{"id":"3bd81f20-b60c-4544-82a7-ea6bfb2fd23c","arxiv_id":"2607.29571","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Extensions by exact groups—semidirect products, abelian-base wreath products, graph wreath products, and certain free-by-cyclic groups—are shown to admit strongly converging finite-dimensional unitary representations.","lead":"Many countable groups built from exact-group extensions—semidirect products, wreath products with abelian base, graph wreath products, and certain free-by-cyclic groups—are shown to admit finite-dimensional unitary representations that converge strongly to the left regular representation. This extends the strong-convergence program to new families and supplies a PFF example (Gersten's group) outside the virtually special groups previously known to be PFF.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.3's key embedding kills the commutator subgroup of F_k, so the ultraproduct map is not injective and the Gersten-group PFF claim is unproven.","rationale":"The reader's weakest_assumption flagged reliance on free exactness and unpublished preprints as the most fragile premise, but did not identify the sharper issue: the specific sequence of homomorphisms in Theorem 1.3 does not separate the commutator subgroup of F_k, so the asserted embedding is false. This invalidates Theorem 1.3, which contains the paper's headline new example (Gersten's group is PFF). Other theorems (1.1, 1.2, 1.4–1.6) may survive with different arguments, but the central claim as written is not established. Since the error is a mathematical falsehood in the proof rather than a missing citation, the manuscript should be rejected in its current form; a revision replacing the abelian quotients with genuinely separating finite quotients (e.g., the permutation representations used in Theorem 1.4) might restore the result, but that is a substantial change.","tokens_in":9780,"tokens_out":26740,"duration_ms":258952,"concrete_test":"Take the nontrivial commutator [a_1,a_2] in F_k ⊂ F_n. For each m, the homomorphism F_n → (Z/mZ)^k * F_{n-k} sends a_1 and a_2 into the abelian group (Z/mZ)^k, so [a_1,a_2] maps to the identity in (Z/mZ)^k * F_{n-k}. Thus the image of λ_{[a_1,a_2]} in the ultraproduct is 1. If the claimed embedding existed, λ_{[a_1,a_2]} would have to map to a non-identity. This direct computation confirms that the ultraproduct map is not injective.","verdict_should_be":"REJECT","load_bearing_attack":"In the proof of Theorem 1.3 (§2.4), the authors claim an embedding C*_r(F_n) ↪ ∏_{m→U} C*_r((Z/mZ)^k * F_{n-k}) induced by homomorphisms sending each of the first k standard free generators of F_n to the corresponding generator of Z/mZ. This is false. Each such homomorphism restricts to F_k as the quotient onto the abelian group (Z/mZ)^k, so every element of the commutator subgroup [F_k,F_k] maps to the identity. Hence, for any nontrivial commutator x ∈ [F_k,F_k] (e.g., [a_1,a_2]), the image of λ_x under the induced ultraproduct map is 1, while λ_x ≠ 1 in C*_r(F_n). The map is not an embedding. Consequently, the subsequent application of Proposition 2.3 to embed C*_r(F_n⋊Z) into the ultraproduct of crossed products does not follow, and the PFF conclusion for Gersten's group and the family in Theorem 1.3 is unsupported. This is a concrete mathematical error, not merely a missing detail.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a 'double crossed product' upgrade principle (Proposition 2.3) and uses it to claim MF/PMF/PFF for several families of groups arising as extensions by exact groups: semidirect products with amenable residually finite groups, generalized wreath products with abelian base, certain free-by-cyclic groups including Gersten's group, Bernoulli-shift crossed products, and graph wreath products. The main advertised new example is that Gersten's group F_3 ⋊ Z is PFF despite not being virtually special. The proofs combine ambient strong convergence, exactness, and free-exactness results, several of which are imported from unpublished companion papers.","tokens_in":9987,"tokens_out":21134,"duration_ms":217559,"significance":"Proposition 2.3 is a clean and potentially useful upgrade argument, and the paper targets a broad and timely family of examples. If the main theorems were correct, they would substantially expand the known classes of MF/PMF/PFF groups. However, the proof of Theorem 1.3 contains a concrete mathematical error in the claimed free-exactness embedding, and Theorem 1.4's final PMF step is not justified. Since the most prominently advertised consequence—PFF for Gersten's group—rests on the