A Classification of Transversal Clifford Gates for Qubit Stabilizer Codes

Shival Dasu; Simon Burton

REVIEW 2 major objections 5 minor 1 cited by

A Classification of Transversal Clifford Gates for Qubit Stabilizer Codes

T0 review · 2 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Every stabilizer code's diagonal transversal Clifford group is one of six matrix families, determined by the code's endomorphism algebra.

desk verdict Real classification result with a sound framework, but the garbled definition of U(ℓ,R8) in Example 7.9 and several unproved identifications need fixing before the main theorem is exact as written. read the letter →

arxiv 2507.10519 v1 pith:6JZTFJVA submitted 2025-07-14 quant-ph

Shival Dasu , Simon Burton This is my paper · ORCID

classification quant-ph MSC 81P7081P68 PACS 03.67.Pp03.67.Lx

keywords stabilizercodestransversalgatesCliffordgroupendomorphismalgebracodeclassificationmagicstatedistillationsymplecticGF(4)-linear

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to close a classification problem: for any qubit stabilizer code $C$, what is the group $G_\ell^C$ of diagonal transversal Clifford gates on $\ell$ copies of $C$? The claimed answer is a finite list of six matrix groups --- $\mathrm{Sp}(2\ell,\mathbb{F}_2)$, $U(\ell,\mathbb{F}_4)$, $\mathrm{GL}(\ell,\mathbb{F}_2)$, $O(\ell,\mathbb{F}_2[x]/(x^2))$, $U(\ell,R_8)$, and $O(\ell,\mathbb{F}_2)$ --- each attached to a specific family of codes determined by the code's algebra of $\mathbb{F}_2$-linear endomorphisms. The classification matters because transversal gates are automatically fault-tolerant, so knowing exactly which logical gates a code family supports tells designers which operations can be done without error propagation. It also yields concrete corollaries, including the absence of nontrivial one-, two-, and three-qubit transversal Cliffords for generic codes, and a transversal protocol for magic state preparation in a gauge-fixed $[[6,2,2]]$ code.

What carries the argument

The carrying object is the endomorphism algebra $A$ of a code: the set of $2\times 2$ matrices over $\mathbb{F}_2$ that preserve the code under the transversal action, which is always one of six isomorphism types $A_0,\ldots,A_5$ (namely $M_2(\mathbb{F}_2)$, $\mathbb{F}_4$, $\mathbb{F}_2\times\mathbb{F}_2$, $\mathbb{F}_2[x]/(x^2)$, $R_8$, and $\mathbb{F}_2$). The two workhorse results are Theorem 5.5, identifying the algebra of $2\ell\times 2\ell$ matrices preserving $C^{(\ell)}$ as the block algebra $M_\ell(A)$, and the symplectic condition $T J_n T^t = J_n$, rewritten through the conjugation $\bar{a}=J a^t J$ as the unitarity condition $T\bar{T}^t=I$. Solving that condition inside each $M_\ell(A_i)$ produces the six matrix groups.

What would settle it

Enumerate, for $\ell=2$, all $4\times 4$ symplectic matrices over $\mathbb{F}_2$ whose $2\times 2$ blocks lie in the upper-triangular algebra $A_4$, and compare the resulting set with $U(2,R_8)$, whose order the paper's table gives as 48; a mismatch would falsify the exactness of the six-group list. Separately, resolve the ambiguity in Example 7.9, where the unitarity condition is written both as $B\bar{B}^t = I$ and as $\bar{B}^t\bar{B} = I$; if these define different subgroups, the $U(\ell,R_8)$ entry is not well defined.

Watch

Extended reading notes

Core claim

The paper's central claim is Theorem 6.1: for every stabilizer code $C$, the group $G_\ell^C$ is exactly one of six families. If $C$ is self-dual CSS, the group is the full symplectic group $\mathrm{Sp}(2\ell,\mathbb{F}_2)$; if $C$ is GF(4)-linear, it is $U(\ell,\mathbb{F}_4)$; up to local-diagonal Clifford equivalence, a non-self-dual CSS code has $\mathrm{GL}(\ell,\mathbb{F}_2)$, a self-dual non-CSS code has $O(\ell,\mathbb{F}_2[x]/(x^2))$, and a semi-self-dual or semi-CSS code has $U(\ell,R_8)$; all remaining codes have only $O(\ell,\mathbb{F}_2)$, the permutations of the $\ell$ blocks. The proof route is to classify possible endomorphism algebras, show that the endomorphism algebra of $C^{(\ell)}$ is exactly the block algebra $M_\ell(A)$, intersect with the symplectic condition to obtain the unitarity equation $T\bar{T}^t = I$, and then identify the solution groups. The paper also strengthens the earlier endomorphism framework so that the classification is exact rather than merely a necessary-condition statement.

