I benchmarked 5 embedding models across 4 datasets
I benchmarked five embedding models across four NanoBEIR datasets and found that bigger embeddings did not always produce better retrieval.
Posts
Everything in one stream so writing stays simple and discoverable. Use tags or search when the archive gets bigger.
Need keyword search? Go to Search.
4 entries shown
Why reranking matters with cross-encoders
Bi-encoders make retrieval fast, but cross-encoders expose why reranking matters when meaning depends on the query.
A beginner-friendly guide to the GGUF model format
GGUF made local LLM inference feel practical by packaging model weights, vocabulary, hyperparameters, and architecture metadata into one runnable format.
Semantic Caching in Production
Repeated user intents can quietly inflate LLM cost and latency. Semantic caching helps, but production use comes with trade-offs.