Evidence

Compare two experiments

Check what changed in the question, wording or readers before comparing the numbers.

This view keeps both results separate. It does not calculate a combined score or decide whether they reproduce each other.

Change the selected experimentsSearch by proposal, select a result or enter an exact identifier

Find experiments by proposal

Choose two experiments by title, measurement and date. The choices include up to 50 newest public completed results, plus your current selections. Historical results stay labelled. For older records, use the evidence explorer or exact entry below.

Use an exact experiment identifier instead

A full attempt UUID entered here replaces the corresponding choice above. Content hashes are not result identifiers.

Clear comparison

Compare the essentials

Read each question across both results. A matching description is not proof that the studies are scientifically comparable.

Question	First result	Second result
Which proposal?	replace(old=…, new=…) — which thing leaves, and which takes its place?2026-09-08 18:36 UTC	replace(old=…, new=…) — which thing leaves, and which takes its place?2026-09-11 21:03 UTC
What was measured?	token costHow does the wording change tokenizer units for the declared tokenizer population?	token costHow does the wording change tokenizer units for the declared tokenizer population?
What was the result?	-0.75 tokens per declared itemReported interval: -2.25 to -0.75Not yet counting in evidence decisions	-0.625 tokens per declared itemReported interval: -2.125 to -0.625Counts in current evidence decisions
Compared with what?	English comparison not recorded as a structured labelCurrent tokenizer cost, not comprehension	English comparison not recorded as a structured labelCurrent tokenizer cost, not comprehension
Independently settled?	DisputedEligible replications disagree and this original does not hold a settlement majority.	Disagrees with the named originalThis eligible row adds one disagreement. An adverse or null direction is a valid result and remains visible.

Inspect first result Inspect second result

What differs in the declared methods?

These are literal recorded values, not a compatibility test. Matching declarations do not prove equivalent inputs or fair scoring. A missing structured field may be described elsewhere in the immutable specification.

What was measured · Same recorded value

First result

token_delta

Second result

token_delta

English comparison declarations · Not recorded on either side

First result

Not recorded in this structured field

Second result

Not recorded in this structured field

Tested population · Same recorded value

First result

Prospectively authored complete replacement mappings: 8 declared domains, 2 distinct old/new reference tuples per domain, each in request/report/proposal/simulation. Both arms share exact slot context, force prefix, and old/new reference bytes.

Second result

Prospectively authored complete replacement mappings: 8 declared domains, 2 distinct old/new reference tuples per domain, each in request/report/proposal/simulation. Both arms share exact slot context, force prefix, and old/new reference bytes.

Unit tested · Same recorded value

First result

one complete meaning-matched utterance pair including all shared contextual text

Second result

one complete meaning-matched utterance pair including all shared contextual text

Named instruments (including order) · Same recorded value

First result

[
    "cl100k_base",
    "o200k_base",
    "p50k_base"
]

Second result

[
    "cl100k_base",
    "o200k_base",
    "p50k_base"
]

Reader exposure · Not recorded on either side

First result

Not recorded in this structured field

Second result

Not recorded in this structured field

Exposure window · Not recorded on either side

First result

Not recorded in this structured field

Second result

Not recorded in this structured field

How results combine · Same recorded value

First result

{
    "reducer": "least_favourable",
    "rule": "Equal-weight form/force strata, equal domain/reference cells within each stratum; maximum tokenizer mean is the least-favourable headline. No rounding."
}

Second result

{
    "reducer": "least_favourable",
    "rule": "Equal-weight form/force strata, equal domain/reference cells within each stratum; maximum tokenizer mean is the least-favourable headline. No rounding."
}

Scoring declaration · Not recorded on either side

First result

Not recorded in this structured field

Second result

Not recorded in this structured field

Conditions (including order) · Same recorded value

First result

[
    {
        "id": "request",
        "weight": 1
    },
    {
        "id": "report",
        "weight": 1
    },
    {
        "id": "proposal",
        "weight": 1
    },
    {
        "id": "simulation",
        "weight": 1
    }
]

Second result

[
    {
        "id": "request",
        "weight": 1
    },
    {
        "id": "report",
        "weight": 1
    },
    {
        "id": "proposal",
        "weight": 1
    },
    {
        "id": "simulation",
        "weight": 1
    }
]

First result: full recordtoken cost · -0.75 tokens per declared item

First result · 2026-09-08 18:36 UTC

replace(old=…, new=…) — which thing leaves, and which takes its place?

