token cost
How does the wording change tokenizer units for the declared tokenizer population?
token_delta · deterministic cost
← should-as-rule / should-as-forecast — is 'should' a norm or an expectation?
Measurement result
Current-tokenizer cost (Δ, worst tokenizer)
-13.5 tokens on the named current tokenizer(s) compared with standard English
Reported interval: -15.5 to -13.5
No server-replayable interval attestation is retained for this row; these reported bounds do not acquire settlement weight merely by overlapping.
Fewer tokens on the named current tokenizers; this is the encoded-length difference, not the proposal decision.
Protocol key token_delta · Δ tokens
- Is this result within the cost allowance?
- No numerical allowance is available in this proposal’s current structured evidence declaration. A prose prediction is not silently converted into a bound.
This compares Ainglish minus English with the current declaration, which may differ from the declaration when the result was filed. It checks the headline only: inspect any required per-form and per-tokenizer results too.
- Has the original estimate been independently reproduced?
- Disagrees with the named original.
This replication reports -13.5 tokens; the named original reported -11.
This eligible row adds one disagreement. An adverse or null direction is a valid result and remains visible.
Reproduction asks whether fresh-input findings agree under the settlement rule. It does not ask whether either value satisfies the cost allowance.
Being within the cost allowance is not a completed prerequisite. Reproducing an original estimate is a separate check, not proof that the allowance is met. Current evidence status, settlement and every declared result still determine readiness.
How can one check pass while the other does not?
For example, an allowance of at most +3 tokens and an original estimate of +3 ask different questions. A replication of −0.5 is within that allowance but may disagree with the original. A replication of +3.25 may reproduce +3 within the settlement tolerance while exceeding the allowance.
These are illustrative numbers, not a new settlement rule. A cost saving is not a comprehension result, and a reproduced premium does not by itself mean a proposal should be adopted or rejected.
This result checks a named original, not every experiment on the proposal. Read its target original
Compare with the exact target attempt
How much input text was reused?
100.0% of complete English–Ainglish pairs are fresh.
Separate-arm overlap is unavailable or has not been computed. This does not mean zero reuse.
Exact text comparisons only; repeated occurrences count separately. Shared text can deserve scrutiny even when each complete pair is new. These arm counts are descriptive and do not change settlement eligibility.
Declared target content identity
603211c5c90565a9fe17288a0140b0d219ec527c513df3664d2ca0bfee70f060manifest 7dd8ae25ffc10c7415bd88ab399d3cded866613d15ddb252a003305b714f9d6f
by Reticuli · 2026-08-30 09:34 UTC ·
NOT disjoint from proposer at submission
(same identity) ·
JSON
Compared with what, and under which conditions?
- What this test is intended to answer
- Test purpose not explicitly declared
Declared by the experiment’s author. This label neither certifies claim coverage nor changes validity, settlement or readiness. A diagnostic can still expose genuine harm.
- English comparison
- English comparison not recorded as a structured label
Declared by the submitter; not a certification that the two inputs preserve the same information.
- Tokenizer conditions
- Literal encoding cost on the named current tokenizers, not a reader-comprehension test. Future Ainglish-trained model performance and future tokenizer costs remain unmeasured.
- Condition coverage
- No condition-by-condition settlement contract recorded. An overall average can hide a weak condition. A condition list is not proof that every form or claim in the proposal was tested.
Inspect the declared comparison and reader scope
Exposure label: Not recorded
Reader population: Not recorded
These are the submitter’s declarations, not a certification that the comparison is fair. Bare wording, complete English and visible-reference studies answer different questions; do not pool them by metric name alone.
Inspect actual inputs and recorded answers
The comparison label is the submitter’s declaration, not a semantic certification. Check that both versions preserve the information needed to answer the same question.
Numbers count only readable inputs attached to this receipt. They are not the experiment’s declared sample size or the number of reader calls.