flawed argument, the central claims are currently unsupported. The paper also depends heavily on unpublished same-group results, which makes verification difficult. The defects appear repairable, so a major revision is appropriate.","major_comments":[{"comment":"The claimed embedding C*_r(F_n) ↪ ∏_{m→U} C*_r((Z/mZ)^k * F_{n-k}) is false. The map F_n → (Z/mZ)^k * F_{n-k} quotients the free factor F_k by its abelianization modulo m, so every commutator [a_i,a_j] with i,j ≤ k lies in the kernel for every m. Thus the ultraproduct map cannot separate λ_{[a_i,a_j]} from 1, and a nontrivial element of C*_r(F_n) is mapped to zero. Moreover, for a non-amenable group a non-injective quotient map does not in general induce a *-hom on the reduced C*-algebra; the abelianization F_2 → Z^2 is a standard counterexample. Consequently the hypothesis of Proposition 2.3 is not established, and the PFF conclusion for Gersten's group and Corollary 2.5 is unsupported.","section":"§2.4, proof of Theorem 1.3"},{"comment":"The final step 'The result follows' is not justified. Even if the free-exactness embedding is granted, the target is an ultraproduct of C*-crossed products of the form ([M_m(C)*C*_r(F_{n-k})]⋊Z/rZ)⊗C*_r(Z), not an ultraproduct of group C*-algebras of groups that are known PMF. The PMF property for F_n⋊Z requires actual group homomorphisms into finite-dimensional unitary groups that strongly converge to the left regular representation. A C*-algebraic embedding into an ultraproduct of matrix algebras gives at best MF, not PMF, unless the embedding is induced by group homomorphisms into PMF groups. The proof does not show that the intermediate crossed products are group C*-algebras or that they admit the required finite-dimensional unitary representations. In addition, the free-exactness step with the non-injective finite-range maps F_k → M_m(C) is not a direct consequence of [Sko15] as sta","section":"§2.5, proof of Theorem 1.4"},{"comment":"The proof relies on [GKEMP26, Corollary 1.3] to assert that every graph product ⋆_Θ G is MF. This is a load-bearing external result that is unpublished and whose precise hypotheses are not stated in the present paper. If the corollary requires additional assumptions not verified here (for example exactness of G or finiteness of the graph), the proof of Theorem 1.6 does not apply. Please state the exact theorem used and either prove it or give a precise reference with the hypotheses.","section":"§2.7, proof of Theorem 1.6"}],"minor_comments":[{"comment":"The notation F_n = F_k * F_{n-k} should be clarified: the integer k (1 ≤ k ≤ n) is fixed but never explicitly introduced in the theorem statements.","section":"Theorems 1.3 and 1.4"},{"comment":"The paper cites [GKE26] and [GKEMP26] as unpublished preprints without arXiv numbers. Since central statements are quoted from them, please add precise theorem/corollary numbers and, ideally, make the statements available in an appendix or public preprint.","section":"References"},{"comment":"The sentence 'It is an embedding because ∩_n H_n = H so the sequence of homomorphisms is eventually separating and thus weakly converging' is compressed. Spell out the trace-preservation argument as done in Theorem 1.1.","section":"Proof of Theorem 1.2"},{"comment":"Example 2.8(4) (any action of a free group on any graph is residually finite) is stated without proof and only with a vague reference to the proof of [GKEP24, Theorem 2.14]. A precise derivation would help the reader.","section":"Definition 2.6 and Examples 2.8"},{"comment":"The phrase 'This is of minor relevance to the pure braid group on four strands' is unexplained; either expand the remark or remove it.","section":"Corollary 2.5"}],"recommendation":"major_revision","confidential_remarks":"The main obstacle is the concrete error in Theorem 1.3's proof; this is not a mere clarity issue. The error appears repairable by replacing the abelian quotients (Z/mZ)^k with finite quotients of F_k that are asymptotically faithful, but the current manuscript does not do this. Theorem 1.4's PMF conclusion also needs a serious rewriting to produce actual unitary representations rather than a C*-algebraic ultraproduct embedding. The heavy reliance on unpublished same-group preprints should be checked carefully by the editor. I recommend major revision rather than rejection because the overall strategy and Proposition 2.3 are plausible and potentially fixable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline: the proof of Theorem 1.3 is broken, and with it the claim that Gersten's group is PFF. The embedding C*_r(F_n) ↪ ∏_U C*_r((Z/mZ)^k ∗ F_{n-k}) is asserted via free exactness, but the maps sending the first k free generators to the standard generators of (Z/mZ)^k factor through the abelianization of F_k. So nontrivial commutators like [a_1,a_2] map to the identity in every coordinate. The induced ultraproduct map annihilates a fixed nonzero element of C*_r(F_n), so no embedding exists. If (Z/mZ)^k were meant as a free product of cyclic groups instead, then the commutators survive, but then the later claim that the Z-action quotients through Z/mZ fails because products of conjugates need not have finite order. Either way, Theorem 1.3 does not go through. The stress-test note is correct.