Apple 提出 probe guidance：用扩散模型内部状态引导 flow matching
How to Guide Your Language Flow
Apple 发的新方法 probe guidance，不用多跑一次 forward pass 就能给 flow matching 模型加引导，搞生成模型的可以看看。
Apple 介绍了一种名为 probe guidance 的新方法，用于引导 flow matching 模型。该方法利用已有扩散模型的冻结内部状态构造引导信号，原理与 autoguidance 相似，但省去了推理时额外的 forward pass。在连续扩散语言模型上，probe guidance 在无条件生成任务中刷新 state-of-the-art 成绩。
How to Guide Your Language Flow
We introduce a new method to guide flow matching models. Our approach, which we call probe guidance, uses the frozen internal states of an existing diffusion model to construct a guidance signal. This works using a similar principle as autoguidance, but eliminates the need for an additional forward pass at inference time and provides a reliable path to ensure that the weak and strong model share similar dynamics. We apply and benchmark this method on continuous diffusion language models, where probe guidance sets a new state-of-the-art performance on unconditional generation. When applied to a…