Not yet counting in evidence decisions. This row remains available for assessment, but does not currently carry a counting evidence result.

What was measured

token cost · token_delta
How does the wording change tokenizer units for the declared tokenizer population?

Reported result

-0.75 tokens per declared item
Reported interval: -2.25 to -0.75.

Fewer tokens on the named current tokenizers; this is the encoded-length difference, not the proposal decision.

English comparison

English comparison not recorded as a structured label

Declared by the submitter; not a certification of equivalent information.

Tokenizer conditions

Literal encoding cost on the named current tokenizers; not comprehension.

Named instruments

cl100k_base, o200k_base, p50k_base

Reader population not separately declared.

Conditions covered

Separate outcomes retained for all 4 declared conditions

request, report, proposal, simulation

An overall average can hide a weak condition. A condition list is not proof that every form or claim in the proposal was tested.

Settlement role

Disputed

Eligible replications disagree and this original does not hold a settlement majority.

Is this result within the cost allowance?

This headline is within the allowance. The reported difference is -0.75 tokens; the current declaration allows at most 0 tokens.

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?

Disputed.

Eligible replications disagree and this original does not hold a settlement majority.

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.

Does the overall result hide differences between conditions?

Every stored condition, without new pooling. Differences and intervals use tokens. Condition names come from the frozen experiment.
Condition	Reported difference	Reported interval
`request`	-0.75	Not recorded
`report`	-0.75	Not recorded
`proposal`	-0.75	Not recorded
`simulation`	-0.75	Not recorded

A missing condition interval is not zero uncertainty. An overall interval cannot substitute for agreement in every load-bearing condition.

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 64 readable, inline study items, in stored order—not a selection of successes. 0 control items are kept separate.

Each result has its own input pages. Positions across the two studies do not imply matched cases.

Input 1 · `credentials-1-request`

English input: For the gateway signing-key slot, please do this: remove key-cedar from that slot and put key-maple there instead.
Ainglish input: For the gateway signing-key slot, please do this: replace(old=key-cedar, new=key-maple).
Condition: request

Input 2 · `credentials-1-report`

English input: For the gateway signing-key slot, the completed change was: remove key-cedar from that slot and put key-maple there instead.
Ainglish input: For the gateway signing-key slot, the completed change was: replace(old=key-cedar, new=key-maple).
Condition: report

Input 3 · `credentials-1-proposal`

English input: For the gateway signing-key slot, the proposed change is: remove key-cedar from that slot and put key-maple there instead.
Ainglish input: For the gateway signing-key slot, the proposed change is: replace(old=key-cedar, new=key-maple).
Condition: proposal

Input 4 · `credentials-1-simulation`

English input: For the gateway signing-key slot, simulate this change: remove key-cedar from that slot and put key-maple there instead.
Ainglish input: For the gateway signing-key slot, simulate this change: replace(old=key-cedar, new=key-maple).
Condition: simulation

Input 5 · `credentials-2-request`

English input: For the archive access-token slot, please do this: remove token-birch from that slot and put token-willow there instead.
Ainglish input: For the archive access-token slot, please do this: replace(old=token-birch, new=token-willow).
Condition: request

Input 6 · `credentials-2-report`

English input: For the archive access-token slot, the completed change was: remove token-birch from that slot and put token-willow there instead.
Ainglish input: For the archive access-token slot, the completed change was: replace(old=token-birch, new=token-willow).
Condition: report

Recorded input digest: ce73ee222026490506cb41b158d185f1d84fb62910a572be0d7d6b2baf7fd202

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.

Declared population, method and retained outcomes

Compared with: Ainglish minus complete careful English in current tokenizer units; negative is fewer tokens, positive is a premium
Tested population: Prospectively authored complete replacement mappings: 8 declared domains, 2 distinct old/new reference tuples per domain, each in request/report/proposal/simulation. Both arms share exact slot context, force prefix, and old/new reference bytes.
Unit tested: one complete meaning-matched utterance pair including all shared contextual text
How results combine: Equal-weight form/force strata, equal domain/reference cells within each stratum; maximum tokenizer mean is the least-favourable headline. No rounding.

These are the study author’s declarations. A finding applies to this tested scope; this summary does not establish that another study is comparable.

Absolute arm results, reader-specific results and condition results below are retained values, not a newly pooled analysis. Accuracy arms use fractions from 0 to 1; their difference uses percentage points.

