Outputs of our check of ai.viXra 2610.0063
Report: https://cybernative.ai/checks/aiv-2610-0063/
Authors' code: https://github.com/zhangsongbo2026/Tunneling-nanotubes
Commit: 5fd7953d1a204b15cd3c32e91033c81172141cff (the code is in a zip inside the repository)

What is here

- seed-20260907/, seed-20260917/, seed-20260927/
  One folder per seeded rerun of the authors' six figure scripts, run in order
  with the edits below. Each folder holds the six script logs and, in outputs/,
  the data files the scripts wrote. The folder name is the seed of parts 1, 2
  and 2b. Parts 3, 4 and 5 set their own seeds in the authors' code (20260910,
  20260911 and 20260912). Each run adds the same offset (0, 10 or 20) to the
  seed of every script.

- as-shipped-run/
  Logs of one run of the six scripts exactly as shipped, with the authors'
  seeds. Parts 2, 2b and 4 stop with errors.

- edits-seed-20260907.diff, edits-seed-20260917.diff, edits-seed-20260927.diff
  The changes we made to the authors' scripts for each run: the three edits
  named in the report, and the seed line of each script. In the 20260907 run
  the seed line adds 0, so the seeds are the authors' own.

- fit_pc_inf.py and fit_pc_inf-output.txt
  The fit behind the report's fitted p_c(inf) values: p_c(N) is where P_inf
  crosses 0.5, and p_c(N) = p_c(inf) - c * N^(-3/8) is fitted by least squares
  over N = 500, 2000 and 8000. Run "python3 fit_pc_inf.py" in this folder
  (standard library only). It reads the three seed folders.

- LICENSE-upstream-MIT.txt
  The licence of the authors' code.

Before each seeded run we removed two saved result files that the package
ships (data/paper_fss_curves.json and data/paper_et5_stats.json), so that the
scripts simulated afresh. The as-shipped run kept them.

Environment: Python with numpy 1.26.4, as pinned in the package's
code/requirements.txt, in a container with no network. Paths that start with
/work are inside that container.

Not here: the authors' package and the saved result files it ships. Get them
from the repository at the commit above.

Licences: the diffs quote lines of the authors' code, which is under the MIT
License, Copyright (c) 2026 general (LICENSE-upstream-MIT.txt). The data files
in the seed folders were written by the authors' scripts. fit_pc_inf.py is
ours, under the Apache License 2.0. This README and fit_pc_inf-output.txt are
ours, under CC BY 4.0.

SHA256SUMS lists every other file in this folder and its subfolders.

CyberNative AI LLC, hello@cybernative.ai