\n\nThat said, the paper has real content. The upgrade mechanism in Proposition 2.3 is clean and I expect it to be reusable. The statements about semidirect products, generalized wreath products, and Bernoulli crossed products are genuinely new families, and Theorem 1.4's strategy—using strong convergence of finite-range representations of F_k combined with free exactness—is more plausible, though it leans on [GKE26] and [GKEMP26], which are unpublished same-group preprints. That dependence is a separate soft spot: several load-bearing facts are black-boxed, so a referee cannot check them without the preprints in hand. Theorem 1.4's final step is also compressed.\n\nThe paper is aimed at the strong-convergence / MF / PMF / PFF community. The Gersten-group example was the most conspicuous new result, and it is unsupported as written. The rest of the paper may still be salvageable, but only after the authors either repair the embedding in Theorem 1.3 or drop that claim. I would not cite the Gersten result in its current form.\n\nFor peer review: yes, this deserves a serious referee—the questions are important and the architecture is promising—but the referee should be asked to focus on the exactness/embedding steps and the unpublished dependencies. Acceptance would require fixing Theorem 1.3 and either proving or clearly importing the black-box results.","headline":"Theorem 1.3's central embedding is invalid, so the Gersten-group PFF claim is unproven; the rest of the paper is promising but leans heavily on unpublished same-group preprints.","tokens_in":10544,"tokens_out":11012,"would_cite":false,"duration_ms":103954,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["46L05","46L55","20F65","20E22"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that many groups formed as extensions of exact groups—including semidirect products, wreath products, and certain free-by-cyclic groups—admit strongly converging finite-dimensional unitary representations.","keywords":["strong convergence","MF","PMF","PFF","exact groups","crossed products","free-by-cyclic groups","wreath products"],"falsifier":"A concrete check: in Theorem 1.3, for the group with action a→a, b→ba, c→ca², verify for each m that the induced automorphism of (Z/mZ)*F_2 has order m; if for some m it has larger order, the embedding into a direct product with Z collapses. Similarly, in Theorem 1.4, explicitly construct the promised finite-dimensional representations for a small example (e.g., F_2⋊Z with a polynomial-growth automorphism) to test whether the 'result follows' step is valid.","tokens_in":9603,"feed_emoji":"🧮","tokens_out":10379,"duration_ms":97930,"temperature":0.7,"pith_summary":"This paper proves that several new families of countable groups—semidirect products with amenable normal subgroups, generalized wreath products with abelian base, graph wreath products, certain free-by-cyclic groups, and Bernoulli-shift crossed products—admit sequences of finite-dimensional unitary representations that converge strongly to the left regular representation. Groups with this property are called MF, PMF, or PFF, depending on whether the approximating maps are approximate homomorphisms, genuine homomorphisms, or homomorphisms with finite range. The result matters because it places the reduced C*-algebra of each such group inside a matrix ultraproduct, giving finite-dimensional approximations that have found uses in random matrix theory, spectral geometry, and von Neumann algebras. A notable new case is a specific free-by-cyclic group (with action a→a, b→ba, c→ca²) that is not virtually special yet is still PFF.","feed_headline":"Extensions by exact groups get strong matrix approximations","feed_subtitle":"A double crossed-product trick extends finite-dimensional unitary models to semidirect, wreath, and free-by-cyclic groups.","key_machinery":"The central mechanism is Proposition 2.3, a 'double crossed product' trick. It starts