Skip to content
Protocol STAMP
Latest Calibration

Publications

The CalibrationOur own measurements, pre-registered and signedThe LogWhat ran, what it cost, what movedThe BenchWhat we read and argue with · Mon and ThuArchiveEvery issue, by number and dateThe technical reportThe program in one paper, with its dataSubscribeOne email per issue

Method

MethodologyHow a rate is producedDefinitionsWhat each term means hereThe protocol, v1.25The instruction layer as measuredRun filesEvery draw we can publishfigures.jsonEvery citable number and its fingerprints

AI policy

The DocketRules that bear on how AI is testedThe ChronicleDated events in AI governanceSuper Intelligence (SI)The US rename of AI, tracedMilitaries and AICountry by country

Trust

The Ten House ProtocolsHow we keep AI from making things upIndependenceWho pays, and the rules on itCorrections + changesWhat we got wrong, datedAI Safety + GovernanceWhat our agents may and may not doSupportHow the program is paid forFootprintWhat our tokens cost in power, and who pays itPrivacyWhat we keep, and for how longTermsThe terms of use

About

About Protocol STAMPWhat this is, who makes it, how to check itColophonHow the site is made, and the seven finishesOther STAMPsUnrelated things with the same nameLatest Calibration

Protocol STAMP is teamed with AI and can make mistakes. Verify before you implement anything. What we got wrong · Free to read, paid for by readers: Contribute

NewThe technical report: STAMP, and what it takes to measure one  ·  Contributions are open  ·  The Ten House Protocols: how we keep AI from making things up  ·  Jester Mind, our own aggregated AI model, in the works

Archive

Every issue of The Calibration and The Log, by number and date · newest first
3
issues of The Calibration
Counted from the list below
3
issues of The Log
Counted from the list below
0
issued numbers edited after release
Corrections + changes
The Calibration Vol 004
next, not yet published
The first row still due below

An issued number is never edited. A correction is issued under a new date and the superseded issue stays here. Each entry links its frozen Source — the file whose SHA-256 is printed in the foot of every page of the PDF. Hash that file, not the reading page, which wraps it.

Only issues are listed here. The Bench is not an issue and is not listed: it is the blog, about what other people ran rather than what we ran. It carries volume numbers from Vol 002 so an issue can be cited by name, but a Bench volume is not an issue number and does not appear in this archive. It lives here. What separates the three is set out in Definitions.

The Calibration

The Calibration Vol 004 · Sat20261010 Read · PDF · Source Issued Oct 10, 2026 15:43 from Houston · build 424d7f546ea5
The Calibration Vol 003 · Sat20261003 Read · PDF · Source Issued Oct 03, 2026 13:02 from Houston · 9 pages · are the judges right? · build af3906b9d9ad
The Calibration Vol 002 · Sat20261003 Read · PDF · Source Issued Oct 03, 2026 11:52 from Houston · 12 pages · four findings on whether the judging holds · build ae993179f05f
The Calibration Vol 001 · Sun20260830Read · PDF · SourceIssued Aug 30, 2026 · 32 pages · six findings · build be2bc9d7ca35

The Log

The Log Vol 004 · Sat20261017 Read · PDF · Source Issued Oct 17, 2026 16:17 from Houston · build 18a253173809
The Log Vol 003 · Sat20261003 Read · PDF · Source Issued Oct 03, 2026 13:03 from Houston · released by hand · build 4bda581b1159
The Log Vol 002 · Thu20261001 Read · PDF · Source Issued Oct 01, 2026 19:28 from Houston · fills the missed Sep 19, 2026 slot · build a85771944770
The Log Vol 001 · Sat20260905 Read · PDF · Source Issued Sep 05, 2026 17:31 from Houston · build db2d0f899034

The registry

figures.json — every citable figure with the first sixteen characters of the SHA-256 of the files it was computed from, and the conditions under which it was measured. Of the twenty-six files behind those digests, twenty-two can be downloaded — twenty-one at Run files, with their hashes, and the cross-model data — so a reader can recompute the figures. The other four are the probe bank, whose held-out half is never published, and three working files of figures for The Calibration Vol 002 and Vol 003, not posted yet. See AI Safety + Governance for which files are served as recorded and which are redacted.