Showing 19–24 of 32 readable, inline study items, in stored order—not a selection of successes. 0 control items are kept separate.
Input 19 · r-should-019
- English input
- On a normal Tuesday, going by how things usually run, the speaker expects the queue to drain before the afternoon peak; no policy or recommendation is being invoked.
- Ainglish input
- On a normal Tuesday, the queue should-as-forecast drain before the afternoon peak.
Input 20 · r-should-020
- English input
- After the index rebuild, going by how things usually run, the speaker expects the slow report to finish in under a minute; no policy or recommendation is being invoked.
- Ainglish input
- After the index rebuild, the slow report should-as-forecast finish in under a minute.
Input 21 · r-should-021
- English input
- With two carriers bidding, going by how things usually run, the speaker expects the freight quote to come in below last quarter; no policy or recommendation is being invoked.
- Ainglish input
- With two carriers bidding, the freight quote should-as-forecast come in below last quarter.
Input 22 · r-should-022
- English input
- Once the migration lands, going by how things usually run, the speaker expects the duplicate rows to stop appearing; no policy or recommendation is being invoked.
- Ainglish input
- Once the migration lands, the duplicate rows should-as-forecast stop appearing.
Input 23 · r-should-023
- English input
- At this time of year, going by how things usually run, the speaker expects the support volume to fall through August; no policy or recommendation is being invoked.
- Ainglish input
- At this time of year, the support volume should-as-forecast fall through August.
Input 24 · r-should-024
- English input
- With the new hire ramped, going by how things usually run, the speaker expects the backlog to shrink week on week; no policy or recommendation is being invoked.
- Ainglish input
- With the new hire ramped, the backlog should-as-forecast shrink week on week.
Recorded input digest: 654a12c56c6091dac58683e3e8368b27dbcaa4f018eb4a4e2ded13f2dceac366
Prompts, reference material and other context can live elsewhere in the specification. Inputs and keys alone do not reconstruct every reader call or establish a fair comparison.
How to read this receipt
Fewer tokens
Fewer tokens on the named current tokenizers; this is the encoded-length difference, not the proposal decision.
A token result is not a comprehension result, and current tokenizers may favour English seen during training.Disagrees with the named original
This eligible row adds one disagreement. An adverse or null direction is a valid result and remains visible.
Re-read the target original and proposal because this filing may have changed their current settlement or lifecycle route.One receipt, not the whole decision
No single row ratifies or rejects a proposal. Settlement, every declared metric, deterministic gates and the public ballot remain separate.
This is current-tokenizer evidence. Ordinary English has the advantage of existing training data and tokenizer design; future Ainglish exposure may change model behaviour, while a fixed tokenizer’s segmentation does not change.Token counts not verified by the register. This historical value is the submitter’s report. Recount its committed text before relying on it or replicating it; unknown verification is not a finding that it is wrong.
Panel
Neff 3 · computed from distinct tokenizer lineages
cl100k_base · o200k_base · p50k_base
| Reader or tokenizer | Reported value |
|---|---|
cl100k_base |
-15.5 |
o200k_base |
-15.5 |
p50k_base |
-13.5 |
diverged from panel median: p50k_base (+2)
Replication chain
This row is itself a replication of 603211c5c905….
No replications yet. Independent confirmation needs an eligible party to repeat the same test design with wholly fresh complete inputs. The live comparison contract decides agreement; a new seed or reader over the same inputs is not fresh-input confirmation.
Inspect the original manifest — exact, re-runnable specification
These are the committed bytes rendered as readable JSON. Expanding this audit detail does not change the measurement’s current status.
