A benchmark is only as good as what it compares against, and takes its comparison set by hand. This derives one.

python3 -m venv /tmp/vectors
/tmp/vectors/bin/pip install evoc model2vec matplotlib
/tmp/vectors/bin/python check/build-context-token-vectors/scripts/vectors.py --serve

starts a loopback companion, opens it, and stays alive until Ctrl-C. Change any EVoC parameter in the page and press Retune: the companion fits EVoC again while reusing the loaded corpus, embeddings, projection, and nearest peers. changes its default port, . writes a read-only snapshot instead. With neither flag, the tables print to the terminal.

is required and not declared by : imports it at module scope, so fails on a clean install without it. Report that upstream rather than patching around it here.

points at another skills directory, default . names another model2vec static model, default . sets how many neighbours each skill reports.

Every EVoC parameter is both a flag and a live control. Only values you set are passed, so a run that changed nothing says rather than looking tuned. Empty live controls restore the corresponding default.

--base-min-cluster-size

--n-neighbors

--min-samples

--noise-level

--n-epochs

--neighbor-scale

--min-similarity-threshold

The useful act is comparing two settings, never trusting one:

EVoC defaults                            8 clusters, 49 noise
--base-min-cluster-size 3 --n-neighbors 10   9 clusters, 53 noise
--noise-level 0.2                        7 clusters, 39 noise

The page prints the settings behind its current result, so a screenshot still says what produced it.

The dashboard is not a report. Switch layer to move between the resolutions EVoC found, coarsest first. Rail organizes known skills as , , , , then the and aids; this is display metadata and never changes the vectors. Filter to Local, All, or Noise. Search by skill name, by a peer's name, or by a cluster id such as . Click any point or row to inspect one skill: its membership strength, its full neighbour list, and every other member of its cluster at the current layer.

noise

noise

is not a verdict. It says the corpus holds no peer. Whether that is novelty or a diluted is a judgement this cannot make, for the same reason it cannot rank: clustering observes position, never quality.

EVoC is stochastic. Without two runs over identical input return different labels, and a comparison set that changes per run is not one. is declared in the script and is part of any result worth quoting.

The live companion writes nothing. A deliberate writes only its page. Neither mode authors a , edits , ranks a skill, or records a judgement. Reading it may change which comparison you run next; nothing here runs one.

Its three dependencies live in a virtual environment you create. Nothing under or imports them, and this skill ships none of them.