WorksMathematics
No infinite cluster at criticality for long-range percolation with fast-decaying interactions
Abstract
We consider independent long-range bond percolation on in which and are joined with probability . Let and let be invariant under coordinate permutations and sign changes, positive on nearest-neighbour pairs, and satisfy . We prove that there is no infinite cluster at the critical parameter. Together with Berger's theorem for heavy tails, this gives continuity of the phase transition for for every except , in every dimension , and for every exponentially decaying kernel satisfying these hypotheses. Previously the fast-decaying case was known for nearest-neighbour and uniform finite-range kernels, in for finite-range and exponential kernels, and in high dimension for sufficiently spread-out kernels. The proof adapts a recent renormalization argument for nearest-neighbour percolation on that is driven by a gluing inequality conjectured by Kozma and Nitzan. Long edges enter in two places: they can jump over the surfaces that a nearest-neighbour path must cross, and they let a cluster outside a box enter it from far away. We control the first with thick outer faces and the tail condition, and the second by measuring entrances through the total edge weight between a cluster and a box rather than through the number of boundary vertices.
Introduction
Long-range percolation and the main result
Let and let be a kernel with and . In long-range percolation at parameter , each unordered pair of distinct vertices of is open independently with probability
and closed otherwise; pairs with are never open. Write for the law, for the probability that the origin lies in an infinite open cluster, and . The function is nondecreasing, and the basic question is whether it is continuous at , that is, whether .
For nearest-neighbour percolation, was proved in by Harris and Kesten [11, 14] and in high dimensions by the lace expansion [10, 7]. Kozma and Nitzan [15] reduced the remaining dimensions to a correlation inequality for independent percolation on finite graphs, which has since been proved in several forms [16, 17, 4, 8]. We use it in the joint form of Theorem 2.1, which holds for arbitrary edge probabilities; all graphs to which we apply it are finite.
For long-range models the answer depends on the tail of . Writing
we use the following hypotheses.
(Sym) is invariant under coordinate permutations and sign changes of coordinates.
(Pos) , where is the first unit vector.
(Tail) .
Theorem 1.1. Let and let satisfy (Sym), (Pos) and (Tail). Then and .
Combined with Berger’s theorem for heavy-tailed kernels [3 Theorem 2.6], which we recall in § 8, this settles the power-law case away from the borderline exponent.
Corollary 1.2. Let , , , and let satisfy (Sym) and for all , with . Then .
Corollary 1.3. Let and let satisfy (Sym), (Pos) and for some . Then . The same holds for every finite-range satisfying (Sym) and (Pos).
Corollary 1.2 covers both regimes of around the exceptional value . For it is Berger’s theorem; for it follows from Theorem 1.1, since then . The value is excluded: there is bounded away from zero, so edges of length comparable to leave the cube with probability of order one at every scale. This is the mechanism behind the discontinuous transition of the one-dimensional model with [1, 5], and Theorem 1.1 says nothing about it.
Earlier results
Berger [3, Theorem 2.6] proved that whenever for some , in every ; Hutchcroft [13] made this quantitative. In the opposite regime the kernel decays fast enough that the model should behave like nearest-neighbour percolation, and Hutchcroft writes that “it is widely believed that the phase transition should be continuous for all and ” [13 p. 2]. The known cases of this regime are the following.
In , Duminil-Copin, Sidoravicius and Tassion [6] proved continuity for finite-range kernels invariant under the symmetries of , and Angel, de la Riva, Panagiotis and Shi [2] extended this to kernels with exponential decay; they state the case of power laws in as a hope [2 §1.4]. Both arguments are planar in spirit: they glue clusters that cross as subsets of .
In high dimensions, the lace expansion gives mean-field behaviour, and in particular , for sufficiently spread-out kernels when [12].
For kernels that are uniform on a finite range, Theorem 1.1’s conclusion follows from the recent preprint on quasi-transitive graphs [18].
Theorem 1.1 treats every and requires neither planarity nor high dimension nor a spread-out kernel. It contains the theorem of [2], whose kernels satisfy (Sym) (in , invariance under -rotations and reflections), (Pos) and exponential decay.
Idea of the proof
We follow the self-contained proof of for nearest-neighbour percolation on in [17], which builds on the renormalization scheme of Kozma and Nitzan [15 §4] and the seed-and-exploration method of Grimmett and Marstrand [9]. Figure 1 shows the structure. Assume, for a contradiction, that at . The argument has four steps.

