Repository navigation

Add Claude Haiku 5.5 and move the references to policyengine-us 2.38.6 (release dashboard-data-20261010, 47 models) - #208

Merged

MaxGhenis merged 78 commits into

mainfrom

release-haiku55

Oct 10, 2026

Merged

Add Claude Haiku 5.5 and move the references to policyengine-us 2.38.6 (release dashboard-data-20261010, 47 models)#208
MaxGhenis merged 78 commits into
mainfrom
release-haiku55

Commits on Oct 9, 2026
- WIP: start the Haiku 5.5 release driver from finish_gpt61sol.py
  
  MaxGhenis
  and
  claude
  committed
  8b9200d
- Pin Claude Haiku 5.5's finished run inputs
  
  MaxGhenis
  and
  claude
  committed
  3495309
- Stage the ten exclusions Max ruled on 2026-10-06 in the Haiku 5.5 driver
  
  MaxGhenis
  and
  claude
  committed
  d0bdb5b
- Name Claude Haiku 5.5 and release 20261006 in the driver's messages
  
  MaxGhenis
  and
  claude
  committed
  bbf5745
- Describe a later-law exclusion's frozen value as the engine's own figure
  
  MaxGhenis
  and
  claude
  committed
  1f97563
- Test the local Claude registration in a fresh interpreter without the remote cost map
  
  MaxGhenis
  and
  claude
  committed
  1944407
- Add the Haiku 5.5 release's freeze driver, tests and design note; fix three driver bugs
  
  MaxGhenis
  and
  claude
  committed
  924a102
- Write the judge provenance record from the stage's sidecars
  
  MaxGhenis
  and
  claude
  committed
  2ccffeb
- Pin the Haiku 5.5 driver to release 20261006 as shipped
  
  MaxGhenis
  and
  claude
  committed
  7ec9c75
- Record the 2026-10-06 wave's decision date and the judge provenance
  
  MaxGhenis
  and
  claude
  committed
  52a0ed2
- Add the reference builder for a second engine upgrade (2.15.17 to a newer policyengine-us)
  
  MaxGhenis
  and
  claude
  committed
  665f3a3
- Let the paper's reference facts follow any number of engine upgrades
  
  MaxGhenis
  and
  claude
  committed
  6074876
- Add --step install-references to the Haiku 5.5 driver
  
  MaxGhenis
  and
  claude
  committed
  9b108a6
- Aim a drifted regeneration at its fix modules' pre-fix value, not a stale record
  
  MaxGhenis
  and
  claude
  committed
  518bab3
- Record the fix modules' pre-fix evidence on policyengine-us 2.37.2
  
  MaxGhenis
  and
  claude
  committed
  4bf7ced
- Add a pre-merge check of a policyengine-us checkout against the fix evidence
  
  MaxGhenis
  and
  claude
  committed
  3613d74
- Decide the outputs an engine upgrade newly excludes
  
  MaxGhenis
  and
  claude
  committed
  5595354
- Point the paper's September upgrade facts at that upgrade
  
  MaxGhenis
  and
  claude
  committed
  6990773
- Name the upstream fixes behind WI 064 and VA 039 state, by bisecting their merges
  
  MaxGhenis
  and
  claude
  committed
  3be304f
- Rule the engine upgrade into the release spec
  
  MaxGhenis
  and
  claude
  committed
  87a2cb9
- Read a regeneration's removed record under the builder's key
  
  MaxGhenis
  and
  claude
  committed
  fa45d4e
- Reuse an earlier build's narrative when its writer inputs are unchanged
  
  MaxGhenis
  and
  claude
  committed
  5588da2
- Name the build's own engine in the drafted Indiana county record
  
  MaxGhenis
  and
  claude
  committed
  174323d
- Decide triage flags from the spec; scope-export the upgrade without the old annotations
  
  MaxGhenis
  and
  claude
  committed
  4c2238d
- Freeze a stage that carries an engine upgrade of the references
  
  MaxGhenis
  and
  claude
  committed
  9d466af
- Answer the machinery review: pin fix-module siblings, bind approvals and tolerances, correct eight narratives
  
  MaxGhenis
  and
  claude
  committed
  fa55444
- Record WI 064's answered flag in the design note, not as a decision
  
  MaxGhenis
  and
  claude
  committed
  78ad5e7
- Stage Claude Haiku 5.5's tool_choice: auto sensitivity run
  
  MaxGhenis
  and
  claude
  committed
  a0b4873
- Draft the Haiku 5.5 release note and paper changes on the 2.37.2 rehearsal
  
  MaxGhenis
  and
  claude
  committed
  39d09fa
- Add the paper's accessors for the 2026-10-06 rulings and the second upgrade's counts
  
  MaxGhenis
  and
  claude
  committed
  0779a3c
- Name each regeneration's upstream fix from a committed map
  
  MaxGhenis
  and
  claude
  committed
  462b27f
- PRACTICE FREEZE on the policyengine-us 2.37.2 rehearsal (replaced at T)
  
  MaxGhenis
  and
  claude
  committed
  41ba29f
- Address the delta review's findings 1-5
  
  MaxGhenis
  and
  claude
  committed
  29cfb13
- Rescore the sensitivity runs on the 47-model board and add Haiku 5.5's
  
  MaxGhenis
  and
  claude
  committed
  2ddecc2
- Carry the release chain's records and tests to release 20261009
  
  MaxGhenis
  and
  claude
  committed
  b5aa080
- Move the paper, notes and cost tests to release 20261009
  
  MaxGhenis
  and
  claude
  committed
  b5869b2
- Hold the snapshot tests to the real builder's setup (delta review finding 6)
  
  MaxGhenis
  and
  claude
  committed
  ed879ca
- Write the release's paper text, card, guide and note
  
