token cost
How does the wording change tokenizer units for the declared tokenizer population?
token_delta · deterministic cost
Measurement result
Current-tokenizer cost (Δ, worst tokenizer)
-6.625 tokens on the named current tokenizer(s) compared with standard English
Reported interval: -6.625 to -6.625
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?
- Target no longer carries evidence.
This replication reports -6.625 tokens; the named original reported -1.5.
The target original was retracted. This replication remains visible, but no longer adds a settlement voice to that target. This does not itself invalidate the replication’s observations.
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?
Complete-pair freshness is not available for this receipt.
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
893510f22c697fc45ab7c073147e90bfcc1a31cf888cb49cb511ed2ceee8e414manifest 8b571f1fa5cbf0591b2d5e9040b56329fce9ff5eace7cdd0bda7c92bf2b9a741
by Dexagon · 2026-08-27 17:04 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 1–6 of 32 readable, inline study items, in stored order—not a selection of successes. 0 control items are kept separate.
Input 1
- English input
- batch numbers from 12 to 48, including both 12 and 48
- Ainglish input
- batch numbers 12 to 48, include-both
Input 2
- English input
- release dates from Tuesday to Saturday, including both Tuesday and Saturday
- Ainglish input
- release dates Tuesday to Saturday, include-both
Input 3
- English input
- supported versions from 3.2 to 5.0, including both 3.2 and 5.0
- Ainglish input
- supported versions 3.2 to 5.0, include-both
Input 4
- English input
- temperature readings from -8 to 14, including both -8 and 14
- Ainglish input
- temperature readings -8 to 14, include-both
Input 5
- English input
- network ports from 4100 to 4199, including both 4100 and 4199
- Ainglish input
- network ports 4100 to 4199, include-both
Input 6
- English input
- priority ranks from 20 to 4, including both 20 and 4
- Ainglish input
- priority ranks 20 to 4, include-both
Recorded input digest: 4fc4eb85fb6b886ebfa6faa8589963fdb8ca30aff9822078cc10fe6fb094556c
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.Target no longer carries evidence
The target original was retracted. This replication remains visible, but no longer adds a settlement voice to that target. This does not itself invalidate the replication’s observations.
Read the target’s retraction reason. Do not repeat a retired instrument or rescore old answers to recover a preferred outcome.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 2 · computed from distinct tokenizer lineages
cl100k_base · o200k_base
| Reader or tokenizer | Reported value |
|---|---|
cl100k_base |
-6.625 |
o200k_base |
-6.625 |
Replication chain
This row is itself a replication of 893510f22c69….
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.
{
"kind": "ainglish.range-endpoint-token-replication-manifest.v1",
"metric": "token_delta",
"formula_version": 1,
"construct": "include-both / include-start-only / include-end-only / exclude-both",
"models": [
"cl100k_base",
"o200k_base"
],
"test_set": [
{
"item_id": "include-both-01",
"form": "include-both",
"ainglish": "batch numbers 12 to 48, include-both",
"english": "batch numbers from 12 to 48, including both 12 and 48"
},
{
"item_id": "include-both-02",
"form": "include-both",
"ainglish": "release dates Tuesday to Saturday, include-both",
"english": "release dates from Tuesday to Saturday, including both Tuesday and Saturday"
},
{
"item_id": "include-both-03",
"form": "include-both",
"ainglish": "supported versions 3.2 to 5.0, include-both",
"english": "supported versions from 3.2 to 5.0, including both 3.2 and 5.0"
},
{
"item_id": "include-both-04",
"form": "include-both",
"ainglish": "temperature readings -8 to 14, include-both",
"english": "temperature readings from -8 to 14, including both -8 and 14"
},
{
"item_id": "include-both-05",
"form": "include-both",
"ainglish": "network ports 4100 to 4199, include-both",
"english": "network ports from 4100 to 4199, including both 4100 and 4199"
},
{
"item_id": "include-both-06",
"form": "include-both",
"ainglish": "priority ranks 20 to 4, include-both",
"english": "priority ranks from 20 to 4, including both 20 and 4"
},
{
"item_id": "include-both-07",
"form": "include-both",
"ainglish": "log offsets 900 to 1200, include-both",
"english": "log offsets from 900 to 1200, including both 900 and 1200"
},
{
"item_id": "include-both-08",
"form": "include-both",
"ainglish": "checkpoint times 18:30 to 06:15, include-both",
"english": "checkpoint times from 18:30 to 06:15, including both 18:30 and 06:15"
},
{
"item_id": "include-start-only-01",
"form": "include-start-only",
"ainglish": "batch numbers 12 to 48, include-start-only",
"english": "batch numbers from 12, including 12, to 48, excluding 48"
},
{
"item_id": "include-start-only-02",
"form": "include-start-only",
"ainglish": "release dates Tuesday to Saturday, include-start-only",
"english": "release dates from Tuesday, including Tuesday, to Saturday, excluding Saturday"
},
{
"item_id": "include-start-only-03",
"form": "include-start-only",
"ainglish": "supported versions 3.2 to 5.0, include-start-only",
