Short Chains, Deep Thoughts: Balancing Reasoning Efficiency and Intra-Segment Capability via Split-Merge Optimization

• URL: http://arxiv.org/abs/2602.03141v2
• Date: Wed, 04 Feb 2026 23:39:47 GMT
• Title: Short Chains, Deep Thoughts: Balancing Reasoning Efficiency and Intra-Segment Capability via Split-Merge Optimization
• Authors: Runquan Gui, Jie Wang, Zhihai Wang, Chi Ma, Jianye Hao, Feng Wu,
• Abstract summary: Large Reasoning Models (LRMs) have demonstrated impressive capabilities in solving complex tasks through the generation of long reasoning chains.<n>We propose textbfCoSMo (textbfSplit-textbfMerge textbfOptimization), a framework designed to eliminate structural redundancy rather than indiscriminately restricting token volume.
• Score: 68.89915707647138
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: While Large Reasoning Models (LRMs) have demonstrated impressive capabilities in solving complex tasks through the generation of long reasoning chains, this reliance on verbose generation results in significant latency and computational overhead. To address these challenges, we propose \textbf{CoSMo} (\textbf{Co}nsistency-Guided \textbf{S}plit-\textbf{M}erge \textbf{O}ptimization), a framework designed to eliminate structural redundancy rather than indiscriminately restricting token volume. Specifically, CoSMo utilizes a split-merge algorithm that dynamically refines reasoning chains by merging redundant segments and splitting logical gaps to ensure coherence. We then employ structure-aligned reinforcement learning with a novel segment-level budget to supervise the model in maintaining efficient reasoning structures throughout training. Extensive experiments across multiple benchmarks and backbones demonstrate that CoSMo achieves superior performance, improving accuracy by \textbf{3.3} points while reducing segment usage by \textbf{28.7\%} on average compared to reasoning efficiency baselines.

