{"id":"3e62b585-d720-4bc0-9809-1d02c147b125","arxiv_id":"2508.12392","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Heterotic one-loop threshold corrections imply upper bounds on the modulus in modular flavor models, ruling out tau near i infinity for typical dilaton and beta-function values and disfavoring tau = i at large volume.","lead":"This paper derives upper bounds on the shape and size parameters, called moduli, that modular flavor models use to explain fermion masses, by requiring that perturbative string corrections stay small. The bounds disfavor the popular large-modulus limit and, in large-volume compactifications, also the self-dual point tau = i, while leaving tau = omega viable.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Threshold-dominance assumption (5.2) is the load-bearing step; if Green-Schwarz terms or other sectors contribute to Delta_a, the claimed modulus bounds are not necessary.","rationale":"The reader's weakest_assumption exactly matches my main concern: Eq. (5.2) with Delta_a truncated to N=2 contributions is not obviously a necessary condition for perturbative control. The manuscript itself flags the dropped Green-Schwarz constants in footnote 2, and the hatted-coefficient identification in Sec. 4.3 is an assumption rather than a derived relation. Since the abstract and conclusions frame the result as a Swampland constraint, the strength of the claim exceeds what the inequality can support. However, the internal algebra from the assumed threshold forms is coherent, the Lambert-W bounds are correctly derived, and the numerical checks are consistent, so the paper is a useful conditional bound rather than being wrong. I keep the reader's CONDITIONAL verdict: if the authors explicitly reframe the result as 'bounds under the stated truncation of threshold corrections' and caveat the Swampland language, the paper could be upgraded; if the Green-Schwarz terms are shown to cancel the N=2 contribution, the central bounds would not hold. The concrete test of adding the complete threshold formula is decisive in principle, since it directly probes whether the omitted terms are numerically relevant at the quoted benchmark parameters.","tokens_in":23842,"tokens_out":1568,"duration_ms":16630,"concrete_test":"Re-derive the one-loop corrected gauge coupling for the T6/Z4 orbifold of Sec. 4.1 including the complete Kaplunovsky-Louis formula, especially the Green-Schwarz contribution Delta_GS and any moduli-independent threshold constants, then evaluate 16 pi^2/g_a^2 at ImS = 2 with the model's b_GS values. If the Green-Schwarz term shifts the effective threshold by order 1-10 relative to the N=2 piece, the bound Im U < 34.3 changes by an order-one factor or disappears entirely for particular b_GS values.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central quantitative claim is that perturbative heterotic vacua require Eq. (5.2), 16 pi^2 k_a ImS > |Delta_a|, and the paper then uses only the moduli-dependent N=2 threshold pieces to bound Im U. The weakest load-bearing assumption is that |Delta_a| can be replaced, for the purpose of a necessary constraint, by the specific threshold correction displayed. Three gaps compound: (i) Footnote 2 explicitly drops numerical constants associated with the Green-Schwarz function, yet the full one-loop threshold correction includes moduli-independent and dilaton-dependent Green-Schwarz terms whose magnitude can be comparable to or larger than the N=2 contribution; (ii) Section 4.3 sets hatted coefficients equal to b^{N=2}_a for groups (4) and (5), which is a choice, not a derivation, and it changes the numerical bounds; (iii) Eq. (5.2) itself is a sufficient condition for weak coupling at the string scale, not a necessary one, since threshold corrections could be cancelled by adjusting other one-loop terms or by non-perturbative effects. Therefore the statement that Im U >~ 34.3 lies in the Swampland, and the abstract's claim about tau ~ i infinity, is a constraint only under a specific model of the threshold correction, not a rigorous string-theoretic necessity.