{
  "schemaVersion": 1,
  "topic": "Embedding Similarity Playground",
  "checked_at": "2026-09-07",
  "claims": [
    {
      "id": "E1",
      "title": "Cosine is normalized dot product",
      "claim": "The scikit-learn cosine similarity definition divides dot product by the two L2 norms; it equals the linear kernel on L2-normalized data.",
      "url": "https://scikit-learn.org/stable/modules/generated/sklearn.metrics.pairwise.cosine_similarity.html",
      "checked_at": "2026-09-07",
      "type": "primary-documentation",
      "classification": "mathematical-definition",
      "boundary": "This demo uses mathematical undefined for zero vectors rather than the library\u2019s zero-fill convention."
    },
    {
      "id": "E2",
      "title": "Unit normalization removes magnitude",
      "claim": "L2 normalization rescales each nonzero vector to unit norm.",
      "url": "https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.Normalizer.html",
      "checked_at": "2026-09-07",
      "type": "primary-documentation",
      "classification": "mathematical-definition",
      "boundary": "Magnitude may contain useful information in a real model; normalization is not universally appropriate."
    }
  ],
  "assumptions": [
    "All examples are geometric teaching coordinates in [-10,10], with no trained encoder or corpus.",
    "Normalization applies to query and all candidates.",
    "Ties use absolute tolerance 1e-10; undefined results are excluded.",
    "Raw zero-vector dot products and Euclidean distances remain defined."
  ],
  "evidenceClasses": {
    "measuredFacts": "No hardware, model-quality, or corpus measurements are claimed.",
    "derivedResults": "Deterministic arithmetic or editorial rule evaluation; see formulas in README.md.",
    "editorialAssumptions": "Dated synthetic examples, stated scope, and storage policies.",
    "primaryFacts": "Each sourced claim carries its own classification and boundary."
  }
}