Pith. sign in

REVIEW 3 major objections 4 minor 28 references

A new open-source pipeline, SET-ANUBIS, aims to make ANUBIS long-lived-particle sensitivity studies fully auditable and model-agnostic.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-01 01:17 UTC pith:MQHWHKVK

load-bearing objection Solid software paper: the reproducibility architecture is real and the engineering is clean, but the physics validation is internal by design, so the acceptance numbers need an external benchmark before they should be trusted for published limits. the 3 major comments →

arxiv 2607.26112 v1 pith:MQHWHKVK submitted 2026-07-28 hep-ph hep-ex

SET-ANUBIS: a modular pipeline for ANUBIS long-lived particle sensitivity studies

Sofie Nordahl Erner , Anna Mullin , Th\'eo Reymermier , Paul Swallow , Oleg Brandt for the ANUBIS Collaboration This is my paper
classification hep-ph hep-ex
keywords long-lived particlesANUBIS detectorsensitivity studiesUFO modelsPythia8MadGraphreproducibilitycutflow
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

SET-ANUBIS is an open-source Python framework that claims to take a user's model of new long-lived particles—specified in a standard UFO format—and produce, with minimal extra code, the full chain of numbers needed for a sensitivity study of the proposed ANUBIS detector: decay widths and lifetimes, generated events, geometric acceptance in the ATLAS cavern, and a fixed sequence of selection cuts. The framework's central design choice is a hexagonal architecture that keeps physics logic separate from external generators and storage, plus a content-addressed repository that records every card, seed, and event bundle. Release v1.0.0 adds five reproducibility scenarios (R1–R5) whose versioned reference outputs let anyone verify that the pipeline behaves deterministically. The paper's claim is that this turns ANUBIS sensitivity studies into an auditable, comparable, and repeatable exercise rather than an ad hoc collection of scripts.

Core claim

The paper's central claim is that SET-ANUBIS v1.0.0 establishes a deterministic, auditable software baseline for ANUBIS sensitivity studies: starting from a UFO model or user-supplied rates, it evaluates decay widths, branching ratios and lifetimes; prepares and ingests events from Pythia8 or MadGraph; reconstructs the ATLAS cavern and ANUBIS RPC geometry; and applies a named cutflow (LLP decay, in-cavern, not-in-ATLAS, geometry, tracker, missing transverse momentum, isolation) to determine which events would be observed. The framework validates its geometry predicates by Monte Carlo volume reconstruction and track-intersection tests, and its R5 reproducibility scenario exercises the full se

What carries the argument

The load-bearing object is the hexagonal (ports-and-adapters) architecture that partitions the workflow into bounded domains—model, branching-ratio, MadGraph/Pythia interfaces, geometry, selection, database—with narrow APIs and replaceable adapters. The selection engine applies configurable geometry predicates ('is_in' functions for the ATLAS cavern, shafts, and ANUBIS tracking layers) and kinematic/isolation cuts, while the content-addressed store (SHA-256) and SQLite catalogue preserve provenance so a campaign can be re-materialised from stored cards, seeds, and metadata. This combination is what lets the framework claim both model-agnosticism and reproducibility.

Load-bearing premise

The load-bearing premise is that the framework's internal consistency—geometry volume reproduction and deterministic reference outputs on a tiny hand-made sample—is sufficient to trust the physics reliability of the resulting sensitivity limits, since the cutflow thresholds and background model are not validated against real collision data or a realistic full Monte Carlo background sample.