"english": "supported versions from 3.2, including 3.2, to 5.0, excluding 5.0"
},
{
"item_id": "include-start-only-04",
"form": "include-start-only",
"ainglish": "temperature readings -8 to 14, include-start-only",
"english": "temperature readings from -8, including -8, to 14, excluding 14"
},
{
"item_id": "include-start-only-05",
"form": "include-start-only",
"ainglish": "network ports 4100 to 4199, include-start-only",
"english": "network ports from 4100, including 4100, to 4199, excluding 4199"
},
{
"item_id": "include-start-only-06",
"form": "include-start-only",
"ainglish": "priority ranks 20 to 4, include-start-only",
"english": "priority ranks from 20, including 20, to 4, excluding 4"
},
{
"item_id": "include-start-only-07",
"form": "include-start-only",
"ainglish": "log offsets 900 to 1200, include-start-only",
"english": "log offsets from 900, including 900, to 1200, excluding 1200"
},
{
"item_id": "include-start-only-08",
"form": "include-start-only",
"ainglish": "checkpoint times 18:30 to 06:15, include-start-only",
"english": "checkpoint times from 18:30, including 18:30, to 06:15, excluding 06:15"
},
{
"item_id": "include-end-only-01",
"form": "include-end-only",
"ainglish": "batch numbers 12 to 48, include-end-only",
"english": "batch numbers from 12, excluding 12, to 48, including 48"
},
{
"item_id": "include-end-only-02",
"form": "include-end-only",
"ainglish": "release dates Tuesday to Saturday, include-end-only",
"english": "release dates from Tuesday, excluding Tuesday, to Saturday, including Saturday"
},
{
"item_id": "include-end-only-03",
"form": "include-end-only",
"ainglish": "supported versions 3.2 to 5.0, include-end-only",
"english": "supported versions from 3.2, excluding 3.2, to 5.0, including 5.0"
},
{
"item_id": "include-end-only-04",
"form": "include-end-only",
"ainglish": "temperature readings -8 to 14, include-end-only",
"english": "temperature readings from -8, excluding -8, to 14, including 14"
},
{
"item_id": "include-end-only-05",
"form": "include-end-only",
"ainglish": "network ports 4100 to 4199, include-end-only",
"english": "network ports from 4100, excluding 4100, to 4199, including 4199"
},
{
"item_id": "include-end-only-06",
"form": "include-end-only",
"ainglish": "priority ranks 20 to 4, include-end-only",
"english": "priority ranks from 20, excluding 20, to 4, including 4"
},
{
"item_id": "include-end-only-07",
"form": "include-end-only",
"ainglish": "log offsets 900 to 1200, include-end-only",
"english": "log offsets from 900, excluding 900, to 1200, including 1200"
},
{
"item_id": "include-end-only-08",
"form": "include-end-only",
"ainglish": "checkpoint times 18:30 to 06:15, include-end-only",
"english": "checkpoint times from 18:30, excluding 18:30, to 06:15, including 06:15"
},
{
"item_id": "exclude-both-01",
"form": "exclude-both",
"ainglish": "batch numbers 12 to 48, exclude-both",
"english": "batch numbers from 12 to 48, excluding both 12 and 48"
},
{
"item_id": "exclude-both-02",
"form": "exclude-both",
"ainglish": "release dates Tuesday to Saturday, exclude-both",
"english": "release dates from Tuesday to Saturday, excluding both Tuesday and Saturday"
},
{
"item_id": "exclude-both-03",
"form": "exclude-both",
"ainglish": "supported versions 3.2 to 5.0, exclude-both",
"english": "supported versions from 3.2 to 5.0, excluding both 3.2 and 5.0"
},
{
"item_id": "exclude-both-04",
"form": "exclude-both",
"ainglish": "temperature readings -8 to 14, exclude-both",
"english": "temperature readings from -8 to 14, excluding both -8 and 14"
},
{
"item_id": "exclude-both-05",
"form": "exclude-both",
"ainglish": "network ports 4100 to 4199, exclude-both",
"english": "network ports from 4100 to 4199, excluding both 4100 and 4199"
},
{
"item_id": "exclude-both-06",
"form": "exclude-both",
"ainglish": "priority ranks 20 to 4, exclude-both",
"english": "priority ranks from 20 to 4, excluding both 20 and 4"
},
{
"item_id": "exclude-both-07",
"form": "exclude-both",
"ainglish": "log offsets 900 to 1200, exclude-both",
"english": "log offsets from 900 to 1200, excluding both 900 and 1200"
},
{
"item_id": "exclude-both-08",
"form": "exclude-both",
"ainglish": "checkpoint times 18:30 to 06:15, exclude-both",
"english": "checkpoint times from 18:30 to 06:15, excluding both 18:30 and 06:15"
}
],
"items_sha256": "4fc4eb85fb6b886ebfa6faa8589963fdb8ca30aff9822078cc10fe6fb094556c",
"test_set_note": "32 fresh complete mappings, eight per form; four numeric/date contexts are descending so start/end remain positions as written rather than lower/upper values",
"estimand": {
"population": "all 32 frozen fresh complete mappings, with equal weight per item and eight items per form",
"aggregation": "mean per tokenizer; headline is the least-favourable maximum mean",
"reference": "same metric and aggregation rule as the target original, using different metric inputs"
},
"environment": {
"library": "tiktoken",
"version": "0.13.0",
"python": "3.12.3"
},
"source": {
"repository": "dexagon-ai/ainglish-evidence",
"commit": "0b93c8446dd90be7ab0282265e419bbecb72e8c0",
"path": "range-endpoint-token-replication-v1-2026-08-27/items.json"
},
"evidentiary_limit": "Price evidence only; no comprehension, correctness, adoption, or flagship-quality claim."
}