MentalThink：在思维SVG世界中塑造思想

摘要

我们提出了MentalThink，这是一种视觉-符号推理范式，为多模态大语言模型（MLLMs）配备了可执行的“心理”可视化机制。MentalThink的核心是一个“思考即SVG”流水线：模型学习生成、渲染和解释可缩放矢量图形（SVG）代码，将其作为多轮推理的中间视觉表示。通过创建结构化的矢量草图，模型能够外化空间假设、通过确定性渲染对其进行检查，并在受限的几何空间内进行推理，从而有效模拟人类的心理意象过程。我们通过一个两阶段训练框架来实例化这一范式：结合用于SVG语法对齐的监督微调（SFT）与多轮强化学习（RL），以鼓励对中间视觉假设进行迭代检查、修正和优化。大量评估表明，MentalThink在空间理解和推理基准上取得了卓越性能（例如，在VSIBench上达到55.1%，在MindCube上达到76.0%），这证明了可执行矢量图形为动态视角转换、视觉反思和组合场景构建提供了一个可验证的视觉工作空间。

English

We introduce MentalThink, a visual-symbolic reasoning paradigm that equips Multimodal LLMs (MLLMs) with an executable mechanism for "mental" visualization. The core of MentalThink is a think-with-SVG pipeline, where the model learns to generate, render, and interpret scalable vector graphics (SVG) code as an intermediate visual representation for multi-turn reasoning. By creating structured vector sketches, the model can externalize spatial hypotheses, inspect them through deterministic rendering, and reason within a constrained geometric space, effectively mimicking the human process of mental imagery. We instantiate this paradigm through a two-stage training framework, combining Supervised Fine-Tuning (SFT) for SVG syntactic alignment with multi-turn Reinforcement Learning (RL) to encourage iterative inspection, revision, and refinement of intermediate visual hypotheses. Extensive evaluations demonstrate that MentalThink achieves superior performance on spatial understanding and reasoning benchmarks (e.g., 55.1% on VSIBench, 76.0% on MindCube), showing that executable vector graphics provide a verifiable visual workspace for dynamic perspective taking, visual reflection, and compositional scene construction.