  MaxGhenis
  and
  claude
  committed
  8d1a6a7
- PRACTICE re-pin of the rendered paper on the 2.37.2 rehearsal (replaced at T)
  
  MaxGhenis
  and
  claude
  committed
  6a758b8
- State the October sweep's timing in the paper from its timing record
  
  MaxGhenis
  and
  claude
  committed
  8561d0b
- Count scored and excluded outputs apart in the publication check
  
  MaxGhenis
  and
  claude
  committed
  01c51ce
Commits on Oct 10, 2026
- Derive the upgrade's day and release-named files in tests
  
  MaxGhenis
  and
  claude
  committed
  e323908
- Count engine-defect records whose defect is now fixed apart
  
  MaxGhenis
  and
  claude
  committed
  06c761f
- Expect the action-binding refusal where it now fires first
  
  MaxGhenis
  and
  claude
  committed
  eaed82d
- Say the earlier models' rates fall, and how restored targets are known
  
  MaxGhenis
  and
  claude
  committed
  4c09f83
- Drop the superseded note and paper drafts
  
  MaxGhenis
  and
  claude
  committed
  534b63c
- Bring the design note up to the engine upgrade
  
  MaxGhenis
  and
  claude
  committed
  38ece52
- Say which restored outputs return to scoring and which stay scored
  
  MaxGhenis
  and
  claude
  committed
  a5e12ed
- Hold the note's facts at the pre-T rehearsal's expected values
  
  MaxGhenis
  and
  claude
  committed
  56d5094
- Tie the publication check to the release's build (pre-T review 1, 6)
  
  MaxGhenis
  and
  claude
  committed
  d87d7fb
- Match the builder's target rule; separate unfixed defects (pre-T review 4, 5, 11)
  
  MaxGhenis
  and
  claude
  committed
  598ae78
- Run the note's claim checks on recomputed facts (pre-T review 2)
  
  MaxGhenis
  and
  claude
  committed
  a54d8db
- Test the card's hand-written counts against the frozen records (pre-T review 3)
  
  MaxGhenis
  and
  claude
  committed
  72dd44f
- Hold the judge evidence to the decisions it names (pre-T review 7)
  
  MaxGhenis
  and
  claude
  committed
  41dc4b4
- Refuse more ways a fix module can reach local files (pre-T review 8)
  
  MaxGhenis
  and
  claude
  committed
  fbe956d
- Pin regeneration targets at the freeze; test the mock against every real build (pre-T review 9, 10)
  
  MaxGhenis
  and
  claude
  committed
  8b12164
- Name later law among the sensitivity doc's exclusions; test its gain claim (pre-T review 11)
  
  MaxGhenis
  and
  claude
  committed
  a09df45
- Run the publication check's sweep itself; require builder and sidecar (pre-T re-review 1-3)
  
  MaxGhenis
  and
  claude
  committed
  99e8ba4
- Close the re-review's nits (pre-T re-review 4-7)
  
  MaxGhenis
  and
  claude
  committed
  7f249bf
- Check the imported engine is the wheel; keep the serving probes (pre-T review round 3)
  
  MaxGhenis
  and
  claude
  committed
  88ba63c
- Revert "PRACTICE re-pin of the rendered paper on the 2.37.2 rehearsal (replaced at T)"
  MaxGhenis
  committed
  202c7e9
- Revert "PRACTICE FREEZE on the policyengine-us 2.37.2 rehearsal (replaced at T)"
  MaxGhenis
  committed
  a08e271
- Roll the release's day to 2026-10-10
  
  MaxGhenis
  and
  claude
  committed
  07395f0
- Check the engine in the process that computes (pre-T review round 4)
  
  MaxGhenis
  and
  claude
  committed
  26ceec6
- Refuse symlinks in the engine; take only fresh receipts and outputs (pre-T review round 5)
  
  MaxGhenis
  and
  claude
  committed
  d06a7f5
- Read the builder's arguments as the builder does (pre-T review round 6)
  
  MaxGhenis
  and
  claude
  committed
  ad81894
- Correct the reference narratives of all 24 outputs the 2.38.6 move changes
  
  MaxGhenis
  and
  claude
  committed
  4ee6443
- Record the judges of the cases the 2.38.6 references re-opened
  
  MaxGhenis
  and
  claude
  committed
  0e291e2
- Keep CA 099's state income tax excluded: California disallows educator expenses
  
  MaxGhenis
  and
  claude
  committed
  a4e76ad
- Record the judge of CA 099's re-opened case
  
  MaxGhenis
  and
  claude
  committed
  0041c76
- Pin policyengine-us 2.38.6, the references' engine
  
  MaxGhenis
  and
  claude
  committed
  0562d77
- Freeze release dashboard-data-20261010: Claude Haiku 5.5 on policyengine-us 2.38.6 references
  
  MaxGhenis
  and
  claude
  committed
  2c4e254
- Date the release's judge verdicts; rescore the sensitivity runs on the 2.38.6 board
  
  MaxGhenis
  and
  claude
  committed
  4f573c1
- Count fixed engine defects by the audit's evidence; separate a further defect
  
  MaxGhenis
  and
  claude
  committed
  b742031
- State the 2.38.6 release in the card, guide, app and note
  
  MaxGhenis
  and
  claude
  committed
  659b2ff
- Render the paper on the 2.38.6 references and pin it
  
  MaxGhenis
  and
  claude
  committed
  d191896
- Pin the tests to the 2.38.6 board and drop a stale judge wave
  
  MaxGhenis
  and
  claude
  committed
  cb42e48
- Answer the final review: CI pins, the 14-of-18 split, and the actions file
  
  MaxGhenis
  and
  claude
  committed
  4e10d12

Commits

Commits on Oct 9, 2026

Commits on Oct 10, 2026