Load-bearing premise

The load-bearing step is the assertion in the proof of Theorem 6.1 that the matrices in $M_\ell(A_i)$ satisfying the unitarity condition are precisely the six listed groups; only three of the six cases are worked out in detail, so the $A_0$, $A_4$, and $A_5$ identifications are carried by a 'reader can check' argument.

Editorial extensions

If this is right

If Theorem 6.1 holds, the CSS property, GF(4)-linearity, self-duality, and semi-self-duality are each operationally characterized by which transversal Clifford gates exist on $\ell$ blocks.
A code has a transversal entangling two-qubit Clifford gate if and only if it is LDC-equivalent to a CSS code or a self-dual code; the gate is CNOT-like in the CSS case and Y-controlled-Y in the self-dual non-CSS case.
Generic codes, those not LDC-equivalent to CSS, GF(4)-linear, or self-dual codes, admit no nontrivial one-, two-, or three-qubit transversal Clifford gates, only permutations of the $\ell$ blocks.
The $[[5,1,3]]$ code has no nontrivial transversal two-qubit Clifford gate: its two-block gates are swaps followed by independent facet gates.
Codes with the $R_8$-type endomorphism algebra support both transversal CNOT and controlled-Z, and the gauge-fixed $[[6,2,2]]$ code supports a transversal magic-state preparation circuit tied to a demonstrated non-Clifford gate.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

One unstated step would be to make the Galois-type duality explicit: the six group families and six code families form matching inclusion lattices, and the lattice in Fig. 2 suggests that inclusions of groups correspond to inclusions of code families under LDC-equivalence.
Since the classification covers diagonal (uniform) transversal gates, a natural testable extension is the non-uniform case where different physical qubits receive different Cliffords; the $M_\ell(A)$ machinery may still constrain that setting.
The unresolved $A_4$/$U(\ell,R_8)$ case could be checked computationally for small $\ell$; any discrepancy would change the six-family list, so this is the most direct place to probe the theorem.
The open question about non-invertible $\mathbb{F}_2$-linear endomorphisms may connect to code switching, since the magic-state example already uses a gauge-fixing endomorphism to turn a $[[6,2,2]]$ code into a $[[6,1,2]]$ code.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Here's my take on arXiv:2507.10519. The headline: this is a real classification result, and the core framework is sound, but the write-up has a load-bearing ambiguity in the definition of U(ℓ,R8), plus a few "reader can check" steps that need to be filled.

What's actually new: the six-family classification of diagonal transversal Clifford groups for qubit stabilizer codes. The paper builds on Rains' endomorphism algebra framework, and credits him for the CSS/GL and trivial/O cases, but the U(ℓ,F4), O(ℓ,F2[x]/(x^2)), and U(ℓ,R8) families, together with the exact correspondence between code families and groups, are new. The proof structure is good: Theorem 5.5 relating multi-block endomorphisms to M_ℓ(A) is clean, and the reduction of G^ℓ_C to M_ℓ(A_i) ∩ Sp(2ℓ,F2) is the right way to approach it. The applications section, including the magic state protocol, is a nice payoff even though the central theorem doesn't depend on it.

The soft spots. In Theorem 6.1, the step 'the reader can check' for three of the six group identifications is not just a formality. For case (4), the paper's own definition of U(ℓ,R8) in Example 7.9 appears to be garbled: it defines the group by \bar B^t \bar B = I, but the symplectic condition from Eq. (11) is T \bar T^t = I. These are different conditions. Counting for ℓ=2, the symplectic condition gives 48 matrices (matching the paper's table), while the \bar B^t \bar B = I condition gives a different number (I get 22; the stress-test said 18, but either way it's not 48). So as written, case (4) is ambiguous and the table disagrees with the example. This is fixable—likely a typo, with the intended condition being \bar B^t B = I—but it's in the central theorem and needs to be corrected and verified. The other two unworked identifications (Sp and O(ℓ,F2)) are probably straightforward, but the authors should supply them. The classification of 3-dimensional subalgebras in Theorem 4.1 is also asserted without proof; it's a finite check, so less concerning, but it should be written out. There are also typos: 'case (5)' where it should be '(4)', and the orthogonality condition written as T^t T = 0 instead of T^t T = I or similar.