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies constraints on the moduli space of modular flavor models arising from one-loop moduli-dependent threshold corrections in perturbative heterotic string theory on toroidal orbifolds. After reviewing how modular flavor symmetries emerge from extra-dimensional wavefunctions, the authors import threshold formulas from Ref. [36] and group them into five classes according to target-space duality symmetries: PSL(2,Z) × PSL(2,Z), Γ0(2)-type groups, and Γ0(3)-type groups. The central step is inequality (5.2), which requires the tree-level dilaton contribution to the gauge coupling to dominate the one-loop threshold correction. Using the large-imaginary-part approximation for Dedekind eta functions, the authors derive analytic upper bounds on the imaginary parts of the complex-structure modulus U and the Kähler modulus T in terms of Lambert W functions, and they supplement these with numerical results in the region Im τ ≲ 1. The main quantitative claims are that for representative inputs (Im S = 2, k_a = 1, |b_a^{N=2}| = 10) the modulus is bounded by Im U ≲ 34.3 when T is fixed at the elliptic point ω, that larger ratios T/U and larger beta-function coefficients strengthen the bounds, and that in the large-volume regime the fixed point τ = i becomes disfavored while τ = ω remains allowed. The paper concludes with phenomenological implications for modular flavor model building and radiative moduli stabilization.","tokens_in":24175,"tokens_out":6166,"duration_ms":66608,"significance":"If the central inequality (5.2) were a necessary condition for perturbative heterotic vacua, the paper would provide a broadly useful top-down constraint on modular flavor model building, linking the modulus values used in bottom-up flavor fits to the validity of the string perturbation expansion. The paper has several genuine strengths: the Lambert W closed-form bounds are derived cleanly, the classification of threshold corrections into five duality-group classes is systematic, and the numerical analysis explicitly covers the region Im τ ≲ 1 where the analytic eta approximation fails. The authors are also transparent that the inputs Im S, k_a, and b_a^{N=2} are treated as scanned parameters rather than fitted to the output, so there is no disguised circularity. The main limitation is that the results are conditional on a specific model of the threshold corrections: the bounds are sufficient conditions for perturbativity under the retained N=2 threshold terms, not necessary string-theoretic constraints once Green-Schwarz constants and other one-loop sectors are included.","major_comments":[{"comment":"The inequality 16π² k_a Im S > |Δ_a| is presented as the condition ensuring weak gauge couplings at the string scale, and it is then used to derive upper bounds such as Eq. (5.10). However, this condition is sufficient rather than necessary: the full one-loop threshold correction contains moduli-independent and dilaton-dependent Green-Schwarz terms, and footnote 2 on page 7 explicitly drops the numerical constants associated with the Green-Schwarz function. Moreover, even within the retained N=2 pieces, other one-loop sectors or additional threshold contributions could partially cancel the moduli-dependent term. Consequently, the statement in the abstract that τ ≃ i∞ 'lies in the Swampland' is stronger than what the derivation supports. The paper should reframe Eq. (5.2) as a sufficient perturbativity constraint under the assumption that the exhibited N=2 threshold terms dominate, and either quantify the dropped Green-Schwarz constants or restrict the claims to cases where those constants are known to be negligible.","section":"Sec. 5, Eq. (5.2)"},{"comment":"For groups (4) and (5), the coefficient Rhat b_a of the (I,θ³) sector is set equal to b_a^{N=2} 'for simplicity'. This is a choice, not a derivation: the orbifold relation (4.7) fixes b_a^{N=2} = 4 b^{(I,θ²)}_a but leaves Rhat b_a independent. Since the constraints (5.17)-(5.20) depend on Rhat b_a through terms with different moduli dependence (T, Rhat T, and the η(ω) contribution in Eq. (4.13)), the reported bounds for these groups are model-dependent. The authors should either derive Rhat b_a for the specific orbifolds listed in Sec. 4.3 or present the bounds as functions of the