为何纠结于连续潜变量？通过渲染压缩实现可解释的离散潜变量推理
Why Struggle with Continuous Latents? Interpretable Discrete Latent Reasoning via Rendered Compression
AI总结：
提出离散潜变量推理（DLR）方法，将连续潜状态转化为离散令牌，通过渲染压缩实现高效、可解释的潜空间推理，在五个基准上优于先前方法并达到20倍压缩。
摘要：
大型语言模型通过显式思维链和强化学习实现了高推理性能，但需要长输出序列和扩展推理时间。潜推理通过将计算转移到潜空间来降低这一成本；然而，连续潜方法难以训练，存在不稳定且不可解释的推理轨迹。我们认为这些问题源于连续空间推理与离散符号监督之间的错位，因为连续状态缺乏用于逐步对齐的显式锚点。为解决这一问题，我们提出了\textbf{离散潜推理（DLR）}，这是第一种将连续潜状态转换为显式离散令牌的方法。受基于渲染的压缩启发，我们将文本思维链渲染为图像，提取视觉特征，并通过基于聚类的微调构建离散潜词汇表。扩展词汇表和输出头使得标准自回归建模能够同时处理自然语言和潜令牌，支持预训练对齐、SFT和RL。在五个推理基准和两个模型系列（Qwen3-VL和LLaMA-3）上的实验证实，\textbf{DLR}优于先前的潜推理基线，压缩比高达\textbf{20倍}。此外，学习到的潜轨迹保留了可解释的语义结构。总体而言，离散潜令牌为高效潜推理提供了可控且可解释的基础。
英文摘要：
Large language models achieve high reasoning performance via explicit chain-of-thought and reinforcement learning, but require long output sequences and extended inference time. Latent reasoning reduces this cost by shifting computation into a latent space; however, continuous latent methods are hard to train, suffering from unstable and uninterpretable reasoning trajectories. We argue these issues stem from a misalignment between continuous-space reasoning and discrete symbolic supervision, as continuous states lack explicit anchors for step-by-step alignment. To resolve this, we propose \textbf{Discrete Latent Reasoning~(DLR)}, the first method that converts continuous latent states into explicit discrete tokens. Inspired by render-based compression, we render textual chains of thought into images, extract visual features, and construct a discrete latent vocabulary via clustering-based fine-tuning. Expanding the vocabulary and output head enables standard autoregressive modeling over both natural language and latent tokens, supporting pretraining alignment, SFT, and RL. Experiments on five reasoning benchmarks and two model series~(Qwen3-VL and LLaMA-3) confirm that \textbf{DLR} outperforms prior latent reasoning baselines with up to \textbf{20$\times$ compression}. Furthermore, the learned latent trajectories retain an interpretable semantic structure. Overall, discrete latent tokens provide a controllable and interpretable basis for efficient latent reasoning.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
