]
  Large language models (LLMs) have rapidly progressed into general-purpose agents capable of solving a broad spectrum of tasks.<n>They apply fixed inference-time compute regardless of task complexity, often overthinking simple problems while underthinking hard ones.<n>This survey presents a comprehensive review of efficient test-time compute strategies, which aim to improve the computational efficiency of LLM reasoning.
  arXiv  Detail & Related papers  (2025-07-02T18:27:42Z)
• AdaptiveLLM: A Framework for Selecting Optimal Cost-Efficient LLM for Code-Generation Based on CoT Length [5.856039862078523]
  We introduce AdaptiveLLM, a framework that dynamically selects optimal Large Language Models (LLMs) for a given coding task by automatically assessing task difficulty.<n>Our framework first estimates task difficulty using Chain-of-Thought lengths generated by reasoning model, clusters these into three difficulty levels via k-means, and fine-tunes CodeBERT to embed difficulty-aware features.<n>Our framework achieves a 7.86% improvement in pass@1 score while reducing resource consumption by 88.9% compared to baseline method ComplexityNet.
  arXiv  Detail & Related papers  (2025-06-12T09:43:48Z)
• TACO: Think-Answer Consistency for Optimized Long-Chain Reasoning and Efficient Data Learning via Reinforcement Learning in LVLMs [50.820065021136024]
  DeepSeek R1 has significantly advanced complex reasoning for large language models (LLMs)<n>Recent methods have attempted to replicate R1's reasoning capabilities in multimodal settings.<n>We propose TACO, a novel reinforcement learning algorithm for visual reasoning.
  arXiv  Detail & Related papers  (2025-05-27T06:30:48Z)
• Incentivizing Dual Process Thinking for Efficient Large Language Model Reasoning [75.04643265875072]
  Large reasoning models (LRMs) have demonstrated strong performance on complex reasoning tasks, but often suffer from overthinking.<n>Inspired by the dual process theory in cognitive science, we propose Adaptive Cognition Policy Optimization.<n>ACPO enables LRMs to achieve efficient reasoning through adaptive cognitive allocation and dynamic system switch.
  arXiv  Detail & Related papers  (2025-05-22T07:15:08Z)
• DISC: Dynamic Decomposition Improves LLM Inference Scaling [54.87338295793453]
  This paper introduces dynamic decomposition, a method that automatically splits solution and reasoning traces into steps during inference.<n>Experiments on coding and math benchmarks show that dynamic decomposition performs better than static methods.
  arXiv  Detail & Related papers  (2025-02-23T20:37:32Z)
• Forest-of-Thought: Scaling Test-Time Compute for Enhancing LLM Reasoning [40.069109287947875]
  We propose a novel reasoning framework called Forest-of-Thought (FoT)<n>FoT integrates multiple reasoning trees to leverage collective decision-making for solving complex logical problems.<n>FoT employs sparse activation strategies to select the most relevant reasoning paths, improving both efficiency and accuracy.
  arXiv  Detail & Related papers  (2024-12-12T09:01:18Z)
• Robust Analysis of Multi-Task Learning Efficiency: New Benchmarks on Light-Weighed Backbones and Effective Measurement of Multi-Task Learning Challenges by Feature Disentanglement [69.51496713076253]
  In this paper, we focus on the aforementioned efficiency aspects of existing MTL methods. We first carry out large-scale experiments of the methods with smaller backbones and on a the MetaGraspNet dataset as a new test ground. We also propose Feature Disentanglement measure as a novel and efficient identifier of the challenges in MTL.
  arXiv  Detail & Related papers  (2024-02-05T22:15:55Z)
• Adaptive-Solver Framework for Dynamic Strategy Selection in Large Language Model Reasoning [31.643337118330944]
  Large Language Models (LLMs) demonstrate impressive ability in handling reasoning tasks.<n>Most LLM-based methods adopt a one-size-fits-all approach.<n>Inflexibility of these methods can bring unnecessary computational overhead or sub-optimal performance.
  arXiv  Detail & Related papers  (2023-10-01T12:28:36Z)
• Reinforcement Learning Methods for Wordle: A POMDP/Adaptive Control Approach [0.3093890460224435]
  We address the solution of the popular Wordle puzzle, using new reinforcement learning methods. For the Wordle puzzle, they yield on-line solution strategies that are very close to optimal at relatively modest computational cost.
  arXiv  Detail & Related papers  (2022-11-15T03:46:41Z)
• Meta-Learning with Neural Tangent Kernels [58.06951624702086]
  We propose the first meta-learning paradigm in the Reproducing Kernel Hilbert Space (RKHS) induced by the meta-model's Neural Tangent Kernel (NTK) Within this paradigm, we introduce two meta-learning algorithms, which no longer need a sub-optimal iterative inner-loop adaptation as in the MAML framework. We achieve this goal by 1) replacing the adaptation with a fast-adaptive regularizer in the RKHS; and 2) solving the adaptation analytically based on the NTK theory.
  arXiv  Detail & Related papers  (2021-02-07T20:53:23Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.