What would settle it

Generate a realistic background Monte Carlo sample (e.g., Pythia8 QCD dijet events with full showering) and pass the reconstructed events through the SET-ANUBIS selection chain, counting how many pass the E_T^miss > 30 GeV, geometric acceptance, and isolation requirements. If the resulting background rate is non-negligible (comparable to or above the N_LLP ≥ 4 threshold assumed for a background-free search) or deviates significantly from the expected near-zero rate, the framework's selection model is wrong.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • If the framework's correctness holds, future ANUBIS sensitivity limits can be produced by independent groups and checked against the R1–R5 reference outputs.
  • Different BSM models can be compared using the same geometry, cutflow and acceptance definitions, removing a common source of inconsistency between studies.
  • The lifetime re-weighting and compact selection-ready bundles make large parameter scans substantially cheaper in both CPU time and storage.
  • The R1–R5 reproducibility contract offers a template for how other detector-sensitivity software could package deterministic validation evidence.
  • Because the selection chain is configurable, the same tooling can be repointed to non-ANUBIS far-detector geometries, widening the framework's use.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The paper's validation is mostly about software consistency—e.g., seven hand-crafted events designed to fail at successive stages—not about whether the chosen cutflow (E_T^miss > 30 GeV, isolation cone 0.5, 50% signal efficiency) correctly models real ANUBIS backgrounds; that physics validation remains open work.
  • If the ANUBIS collaboration adopts SET-ANUBIS as its default sensitivity tool, as the paper indicates, the framework's approximations for hadronic decays and simplified geometry will directly shape all published limits, raising the stakes for a dedicated background-validation campaign.
  • A concrete testable extension would be to run the full chain on realistic background samples (e.g., QCD dijets producing K_L/n punch-through) and compare the efficiency and background rejection against the current parameter choices.
  • The architecture is generic enough that the same pipeline could be adapted to other proposed displaced-decay detectors, provided their geometry is encoded in the same predicate language.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The paper describes SET-ANUBIS, an open-source Python framework for end-to-end long-lived-particle sensitivity studies for the proposed ANUBIS detector. It connects UFO model handling, decay-width and branching-ratio calculation, Pythia/MadGraph event-generation preparation, HepMC ingestion, ATLAS-cavern and ANUBIS-RPC geometry, a named selection cutflow, and SQLite/content-addressed provenance storage. Release v1.0.0 introduces five reproducibility scenarios R1–R5, machine-readable JSON and HTML selection traces, optional Dash inspection tools, and 247 automated tests with 67.10% line coverage. The paper's central claim is that the framework provides an auditable, reproducible path from BSM model parameters to ANUBIS acceptance and sensitivity projections with minimal bespoke user code.

Significance. If the geometry and selection implementation are physically correct, SET-ANUBIS would be a valuable community tool: it is open source, has a clear ports-and-adapters architecture, ships deterministic card generation, versioned reproducibility scenarios, content-addressed storage, and machine-readable selection traces. The reproducibility contract R1–R5 is a genuine strength and makes the software's internal consistency machine-checked. However, the paper currently validates only internal consistency: the geometry test compares predicates with volumes derived from the same geometry definition, and R5 is explicitly a seven-event regression fixture rather than a physical signal sample. No external benchmark against a published ANUBIS acceptance, a full realistic Monte Carlo sample, or an independent selection implementation is provided. The central physics output, the acceptance A in Eq. (1), is therefore not yet anchored to an external authority. This is a load-bearing gap for a paper whose stated purpose is to support auditable sensitivity studies.

major comments (3)
  1. [§5.4] The Monte-Carlo volume-reconstruction test compares the geometry predicates against 'reference volumes computed directly from the geometry definition.' This is an internal-consistency check of coordinate transforms and region classification, not a validation of the geometry parameters or of the resulting acceptance. A systematic error in the cavern dimensions, ceiling curvature, IP offset, or RPC placement would pass both sides of the test. To support the claim that the geometry acceptance is 'correctly implemented and validated,' please add an external anchor: reproduce the ANUBIS acceptance or sensitivity contour of Ref. [2] for a benchmark model, or compare the implemented geometry against an independent CAD/analytic geometry model.
  2. [§6.2 and §4.5–4.6] Scenario R5 uses a seven-event HepMC2 sample that the manuscript itself states 'is not intended to represent a physical signal distribution'; the expected cutflow counts are versioned references produced by the same pipeline. This validates cutflow bookkeeping, determinism, and trace generation, but it does not establish that the implemented selection (E_T^miss > 30 GeV, ΔR = 0.5 isolation, d_IP < 10 mm, tracker-intersection logic) yields physically correct acceptances. No comparison is made against a full realistic Monte Carlo sample, an independent implementation of the ANUBIS selection, or the published sensitivity study [22] that used the framework. Because A in Eq. (1) is the central physics output, please add a closure test: generate a realistic LLP sample (or use a public sample) and compare the resulting acceptance/cutflow/N_LLP against Ref. [2] or an independent calculation. Not
  3. [Abstract, §3.1, §8.1] The paper describes the framework as supporting 'an arbitrary BSM model' with 'minimal additional bespoke code,' but Section 8.1 states that hadronic decay channels are not implemented and rely on user-supplied calculators or tabulated rates, and that MARTY coverage of hadronic matrix elements is not guaranteed. For models whose dominant decays are hadronic, the automated width/BR and lifetime chain is therefore incomplete. Since the framework's stated goal is direct comparison of alternative LLP models, please qualify the scope in the abstract and introduction and state explicitly which model classes are supported without additional user input.
minor comments (4)
  1. [§3.1, §4.6, §6.2] The manuscript cites HepMC3 [15] and lists HepMC3 as an external dependency, but R5 and Section 6.2 use a 'HepMC2' sample. Please clarify which HepMC versions are supported for ingestion and which are covered by the validation scenarios.
  2. [§7.2, Eq. (14)] The indicator I_HepMC in Eq. (14) is not defined in the text. Please define it explicitly, e.g. as 1 when raw HepMC retention is requested and 0 otherwise.
  3. [§5.2] In the selection example, the code prints cutflow["nLLP_Final"], while the text's stage list uses 'Final' and earlier mentions 'nLLP_original'. Please ensure the key names in the printed example match the documented JSON output exactly.
  4. [References [2], [22]] The paper relies on Ref. [2] for the background model and selection thresholds and on Ref. [22] as an application of the framework, but both are preprint citations dated 2025/2026. Please confirm that they are publicly available and include the exact arXiv identifiers in the bibliography.

