OpenAI preprint, September 2026 · §6. Overlap conventions and dimension of the attractor
§6 (setting): overlap conventions:Injectivity on finite words agrees with the nonempty-word and fixed-length conditions
The overlap conventions are equivalent.
Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its
declaration both in Lean and in words.
The Setting
The overlap conventions concern whether different finite addresses determine the same complete affine map. A weighted affine iterated function system on a finite alphabet consists of real signed slopes , real translations and strictly positive normalized weights , with (1)✦
Let be the set of all finite words over , including the empty word .For the maps in (1), the word affine composition associated with is the complete affine map , composed in address order; its word map records the signed slope and translation: (2)✦For , the length complete-map function assigns to an indexed word the pair in (2): (3)✦
The Claim
Definition 1. The no-exact-overlap property of is injectivity of the word map in (2): (4)✦The same-length no-exact-overlap property of is injectivity of every complete-map function in (3): (5)✦
Theorem 1. Let be a weighted affine iterated function system on a finite alphabet, with the properties and defined in (4) and (5). Then (6)✦ and (7)✦
In the paper, the result reads: an exact overlap is equality of complete affine maps for distinct positive-length words of the same length; one may instead allow distinct nonempty words of arbitrary lengths, and Lemma 6.1 reconciles these conventions.
What Is Checked
The equivalences (6) and (7) are stated over every weighted affine iterated function system on a finite alphabet, with word compositions valued in pairs of real signed slopes and translations.