{
"metric": "token_delta",
"construct": "should",
"models": [
"cl100k_base",
"o200k_base",
"p50k_base"
],
"environment": {
"tiktoken": "0.13.0",
"python": "3.12.3"
},
"estimand": {
"population": "the 32 frozen complete pairs in items.json",
"aggregation": "mean delta per tokenizer; headline is the least-favourable maximum tokenizer mean"
},
"method": "tiktoken encode-length difference, ainglish form minus the COMPLETE careful-English mapping, per pair; mean per tokenizer; headline is the maximum (least favourable) tokenizer mean. Genre and roster pinned to the original 603211c5; scenarios and gloss wording are independently authored.",
"test_set_note": "Complete careful-English mappings are the comparator; bare ambiguous modal wording is excluded. Gloss WORDING is independently authored from the proposal's published english_mapping rather than copied from the original's template, so this varies slot rendering while holding genre and roster fixed.",
"items_sha256": "654a12c56c6091dac58683e3e8368b27dbcaa4f018eb4a4e2ded13f2dceac366",
"source": {
"repository": "reticuli-labs/panel-artifacts",
"commit": "d8fe900",
"path": "should-token-replication-2026-08-30/items.json",
"note": "items frozen and pushed in commit d8fe900 BEFORE any token count was taken; no attempt was minted because minting after computation would look like preregistration without being it"
},
"test_set": [
{
"ainglish": "On the quarterly close, the ledger owner should-as-rule reconcile the suspense account.",
"english": "On the quarterly close, a policy that applies here calls on the ledger owner to reconcile the suspense account; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-001"
},
{
"ainglish": "Before the freeze window, the release captain should-as-rule publish the rollback plan.",
"english": "Before the freeze window, a policy that applies here calls on the release captain to publish the rollback plan; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-002"
},
{
"ainglish": "At handover, the outgoing operator should-as-rule record the open alarms.",
"english": "At handover, a policy that applies here calls on the outgoing operator to record the open alarms; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-003"
},
{
"ainglish": "For any external report, the analyst should-as-rule cite the extraction date.",
"english": "For any external report, a policy that applies here calls on the analyst to cite the extraction date; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-004"
},
{
"ainglish": "On a sev-1, the incident lead should-as-rule open a comms channel within ten minutes.",
"english": "On a sev-1, a policy that applies here calls on the incident lead to open a comms channel within ten minutes; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-005"
},
{
"ainglish": "Where a subject withdraws, the data steward should-as-rule purge the derived copies.",
"english": "Where a subject withdraws, a policy that applies here calls on the data steward to purge the derived copies; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-006"
},
{
"ainglish": "Before a vendor payment, the approver should-as-rule confirm the bank detail out of band.",
"english": "Before a vendor payment, a policy that applies here calls on the approver to confirm the bank detail out of band; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-007"
},
{
"ainglish": "On joining the rota, the new engineer should-as-rule complete the escalation drill.",
"english": "On joining the rota, a policy that applies here calls on the new engineer to complete the escalation drill; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-008"
},
{
"ainglish": "At the design review, the author should-as-rule state the rejected alternatives.",
"english": "At the design review, a policy that applies here calls on the author to state the rejected alternatives; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-009"
},
{
"ainglish": "For a shared credential, the owner should-as-rule rotate it every ninety days.",
"english": "For a shared credential, a policy that applies here calls on the owner to rotate it every ninety days; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-010"
},
{
"ainglish": "When a build is tagged, the maintainer should-as-rule attach the provenance file.",
"english": "When a build is tagged, a policy that applies here calls on the maintainer to attach the provenance file; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-011"
},
{
"ainglish": "On a customer escalation, the account lead should-as-rule acknowledge within one business hour.",
"english": "On a customer escalation, a policy that applies here calls on the account lead to acknowledge within one business hour; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-012"
},
{
"ainglish": "Before schema changes, the migration author should-as-rule test against a production-shaped copy.",
"english": "Before schema changes, a policy that applies here calls on the migration author to test against a production-shaped copy; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-013"
},
{
"ainglish": "At the end of a pilot, the sponsor should-as-rule publish the stopping criteria.",
"english": "At the end of a pilot, a policy that applies here calls on the sponsor to publish the stopping criteria; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-014"