Circularity Check

0 steps flagged

No significant circularity: the paper is a software/reproducibility report, and its validation checks are internal-consistency tests rather than physics predictions derived from their own inputs.

full rationale

SET-ANUBIS is a software framework paper, not a physics-derivation paper. Its central deliverables are the pipeline architecture, the R1–R5 reproducibility scenarios, and the versioned reference outputs. The R5 cutflow fixture is explicitly described as 'not intended to represent a physical signal distribution' and as 'a transparent regression fixture for the interpretation and ordering of the production cuts'; the expected cutflow counts are constructed by hand, not fitted from data and then renamed as a prediction. The geometry validation in Sec. 5.4 compares Monte Carlo volume estimates with 'reference volumes computed directly from the geometry definition'; this is a check that the predicates are implemented consistently with the parameterized geometry, i.e. a software-consistency test, not an independent physics prediction that reduces to its own input. The self-citation of the ANUBIS HNL sensitivity study [22] is presented as a use of the framework, not as evidence for the correctness of its physics outputs. The selection thresholds (E_T^miss > 30 GeV, isolation cone, 50% efficiency) are taken from the separate ANUBIS sensitivity study [2] and are not derived in this paper. The paper also flags its own limitation in Sec. 8.1 that hadronic decays and form factors remain a limiting factor. None of the load-bearing claims—deterministic card generation, reproducible cutflow execution, geometry-predicate consistency, provenance tracking—are obtained by defining a variable in terms of the result, by fitting a parameter and calling it a prediction, or by invoking a self-citation as the sole justification. A weaker external physics benchmark would strengthen the physics validation, but that is a correctness/validation gap, not circularity under the specified criteria.

Axiom & Free-Parameter Ledger

3 free parameters · 5 axioms · 1 invented entities

The central claim is a software framework, so most assumptions are domain assumptions about the validity of the physics inputs and the adequacy of the simplified detector model. The free parameters are selection thresholds and geometry choices inherited from prior ANUBIS studies. There are no invented entities. The QCD running formulae are standard background.

