OctoInformBeta

OpenAI preprint, September 2026 · Section 9. Exceptional terms and the power-saving induction

Section 9 (setting): exceptional products and child scales: Ceiling lengths accompany the progression transfer factor

Exceptional coordinates in the induction on scale.

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 induction on length assigns an exceptional mass to selected prime coordinates and a child scale to a progression. The mass at a prime comes from the terminal part of a mixture of permutation tuple configurations. Write [n]={1,…,n}[n]=\{1,\ldots,n\}; the natural range {0,…,n−1}\{0,\ldots,n-1\} indexes this interval by the translation x↦x−1x\mapsto x-1.

The dimension parameter dd is a nonnegative integer. The fixed real parameters β\beta, γ\gamma, and τ(d)\tau(d) are given in (1). β=1100000,γ=164,τ(d)=βγ1000(d+1).\beta=\frac{1}{100000},\qquad \gamma=\frac{1}{64},\qquad \tau(d)=\frac{\beta\gamma}{1000(d+1)}.(1)

The Exceptional Tuple Mass

For a block set KK and a natural height kk, a permutation word is a map v:K→𝔖kv:K\to\mathfrak S_k, where 𝔖k\mathfrak S_k is the permutation group of {0,…,k−1}\{0,\ldots,k-1\}.✦ For a natural height kk, an ordered position pair (a,b)(a,b) is a pair of positions with 0≤a<b<k0\leq a<b<k.✦ For a block c∈Kc\in K and an ordered position pair (a,b)(a,b), the position transposition vc;a,bv^{c;a,b} is the word obtained by replacing v(c)v(c) by v(c)(ab)v(c)(a\ b) and keeping every other block unchanged.✦ The transposition interaction HS,t(v,w)H_{S,t}(v,w), for a finite block set KK, a subset S⊆KS\subseteq K, a natural height kk, real tt, and permutation words v,w:K→𝔖kv,w:K\to\mathfrak S_k, is the counting sum in (2). HS,t(v,w)=∑c∈S∑0≤a<b<ktb−a𝟏{w=vc;a,b}.H_{S,t}(v,w)= \sum_{c\in S}\ \sum_{0\leq a<b<k} t^{b-a}\mathbf{1}_{\{w=v^{c;a,b}\}}.(2)✦

For a finite index set II, a label set Λ\Lambda, a real matrix AA on II, and labelings j,k:I→Λj,k:I\to\Lambda, the label coincidence pairing CA(j,k)C_A(j,k) is the counting sum in (3). The index energy EA(j)E_A(j) and index Boltzmann weight wA(j)w_A(j) associated with a finite index set II, a label set Λ\Lambda, a real matrix AA on II, and a labeling j:I→Λj:I\to\Lambda are the real numbers given in (3). CA(j,k)=∑x∈I∑y∈IAx,y𝟏{j(x)=k(y)},EA(j)=12CA(j,j),wA(j)=exp⁡(EA(j)).\begin{aligned} C_A(j,k)&=\sum_{x\in I}\sum_{y\in I} A_{x,y}\mathbf{1}_{\{j(x)=k(y)\}},\\ E_A(j)&=\tfrac12 C_A(j,j),\qquad w_A(j)=\exp(E_A(j)). \end{aligned}(3)

The tuple height hh, block cutoff TT, block count mm, list base rr, and interaction parameter δ\delta have the values in (4). h=24,T=80,m=2T+1=161,r=h!2,δ=110h2.h=24,\qquad T=80,\qquad m=2T+1=161,\qquad r=\frac{h!}{2},\qquad \delta=\frac{1}{10h^2}.(4) The block set is 𝒦={0,…,m−1}\mathcal K=\{0,\ldots,m-1\}, and the permutation tuple vertices form 𝒱=𝔖h𝒦\mathcal V=\mathfrak S_h^{\mathcal K}.✦ The cyclic shift ρ∈𝔖h\rho\in\mathfrak S_h sends each position to the next and the last to zero.✦ For a finite block set KK, a group GG, an element c∈Gc\in G, and a natural threshold tt, functions u,v:K→Gu,v:K\to G satisfy threshold shift adjacency when |{b∈K:v(b)=u(b)c}|≥t|\{b\in K:v(b)=u(b)c\}|\geq t.✦

