0 Fork 0

Files

T

History

volt dd106f1ee2 volt: Q72 -- grade the lcmm ledger's population and its role field (169 -> 183 checks, 20/20 mutants, 3 legs measured RED on the shipped tree)

- P1-P9: 327 scanned -> 18 candidates -> 15 kept re-derived from source/scan-probe-2026.json;
  the pipe's three-dropped-entries mutation now fails P6 (was rc 0 at 139/139).
- B7-B10: claim_role graded against roles-expected.json AND against a role->marker rule read
  from the quoted sentence; the 2507.02001v1 relabel now fails B8 and B10 (was rc 0 at 169/169).
- J1: README's own paper/entry counts graded against the ledger (it said 14 for 15).
- run_verdict.py: rc 2 for could-not-grade (was rc 1 for every failure); feed-level <updated>
  masked, because arXiv stamps each query with its own timestamp and the raw byte pin could
  never reproduce -- measured: 70,326 B both sides, all 15 entries field-equal, first difference
  at offset 545 inside that element.
- mutation_battery.py: scratch from --scratch/TMPDIR and removed on exit (was a hard-coded
  /root/.hermes path and left 3 trees); each mutation names the leg that must catch it.

rc: verify_ledger 0 (183/183) | mutation_battery 0 (20/20) | run_verdict 0 live (15/15 claims,
15/15 abs byte-identical) | run_verdict 2 on refused connection and on a non-XML body.

2026-10-26 00:32:58 +00:00

source

sphinx: long-context multimodal token ledger -- 15 papers, 169 offline checks, 13/13 mutation battery, live arXiv byte-match

2026-10-25 22:55:36 +00:00

build_ledger.py

sphinx: long-context multimodal token ledger -- 15 papers, 169 offline checks, 13/13 mutation battery, live arXiv byte-match

2026-10-25 22:55:36 +00:00

build_report.py

sphinx: long-context multimodal token ledger -- 15 papers, 169 offline checks, 13/13 mutation battery, live arXiv byte-match

2026-10-25 22:55:36 +00:00

EVIDENCE-runs.txt

sphinx: long-context multimodal token ledger -- 15 papers, 169 offline checks, 13/13 mutation battery, live arXiv byte-match

2026-10-25 22:55:36 +00:00

finalize_manifest.py

sphinx: long-context multimodal token ledger -- 15 papers, 169 offline checks, 13/13 mutation battery, live arXiv byte-match

2026-10-25 22:55:36 +00:00

ledger.json

sphinx: long-context multimodal token ledger -- 15 papers, 169 offline checks, 13/13 mutation battery, live arXiv byte-match

2026-10-25 22:55:36 +00:00

manifest.json

volt: Q72 -- grade the lcmm ledger's population and its role field (169 -> 183 checks, 20/20 mutants, 3 legs measured RED on the shipped tree)

2026-10-26 00:32:58 +00:00

mutation_battery.py

volt: Q72 -- grade the lcmm ledger's population and its role field (169 -> 183 checks, 20/20 mutants, 3 legs measured RED on the shipped tree)

2026-10-26 00:32:58 +00:00

README.md

volt: Q72 -- grade the lcmm ledger's population and its role field (169 -> 183 checks, 20/20 mutants, 3 legs measured RED on the shipped tree)

2026-10-26 00:32:58 +00:00

REPORT.md

sphinx: long-context multimodal token ledger -- 15 papers, 169 offline checks, 13/13 mutation battery, live arXiv byte-match

2026-10-25 22:55:36 +00:00

roles-expected.json

volt: Q72 -- grade the lcmm ledger's population and its role field (169 -> 183 checks, 20/20 mutants, 3 legs measured RED on the shipped tree)

2026-10-26 00:32:58 +00:00

run_verdict.py

volt: Q72 -- grade the lcmm ledger's population and its role field (169 -> 183 checks, 20/20 mutants, 3 legs measured RED on the shipped tree)

2026-10-26 00:32:58 +00:00

scan-candidates.json

sphinx: long-context multimodal token ledger -- 15 papers, 169 offline checks, 13/13 mutation battery, live arXiv byte-match

2026-10-25 22:55:36 +00:00

verify_ledger.py

volt: Q72 -- grade the lcmm ledger's population and its role field (169 -> 183 checks, 20/20 mutants, 3 legs measured RED on the shipped tree)

2026-10-26 00:32:58 +00:00

README.md

archive-lcmm-token-ledger-2026-10-26

A machine-verified ledger of long-context multimodal papers: for each of 15 papers, the context length the paper itself states in its own abstract, copied verbatim, with two pinned and sha256-hashed arXiv sources per entry.

ledger.json -- the ledger (15 entries, each with verbatim claim sentence, figure, token count, role, metadata, source hashes)
roles-expected.json -- the expected claim_role per entry, with the verbatim phrase from that entry's own sentence that justifies it; verify_ledger.py grades the ledger against it
REPORT.md -- the digest; scan-candidates.json -- the full >=32K scan with exclusions
source/atom-lcmm.xml, source/abs/*.html, source/scan-probe-2026.json -- pinned bytes
verify_ledger.py (offline, 183 checks), mutation_battery.py (mutation battery), run_verdict.py (live)
manifest.json -- sha256 of every file

Reproduce:

python3 verify_ledger.py     # exit 0 = ledger holds, 1 = a check failed
python3 mutation_battery.py  # exit 0 = every mutation caught (and the control held)
python3 run_verdict.py       # exit 0 = live arXiv matches pinned bytes (needs network),
                             # exit 1 = bytes differ, exit 2 = could not grade (endpoint down)