free parameters (3)
  • Selection thresholds (E_T^miss > 30 GeV, Delta R = 0.5, pT > 10 GeV, d_IP < 10 mm, 50% signal efficiency, N_LLP >= 4 or = E_T^miss > 30 GeV; DeltaR = 0.5; pT > 10 GeV; d_IP < 10 mm; epsilon_sig = 50%
    These constants define the central sensitivity calculation (Eq. 1 and Section 3.2). They are not derived in the paper; they are stated as nominal values inherited from Ref. [2] or user configurations. They are not fitted to data in this paper, but they are input parameters from prior ANUBIS analyses.
  • Geometry configuration (2 tracking layers, 1 m separation, first layer 20 cm below ceiling, RPC efficiency 100%) = 2 layers, 1 m separation, 20 cm clearance, 100% efficiency
    The ANUBIS geometry predicates and the acceptance calculation depend on these choices; they are configurable but the nominal values are used in the validation and cutflow. They are not fitted here, but they are free choices that affect the claimed acceptance.
  • Luminosity L = 3 ab^-1 = 3 ab^-1
    Used in Eq. (1) for the sensitivity projection; taken from the HL-LHC assumption. Standard input, not fitted.
axioms (5)
  • domain assumption HepMC event records faithfully represent generator output and the dataframe conversion preserves the information needed for the selection.
    The entire selection chain depends on the HepMC-to-dataframe conversion (Section 3.2 and 4.2). The paper does not validate the conversion against known results.
  • domain assumption The simplified cavern/RPC geometry (cylindrical shells, vaulted ceiling, shafts) is an adequate model of the real ATLAS cavern and ANUBIS stations for acceptance purposes.
    The Geometry section builds a simplified model; the Monte Carlo validation shows the code matches its own model, not the real cavern. This is acknowledged by the authors in the limitations (Section 8.1 only mentions hadronic decays, but the geometry simplification is inherent).
  • domain assumption The selection thresholds and the signal-efficiency factor epsilon_sig = 50% correctly represent ANUBIS background rejection and detector efficiency.
    These are taken from Ref. [2] and stated in Section 3.2-3.3. The paper does not derive them or validate them with data; they are assumptions inherited from an earlier ANUBIS study.
  • domain assumption External generators (Pythia8, MadGraph 3.5.8, MARTY) and UFO models correctly compute the physics input (cross sections, branching ratios, lifetimes).
    The framework delegates physics computation to external tools. The R3/R4 scenarios intentionally do not run them. The paper states this in Section 4.6 and Appendix A.3.
  • standard math The QCDRunner's three-loop alpha_s and mass running formulae are correct implementations of the standard MS-bar beta functions.
    Appendix B presents the standard NNLO QCD running expressions. These are unproved background results from the literature, but standard and presumably correct.
invented entities (1)
  • No new physical entities are invented no independent evidence
    purpose: The paper does not introduce new particles, forces, or dimensions; it simulates known BSM models such as HNLs.
    No new physical entities are proposed.

pith-pipeline@v1.3.0-alltime-deepseek · 22075 in / 8562 out tokens · 62421 ms · 2026-08-01T01:17:44.578486+00:00 · methodology

0 comments
read the original abstract

SET-ANUBIS (Simulation, accEptance and sensiTivity studies framework for ANUBIS) is an open-source Python framework for end-to-end studies of long-lived particle signatures in the proposed ANUBIS detector. Starting from a Universal FeynRules Output model or user-supplied rates, it exposes model parameters and particle content, evaluates decay widths, branching ratios and lifetimes, prepares or runs event generation through Pythia8 or MadGraph, ingests HepMC event records, models the ATLAS cavern and ANUBIS tracking stations, and applies a configurable sequence of geometric, kinematic and isolation requirements. The implementation follows a ports-and-adapters architecture so that physics-domain logic remains separate from external generators, persistent storage and visualisation. A scan-aware SQLite catalogue and content-addressed store preserve cards, metadata and compact selection-ready event bundles while avoiding duplicate artefacts. Release v1.0.0 introduces a deterministic reproducibility suite with five independent scenarios (R1--R5) covering model access, branching-ratio interpolation, Pythia command generation, MadGraph card generation and the complete selection chain on a compact HepMC2 sample. The final scenario produces machine-readable JSON and a standalone HTML selection trace that records cumulative cutflows and event-level pass/fail transitions. Optional scientific dashboards provide event, geometry, campaign and provenance inspection without changing the physics results. The framework, documentation and reference inputs are distributed under the GNU GPL version 3 or later and are intended to support auditable ANUBIS sensitivity studies and direct comparison of alternative long-lived particle models.

Figures

Figures reproduced from arXiv: 2607.26112 by Anna Mullin, Oleg Brandt for the ANUBIS Collaboration, Paul Swallow, Sofie Nordahl Erner, Th\'eo Reymermier.