Figure 1. How Theorem 1.1 and its corollaries are proved. A solid arrow means that its source is used in the proof of its target; only the main dependencies are drawn. The dashed arrow means that the theorem of [2] is contained in the case of Corollary 1.3. Box styles give provenance; grey labels give the source of each adapted step and where it is proved here. The modifications are listed in § 1.3.
Faces at . A large box connects with high probability, inside a surrounding cube , to each of the quarter-faces of that cube (a face cut in half in every transverse coordinate). Only three radii are needed, so these finitely many estimates persist at some .
Extension. If a vertex outside a box reaches , then it reaches, with almost the same probability, any target set that every point near can hit through a quarter-face. The cluster of enters a neighbourhood of at many possible points; at each one a small seed box, if all its edges are open, gives a relay point that reaches with high probability. The gluing inequality (Theorem 2.1) turns “every relay point reaches with high probability” into “ reaches if it reaches a relay point”.
Relay. A fixed sequence of extensions moves a connection from the centre of one coarse cube
to the centre of a neighbouring one.
Exploration. Cubes indexed by are explored one at a time, keeping a high conditional connection probability for each pending cube; a planar contour count then yields an infinite cluster at , a contradiction.
Long edges break two nearest-neighbour features of this argument.
First, a path that leaves a box need not pass through its boundary: its first step outside can be a long jump. In Step 1 we therefore use outer thick quarter-faces, slabs of width just outside a face of the cube (Figure 2), and we require the last edge of the connection to have length at most . Jumps longer than out of have total probability at most . The tail condition makes this small for a that is a small fixed fraction of , so the thick faces still fit inside the relay geometry of Step 3.

Figure 2. Lemma 3.2 in a planar cross-section. An open path from that leaves by an edge of length at most lands in the shell (light grey), which the outer thick quarter-faces cover; one is shown in dark grey. A path can also leave by a long edge (dashed), which a nearest-neighbour argument would not see; such edges are rare when is small.
Second, in Step 2 the nearest-neighbour proof selects an expansion of such that the cluster of outside has many exterior neighbours, using the fact that closing the single edge from each exterior neighbour into disconnects from . With long edges, every vertex of the outside cluster has edges into . We replace the count of exterior neighbours by the crossing weight
where is the cluster of in the complement of ; the probability that all crossing edges are closed is exactly (Figure 3). If is small this disconnection is cheap, and disjointness of the corresponding events over well-separated layers shows that a small weight is rare at some layer. If is large, the landing points of the crossing edges support many independent attempts to open a seed box, because a single seed box can absorb at most of the weight. Entrances that jump over a shell of width around are again controlled by the tail condition. The other steps of [17] need only bookkeeping: contraction steps instead of three in the relay between neighbouring cubes, margins enlarged by , and an exploration whose root cube is opened without fixing the edges that leave it.

Figure 3. The extension lemma (Lemma 4.1) in a planar cross-section. The cluster of outside reaches the shell . All crossing edges from into are closed with probability exactly , which is cheap when is small; when is large, the landing points in one residue class carry disjoint seed boxes , which give many independent chances to reach a good relay centre. Entrances that jump over the shell are rare by the tail condition.
The scales are chosen in the order , where and are large constant multiples of and the tail condition makes small (Lemma 3.3). Only one such choice is needed; this is why a lim inf suffices.
Adapted results. Lemma 3.2 and Lemma 3.3 adapt [17 Lemma 3.3] (outer thick faces and short exits); Lemma 4.1 adapts [17 Lemma 4.1], itself adapted from [15 Lemma 10] (crossing weight, layers spaced by , seeds at landing points); Lemma 5.1 adapts [17 Lemma 5.1] to and to the margins ; Section 6 adapts [17 §6, Lemmas 6.1–6.3] (long-range edges between neighbouring cubes, a hypothesis on edges entering the union of two cubes, and the root). The planar contour count is used as in [17 Lemma 3.1], and we include its proof.
Organization
Section 2 collects the inputs. Sections 3–6 carry out Steps 1–4, and Section 7 proves the corollaries and records where each hypothesis is used.
Preliminaries
Notation. and . A rectangle is a nonempty product of integer intervals, and denotes the expansion of a rectangle by in every coordinate direction. Throughout, ; it is finite under (Tail), since for some and the remaining sum is finite.
