How I Reproduced BM25, Dense Retrieval, and SPLADE on a 16GB MacBook
Large Language Models How I Reproduced BM25, Dense Retrieval, and SPLADE on a 16GB MacBook A practical reproduction of three retrieval baselines, including the crashes, fixes, and score checks that matter for RAG systems. BM25 is a keyword scoring method that has been used for decades and still shows up as the baseline in nearly every retrieval paper published today. Dense retrieval is now standard in production search.
- ▪Large Language Models How I Reproduced BM25, Dense Retrieval, and SPLADE on a 16GB MacBook A practical reproduction of three retrieval baselines, including the crashes, fixes, and score checks that matter for RAG systems.
- ▪BM25 is a keyword scoring method that has been used for decades and still shows up as the baseline in nearly every retrieval paper published today.
- ▪Dense retrieval is now standard in production search.
Towards Data Science files mainly under ai. We currently carry 82 of its stories.
Story provenance
Source · retrieval · rights · ranking — open for full record
inspect →
Attribution is not the same as permission. This drawer separates discovery metadata, excerpts, WeSearch-generated summaries, reuse status, and whether the publisher receives the visit. Nothing here claims a legal grant the publisher has not made.
Record
| Original publisher | Towards Data Science |
| Canonical URL | https://towardsdatascience.com/how-i-reproduced-bm25-dense-retrieval-and-splade-on-a-16gb-macbook/ |
| Publication time | Mon, 27 Jul 2026 12:00:00 +0000 |
| Retrieval time | 2026-07-27T12:12:32.939Z |
| Last seen | 2026-07-27T12:12:32.939Z |
| Headline source | Publisher (no WeSearch rewrite) |
| Excerpt source | publisher body |
| Excerpt method | First ~120 words (~800 chars) of extracted publisher body, fair-use limited. |
| Summary | WeSearch · cerebras-chat (WeSearch summarizer) |
| Summary source text | contentText |
| Citation coverage | Summary is a WeSearch-generated derivative; primary citation is the original publisher URL. |
| Cluster | ItRr_l7dby81 · 1 stories |
| Cluster logic | Grouped by semantic title/content similarity across sources within a rolling window. Same-publisher template collisions are excluded from coverage comparison. |
| Ranking reason | Story pages are not engagement-ranked. Hub feeds use recency, with optional source-diversified chronological ordering (cap consecutive stories per source). No personalized ranking. |
| Publisher visit | Yes — open original |
| Substitutes article? | No — link-out required for full text |
Rights status (four layers)
WeSearch handling by dimension
| Indexing | May the item be indexed (stored, ranked, made findable)? | Allowed |
| Snippet | May a short excerpt of the publisher's text be shown? | Allowed |
| AI summary | May WeSearch generate its own short summary of the article? | Limited |
| Retrieval / RAG | May the content be exposed for third-party retrieval-augmented generation? | Not asserted |
| Model training | May the content be used to train AI models? | Not asserted |
| Commercial reuse | May the content be reused commercially? | Not permitted |
Basis: Derived from the published RSS/Atom feed. Contact: [email protected]. Reviewed: 2026-07-24.
Opening excerpt (first ~120 words) tap to expand
Large Language Models How I Reproduced BM25, Dense Retrieval, and SPLADE on a 16GB MacBook A practical reproduction of three retrieval baselines, including the crashes, fixes, and score checks that matter for RAG systems. Abdullahi Dattijo Jul 27, 2026 16 min read Share Three distinct retrieval pipelines—keyword, dense embedding, and learned sparse search—converge into ranked documents Teams building retrieval-augmented generation (RAG) systems constantly say things like “dense retrieval beats BM25” or “hybrid search works better,” and most of the time, nobody involved has actually reproduced the baseline being compared against. BM25 is a keyword scoring method that has been used for decades and still shows up as the baseline in nearly every retrieval paper published today.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Towards Data Science.
Discussion
0 commentsMore from Towards Data Science