Figure 1
Figure 1. Figure 1: (a) The layout of the underground cavern at LHC Point 1, UX15, featuring the ATLAS experiment and the PX14 and PX16 service shafts above it. The proposed configuration of the ANUBIS detector is shown in orange. This includes two circular stations at the bottom of the two access shafts. (b) Layout of the RPC tracking layers for the ANUBIS detector, with two layers of triplets separated by ∼ 1 m and an optio… view at source ↗
Figure 2
Figure 2. Figure 2: Global architecture of SET-ANUBIS. Domain kernels are in the centre and adapters connect to external tools such as Pythia8, MadGraph, MARTY, the event database and geometry representation. Adapters: Pythia, MadGraph, Database, Geometry, Selection, . . . Domain: Ports, Use cases, Entities Common: logging, decorators, MultiSet, utilities [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Place of the Common layer in the General architecture outside the hexagonal boundaries, used by all other layers. Adapters (input): CardGetter, DecayProvider, FileCopyBuilder JSONExtractor, ParamCardGeneratorAdapter, ParamProvider ParticleProvider, UFOInterface, UFOParser, YAMLParser Domain: EventDataBaseManager, UFOManager, CardGenerator Objects: UFO tree Input port → MadGraph (ICardGetter, ICardGetter) I… view at source ↗
Figure 4
Figure 4. Figure 4: Architecture of the DataBase layer. 4.3. ModelCore This domain offers the user-facing entry point to model parameters and particles through SetAnubisInter￾face. Internally, it builds and partially evaluates a pa￾rameter tree (using a proxy to the DataBase part), pre￾serves LHA formatted metadata, normalises particle in￾formation to a common format and can enrich it from a JSON file when needed during the s… view at source ↗
Figure 8
Figure 8. Figure 8: Architecture of the Pythia layer. In the .cmnd file, all of the sections presented in Fig￾ure 9 can be modified using the interface. However, we strongly advise the user not to modify general parame￾ters and only adjust those associated with the introduced LLP model, unless they have a strong understanding of the underlying Pythia generator. The interface can add all the decays of a new particle (and its p… view at source ↗
Figure 9
Figure 9. Figure 9: Decomposition of the .cmnd file within the Pythia Inter￾face. The binding with the Pythia8 software is achieved using pybind11, allowing control within a Python envi￾ronment and also generator level cuts. 4.7. Geometry The Geometry section builds a representation of the ATLAS cavern and ANUBIS RPC stations, including 9 [PITH_FULL_IMAGE:figures/full_fig_p009_9.png] view at source ↗
Figure 11
Figure 11. Figure 11: Architecture of the Selection layer. The HepMC selection explorer starts from the com￾pact seven-event HepMC2 reference sample used by scenario R5 when no user file is supplied. Its interface follows the canonical selection sequence—Original, LLPDecay, InCavern, NotInATLAS, Geometry, Tracker, MET, IsoJets, IsoCharged, IsoAll and Final—and reports the cumulative cutflow together with the last passed and fi… view at source ↗
Figure 10
Figure 10. Figure 10: Architecture of the Geometry layer. 4.8. Selection Applies a fixed-order set of selections to recon￾structed event bundles. This part’s purpose is to check whether the LLPs have decayed; the geometric acceptance of the LLP vertex, i.e. whether it is within the ATLAS cavern, and out￾side the bounds of ATLAS; whether the LLP tracks would intersect with ANUBIS RPC layers and thus be recorded; the charged-tra… view at source ↗
Figure 12
Figure 12. Figure 12: Example of the Decay Width calculation of LLP (HNL) using MARTY frameworks for three-body decays. Here the HNL has a mass of 10 GeV and is decaying into one neutrino (νµ) and a pair of leptons. could have J µ V = u¯γ µu − d¯γ µd related to the π +π − cur￾rent, which gives us: ⟨π + (p1)π − (p2)| J µ V |0⟩ = (p1 − p2) µF ππ V (s) (8) where F ππ V is a vectorial form factor which can be taken from experiment… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

28 extracted references · 2 canonical work pages

  1. [1]

    Bauer, O

    M. Bauer, O. Brandt, L. Lee, C. Ohm, ANUBIS: Proposal to search for long-lived neutral particles in CERN service shafts (9 2019).arXiv:1909.1 3022

    2019

  2. [2]

    Brandt, et al., The ANUBIS detector and its sensitivity to neutral long-lived particles (10 2025).arXiv:2510.26932

    O. Brandt, et al., The ANUBIS detector and its sensitivity to neutral long-lived particles (10 2025).arXiv:2510.26932

    arXiv 2025

  3. [3]

    Sjöstrand, S

    T. Sjöstrand, S. Ask, J. R. Christiansen, R. Corke, N. Desai, P. Ilten, S. Mrenna, S. Prestel, C. O. Ras- mussen, P. Z. Skands, An introduction to PYTHIA 8.2, Comput. Phys. Commun. 191 (2015) 159– 177.arXiv:1410.3012,doi:10.1016/j.cp c.2015.01.024