Product laws on finite graphs. In Sections 4–6 we work on finite graphs with , whose edges are pairs with , with independent bits of individual probabilities . We put , so that and the probability that all edges of a set are closed is . A bit is internal to a vertex set if both endpoints lie in . For we say that holds if every pair with $J(x-y) > 0 is an edge of with .
The gluing inequality
Theorem 2.1 [17, 4, 8]. Consider independent bond percolation on a finite graph with arbitrary edge probabilities in . For vertices and a nonempty vertex set ,
This is a joint form of Conjecture 1 of Kozma and Nitzan [15], in which the event alone appears on the left. Theorem 2.1 is proved for hypergraphs, with replaced by a vertex set, in [17 Proposition 2.1], and in the form above in [8 Theorem 1], where the extension from to is by continuity; it also follows from [4 Theorem 1] together with Harris’s inequality. The Lean formalization of Leder [16] proves the additive form: if for every , then [16 Theorem 2.2]. Either form gives the following corollary, which is [17 Corollary 2.3]; we include the short proof from Theorem 2.1.
Corollary 2.2 (failure comparison). Consider independent bond percolation on a finite graph with arbitrary edge probabilities in , let be vertex sets and . If for every , then for every vertex . The same holds after any deterministic set of bits has been fixed.
Proof. If , or (which forces when ) there is nothing to prove. Otherwise let be obtained from by identifying the vertices of to a single vertex , removing loops, and merging the parallel edges between two vertices into one edge, open if and only if one of the is open; this couples the two models, and the merged edge has probability . Under the coupling, in if and only if in , and then the open clusters of in and coincide. Let be the image of in . Hence . In every satisfies : for this is , and always. Theorem 2.1 in gives . Fixing a bit means giving it probability 0 or 1.
We also use Harris’s inequality [11]: increasing events of a finite product of independent bits are positively correlated, and so are decreasing events.
A planar boundary count. The following lemma is [17 Lemma 3.1]; we include the short proof.
Lemma 2.3. There are families of simple cycles of length in the dual lattice , with , such that the dual edge boundary of every finite containing 0 contains a cycle in . For each cycle in one can fix at least of its crossed primal edges with pairwise disjoint endpoints.
Proof. Each dual vertex has even degree in the dual boundary of (it counts membership changes around a primal square), so the boundary decomposes into edge-disjoint simple cycles. Along the positive horizontal ray from the number of boundary crossings is odd, so some cycle of the decomposition crosses it an odd number of times; a closed cycle crosses the whole horizontal line an even number of times, so this cycle also crosses the negative ray. If its length is , some positive crossing has horizontal coordinate with . Let be the set of simple cycles of length with such an anchored crossing and an odd number of positive crossings: there are at most anchors, two orientations and non-backtracking continuations. Finally select crossed primal edges greedily, discarding those incident to an endpoint of a selected edge; each selection discards at most seven edges, since primal degrees are four.
Elementary facts.
Lemma 2.4. Under (Sym), (Pos) and , if then . If , then an infinite cluster exists almost surely, and
Proof. The nearest-neighbour pairs in a coordinate plane are open independently with probability , by (Sym) and (Pos), and planar bond percolation with parameter close to percolates by Lemma 2.3 and a union bound over dual cycles; so . The cluster of is dominated by a Galton–Watson tree with mean offspring , so . Removing an edge splits a connected component into at most two components, so the event that an infinite cluster exists is unchanged by modifying finitely many bits and is a tail event; by Kolmogorov’s 0–1 law it has probability 1 when . The events that meets an infinite cluster increase to this event.
Faces at a percolating parameter, and the choice of scales
Definition 3.1. Let and . For , and , the outer thick quarter-face is
There are of them. For and such a face , let be the event that there is an open path with , , and . The event is increasing. It depends only on the bits internal to and on the pairs between and , all of which are internal to .
Lemma 3.2 (faces at a percolating parameter). Assume (Sym) and . For , , every and every outer thick quarter-face of ,
Proof. Take . Let be the event that some open edge has and ; then . If meets an infinite cluster, take a finite open path from to a vertex outside , and let be its first vertex outside and its predecessor. Off we have , so . Choose with , and
(either sign if ); then and the path up to witnesses . Hence . The events are decreasing and depend on finitely many bits. The signed coordinate permutations fixing act transitively on the faces and, by (Sym), preserve the law, and ; so all complements have the same probability . Harris’s inequality gives . □
Lemma 3.3 (choice of scales). Assume (Sym), (Tail) and . There are , positive integers , constants and such that:
(S1) and ;
(S2) , , and ;
(S3) ;
(S4) , the product over the pairs internal to with ; ; ; ; and ;
(S5) ;
(S6) , , and ;
(S7) for , every and every outer thick quarter-face of , .
