用机制可解释性技术探测并控制 LLM 中的隐喻表征
The Geometry of Metaphor: Probing and Steering Metaphor Representations in Large Language Models
有人用机制可解释性把模型里的“隐喻开关”找到了，能让输出变比喻或变直白，研究文本生成风格的可以看看方法细节。
一篇 arXiv 论文用机制可解释性技术在多种不同架构和规模的 LLM 中定位隐喻的因果子空间。研究发现可以按这些子空间对分类任务的输出进行转向，覆盖不同词性。在生成任务上也能控制输出的隐喻程度，说明 LLM 内部存在可操纵的隐喻表征方向。
The Geometry of Metaphor: Probing and Steering Metaphor Representations in Large Language Models
Metaphor is a crucial aspect of human cognition. As such, we would like to understand how it is represented in large language models (LLMs) and how it might be controlled. We use mechanistic interpretability techniques to identify causal subspaces in a range of LLMs covering different architectures and sizes. We show that we are able to steer output on classification tasks across different parts of speech. We are also able to steer output on generation tasks, indicating that these techniques can be used to control the metaphoricity of LLM output.