None of this makes me doubt the underlying math. The framework is sound, and the gaps are presentation issues rather than fatal flaws. If the authors clean up Example 7.9, provide the missing identifications, and correct the typos, this will be a solid paper. Who's it for? People working on fault-tolerant gate sets and code classification will want to read it. I'd send it to review, with the expectation of revision rather than desk rejection.

Referee Report

2 major / 5 minor

Summary. The paper claims to classify all possible groups of diagonal transversal Clifford gates on ℓ codeblocks of any qubit stabilizer code. It proves a classification of the F2-linear endomorphism algebras of stabilizer codes up to local diagonal Clifford equivalence (Theorem 4.1), shows that the multi-block endomorphism algebra of C^(ℓ) is M_ℓ(A) (Theorem 5.5), and then states that the group G_C^ℓ of transversal Cliffords is always one of six matrix groups: Sp(2ℓ,F2), U(ℓ,F4), GL(ℓ,F2), O(ℓ,F2[x]/(x^2)), U(ℓ,R8), or O(ℓ,F2) (Theorem 6.1). The final section applies the classification to two-qubit entangling gates and to a magic-state protocol, and includes tables of group orders for small ℓ.

Significance. If the main theorem is correct after the gaps are repaired, this is a valuable and surprisingly complete classification: it unifies known operational characterizations (CSS codes, GF(4)-linear codes, self-dual CSS codes) and adds new cases, in particular the self-dual non-CSS family with group O(ℓ,F2[x]/(x^2)) and the semi-self-dual family with group U(ℓ,R8). The proof strategy via endomorphism algebras is elegant, and the two foundational theorems (4.1 and 5.5) are proved in substantial detail. The paper also gives explicit group orders for small ℓ and derives a useful corollary about entangling two-qubit gates, with an interesting application to magic-state preparation. These strengths make the manuscript potentially publishable, but the main theorem currently rests on an unproved 'reader can check' identification for several algebras, and one of those identifications (U(ℓ,R8)) is internally inconsistent as written.

major comments (2)

[Theorem 6.1 proof and Example 7.9] The step 'The reader can check that the matrices in Mℓ(Ai) satisfying this unitarity condition are precisely the matrix groups in the theorem' is not carried out for A0, A4, and A5, and for A4 it is actually inconsistent with the text. Applying the symplectic condition (11), T \bar T^t = I, to T = [[A,B],[0,D]] with ℓ×ℓ blocks gives A D^t = I, D A^t = I, and A B^t + B A^t = 0, yielding |GL(ℓ,F2)|·2^{ℓ(ℓ+1)/2} elements (48 for ℓ=2, matching Figure 4). Example 7.9 instead defines U(ℓ,R8) by \bar B^t \bar B = I, which forces A^t A = I and D^t D = I and gives a strictly smaller group for ℓ=2. Thus the exact content of case (4) is ambiguous, and the six-group list is not established as stated. The authors should specify which unitarity convention defines U(ℓ,R8), prove the matrix identification for all A_i, and reconcile Example 7.9 with the counts in Figure 4.
[Theorem 4.1(4) and Figure 3] The definition of A4 is internally inconsistent. Theorem 4.1(4) and Figure 3 define A4 = F2⟨[[1,0],[0,0]], [[1,1],[0,0]]⟩, but these two matrices generate only the 2-dimensional algebra of matrices of the form [[a,b],[0,0]], which does not contain the 2×2 identity matrix. This contradicts Definition 3.5 and Lemma 3.4, which require the endomorphism algebra to contain the identity, and it also contradicts the proof of Theorem 4.1 and Example 7.9, where A4 is the 3-dimensional unital ring of upper-triangular matrices. Please correct the generators (e.g., replace the second generator with [[0,0],[0,1]]) so that A4 is the intended unital algebra R8.

minor comments (5)

[Section 6, Eq. (7)] In the proof of Theorem 6.1, the symplectic condition is written as T J_n T^t = J_n, but for ℓ codeblocks the relevant form is J_ℓ on F2^{2ℓ}; the notation should distinguish the two.
[Example 7.5] The text says the symplectic condition 'becomes T^t T = 0' for A3; it should be T^t T = I, since the conjugation action fixes all elements of A3.
[Example 7.9] The displayed identity 'BC^t = \bar C^t \bar B^t' is garbled; it should presumably be \overline{BC}^t = \bar C^t \bar B^t.
[Proof of Theorem 4.1] The assertion that 'there are three 3-dimensional subalgebras' of M2(F2) is made without proof; a short justification of completeness would be helpful.
[Abstract and references] The abstract contains the typo 'classifying stabilizer codes by via matrix algebras', and references [8] and [9] appear to be the same Gottesman paper and should be merged.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the classification of transversal Clifford gates is derived from the endomorphism algebra and the symplectic condition, not assumed or fitted.