Proof. Choose , then in this order; then by (1); then . None of these depends on or . By (Tail) there are arbitrarily large with as small as we like. For such put and ; then (S6) holds and with depending only on , so and (S5) holds for a suitable such . Every satisfies and , so Lemma 3.2, (S3) and (S5)
::: give . These finitely many events (up to translation) depend on finitely many bits whose probabilities are continuous in the parameter, so (S7) holds for some . □
We fix these choices from now on, and write (over the same pairs) for the probability that all pairs internal to a translate of with are open.
Extension through a fresh region
Lemma 4.1 (extension). Let be a finite graph as in § 2, a rectangle with and , a rectangle with , and . Assume:
every edge of with and has ;
for every there are and an outer thick quarter-face of with and .
Then .
Proof. By and (E2), every edge with and has ; hence for .
Step 1: a layer. For an integer let be the bits internal to , let be the open cluster of in the graph on with these bits, and set
Here . The crossing bits between and are independent of , so on . On the open cluster of in equals , because every other edge leaving is internal to and therefore closed by maximality.
Let . If both lie in , then ; on we have , hence , while , so and fails. The events , , are therefore pairwise disjoint, and
Since , we can fix with , by (S6).
Let be the event that some open edge has and . Then
by (S5). If , and , then an open path from enters for the first time at some from a predecessor , so and occurs. Hence
Step 2: relay quality. Put and let be the bits internal to . Every side of has at least points. For let be obtained from by clamping each coordinate to the corresponding side of shrunk by at both ends, let (so ), let be the event that all bits internal to are open, so , and put
is a function of . Fix . By (E3) at there are and with and ; the event uses only bits internal to , so by and (S7) its probability exceeds . By (Pos) and (Sym) an open box is connected through its nearest-neighbour pairs, so on we have . Harris’s inequality gives , and since is -measurable,
Step 3: entrances. Condition on ; write and for , so that and . Among the residue classes of modulo , choose (as a function of ) a class with . Distinct satisfy , and clamping moves each coordinate by at most , so and the boxes , , are disjoint.
Let be the event that some edge between and is open and holds. Given , the crossing bits are independent of , edges into distinct are distinct, and the events , , use disjoint bits of ; hence the events , , are independent with . Let and . Since is decreasing, and , which is at least on . By Chebyshev’s inequality, on ,
Let . By the independence of the crossing bits from and (3),
so . If , some has and : then reaches some , the open edge , and through the open box . Together with (2),
Step 4: the target. Condition on . Then is fixed, the remaining bits keep their product law, and for every . Corollary 2.2, applied with the bits of fixed, gives . Splitting according to whether gives .
Relay between neighbouring coarse cubes
For let
.
The cubes are vertex-disjoint. For neighbours in , let be the finite set of pairs with , and . The union is a rectangle with , whose internal bits are those internal to , those internal to , and . Let be the set of points of at distance more than from .
Lemma 5.1 (relay). Let be neighbours in and a finite graph as in § 2 with and . Assume
every edge of with and has .
If , then .
Proof. By (Sym) we may assume and , so that . Put . We define rectangles and apply Lemma 4.1 with and ; the radius is for and for . Let . For let be centred at with half-width in directions and in directions . For put (), , , , and .
Contraction steps. Let and . Then for and for . Take with and (with ). A point has with , so if and otherwise. For , lies between and , so , which is at most the half-width of in direction . Thus . The coordinates of are at most in absolute value, by (S6), so and . After the last contraction, .
Translation steps. Let and , so with . Take with . For , with , so , and for . Thus . The first coordinates of lie in and the others have absolute value at most ; by (S6) these lie in at distance more than from its complement. Finally and , so .
Each lies in , so gives (E2), and (E3) was just verified; also . If but , then and for some , and the union bound over the applications of Lemma 4.1 proves the claim.
