Your files stay on your device

# Vector similarity explorer

Rank embeddings against a query by cosine, dot product, or distance, with dimension and normalization checks.

Paste a query vector and some candidates, then press Rank candidates.

All four metrics are computed for every candidate no matter which one ranks, so you can see where the orders disagree. Cosine and dot product rank higher-is-closer; the two distances rank lower-is-closer. Everything runs in this tab and nothing is uploaded.
OptionsRank candidates

## More AI tools

- [Tokenizer playground](/ai/tokenizer)
- [Context fit](/ai/context-fit)
- [Fine-tune JSONL formatter](/ai/jsonl-formatter)
- [Prompt diff](/ai/prompt-diff)
[All AI tools](/ai)

Was this tool useful?
YesNo

---

Canonical HTML: https://nutter.tools/ai/embedding-similarity-explorer
Markdown version: https://nutter.tools/ai/embedding-similarity-explorer.md
Plain-text version: https://nutter.tools/ai/embedding-similarity-explorer.txt
Agent index: https://nutter.tools/llms.txt