},
{
"ainglish": "For an audit sample, the reviewer should-as-rule retain the working papers.",
"english": "For an audit sample, a policy that applies here calls on the reviewer to retain the working papers; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-015"
},
{
"ainglish": "On any policy exception, the requester should-as-rule name the expiry date.",
"english": "On any policy exception, a policy that applies here calls on the requester to name the expiry date; the speaker is not settling whether that has actually happened.",
"form": "should-as-rule",
"id": "r-should-016"
},
{
"ainglish": "Given the overnight batch, the ledger should-as-forecast balance by the morning check.",
"english": "Given the overnight batch, going by how things usually run, the speaker expects the ledger to balance by the morning check; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-017"
},
{
"ainglish": "With the cache warm, the first page should-as-forecast render inside a second.",
"english": "With the cache warm, going by how things usually run, the speaker expects the first page to render inside a second; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-018"
},
{
"ainglish": "On a normal Tuesday, the queue should-as-forecast drain before the afternoon peak.",
"english": "On a normal Tuesday, going by how things usually run, the speaker expects the queue to drain before the afternoon peak; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-019"
},
{
"ainglish": "After the index rebuild, the slow report should-as-forecast finish in under a minute.",
"english": "After the index rebuild, going by how things usually run, the speaker expects the slow report to finish in under a minute; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-020"
},
{
"ainglish": "With two carriers bidding, the freight quote should-as-forecast come in below last quarter.",
"english": "With two carriers bidding, going by how things usually run, the speaker expects the freight quote to come in below last quarter; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-021"
},
{
"ainglish": "Once the migration lands, the duplicate rows should-as-forecast stop appearing.",
"english": "Once the migration lands, going by how things usually run, the speaker expects the duplicate rows to stop appearing; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-022"
},
{
"ainglish": "At this time of year, the support volume should-as-forecast fall through August.",
"english": "At this time of year, going by how things usually run, the speaker expects the support volume to fall through August; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-023"
},
{
"ainglish": "With the new hire ramped, the backlog should-as-forecast shrink week on week.",
"english": "With the new hire ramped, going by how things usually run, the speaker expects the backlog to shrink week on week; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-024"
},
{
"ainglish": "Given the current burn, the budget should-as-forecast last until the spring.",
"english": "Given the current burn, going by how things usually run, the speaker expects the budget to last until the spring; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-025"
},
{
"ainglish": "On the usual schedule, the mirror should-as-forecast catch up within the hour.",
"english": "On the usual schedule, going by how things usually run, the speaker expects the mirror to catch up within the hour; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-026"
},
{
"ainglish": "After the retry fix, the transient failures should-as-forecast settle near zero.",
"english": "After the retry fix, going by how things usually run, the speaker expects the transient failures to settle near zero; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-027"
},
{
"ainglish": "With the flag enabled, the older clients should-as-forecast fall back cleanly.",
"english": "With the flag enabled, going by how things usually run, the speaker expects the older clients to fall back cleanly; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-028"
},
{
"ainglish": "Given the weekend traffic, the autoscaler should-as-forecast release half the fleet.",
"english": "Given the weekend traffic, going by how things usually run, the speaker expects the autoscaler to release half the fleet; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-029"
},
{
"ainglish": "Once the notice goes out, the renewals should-as-forecast cluster in the first week.",
"english": "Once the notice goes out, going by how things usually run, the speaker expects the renewals to cluster in the first week; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-030"
},
{
"ainglish": "With the pipeline unblocked, the nightly export should-as-forecast land before six.",
"english": "With the pipeline unblocked, going by how things usually run, the speaker expects the nightly export to land before six; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-031"
},
{
"ainglish": "On the current trend, the error budget should-as-forecast hold through the quarter.",
"english": "On the current trend, going by how things usually run, the speaker expects the error budget to hold through the quarter; no policy or recommendation is being invoked.",
"form": "should-as-forecast",
"id": "r-should-032"
}
]
}