Related papers

• Self-Compression of Chain-of-Thought via Multi-Agent Reinforcement Learning [34.10133693878611]
  We propose a multi-agent RL framework that selectively penalizes redundant chunks, while preserving essential reasoning logic.<n>Our framework, Self-Compression via MARL (SCMA), instantiates redundancy detection and evaluation through two specialized agents.<n> Empirical evaluations across model scales demonstrate that SCMA reduces response length by 11.1% to 39.0% while boosting accuracy by 4.33% to 10.02%.
  arXiv  Detail & Related papers  (2026-01-29T16:13:10Z)
• Accelerate Speculative Decoding with Sparse Computation in Verification [49.74839681322316]
  Speculative decoding accelerates autoregressive language model inference by verifying multiple draft tokens in parallel.<n>Existing sparsification methods are designed primarily for standard token-by-token autoregressive decoding.<n>We propose a sparse verification framework that jointly sparsifies attention, FFN, and MoE components during the verification stage to reduce the dominant computation cost.
  arXiv  Detail & Related papers  (2025-12-26T07:53:41Z)
• Efficient Thought Space Exploration through Strategic Intervention [54.35208611253168]
  We propose a novel Hint-Practice Reasoning (HPR) framework that operationalizes this insight through two synergistic components.<n>The framework's core innovation lies in Distributional Inconsistency Reduction (DIR), which dynamically identifies intervention points.<n> Experiments across arithmetic and commonsense reasoning benchmarks demonstrate HPR's state-of-the-art efficiency-accuracy tradeoffs.
  arXiv  Detail & Related papers  (2025-11-13T07:26:01Z)
• SalaMAnder: Shapley-based Mathematical Expression Attribution and Metric for Chain-of-Thought Reasoning [45.78228118909098]
  Chain-of-Thought (CoT) prompting enhances the math reasoning capability of large language models (LLMs) to a large margin.<n>We present textbfSalaMAnder (textbfShtextbfaptextbfley-btextbfased textbfMathematical Expression textbfAttribution atextbfnd Mtextbfettextbfric), a theoretically grounded methodology.<n>We
  arXiv  Detail & Related papers  (2025-09-20T07:38:58Z)
• CCF: A Context Compression Framework for Efficient Long-Sequence Language Modeling [52.05149789178508]
  CCF is a novel context compression framework designed to enable efficient long-context modeling.<n>CCF integrates segment-wise semantic aggregation with key-value memory encoding, forming compact representations.<n> Empirical results on multiple long-context language modeling benchmarks demonstrate that CCF achieves competitive perplexity under high compression ratios.
  arXiv  Detail & Related papers  (2025-09-11T07:13:49Z)
• READER: Retrieval-Assisted Drafter for Efficient LLM Inference [0.0386965802948046]
  Autoregressive Language Models instantiate a factorized likelihood over token sequences, yet their strictly sequential decoding process imposes an intrinsic lower bound on latency inference.<n>This bottleneck has emerged as a central obstacle to the scalable deployment of large-scale generative models.<n>We present READER, a speculative decoding framework that bypasses the training of the auxiliary draft model.
  arXiv  Detail & Related papers  (2025-08-12T16:47:48Z)
• Effects of structure on reasoning in instance-level Self-Discover [0.0]
  This paper introduces iSelf-Discover, an instance-level adaptation of the Self-Discover framework, and using it compares dynamically generated structured reasoning with its unstructured counterpart.<n>Our empirical evaluation across diverse benchmarks using state-of-the-art open-source models supports a consistent advantage for unstructured reasoning.
  arXiv  Detail & Related papers  (2025-07-04T07:28:42Z)
• ConciseHint: Boosting Efficient Reasoning via Continuous Concise Hints during Generation [74.37307916314407]
  We propose a framework dubbed ConciseHint, which continuously encourages the reasoning model to speak concisely.<n>Experiments on the state-of-the-art LRMs, including DeepSeek-R1 and Qwen-3 series, demonstrate that our method can effectively produce concise reasoning.
  arXiv  Detail & Related papers  (2025-06-23T16:20:44Z)
• PixelThink: Towards Efficient Chain-of-Pixel Reasoning [70.32510083790069]
  PixelThink is a simple yet effective scheme that integrates externally estimated task difficulty and internally measured model uncertainty.<n>It learns to compress reasoning length in accordance with scene complexity and predictive confidence.<n> Experimental results demonstrate that the proposed approach improves both reasoning efficiency and overall segmentation performance.
  arXiv  Detail & Related papers  (2025-05-29T17:55:49Z)
• Fractured Chain-of-Thought Reasoning [61.647243580650446]
  We introduce Fractured Sampling, a unified inference-time strategy that interpolates between full CoT and solution-only sampling.<n>We show that Fractured Sampling consistently achieves superior accuracy-cost trade-offs, yielding steep log-linear scaling gains in Pass@k versus token budget.
  arXiv  Detail & Related papers  (2025-05-19T11:30:41Z)
• Scalable Chain of Thoughts via Elastic Reasoning [61.75753924952059]
  Elastic Reasoning is a novel framework for scalable chain of thoughts.<n>It separates reasoning into two phases--thinking and solution--with independently allocated budgets.<n>Our approach produces more concise and efficient reasoning even in unconstrained settings.
  arXiv  Detail & Related papers  (2025-05-08T15:01:06Z)
• Framework for Progressive Knowledge Fusion in Large Language Models Through Structured Conceptual Redundancy Analysis [0.0]
  The organization of latent knowledge within large-scale models poses unique challenges when addressing overlapping representations and optimizing contextual accuracy.<n>A framework was proposed to restructure these redundancies through advanced clustering techniques and dynamic thresholding.<n> Evaluations revealed improved memory efficiency and faster inference times, alongside better alignment in latent knowledge clusters that enhanced interpretability.
  arXiv  Detail & Related papers  (2025-01-23T11:34:04Z)
• HELPNet: Hierarchical Perturbations Consistency and Entropy-guided Ensemble for Scribble Supervised Medical Image Segmentation [4.034121387622003]
  We propose HELPNet, a novel scribble-based weakly supervised segmentation framework.<n>HELPNet integrates three modules to bridge the gap between annotation efficiency and segmentation performance.<n>HELPNet significantly outperforms state-of-the-art methods for scribble-based weakly supervised segmentation.
  arXiv  Detail & Related papers  (2024-12-25T01:52:01Z)
• Rethinking Chain-of-Thought from the Perspective of Self-Training [10.722453877596998]
  Chain-of-thought (CoT) reasoning has emerged as an effective approach for activating latent capabilities in LLMs.<n>We propose a novel CoT framework to improve reasoning performance.<n>Our framework integrates two key components: (i) a task-specific prompt module that optimize the initial reasoning process, and (ii) an adaptive reasoning module that dynamically refines the reasoning process.
  arXiv  Detail & Related papers  (2024-12-14T13:12:50Z)
• Retraining-Free Merging of Sparse MoE via Hierarchical Clustering [24.28646376876676]
  This paper introduces Hierarchical Clustering for Sparsely activated Mixture of Experts (HC-SMoE)<n>HC-SMoE is a task-agnostic expert merging framework for parameter reduction without retraining.<n>We provide theoretical analysis and evaluations across multiple zero-shot language tasks to demonstrate HC-SMoE's effectiveness in state-of-the-art models including Qwen and Mixtral.
  arXiv  Detail & Related papers  (2024-10-11T07:36:14Z)

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.