让大模型推理更高效，通过模块化树搜索框架提升逻辑、规划与工具使用能力。
LiTS: A Modular Framework for LLM Tree Search
- 将树搜索拆分为策略、转移、奖励三个可复用模块，支持灵活组合。
- 在MATH500等任务中验证组件与算法正交，跨域通用且效果稳定。
- 发现无限动作空间下策略多样性是关键瓶颈，而非奖励质量。
LiTS 是一个基于 Python 的模块化框架，用于通过树搜索实现大语言模型的推理。它将树搜索分解为三个可复用组件：策略（Policy）、转移（Transition）和奖励模型（RewardModel），可集成至 MCTS、BFS 等算法中。通过装饰器注册机制，领域专家可扩展新任务，算法研究者可自定义搜索算法。我们在 MATH500（语言推理）、Crosswords（环境规划）和 MapEval（工具使用）上验证了其可组合性，结果显示组件与算法在任务类型内正交：组件可在不同算法间复用，算法也可适配所有组件与领域。此外，我们发现：在无限动作空间下，大模型策略多样性（非奖励质量）成为有效树搜索的主要瓶颈。演示视频见 https://youtu.be/nRGX43YrR3I。代码以 Apache 2.0 许可发布于 https://github.com/xinzhel/lits-llm，含安装说明与可运行示例，支持用户复现演示流程。
原文摘要 · Abstract (English)
LiTS is a modular Python framework for LLM reasoning via tree search. It decomposes tree search into three reusable components (Policy, Transition, and RewardModel) that plug into algorithms like MCTS and BFS. A decorator-based registry enables domain experts to extend to new domains by registering components, and algorithmic researchers to implement custom search algorithms. We demonstrate composability on MATH500 (language reasoning), Crosswords (environment planning), and MapEval (tool use), showing that components and algorithms are orthogonal: components are reusable across algorithms within each task type, and algorithms work across all components and domains. We also report a mode-collapse finding: in infinite action spaces, LLM policy diversity (not reward quality) is the bottleneck for effective tree search. A demonstration video is available at https://youtu.be/nRGX43YrR3I. The package is released under the Apache 2.0 license at https://github.com/xinzhel/lits-llm, including installation instructions and runnable examples that enable users to reproduce the demonstrated workflows.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。