---
title: "Comment by Ahmet Özel on \"Exploring GenAI Retrieval with a Resume: Keyword, Semantic, and Hybrid Search\" | Hashnode"
description: "The 0.7 semantic / 0.3 keyword blend is a useful experiment, but the scales determine what those weights actually mean. Your keyword score counts matching query"
url: "https://hashnode.com/posts/exploring-genai-retrieval-with-a-resume-keyword-semantic-and-hybrid-search/6ac2492734776c6712e86346/comment/6ac2918225455dedc00c7ccf"
---

[Discussion](https://hashnode.com/posts/exploring-genai-retrieval-with-a-resume-keyword-semantic-and-hybrid-search/6ac2492734776c6712e86346)

Comment [Exploring GenAI Retrieval with a Resume: Keyword, Semantic, and Hybrid Search](https://hashnode.com/posts/exploring-genai-retrieval-with-a-resume-keyword-semantic-and-hybrid-search/6ac2492734776c6712e86346)

[Ahmet Özel](https://hashnode.com/@ahmetozel)

[Oct 4](https://hashnode.com/posts/exploring-genai-retrieval-with-a-resume-keyword-semantic-and-hybrid-search/6ac2492734776c6712e86346/comment/6ac2918225455dedc00c7ccf)

The 0.7 semantic / 0.3 keyword blend is a useful experiment, but the scales determine what those weights actually mean. Your keyword score counts matching query tokens, while cosine similarity lives on a different scale; a longer query can therefore increase the keyword contribution without any change in relevance.

I would print both raw components beside the combined score and try the same question with a few unnecessary words added. Normalizing the keyword component or comparing rank fusion would help separate a deliberate weighting choice from an accidental consequence of query length.