full rationale

The paper's central derivation is self-contained. Theorem 5.5 proves that the algebra of transversal endomorphisms of C^(ℓ) is exactly M_ℓ(A), where A is the endomorphism algebra of C. Theorem 6.1 then characterizes G_ℓ^C as M_ℓ(A_i) ∩ Sp(2ℓ, F_2), an intersection of a proven algebraic condition with the symplectic condition that defines Clifford tableaus. The six group families are not introduced as the answer and then used to define the condition; rather, the condition T·Tbar^t = I is derived from Eq. (7), and the subsequent identifications for A_1, A_2, A_3 are worked out explicitly in Section 7. The A_4 case is asserted with a 'reader can check' step, and Example 7.9's displayed unitarity condition is internally inconsistent with Eq. (11) as written, but that is a rigor or correctness gap, not circularity: the theorem's claim does not reduce to its own input by construction. The only self-citation is [4], which appears in the application to magic state distillation and is not load-bearing for the classification theorem. No parameters are fitted to data, and no prediction is statistically forced. Accordingly, no circular step can be exhibited, and the appropriate score is 0.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The classification theorem depends on the stabilizer formalism, Rains' prior framework, a finite algebra enumeration, and an unproved group-identification step. No free parameters are fitted and no new physical entities are introduced.

assumptions (5)

standard math The Clifford group modulo Pauli phases is isomorphic to Sp(2ℓ,F2)
Used in Section 2.2 to represent Clifford gates by symplectic tableaus; standard result cited to [13].
standard math Stabilizer codes correspond to symplectic subspaces of F2^{2n}
Foundational identification used throughout; standard (Definition 2.5).
domain assumption Rains' endomorphism algebra framework, including the strengthening asserted in the introduction
The paper states it strengthens Rains' result so that non-transversal operators fail for all codes with endomorphism algebra A, but no proof appears in the text; the exactness of the classification depends on it.
domain assumption The enumeration of 2- and 3-dimensional subalgebras of M2(F2) is complete
In the proof of Theorem 4.1, the 3-dimensional algebras U, L, E are asserted to be the only ones; a missing algebra would add a code family.
ad hoc to paper For each A_i, the conjugation bar(a)=J a^t J maps A_i to itself, and the unitarity condition T bar(T)^t = I yields exactly the listed matrix groups
This is the load-bearing step of Theorem 6.1, stated as 'the reader can check' and only partially verified in Section 7.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A Classification of Transversal Clifford Gates for Qubit Stabilizer Codes." pith.science (2026). https://pith.science/paper/6JZTFJVA

@misc{pith2026250710519,
  author       = {Pith},
  title        = {Pith review of: A Classification of Transversal Clifford Gates for Qubit Stabilizer Codes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6JZTFJVA}},
  note         = {Machine review of arXiv:2507.10519}
}

abstract

This work classifies stabilizer codes by the set of diagonal Clifford gates that can be implemented transversally on them. We show that, for any stabilizer code, its group of diagonal transversal Clifford gates on $\ell$ code blocks must be one of six distinct families of matrix groups. We further develop the theory of classifying stabilizer codes by via matrix algebras of endomorphisms first introduced by Rains, and give a complete classification of the diagonal Clifford symmetries of $\ell$ code blocks. A number of corollaries are given in the final section.

Figures

Figures reproduced from arXiv: 2507.10519 by the authors.

**Figure 1.** We consider ℓ copies of an n-qubit code. For clarity here we choose n = 4. A diagonal Clifford operator uses n copies of the same Clifford gate U. In this work we completely solve this problem, finding that the group of transversal diagonal ℓ-qubit Clifford gates on ℓ code blocks of a stabilizer code must be one of six families of matrix groups, together with their corresponding code families. See [PITH_FULL_IMAGE… view at source ↗

**Figure 2.** (i) Each family of stabilizer code is characterized by (ii) a corresponding group of [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** The endomorphism algebras used in the classification theorem 4.1, as (i) concrete [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: The entries in this table show the orders of six families of matrix groups which are [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Beyond transversality: structure of Clifford circuits for CSS codes
quant-ph 2026-08 conditional novelty 8.0 of 10

Every code-preserving Clifford circuit for a CSS code is a product of Z-diagonal and X-diagonal circuits, and two-fold transversal circuits realize the full logical Clifford group for 78 codes.

