Show HN: An interactive visualization of a semantic cache

semanticcachehit.com

This uses the new text-embedding-3-large model from OpenAI and UMAP for the clustering. All the prompts in the visual were generated using GPT-4.

Open articleView on HN

Hey. Its a good practical application, especially to reduce cost. But at about 0.6 similarity, I get some cache hit. Maybe with more examples and for a high use app, the cache hit would increase based on a higher similarity scores.

Still early days on this I guess, but any observations that can help improve the hit rates?

Reply on news.ycombinator.com