Fix an active block subset S⊆𝒦S\subseteq\mathcal K. An inactive-block address is an assignment of a permutation in 𝔖h\mathfrak S_h to each block outside SS; the address set is 𝒜S=𝔖h𝒦\S\mathcal A_S=\mathfrak S_h^{\mathcal K\setminus S}.✦ The addressed list indices form IS={0,…,r|S|−1}I_S=\{0,\ldots,r^{|S|}-1\}✦, and an addressed list entry is a pair in 𝒜S×IS\mathcal A_S\times I_S.✦ The inactive-block projection πS(v)\pi_S(v) of a vertex vv is its restriction to 𝒦\S\mathcal K\setminus S.✦

For a natural modulus pp, list values Y:𝒜S×IS→ℤ/pℤY:\mathcal A_S\times I_S\to\mathbb Z/p\mathbb Z, and a labeling j:𝒱→ISj:\mathcal V\to I_S, the sampled tuple configuration XS,Y,j∈(ℤ/pℤ)𝒱X_{S,Y,j}\in(\mathbb Z/p\mathbb Z)^{\mathcal V} is defined by (5). For prime pp, the residue ring is the field 𝔽p\mathbb F_p. XS,Y,j(v)=Y(πS(v),j(v)).X_{S,Y,j}(v)=Y(\pi_S(v),j(v)).(5)✦

Let ℰS\mathcal E_S be the finite set of list constraint edges, whose ordered pairs of addressed list entries prescribe the square-difference constraints between list values. For a field FF, the nonzero square indicator sF:F→ℝs_F:F\to\mathbb R is one on the nonzero squares and zero elsewhere.✦ In an odd prime field, the strict squares are the nonzero squares. For a prime pp and an active block subset S⊆𝒦S\subseteq\mathcal K, the square-constraint indicator Vp,SV_{p,S} is the real-valued function of list values Y:𝒜S×IS→𝔽pY:\mathcal A_S\times I_S\to\mathbb F_p given by the product in (6). Vp,S(Y)=∏(a,b)∈ℰSs𝔽p(Y(b)−Y(a)).V_{p,S}(Y)=\prod_{(a,b)\in\mathcal E_S} s_{\mathbb F_p}\bigl(Y(b)-Y(a)\bigr).(6)✦

The tuple Boltzmann weight WS(j)W_S(j) of a labeling j:𝒱→ISj:\mathcal V\to I_S is the index Boltzmann weight wHS,δ(j)w_{H_{S,\delta}}(j).✦ For a real observable G:𝔽p𝒱→ℝG:\mathbb F_p^{\mathcal V}\to\mathbb R, the tuple component average Lp,S(G)L_{p,S}(G) is given by (7), where each expectation is a uniform probability average over the displayed finite function space. Lp,S(G)=𝔼Y∈𝔽p𝒜S×IS𝔼j∈IS𝒱Vp,S(Y)WS(j)G(XS,Y,j).L_{p,S}(G)= \mathbb E_{Y\in\mathbb F_p^{\mathcal A_S\times I_S}} \mathbb E_{j\in I_S^{\mathcal V}} V_{p,S}(Y)W_S(j)G(X_{S,Y,j}).(7)✦