ratio Rhat b_a / b_a^{N=2} and explicitly state which ranges are covered by the figures.","section":"Sec. 4.3, Eqs. (4.12)-(4.13)"},{"comment":"The headline numerical result Im U ≲ 34.3 is obtained for the representative inputs Im S = 2, k_a = 1, and |b_a^{N=2}| = 10. These values are taken from MSSM gauge-coupling unification and typical beta-function magnitudes, but they are not derived for any concrete heterotic vacuum; in a given orbifold the beta-function coefficients, modular levels, and dilaton VEV are correlated, and the relevant threshold formula may not be the group (1) form. The paper should therefore label the quantitative bounds and the associated phenomenological conclusions in Sec. 6 as illustrative rather than as universal string-theoretic constraints, and it should indicate how the bounds change when the parameters are varied within ranges actually realized in the orbifold models of Sec. 4.3.","section":"Sec. 5.1, Eq. (5.10) and Figs. 2-5"}],"minor_comments":[{"comment":"The name 'Green-Shwarz' is a typo; it should be 'Green-Schwarz'.","section":"Footnote 2, p. 7"},{"comment":"The same symbol Rbar Γ0(n) is defined in two different ways, once with c ≡ 0 mod n and once with b ≡ 0 mod n; the notation should distinguish Rbar Γ₀(n) from Rbar Γ⁰(n) or use different symbols to avoid confusion.","section":"Sec. 3, Eqs. (3.6)-(3.7)"},{"comment":"The target-space duality group is written as Γ0_T(2) × (Γ_U)0(2) in Sec. 4.1 and as Γ0_T(2) × (Γ_U)0(2) in Eq. (4.10); please unify the notation for the upper and lower congruence subgroups.","section":"Sec. 4.1 and Sec. 4.3"},{"comment":"The sentence 'the RHS can be calculated as follows' presents the numerical value 0.680... without specifying the minimizing point of T₂|η(T)|⁴ U₂|η(U)|⁴; adding that the minimum occurs at T = U = ω would make the derivation easier to follow.","section":"Sec. 5.1, Eq. (5.4)"},{"comment":"The captions 'These are the boundaries concerning 7T = U' and 'concerning 7T = U, Rhat T = T, Rhat U = T' are terse; please state explicitly that Im S = 2 is fixed and indicate which axis corresponds to the plotted modulus.","section":"Sec. 5.2, captions of Figs. 10 and 13"},{"comment":"The claim that 'the larger the volume of extra dimensional space, the stronger the stringy constraints' is inferred from a few values T = 10i, 10.02i, 10.04i, 10.06i; please clarify whether this is a numerical trend observed in the plotted examples or an analytic property of Eq. (5.3).","section":"Sec. 6, Fig. 16"}],"recommendation":"major_revision","confidential_remarks":"The paper is a reasonable application of known heterotic threshold formulas to modular flavor model building, and the central derivation is internally consistent once the assumptions are accepted. The main concern is not technical error but the strength of the claim: inequality (5.2) yields sufficient, not necessary, conditions for perturbativity, and the Green-Schwarz terms and the choice Rhat b_a = b_a^{N=2} are dropped without a justification that they are small in the cases of interest. This is fixable by rewriting the abstract and conclusions in conditional language and by presenting the bounds as model-dependent sufficient constraints. I would also suggest that the authors clarify in Sec. 6 that the phrase 'Swampland' is used in the loose sense of 'excluded by a perturbative string-theory consistency requirement under stated assumptions,' rather than in the strict sense of the Swampland program."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this if you care about modular flavor models with a heterotic UV completion. The new thing is the map, not the ingredients: using old threshold formulas, the authors derive explicit Lambert-W bounds on ImU and scan them over five duality-group structures. The numbers are concrete — with ImS=2, k_a=1, |b^{N=2}_a|=10, they get ImU ≲ 34.3 when T is fixed at the omega fixed point — and the low-Im region is checked numerically rather than left to the asymptotic eta expansion. That part is honest, reproducible, and useful.\n\nCredit where it's due: the internal algebra is consistent, the approximation η(U) ~ exp(2πiU/24) for ImU>1 is legitimate, and the authors do not dress up scanned inputs as predictions. The classification of threshold correction structures in Sec. 4.3, following Refs. [34,36], is convenient, and the figures are readable.