不要分类，要幻觉：LLM 标签生成新思路

AI 摘要

Simon Willison 在其博客中介绍了一种利用 LLM 进行内容标签化的新方法。该方法不直接让模型从现有标签列表中选择，而是先让模型自由生成可能的标签，再通过向量嵌入与现有标签库匹配，找到最接近的具体标签。Doug Turnbull 提出了这一解决方案，并建议在提示中提供标签形状示例以提升效果。

正文节选

14th August 2026 - Link Blog Don't classify. Hallucinate! I still have quite a bit of older content on my blog that I never got round to tagging. My blog has 1,856 tags - likely too many to feed to an LLM in one go and say "which of these tags match the following content". Doug Turnbull has a neat solution. Tell the model to output tags without any details of the existing vocabulary, then use vector embeddings against the existing corpus to find the concrete tags that are closest to the ones the

发布时间：2026-08-15 05:54

抓取时间：2026-08-15 06:01

来源机构：Simon Willison