from a sequence of C*-dynamical systems (A^(k), α^(k), G) converging strongly to (A^(∞), α^(∞), G), and shows that the crossed products A^(k) ⋊ G converge strongly to A^(∞) ⋊ G. The proof embeds the crossed product into a double crossed product where one crossed product by G is inner, then uses exactness of G to peel it off as a tensor factor with C*_r(G) and applies Fell's absorption. A secondary tool is free exactness for free products, used in the free-by-cyclic cases to reduce the free group modulo finite quotients.","core_discovery":"The paper's central claim is that extension constructions by exact groups preserve the existence of strongly converging unitary representations. Concretely, it establishes that if L is an exact group with MF/PMF/PFF and G is a finitely generated residually finite amenable group, then the semidirect product G⋊L has the same property; that generalized wreath products ⊕_I G ⋊ L are PMF/PFF when G is residually finite abelian and L exact PMF/PFF; that free-by-cyclic groups of the form (F_k * F_{n-k}) ⋊ Z with certain 'multiplying by conjugates' actions are PFF or PMF; that Bernoulli-shift crossed products (⊗_I A)⋊L are MF under exactness assumptions; and that graph wreath products ⋆_Γ G ⋊ H are","pith_inferences":["The paper leaves open whether every free-by-cyclic group is PFF; the same double-crossed-product strategy might be pushed further if free exactness holds broadly enough.","If free exactness fails in any of the imported settings, the free-by-cyclic and graph-wreath theorems would need repair, but the semidirect and wreath-product theorems, which do not use free exactness, would survive.","The upgrade from a C*-algebraic embedding to explicit matrix homomorphisms in Theorem 1.4 is asserted rather than fully shown; making it explicit for small examples would either verify or reveal a gap.","The residually finite action condition in Theorem 1.6 is strong; weakening it (e.g., to sofic actions) would likely require new ideas, but might extend the result."],"forward_implications":["The reduced C*-algebra of each group treated embeds into a matrix ultraproduct, so the groups are MF (or PMF/PFF where stated).","The specific free-by-cyclic group with action a→a, b→ba, c→ca² is PFF despite not being virtually special, extending the known class of PFF groups.","Generalized wreath products with abelian base and exact PMF/PFF groups are PMF/PFF, a new result for this previously unaddressed family.","Bernoulli shift crossed products need not have amenable base algebras to be MF, as long as the acting group is exact and MF.","Graph wreath products of residually finite exact MF groups are MF, provided the action on the graph is residually finite."],"fun_headline_variants":["Exact extensions keep strong matrix approximations","Group extensions get unitary models via exactness","Semidirect and wreath products retain matrix approximations","Free-by-cyclic groups inherit strong unitary approximations","Exactness expands groups with matrix approximations"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The free-by-cyclic and graph-wreath theorems rely on unpublished 'free exactness' results and on a graph-product MF theorem imported from other work; if those statements require extra hypotheses not satisfied here, the corresponding conclusions fail.","fun_headline_variants_meta":{"raw":{"variants":["Exact extensions keep strong matrix approximations","Group extensions get unitary models via exactness","Semidirect and wreath products retain matrix approximations","Free-by-cyclic groups inherit strong unitary approximations","Exactness expands groups with matrix approximations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000148,"raw_usage":{"total_tokens":956,"prompt_tokens":607,"completion_tokens":349,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":351,"completion_tokens_details":{"reasoning_tokens":281}},"tokens_in":351,"tokens_out":349,"duration_ms":4434,"temperature":1.0,"reasoning_tokens":281,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T04:23:31.044909+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete check: in Theorem 1.3, for the group with action a→a, b→ba, c→ca², verify for each m that the induced automorphism of (Z/mZ)*F_2 has order m; if for some m it has larger order, the embedding into a direct product with Z collapses. Similarly, in Theorem 1.4, explicitly construct the promised finite-dimensional representations for a small example (e.g., F_2⋊Z with a polynomial-growth automorphism) to test whether the 'result follows' step is valid.","supporting_citations":[],"review_version":1}