What I Learned Switching My Blog's Embedding Model for Recommendations, With an LLM as a Research Partner

Intro This site’s related-article recommendations run on a tool I built myself: , a CLI wrapper around , which enriches Hugo front matter with various metadata. itself only computes related articles via TF-IDF keyword similarity, but prelims-cli adds embedding-based recommendation on top. It’s built to run on GitHub Actions on every new post: no GPU, limited memory. I’d been using for Japanese and for English, the lightest models that still gave decent quality. Then , a new small, lightweight embedding model, came out, and I tried it to see whether I could unify the Japanese and English models under one that also covers other languages. The related articles on this site now run on Bekko a25m. This post covers how I got there, and in particular the trouble I ran into running the experiments with Claude. How to use it for embedding-based recommendation with prelims-cli Since prelims-cli v0.0.11, you can use Bekko embedding by setting the multilingual option. Write a file like the one…