\n\nThe soft spot is the load-bearing inequality (5.2). It is a sufficient condition for perturbativity — tree-level dilaton beats the one-loop correction — but not a necessary one. Threshold corrections could in principle be offset by Green-Schwarz terms or other one-loop sectors, and the paper's own footnote 2 drops the numerical constants attached to the Green-Schwarz function. Section 4.3 also sets the hatted beta coefficients equal to b^{N=2}_a by choice, not derivation. So the claimed exclusion of τ ~ i∞ and τ=i in large volume is conditional on a particular model of the threshold correction. The abstract's \"lies in the Swampland\" wording overstates it: the paper shows that these regions are disfavored under the stated assumptions, not that they are impossible in any consistent string vacuum. That distinction matters for someone who wants to build a model at τ=i.\n\nWho is this for? Modular flavor model builders, especially those using heterotic orbifolds or relying on large moduli. For that community it is a useful map and a reminder that perturbativity is a real constraint. It does not reshape a major field, and it would be better if the title and abstract said \"perturbativity constraints\" rather than \"Swampland constraints.\"\n\nI would send it to a serious referee. The referee should ask the authors to quantify the Green-Schwarz contributions, or state conditions under which those terms are negligible, and to soften the abstract accordingly. If they address that, it becomes a solid paper for the flavor-model literature.","headline":"Systematic, honestly-labeled perturbativity bounds on heterotic moduli; the Swampland wording outruns the evidence.","tokens_in":24651,"tokens_out":2911,"would_cite":true,"duration_ms":30721,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Perturbative heterotic string theory excludes the large modulus values modular flavor models use and disfavors tau = i at large volume.","keywords":["modular flavor models","heterotic string theory","threshold corrections","target-space modular symmetry","Lambert W function","Swampland","moduli stabilization","orbifold compactification"],"falsifier":"Compute the full one-loop threshold correction $\\Delta_a$ from Eq. (3.3), including all twisted sectors and Green-Schwarz contributions, for a concrete $T^6/\\mathbb{Z}_4$ or $T^6/\\mathbb{Z}_{6-II}$ vacuum with $\\operatorname{Im} S = 2$, $k_a = 1$, $|b_a^{N=2}| = 10$, $T = \\omega$, and $\\operatorname{Im} U = 40$. If the resulting corrected gauge coupling $16\\pi^2 k_a \\operatorname{Im} S - |\\Delta_a|$ remains positive, then inequality (5.2) is not necessary and the bound $\\operatorname{Im} U \\lesssim 34.3$ is evaded.","tokens_in":23643,"feed_emoji":"🌀","tokens_out":7515,"duration_ms":68580,"temperature":0.7,"pith_summary":"The paper tries to establish that string theory itself restricts which values of the moduli in modular flavor models are physically allowed, by demanding that one-loop threshold corrections do not overwhelm the tree-level gauge coupling in perturbative heterotic vacua. The resulting inequality, $16\\pi^2 k_a \\operatorname{Im} S > |\\Delta_a|$, turns into explicit upper bounds on the imaginary part of the complex-structure modulus, solved with the Lambert W function. With the dilaton at $\\operatorname{Im} S \\sim 2$, level $k_a = 1$, and $\\beta$-function coefficient $|b_a^{N=2}| \\sim 10$, the bound is $\\operatorname{Im} U \\lesssim 34.3$ when the Kähler modulus is fixed at $\\omega$. This matters because modular flavor models often place the modulus at $\\tau \\simeq i\\infty$ or near $\\tau = i$ to produce hierarchical Yukawa couplings; the paper argues those points lie in the Swampland or are disfavored in the large-volume regime.","feed_headline":"Heterotic string thresholds cap the flavor modulus to about 34","feed_subtitle":"One-loop gauge corrections rule out the large moduli behind fermion mass hierarchies and disfavor tau = i.","key_machinery":"The load-bearing object is the perturbativity inequality (5.2), $16\\pi^2 k_a \\operatorname{Im} S > |\\Delta_a|$, evaluated at $\\mu^2 \\sim M_{\\text{string}}^2$, where $\\Delta_a$ is the one-loop moduli-dependent threshold correction to the gauge coupling. The paper uses the classification of these threshold corrections by target-space duality symmetries, $PSL(2,\\mathbb{Z})$, $\\Gamma_0(n)$, and $\\Gamma^0(n)$, together with explicit forms built from Dedekind eta functions such as $\\operatorname{Im} T\\,|\\eta(T)|^4 \\operatorname{Im} U\\,|\\eta(U)|^4$. In the region $\\operatorname{Im} U > 1$, the eta function is approximated by $q^{1/24}$, which turns the inequality into an equation solved by the Lambert W function, yielding closed-form bounds on $\\operatorname{Im} U$ and $\\operatorname{Im} T$ as functions of $\\operatorname{Im} S$, $b_a^{N=2}$, and the ratio $T/U$.","core_discovery":"Within perturbative heterotic string theory on toroidal orbifolds, the paper's central claim is that perturbativity, meaning that the tree-level gauge coupling exceeds the one-loop moduli-dependent threshold correction, imposes necessary constraints on the modulus space of modular flavor models. Using threshold corrections classified by target-space duality symmetries $PSL(2,\\mathbb{Z})$, $\\Gamma_0(n)$, and $\\Gamma^0(n)$, it derives bounds on $\\operatorname{Im} U$ and $\\operatorname{Im} T$ whose boundary is given by the Lambert W function. For representative values $\\operatorname{Im} S \\sim 2$, $k_a = 1$, and $|b_a^{N=2}| \\sim 10$, fixing $T = \\omega$ gives $\\operatorname{Im} U \\lesssim 34.3$; the constraints tighten as the dilaton decreases, the $\\beta$-function coefficient increases, or the ratio $T/U$ grows. In the large-volume regime $T \\gg 1$, almost all of the fundamental region is excluded and the allowed region concentrates near duality fixed points, with $\\tau = i$ more strongly constrained than $\\tau = \\omega$; hence the paper concludes that $\\tau \\simeq i\\infty$ lies in the Swampland and $\\tau = i$ is disfavored.","pith_inferences":["Editorial inference: If the bounds hold, they give a top-down selection rule for modular flavor model building: among the fixed points $\\tau = i\\infty$, $i$, and $\\omega$, only $\\omega$ survives the large-volume limit, so models built near $\\omega$ become the prime string-embedding candidates.","Editorial inference: The same perturbativity argument could be turned into a scan over heterotic orbifold gauge groups, since for each model's $b_a^{N=2}$ and dilaton value the Lambert-W bound gives the full allowed region in $(T,U)$, which could be combined with moduli stabilization to test specific vacua.","Editorial inference: A direct test would be to compute the exact one-loop threshold correction, not just the leading $N=2$ beta-function term, for a concrete orbifold near $\\operatorname{Im} U \\approx 34$; if the exact $\\Delta_a$ stays below the tree-level coupling, the numerical cap is an artifact of the approximation."],"forward_implications":["The fixed point $\\tau \\simeq i\\infty$, often used to generate hierarchical fermion masses, is not compatible with perturbative heterotic vacua at $\\operatorname{Im} S \\sim O(1)$ and $|b_a^{N=2}| \\sim O(10)$.","Radiative moduli stabilization scenarios that fix the modulus at $\\operatorname{Im} \\tau \\sim O(10)$ are difficult to realize inside the perturbative regime.","In the large-volume regime $T \\gg 1$, almost all of the moduli space is excluded and the surviving region concentrates near an elliptic point of the duality group, with $\\tau = \\omega$ allowed and $\\tau = i$ excluded for large enough volume.","Increasing $|b_a^{N=2}|$ strengthens the bounds; for $b_a^{N=2} \\gtrsim 90$ with $T = 7U$ and $\\operatorname{Im} S = 2$, no point of the $PSL(2,\\mathbb{Z})$ fundamental region remains perturbatively valid.","The constraints depend on the ratio $T:U$; larger ratios of complex-structure to Kähler modulus give stronger bounds."],"supporting_citations":[{"why":"Supplies the general one-loop threshold-correction formula for heterotic string vacua used in Eq. (3.2).","marker":"[33]"},{"why":"Derives the moduli dependence of string loop corrections, giving the eta-function form used for the $PSL(2,\\mathbb{Z}) \\times