Reference graph

Works this paper leans on

15 extracted references · 14 canonical work pages · cited by 1 Pith paper

[1]

Aaronson and D

S. Aaronson and D. Gottesman. Improved simulation of stabilizer circuits.Phys. Rev. A, 70:052328, Nov 2004

work page 2004
[2]

Calderbank, E

A. Calderbank, E. Rains, P. Shor, and N. Sloane. Quantum error correction via codes over GF(4).IEEE Transactions on Information Theory, 44(4):1369–1387, 1998

work page 1998
[3]

A. R. Calderbank and P. W. Shor. Good quantum error-correcting codes exist.Physical Review A, 54(2):1098, 1996

1996
[4]

S. Dasu, S. Burton, K. Mayer, D. Amaro, J. A. Gerber, K. Gilmore, D. Gresh, D. Del- Vento, A. C. Potter, and D. Hayes. Breaking even with magic: demonstration of a high-fidelity logical non-clifford gate, 2025

work page 2025
[5]

Gidney, M

C. Gidney, M. Newman, and M. McEwen. Benchmarking the planar honeycomb code. Quantum, 6:813, 2022

work page 2022
[6]

H. Goto. Step-by-step magic state encoding for efficient fault-tolerant quantum compu- tation.Scientific reports, 4:7501, 12 2014

work page 2014
[7]

Gottesman.Stabilizer codes and quantum error correction

D. Gottesman.Stabilizer codes and quantum error correction. California Institute of Technology, 1997

work page 1997
[8]

Gottesman

D. Gottesman. Theory of fault-tolerant quantum computation.Physical Review A, 57(1):127–137, Jan. 1998. 18

work page 1998

Show all 15 references

[9]

Gottesman

D. Gottesman. Theory of fault-tolerant quantum computation.Phys. Rev. A, 57:127–137, Jan 1998

1998
[10]

C. Jones. Multilevel distillation of magic states for quantum computing.Physical Review A, 87(4), Apr. 2013

2013
[11]

F. Klein. A comparative review of recent researches in geometry.Bulletin of the American Mathematical Society, 2(10):215–249, 1893
[12]

E. Knill. Quantum computing with realistically noisy devices.Nature, 434(7029):39–44, Mar. 2005

2005
[13]

K. Mastel. The clifford theory of then-qubit clifford group, 2023

2023
[14]

E. Rains. Nonbinary quantum codes.IEEE Transactions on Information Theory, 45(6):1827–1832, 1999

1999
[15]

A. Steane. Multiple-particle interference and quantum error correction.Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences, 452(1954):2551–2577, 1996. 19

1954

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

Aaronson and D

S. Aaronson and D. Gottesman. Improved simulation of stabilizer circuits.Phys. Rev. A, 70:052328, Nov 2004

work page 2004

[2] [2]

Calderbank, E

A. Calderbank, E. Rains, P. Shor, and N. Sloane. Quantum error correction via codes over GF(4).IEEE Transactions on Information Theory, 44(4):1369–1387, 1998

work page 1998

[3] [3]

A. R. Calderbank and P. W. Shor. Good quantum error-correcting codes exist.Physical Review A, 54(2):1098, 1996

1996

[4] [4]

S. Dasu, S. Burton, K. Mayer, D. Amaro, J. A. Gerber, K. Gilmore, D. Gresh, D. Del- Vento, A. C. Potter, and D. Hayes. Breaking even with magic: demonstration of a high-fidelity logical non-clifford gate, 2025

work page 2025

[5] [5]

Gidney, M

C. Gidney, M. Newman, and M. McEwen. Benchmarking the planar honeycomb code. Quantum, 6:813, 2022

work page 2022

[6] [6]

H. Goto. Step-by-step magic state encoding for efficient fault-tolerant quantum compu- tation.Scientific reports, 4:7501, 12 2014

work page 2014

[7] [7]

Gottesman.Stabilizer codes and quantum error correction

D. Gottesman.Stabilizer codes and quantum error correction. California Institute of Technology, 1997

work page 1997

[8] [8]

Gottesman

D. Gottesman. Theory of fault-tolerant quantum computation.Physical Review A, 57(1):127–137, Jan. 1998. 18

work page 1998