Sometimes a page notes, quietly: Numbered 2.3.15 in this text. This is a bibliographic fact, not an error. The commentarial traditions do not divide and number the sūtras identically: one commentator reads a cluster of sūtras as a single continuous argument where another divides it in two; one treats a short aphorism as complete where another reads it as the close of the previous point. Over a long text, small differences accumulate into offset numbering.
The Atlas gives every sūtra one shared reference number so that Browse, the sūtra pages, and Compare can speak about the same place in the text. Wherever an edition’s own printed number differs, the page says so, in exactly those quiet notes. A handful of correspondences between the texts are genuinely uncertain, and the Atlas says that plainly too, rather than forcing a tidy match. And where one commentator simply has no commentary at a shared sūtra, that absence is a fact about the sources, not missing data.
The full story of how the editions are compared lives in How Vedanta Atlas Works.