PSL(2,\\mathbb{Z})$ threshold correction.","marker":"[34]"},{"why":"Establishes the field-dependent gauge-coupling relation and the role of the dilaton $\\operatorname{Im} S$ in inequality (5.2).","marker":"[35]"},{"why":"Provides the explicit threshold corrections on $\\mathbb{Z}_N$ Coxeter orbifolds and the target-space duality classification into the five groups used throughout.","marker":"[36]"},{"why":"Rewrites the threshold correction over the enlarged region $\\tilde{F}$, giving the eta-function expressions the bounds are built on.","marker":"[62]"},{"why":"Sets the radiative moduli stabilization scenario at $\\operatorname{Im} \\tau \\sim O(10)$ that the derived constraints rule out in general.","marker":"[46]"},{"why":"Shows universal predictions of modular flavor models near the self-dual point $\\tau = i$, the point the paper finds disfavored at large volume.","marker":"[74]"},{"why":"Analyzes fermion mass hierarchies around $\\tau = i$, a phenomenological target that the stringy constraints restrict.","marker":"[75]"}],"fun_headline_variants":["Heterotic thresholds send large flavor moduli to Swampland","String theory limits modular flavor modulus to ~34","Perturbativity rules out tau=i infinity in modular flavor models","Heterotic strings exclude infinite tau and disfavor tau=i"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole derivation assumes that a heterotic vacuum is perturbatively valid exactly when the tree-level gauge coupling $16\\pi^2 k_a \\operatorname{Im} S$ exceeds the absolute value of the one-loop threshold correction $\\Delta_a$, with $\\Delta_a$ taken from the $N=2$ $\\beta$-function sector only; if Green-Schwarz terms, other threshold sectors, or non-perturbative effects can cancel a large correction, the upper bounds do not follow.","fun_headline_variants_meta":{"raw":{"variants":["Heterotic thresholds send large flavor moduli to Swampland","String theory limits modular flavor modulus to ~34","Perturbativity rules out tau=i infinity in modular flavor models","Heterotic strings exclude infinite tau and disfavor tau=i"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000339,"raw_usage":{"total_tokens":1895,"prompt_tokens":989,"completion_tokens":906,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":605,"completion_tokens_details":{"reasoning_tokens":835}},"tokens_in":605,"tokens_out":906,"duration_ms":10046,"temperature":1.0,"reasoning_tokens":835,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T17:23:29.004304+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the full one-loop threshold correction $\\Delta_a$ from Eq. (3.3), including all twisted sectors and Green-Schwarz contributions, for a concrete $T^6/\\mathbb{Z}_4$ or $T^6/\\mathbb{Z}_{6-II}$ vacuum with $\\operatorname{Im} S = 2$, $k_a = 1$, $|b_a^{N=2}| = 10$, $T = \\omega$, and $\\operatorname{Im} U = 40$. If the resulting corrected gauge coupling $16\\pi^2 k_a \\operatorname{Im} S - |\\Delta_a|$ remains positive, then inequality (5.2) is not necessary and the bound $\\operatorname{Im} U \\lesssim 34.3$ is evaded.","supporting_citations":[{"cited_title":"Kaplunovsky,One-loop threshold effects in string unification, 1992","cited_arxiv_id":null,"evidence_quote":"Supplies the general one-loop threshold-correction formula for heterotic string vacua used in Eq. (3.2)."},{"cited_title":"Dixon, V","cited_arxiv_id":null,"evidence_quote":"Derives the moduli dependence of string loop corrections, giving the eta-function form used for the $PSL(2,\\mathbb{Z}) \\times PSL(2,\\mathbb{Z})$ threshold correction."},{"cited_title":"Kaplunovsky and J","cited_arxiv_id":null,"evidence_quote":"Establishes the field-dependent gauge-coupling relation and the role of the dilaton $\\operatorname{Im} S$ in inequality (5.2)."},{"cited_title":"BAILIN, A","cited_arxiv_id":null,"evidence_quote":"Provides the explicit threshold corrections on $\\mathbb{Z}_N$ Coxeter orbifolds and the target-space duality classification into the five groups used throughout."},{"cited_title":"Feruglio, V","cited_arxiv_id":null,"evidence_quote":"Analyzes fermion mass hierarchies around $\\tau = i$, a phenomenological target that the stringy constraints restrict."}],"review_version":2}