For a natural modulus qq and S⊆𝒦S\subseteq\mathcal K, the block subset mixture weight is aq(S)=q−|S|/16a_q(S)=q^{-|S|/16}.✦ The unnormalized tuple mixture ℳp\mathcal M_p and the restricted tuple mixture ℳp,𝒟\mathcal M_{p,\mathscr D}, for a family 𝒟⊆2𝒦\mathscr D\subseteq 2^{\mathcal K}, are the counting sums in (8). ℳp(G)=∑S⊆𝒦|S|≤T+1ap(S)Lp,S(G),ℳp,𝒟(G)=∑S∈𝒟ap(S)Lp,S(G).\mathcal M_p(G)= \sum_{\substack{S\subseteq\mathcal K\\|S|\leq T+1}}a_p(S)L_{p,S}(G), \qquad \mathcal M_{p,\mathscr D}(G)= \sum_{S\in\mathscr D}a_p(S)L_{p,S}(G).(8) The terminal block family is 𝒯={S⊆𝒦:|S|=T+1}\mathscr T=\{S\subseteq\mathcal K:|S|=T+1\}.✦ The exceptional tuple sublaw εp\varepsilon_p is the restricted mixture on 𝒯\mathscr T normalized by the total mass of ℳp\mathcal M_p✦; its exceptional mass ηp\eta_p is its value on the constant observable 11, as in (9). εp(G)=ℳp,𝒯(G)ℳp(1),ηp=εp(1).\varepsilon_p(G)=\frac{\mathcal M_{p,\mathscr T}(G)}{\mathcal M_p(1)}, \qquad \eta_p=\varepsilon_p(1).(9)

The Definition

Definition 1. Let (pj)j∈J(p_j)_{j\in J} be a fixed indexed family of primes and let B⊆JB\subseteq J be finite. The exceptional mass product η(B)∈ℝ\eta(B)\in\mathbb R is defined by (10). η(B)=∏j∈Bηpj.\eta(B)=\prod_{j\in B}\eta_{p_j}.(10)✦ For natural numbers N,DN,D, the child length ℓ(N,D)∈ℕ\ell(N,D)\in\mathbb N is the natural ceiling in (11); here ⌈x⌉+\lceil x\rceil_+ is the least nonnegative integer at least xx, and the quotient is taken to be zero when D=0D=0. ℓ(N,D)=⌈ND2⌉+.\ell(N,D)=\left\lceil\frac{N}{D^2}\right\rceil_+.(11)✦ For natural d,M,Nd,M,N, the progression transfer factor Fd(M,N)∈ℝF_d(M,N)\in\mathbb R is defined by (12). Fd(M,N)=(1+M2N2τ(d)−1)d.F_d(M,N)=\left(1+M^2N^{2\tau(d)-1}\right)^d.(12)✦

In the paper, the result reads: for the fixed parameters, including the modulus M0M_0 and τ=τ(d)\tau=\tau(d), and the paper’s functional Y(n,A)Y(n,A), define 𝒴(n)=max⁡A⊆[n] square-difference-freeY(n,A)\mathcal Y(n)=\max_{A\subseteq[n]\text{ square-difference-free}}Y(n,A), and, for B⊆𝒫NB\subseteq\mathcal P_N with 𝒫N\mathcal P_N the prime family, set qB=∏p∈Bpq_B=\prod_{p\in B}p, ηB=∏p∈Bηp\eta_B=\prod_{p\in B}\eta_p, nB=⌈N/(M0qB)2⌉n_B=\lceil N/(M_0q_B)^2\rceil, and uN=(1+M02N2τ−1)d−1u_N=(1+M_0^2N^{2\tau-1})^d-1, with empty products equal to one and n⌀=⌈N/M02⌉n_{\varnothing}=\lceil N/M_0^2\rceil.

What Is Checked

The definitions are stated over a fixed indexed family of prime moduli and a finite subset of its coordinates, with d,M,N,Dd,M,N,D natural numbers. For a prime pp and S⊆𝒦S\subseteq\mathcal K, the tuple component functional G↦Lp,S(G)G\mapsto L_{p,S}(G) is real linear on real observables of 𝔽p𝒱\mathbb F_p^{\mathcal V}✦; the restricted tuple mixture and exceptional tuple sublaw are real linear functionals on the same space.