The Conductor and the Engine: A Path Towards Co-Designed Reasoning

• URL: http://arxiv.org/abs/2509.19762v1
• Date: Wed, 24 Sep 2025 05:09:43 GMT
• Title: The Conductor and the Engine: A Path Towards Co-Designed Reasoning
• Abstract summary: LLM reasoning relies on extensive test-time computation, driven by internal model training and external agentic orchestration.<n>We analyze this capability-cost trade-off and introduce an optimized reasoning workflow (cepo) that empowers smaller open-source models to outperform models multiple times their size.
• Score: 2.112973599034511
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Modern LLM reasoning relies on extensive test-time computation, driven by internal model training and external agentic orchestration. However, this synergy is often inefficient, as model verbosity and poor instruction following lead to wasted compute. We analyze this capability-cost trade-off and introduce an optimized reasoning workflow (\cepo) that empowers smaller open-source models to outperform models multiple times their size. We will open-source this workflow to enable further research. Our work demonstrates a clear path toward co-designing orchestration frameworks with the underlying model capabilities to unlock powerful reasoning in small-to-medium sized models.

Related papers

• AHD Agent: Agentic Reinforcement Learning for Automatic Heuristic Design [5.7547035595428255]
  Automatic design (AHD) has emerged as a promising paradigm for solving NP-hard optimization problems.<n>Existing AHD frameworks typically treat large language models (LLMs) as passive generators within fixed contexts.<n>We propose AHD Agent, a novel tool-integrated, multi-turn framework that empowers LLMs to proactively decide whether to generate tools or retrieve targeted evidence.
  arXiv  Detail & Related papers  (2026-05-09T07:36:45Z)
• AgentCollab: A Self-Evaluation-Driven Collaboration Paradigm for Efficient LLM Agents [37.232397795331444]
  We present AgentCollab, a self-driven collaborative inference framework that dynamically coordinates models with different reasoning capacities during agent execution.<n>Instead of relying on external routing modules, the framework uses the agent's own self-reflection signal to determine whether the current reasoning trajectory is making meaningful progress.
  arXiv  Detail & Related papers  (2026-03-27T03:07:34Z)
• O-Researcher: An Open Ended Deep Research Model via Multi-Agent Distillation and Agentic RL [28.10102994309489]
  We introduce a novel framework for the automated synthesis of sophisticated, research-grade instructional data.<n>Our approach centers on a multi-agent workflow where collaborative AI agents simulate complex tool-integrated reasoning.<n>We develop a two-stage training strategy that integrates supervised fine-tuning with a novel reinforcement learning method.
  arXiv  Detail & Related papers  (2026-01-07T09:31:10Z)
• Revisiting Model Interpolation for Efficient Reasoning [27.32667995137936]
  We revisit the simplest merging method that interpolates two weights directly.<n>We observe that model follows a three-stage evolutionary paradigm with distinct behaviors on the reasoning trajectory.
  arXiv  Detail & Related papers  (2025-10-13T03:30:01Z)
• MCP: A Control-Theoretic Orchestration Framework for Synergistic Efficiency and Interpretability in Multimodal Large Language Models [0.0]
  This study proposes a three-layer collaboration framework based on model-controller-task adaptation (MCP)<n> Experiments show that the MCP framework improves the performance of cross-modal benchmarking tasks, such as GLUE, COCO, ScienceQA, etc., by 15-30% compared with the baseline model, improves the reasoning efficiency by 40%, and generates the interpretable intermediate results through the Presenter layer, obtaining 90% of the manual interpretability scores.
  arXiv  Detail & Related papers  (2025-09-20T09:44:11Z)
• Don't Overthink It: A Survey of Efficient R1-style Large Reasoning Models [49.598776427454176]
  Large Reasoning Models (LRMs) have gradually become a research hotspot due to their outstanding performance in handling complex tasks.<n>However, with the widespread application of these models, the problem of overthinking has gradually emerged.<n>Various efficient reasoning methods have been proposed, aiming to reduce the length of reasoning paths without compromising model performance and reasoning capability.
  arXiv  Detail & Related papers  (2025-08-04T06:54:31Z)
• Synergistic Weak-Strong Collaboration by Aligning Preferences [53.47675666475273]
  Current Large Language Models (LLMs) excel in general reasoning yet struggle with specialized tasks requiring proprietary or domain-specific knowledge.<n>We propose a collaborative framework that pairs a specialized weak model with a general strong model.<n>We find that the collaboration significantly outperforms each model alone by leveraging complementary strengths.
  arXiv  Detail & Related papers  (2025-04-21T15:57:33Z)
• LoRE-Merging: Exploring Low-Rank Estimation For Large Language Model Merging [10.33844295243509]
  We propose a unified framework for model merging based on low-rank estimation of task vectors without the need for access to the base model, named textscLoRE-Merging.<n>Our approach is motivated by the observation that task vectors from fine-tuned models frequently exhibit a limited number of dominant singular values, making low-rank estimations less prone to interference.
  arXiv  Detail & Related papers  (2025-02-15T10:18:46Z)
• Satori: Reinforcement Learning with Chain-of-Action-Thought Enhances LLM Reasoning via Autoregressive Search [57.28671084993782]
  Large language models (LLMs) have demonstrated remarkable reasoning capabilities across diverse domains.<n>Recent studies have shown that increasing test-time computation enhances LLMs' reasoning capabilities.<n>We propose a two-stage training paradigm: 1) a small-scale format tuning stage to internalize the COAT reasoning format and 2) a large-scale self-improvement stage leveraging reinforcement learning.
  arXiv  Detail & Related papers  (2025-02-04T17:26:58Z)
• RAVEN: In-Context Learning with Retrieval-Augmented Encoder-Decoder Language Models [57.12888828853409]
  RAVEN is a model that combines retrieval-augmented masked language modeling and prefix language modeling. Fusion-in-Context Learning enables the model to leverage more in-context examples without requiring additional training. Our work underscores the potential of retrieval-augmented encoder-decoder language models for in-context learning.
  arXiv  Detail & Related papers  (2023-08-15T17:59:18Z)
• Scaling Vision-Language Models with Sparse Mixture of Experts [128.0882767889029]
  We show that mixture-of-experts (MoE) techniques can achieve state-of-the-art performance on a range of benchmarks over dense models of equivalent computational cost. Our research offers valuable insights into stabilizing the training of MoE models, understanding the impact of MoE on model interpretability, and balancing the trade-offs between compute performance when scaling vision-language models.
  arXiv  Detail & Related papers  (2023-03-13T16:00:31Z)
• When to Update Your Model: Constrained Model-based Reinforcement Learning [50.74369835934703]
  We propose a novel and general theoretical scheme for a non-decreasing performance guarantee of model-based RL (MBRL) Our follow-up derived bounds reveal the relationship between model shifts and performance improvement. A further example demonstrates that learning models from a dynamically-varying number of explorations benefit the eventual returns.
  arXiv  Detail & Related papers  (2022-10-15T17:57:43Z)

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.