term frequency
technique judged 85% likely a named technique or architecture
See how it has spread across media tiers
Mentions over time
Table view
| Week | Articles | Sources |
|---|---|---|
| 2024-W26 | 1 | 1 |
| 2026-W07 | 1 | 1 |
| 2026-W09 | 1 | 1 |
| 2026-W11 | 2 | 1 |
| 2026-W24 | 1 | 1 |
In their words
The scoring formula has two key mechanisms: - Term frequency saturation: Additional occurrences of a term contribute less and less to the score.
Hybrid search combines vector search with keyword search (typically BM25, a classical term-frequency algorithm).
Where Keyword Search Fails Section titled “Where Keyword Search Fails” Keyword search relies on term frequency and inverse document frequency (TF-IDF) or BM25 scoring.
BM25 is a term-frequency algorithm that scores documents by the presence and frequency of exact query terms, adjusted for document length.
Under the hood, a semantic search API typically: - Converts text into vector embeddings using transformer-based ML models - Stores embeddings in a vector index for fast similarity search - Processes queries by converting them into the same vector space - Ranks results using a combination of vector similarity and neural re-ranking This is fundamentally different from traditional keyword search, whi
There are three main retrieval strategies: | Strategy | How It Works | Strengths | Weaknesses | |---|---|---|---| | Dense (vector) | Embeds query and chunks; finds nearest neighbors by cosine similarity | Captures semantic meaning; handles paraphrases | Misses exact keyword matches | | Sparse (BM25) | Keyword-based search using term frequency | Exact match on names, IDs, technical terms | No seman