Exploration and the proof of Theorem 1.1
We follow the exploration of [17 §6], with the long-range interfaces . Let , and let be the event that all bits internal to are open. It prescribes finitely many bits, each of probability in , so , and under all other bits are independent with their parameter- law.
The exploration. The coarse root is declared good and visited, but is never queued or processed; its four neighbours are put in a queue with parent . A coarse vertex is visited once it has been queued, and pending until it is processed; vertices are processed in queue order, and none is queued twice. For a pending , let be the graph consisting of the cubes on the parent chain of back to the root, the bits internal to these cubes, and the interfaces between successive cubes of the chain, and no other pairs. Put . To process , reveal the bits internal to and those of , where is the parent of . For each unvisited neighbour of , let (with the bits internal to ) and compute
which reveals nothing. Declare good if holds and for every such , and in that case queue these with parent ; otherwise declare bad. We maintain
Lemma 6.1 (freshness and the initial promise). Conditional on any revealed history of positive probability, the bits that are neither prescribed by nor revealed are independent with their parameter- law, and (4) holds at every step.
Proof. Each step reveals a finite set of bits determined by earlier revealed values, and all decisions are functions of revealed values; induction gives the first claim. The unknown bits of for a pending with parent are internal to or lie in . The cubes are vertex-disjoint, each step reveals the internal bits of one cube and its own parent interface, and contains no other pairs, so these bits stay unrevealed while is pending; together with the definition of good, this keeps (4) for every queued once it holds when is queued.
It remains to check (4) for the root’s neighbours. Let be one; by symmetry lies in direction from . Let , and . Then ; since ; and , since its first coordinates lie in and its other coordinates in . On , the box is joined inside to a vertex with an open edge of length at most into , and this edge lies in , so (the cube is open and connected by (Pos)). The event depends only on bits internal to and on pairs of , which are independent of ; hence by (S7) and (S2).
Lemma 6.2. Immediately before a pending is processed, the conditional probability that is declared bad is less than .
Proof. Fix an unvisited neighbour of and let . Neither (pending) nor (unvisited) has been processed and is not a parent interface yet, so every bit internal to is fresh with parameter , and holds in . The pairs of with one endpoint in and the other outside are those of , which is still unrevealed, with ; so (E2) holds. Since , Lemma 5.1 gives , and the tower property gives . With at most four such and (4), by (S2).
Lemma 6.3. For every deterministic set of distinct coarse vertices, the -probability that all of them are processed and declared bad is at most .
Proof. The root is good, so assume it is not in the set . Let be the time at which the th member of is processed ( if there is none) and the event that and the first processed members are bad. For each finite the event is determined before the reveal at time , so Lemma 6.2 gives , and induction finishes the proof.
Proof of Theorem 1.1. By Lemma 2.4, . Suppose that and choose the parameters of Lemma 3.3 with . Let be the set of good coarse vertices. If is finite, the queue empties after finitely many steps, and every coarse vertex outside adjacent to has been processed and declared bad. Apply Lemma 2.3 to : its boundary contains a cycle of some length from a family of at most candidates, each with crossed edges with disjoint endpoints, each of which has a bad endpoint outside . Choosing one endpoint per edge in ways, Lemma 6.3 bounds the probability of a given candidate by . By (S1), . Every good non-root vertex satisfies , and the boxes are disjoint, so with positive -probability the open cluster of in the long-range model is infinite. Hence with , a contradiction.
The corollaries, and where the hypotheses enter
Proof of Corollary 1.2. If then , so (Tail) holds, and (Pos) holds by the lower bound; apply Theorem 1.1. If , put and at a parameter with . Berger’s theorem [3 Theorem 2.6] states: if is percolating and for some , then is percolating for some (his standing assumption is , which holds here). Monotonicity and on give for large . Concavity of gives , so . Thus forces .
Proof of Corollary 1.3. Exponential decay and finite range both give .
Remark 7.1 (where the hypotheses enter). The tail condition is used only in (S5), at one pair of scales with a large constant multiple of fixed by (S6): in Lemma 3.2, for exits from by edges longer than , and in the event of Lemma 4.1, for entrances that jump over the shell. The layer argument needs and the relay needs ; both hold with , and (Tail) then makes small. Summability gives the uniform bound , which turns a large crossing weight into many seed attempts. (Sym) gives the equal probabilities of the faces in Lemma 3.2 and the orientation reductions. (Pos) gives and the connectivity of open boxes. The dimension enters through the coordinate plane in Lemma 2.4 and the coarse lattice with its contour count.
