← Explore all corpora

knowledge corpus

C2 Atlas

Pinned English Wikipedia prose that expands every C1 topic into a training-scale text corpus.

v0.1.0Locked releaseProof included
Size6,675 chunks
Tokens11.9 million
Domains11
Splits5,317 / 717 / 641

Why teams choose it

Atlas deepens the full C1 knowledge map with coherent article text while keeping each entity, split, source revision, and rights boundary visible.

Use it to

  • Continue training on coherent prose
  • Build source-aware retrieval
  • Compare compact facts with long-form context

Know before you use it

  • Covers the same 999 topic families as C1
  • Wikipedia coverage and writing biases remain
  • It is not a general web-scale corpus
Lineage

C1 → C2

Clear rights notes

Atlas uses pinned English Wikipedia article revisions with attribution and licence evidence kept in the release.

Put better data behind your next model.

Get access→