Large Language Models (LLMs) have demonstrated impressive reasoning capabilities, yet their performance is highly dependent on the prompting strategy and model scale. While reinforcement learning and fine-tuning have been deployed to boost reasoning, these approaches incur substantial computational and data overhead. In this work, we introduce Adaptive Graph of Thoughts (AGoT), a dynamic, graph-based inference framework that enhances LLM reasoning solely at test time. Rather than relying on fixed-step methods like Chain of Thought (CoT) or Tree of Thoughts (ToT), AGoT recursively decomposes complex queries into structured subproblems, forming an dynamic directed acyclic graph (DAG) of interdependent reasoning steps. By selectively expanding only those subproblems that require further analysis, AGoT unifies the strengths of chain, tree, and graph paradigms into a cohesive framework that allocates computation where it is most needed. We validate our approach on diverse benchmarks spanning multi-hop retrieval, scientific reasoning, and mathematical problem-solving, achieving up to 46.2% improvement on scientific reasoning tasks (GPQA) - comparable to gains achieved through computationally intensive reinforcement learning approaches and outperforming state-of-the-art iterative approaches. These results suggest that dynamic decomposition and structured recursion offer a scalable, cost-effective alternative to post-training modifications, paving the way for more robust, general-purpose reasoning in LLMs.

翻译：大型语言模型（LLMs）已展现出卓越的推理能力，但其性能高度依赖于提示策略与模型规模。尽管强化学习与微调技术已被用于提升推理能力，这些方法仍需耗费大量计算与数据资源。本研究提出自适应思维图谱（AGoT）——一种基于动态图结构的推理框架，该框架仅在测试阶段即可增强LLM的推理能力。与思维链（CoT）或思维树（ToT）等固定步骤方法不同，AGoT通过递归方式将复杂查询分解为结构化子问题，形成相互依赖推理步骤的动态有向无环图（DAG）。通过选择性地扩展需要深入分析的子问题，AGoT将链式、树状与图式范式的优势统一于协同框架中，从而将计算资源精准分配至最需要的环节。我们在涵盖多跳检索、科学推理与数学问题求解的多样化基准测试中验证了该方法，其中科学推理任务（GPQA）性能最高提升达46.2%——该提升幅度与计算密集型强化学习方法相当，且优于当前最先进的迭代方法。这些结果表明，动态分解与结构化递归为训练后模型优化提供了一种可扩展、高性价比的替代方案，为LLM实现更鲁棒、更通用的推理能力开辟了新路径。