Remark 7.2 (the borderline). For the quantity stays bounded away from zero, and nothing in the argument guarantees (S5) and (S6) together: long edges at scale leave the cube with probability of order one at every scale. In and the transition is discontinuous [1, 5], and the argument above also lacks the coarse plane. Apart from sufficiently spread-out kernels in , where and [12] applies, whether for remains open.
References
- [1]M. Aizenman and C. M. Newman. Discontinuity of the percolation density in one-dimensional 1/|x − y|² percolation models. Comm. Math. Phys., 107(4):611–647, 1986.
- [2]O. Angel, D. de la Riva, C. Panagiotis, and Y. Shi. Continuity of the phase transition for long-range Bernoulli percolation with exponential decay on ℤ². Preprint, arXiv:2610.35233, 2026.arxiv.org/abs/2610.35233
- [3]N. Berger. Transience, recurrence and critical behavior for long-range percolation. Comm. Math. Phys., 226:531–558, 2002. doi:10.1007/s002200200617, arXiv:math/0110296.DOI
- [4]A. Blanc-Renaudie. Notes on Kozma and Nitzan’s proof of θ(p_c) = 0. Preprint, arXiv:2610.10502, 2026.
- [5]H. Duminil-Copin, C. Garban, and V. Tassion. Long-range models in 1D revisited. Preprint, arXiv:2011.04642, 2020.
- [6]H. Duminil-Copin, V. Sidoravicius, and V. Tassion. Absence of infinite cluster for critical Bernoulli percolation on slabs. Comm. Pure Appl. Math., 69(7):1397–1411, 2016.
- [7]R. Fitzner and R. van der Hofstad. Mean-field behavior for nearest-neighbor percolation in d > 10. Electron. J. Probab., 22:43, 2017. doi:10.1214/17-EJP56.DOI
- [8]J.-B. Goûéré. On the proof of the Kozma–Nitzan conjecture. Preprint, arXiv:2610.30063, 2026.arxiv.org/abs/2610.30063
- [9]G. R. Grimmett and J. M. Marstrand. The supercritical phase of percolation is well behaved. Proc. Roy. Soc. London Ser. A, 430(1879):439–457, 1990.DOI
- [10]T. Hara and G. Slade. Mean-field critical behaviour for percolation in high dimensions. Comm. Math. Phys., 128(2):333–391, 1990.
- [11]T. E. Harris. A lower bound for the critical probability in a certain percolation process. Proc. Cambridge Philos. Soc., 56:13–20, 1960.DOI
- [12]M. Heydenreich, R. van der Hofstad, and A. Sakai. Mean-field behavior for long- and finite range Ising model, percolation and self-avoiding walk. J. Stat. Phys., 132:1001–1049, 2008. doi:10.1007/s10955-008-9580-5.DOI
- [13]T. Hutchcroft. Power-law bounds for critical long-range percolation below the upper-critical dimension. Probab. Theory Related Fields, 181:533–570, 2021. arXiv:2008.11197.arxiv.org/abs/2008.11197
- [14]H. Kesten. The critical probability of bond percolation on the square lattice equals 1/2. Comm. Math. Phys., 74:41–59, 1980.DOI
- [15]G. Kozma and S. Nitzan. A reduction of the θ(p_c) = 0 problem to a conjectured inequality. Preprint, arXiv:2401.12397, 2024.arxiv.org/abs/2401.12397
- [16]J. Leder. θ(p_c) = 0 for Bernoulli bond percolation on ℤ^d in all dimensions d ≥ 2, in Lean 4/Mathlib, 2026. https://github.com/anthropics/formal-math/tree/795efb86f191735c5481675763537cfb4ff37e55/percolation.
- [17]OpenAI. Critical bond and site percolation on the cubic lattice, 2026. https://github.com/openai/math/tree/adc7f1241b42e322a6451854ab7e4b4c146bf78a/preprints/Critical-bond-and-site-percolation-on-the-cubic-lattice-September-24-2026.DOI
- [18]OpenAI. No percolation at criticality on quasi-transitive graphs, 2026. https://github.com/openai/math/tree/adc7f1241b42e322a6451854ab7e4b4c146bf78a/preprints/No-percolation-at-criticality-on-quasi-transitive-graphs-September-24-2026.DOI