    Pith/arXiv arXiv 2015

  4. [4]

    Alwall, R

    J. Alwall, R. Frederix, S. Frixione, V . Hirschi, F. Maltoni, O. Mattelaer, H. S. Shao, T. Stelzer, P. Torrielli, M. Zaro, The automated computation of tree-level and next-to-leading order differen- tial cross sections, and their matching to parton shower simulations, JHEP 07 (2014) 079.arXiv: 1405.0301,doi:10.1007/JHEP07(2014)079

    Pith/arXiv arXiv 2014

  5. [5]

    Uhlrich, F

    G. Uhlrich, F. Mahmoudi, A. Arbey, MARTY , a new C++framework for automated symbolic cal- culations in Beyond the Standard Model physics, PoS ICHEP2020 (2021) 928.arXiv:2011.065 58,doi:10.22323/1.390.0928

    work page doi:10.22323/1.390.0928 2021

  6. [6]

    Alimena, et al., Searching for long-lived par- ticles beyond the Standard Model at the Large Hadron Collider, J

    J. Alimena, et al., Searching for long-lived par- ticles beyond the Standard Model at the Large Hadron Collider, J. Phys. G 47 (9) (2020) 090501. arXiv:1903.04497,doi:10.1088/1361-647 1/ab4574

    arXiv 2020

  7. [7]

    Aaboud, et al., Search for long-lived particles produced inppcollisions at √s=13 TeV that decay into displaced hadronic jets in the ATLAS muon spectrometer, Phys

    M. Aaboud, et al., Search for long-lived particles produced inppcollisions at √s=13 TeV that decay into displaced hadronic jets in the ATLAS muon spectrometer, Phys. Rev. D 99 (5) (2019) 052005.arXiv:1811.07370,doi:10.1103/Ph ysRevD.99.052005

    Pith/arXiv arXiv 2019

  8. [8]

    Khachatryan, et al., Search for long-lived par- ticles that decay into final states containing two electrons or two muons in proton-proton colli- sions at √s=8 TeV, Phys

    V . Khachatryan, et al., Search for long-lived par- ticles that decay into final states containing two electrons or two muons in proton-proton colli- sions at √s=8 TeV, Phys. Rev. D 91 (5) (2015) 052012.arXiv:1411.6977,doi:10.1103/Ph ysRevD.91.052012

    Pith/arXiv arXiv 2015

  9. [9]

    A. Tumasyan, et al., Search for long-lived par- ticles decaying into muon pairs in proton-proton collisions at √s=13 TeV collected with a dedi- cated high-rate data stream, JHEP 04 (2022) 062. arXiv:2112.13769,doi:10.1007/JHEP04(20 22)062

    Pith/arXiv arXiv 2022

  10. [10]

    V . V . Gligorov, S. Knapen, M. Papucci, D. J. Robinson, Searching for Long-lived Particles: A Compact Detector for Exotics at LHCb, Phys. Rev. D 97 (1) (2018) 015023.arXiv:1708.09395, doi:10.1103/PhysRevD.97.015023

    Pith/arXiv arXiv 2018

  11. [11]

    Curtin, et al., Long-Lived Particles at the En- ergy Frontier: The MATHUSLA Physics Case, Rept

    D. Curtin, et al., Long-Lived Particles at the En- ergy Frontier: The MATHUSLA Physics Case, Rept. Prog. Phys. 82 (11) (2019) 116201.arXiv: 1806.07396,doi:10.1088/1361-6633/ab28 d6

    Pith/arXiv arXiv 2019

  12. [12]

    Ahdida, et al., The SHiP experiment at the pro- posed CERN SPS Beam Dump Facility, Eur

    C. Ahdida, et al., The SHiP experiment at the pro- posed CERN SPS Beam Dump Facility, Eur. Phys. J. C 82 (5) (2022) 486.arXiv:2112.01487, doi:10.1140/epjc/s10052-022-10346-5

    Pith/arXiv arXiv 2022

  13. [13]

    Darmé, et al., UFO 2.0: the ‘Universal Feyn- man Output’ format, Eur

    L. Darmé, et al., UFO 2.0: the ‘Universal Feyn- man Output’ format, Eur. Phys. J. C 83 (7) (2023) 631.arXiv:2304.09883,doi:10.1140/epjc /s10052-023-11780-9

    Pith/arXiv arXiv 2023

  14. [14]

