Visualizers

Learn by moving things

Interactive explainers for the ideas behind retrieval, embeddings, and agents. A formula tells you what is true; dragging a vector shows you why. Each one also appears inside the article it belongs to, and runs entirely in your browser, so drag, poke, and break it.

Portrait of Sachin Gupta rendered in binary

GraphRAG global search

From: Beyond the Vector →

Global search clusters the graph into communities, summarizes each, maps the question over every community summary to get partial answers, then reduces them into one. Ideal for broad, whole-corpus questions flat RAG cannot touch.

Global search: summarize every community, then combineinteractive
Finance clustersummarypartial answerProduct clustersummarypartial answerPeople clustersummarypartial answerreduceone answer

The graph is clustered into communities, each with its own summary.

Multimodal: shared embedding space

From: Beyond the Vector →

Text and images embed into one shared space (CLIP-style), so a text query lands near the matching image and retrieves it directly, with no words on the image required.

Shared space: a text query points straight at an imageinteractive
one shared embedding spacetext query"revenue chart"imgchart.pngimgcat.pngimgmap.png

Everything, words and pictures, is embedded into the same space by a CLIP-style model.

Multimodal: translate to text first

From: Beyond the Vector →

Turn each non-text item into text first, captioning images, parsing tables, transcribing audio, then retrieve over the descriptions with ordinary text RAG.

Translate first: describe the picture, then search the wordsinteractive
imagecaption...text indexordinary text RAG

Turn each non-text item into text first: caption images, parse tables, transcribe audio. Then retrieve over the descriptions with ordinary text RAG.

Visualizers — Sachin Gupta