    Degrande, C

    C. Degrande, C. Duhr, B. Fuks, D. Grellscheid, O. Mattelaer, T. Reiter, UFO - The Universal FeynRules Output, Comput. Phys. Commun. 183 (2012) 1201–1214.arXiv:1108.2040,doi: 10.1016/j.cpc.2012.01.022

    Pith/arXiv arXiv 2012

  15. [15]

    Buckley, P

    A. Buckley, P. Ilten, D. Konstantinov, L. Lönnblad, J. Monk, W. Pokorski, T. Przedzin- ski, A. Verbytskyi, The HepMC3 event record library for Monte Carlo event gen- erators, Comput. Phys. Commun. 260 (2021) 107310.arXiv:1912.08005, doi:10.1016/j.cpc.2020.107310

    Pith/arXiv arXiv 2021

  16. [16]

    Bierlich, et al., A comprehensive guide to the physics and usage of PYTHIA 8.3, SciPost Phys

    C. Bierlich, et al., A comprehensive guide to the physics and usage of PYTHIA 8.3, SciPost Phys. Codeb. 2022 (2022) 8.arXiv:2203.11601,doi: 10.21468/SciPostPhysCodeb.8

    Pith/arXiv arXiv 2022

  17. [17]

    Alloul, N

    A. Alloul, N. D. Christensen, C. Degrande, C. Duhr, B. Fuks, FeynRules 2.0 - A complete toolbox for tree-level phenomenology, Comput. Phys. Commun. 185 (2014) 2250–2300.arXiv: 1310.1921,doi:10.1016/j.cpc.2014.04.01 2

    Pith/arXiv arXiv 2014

  18. [18]

    HTCondor Team, HTCondor, Zenodo (2024).do i:10.5281/zenodo.11397217. 18

    work page doi:10.5281/zenodo.11397217 2024

  19. [19]

    Cacciari, G

    M. Cacciari, G. P. Salam, G. Soyez, The anti-kt jet clustering algorithm, JHEP 04 (2008) 063.arXi v:0802.1189,doi:10.1088/1126-6708/2008 /04/063

    Pith/arXiv arXiv 2008

  20. [20]

    Navas, et al., Review of particle physics, Phys

    S. Navas, et al., Review of particle physics, Phys. Rev. D 110 (3) (2024) 030001.doi:10.1103/Ph ysRevD.110.030001

    doi:10.1103/ph 2024

  21. [21]

    Meurer, et al., Sympy: symbolic computing in python, PeerJ Computer Science 3 (2017) e103

    A. Meurer, et al., Sympy: symbolic computing in python, PeerJ Computer Science 3 (2017) e103

    2017

  22. [22]

    ANUBIS Collaboration, Projected sensitivity of the ANUBIS detector to heavy neutral leptons (6 2026).arXiv:2606.26862

    Pith/arXiv arXiv 2026

  23. [23]

    UKRI under the Future Leaders Fellowship scheme (Grant number G107408)

    T. Reymermier, P. Swallow, SET-ANUBIS: a modular pipeline for ANUBIS long-lived particle sensitivity studies, Software release v1.0.0 (2026). URLhttps://github.com/SET-ANUBIS/se t-anubis CRediT authorship contribution statement Anna Mullin:Methodology, Software (first imple- mentation of the event-selection component).ANU- BIS Collaboration:provision of c...

    2026

  24. [24]

    FixΛ 5 from the inputα s(MZ) by inverting Equa- tion (2)

  25. [25]

    Determine the ordered thresholds (mu,m d,m s,m c,m (type) b ,m (type) t )

  26. [26]

    To evaluateα s(Q): step the flavour number from 5 ton f (Q) using Equation (B.1), then use Equa- tion (2) atn f (Q)

  27. [27]

    To run a quark massm(Q i)→m(Q f ): decom- pose the path into constant-n f segments and apply Equation (B.3) on each

  28. [28]

    Practical settings and caveats •Inputs:α s(MZ),M Z,m pole t , ¯mb, and ¯mu,d,s,c (all in GeV)

    Convertm b orm t when needed using Equa- tions (B.4) and (B.5). Practical settings and caveats •Inputs:α s(MZ),M Z,m pole t , ¯mb, and ¯mu,d,s,c (all in GeV). •Threshold choice forb,t(pole or running) is con- figurable and only affects wheren f changes. •Matching usescontinuityofα s atQ=m h with- out higher–order decoupling constants; this is ade- quate f...