Absolute arm results

Not recorded

Reader or tokenizer results

[
    {
        "model": "cl100k_base",
        "value": -2.25
    },
    {
        "model": "o200k_base",
        "value": -2
    },
    {
        "model": "p50k_base",
        "value": -0.75
    }
]

Condition results

[
    {
        "id": "request",
        "weight": 1,
        "share": 0.25,
        "value": -0.75,
        "value_lo": null,
        "value_hi": null,
        "arms": null,
        "resolution_bound": "not_applicable"
    },
    {
        "id": "report",
        "weight": 1,
        "share": 0.25,
        "value": -0.75,
        "value_lo": null,
        "value_hi": null,
        "arms": null,
        "resolution_bound": "not_applicable"
    },
    {
        "id": "proposal",
        "weight": 1,
        "share": 0.25,
        "value": -0.75,
        "value_lo": null,
        "value_hi": null,
        "arms": null,
        "resolution_bound": "not_applicable"
    },
    {
        "id": "simulation",
        "weight": 1,
        "share": 0.25,
        "value": -0.75,
        "value_lo": null,
        "value_hi": null,
        "arms": null,
        "resolution_bound": "not_applicable"
    }
]

Exact result and immutable specification Experiment history

Attempt 8f2292a3-daec-4fd1-b789-82fed2aca03f
Content e2ff808e72df863f2c403344843ac1f8e81cd6ae3b55ed3150e05ff922de5842

Second result: full recordtoken cost · -0.625 tokens per declared item

Second result · 2026-09-11 21:03 UTC

replace(old=…, new=…) — which thing leaves, and which takes its place?

Counts in current evidence decisions. This row currently contributes to evidence decisions. Its direction is separate from whether the proposal is ready for adoption.

What was measured

token cost · token_delta
How does the wording change tokenizer units for the declared tokenizer population?

Reported result

-0.625 tokens per declared item
Reported interval: -2.125 to -0.625.

Fewer tokens on the named current tokenizers; this is the encoded-length difference, not the proposal decision.

English comparison

English comparison not recorded as a structured label

Declared by the submitter; not a certification of equivalent information.

Tokenizer conditions

Literal encoding cost on the named current tokenizers; not comprehension.

Named instruments

cl100k_base, o200k_base, p50k_base

Reader population not separately declared.

Conditions covered

Separate outcomes retained for all 4 declared conditions

request, report, proposal, simulation

An overall average can hide a weak condition. A condition list is not proof that every form or claim in the proposal was tested.

Settlement role

Disagrees with the named original

This eligible row adds one disagreement. An adverse or null direction is a valid result and remains visible.

Is this result within the cost allowance?

This headline is within the allowance. The reported difference is -0.625 tokens; the current declaration allows at most 0 tokens.

Has the original estimate been independently reproduced?

Disagrees with the named original. This replication reports -0.625 tokens; the named original reported -0.75.

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.

How can one check pass while the other does not?

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

Every declared condition must agree. Overlapping overall intervals alone do not confirm this original.

How much input text was reused?

100.0% of complete English–Ainglish pairs are fresh.

0 of 64 English inputs reuse text from either side of the original.
0 of 64 Ainglish inputs reuse text from either side of the original.

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

e2ff808e72df863f2c403344843ac1f8e81cd6ae3b55ed3150e05ff922de5842

Does the overall result hide differences between conditions?

Every stored condition, without new pooling. Differences and intervals use tokens. Condition names come from the frozen experiment.
Condition	Reported difference	Reported interval
`request`	-0.625	Not recorded
`report`	-0.625	Not recorded
`proposal`	-0.625	Not recorded
`simulation`	-0.625	Not recorded

A missing condition interval is not zero uncertainty. An overall interval cannot substitute for agreement in every load-bearing condition.

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 64 readable, inline study items, in stored order—not a selection of successes. 0 control items are kept separate.

Each result has its own input pages. Positions across the two studies do not imply matched cases.

Input 1 · `sat-credentials-1-request`

English input: For the relay signing-key slot, please do this: remove key-alder from that slot and put key-rowan there instead.
Ainglish input: For the relay signing-key slot, please do this: replace(old=key-alder, new=key-rowan).
Condition: request

Input 2 · `sat-credentials-1-report`

English input: For the relay signing-key slot, the completed change was: remove key-alder from that slot and put key-rowan there instead.
Ainglish input: For the relay signing-key slot, the completed change was: replace(old=key-alder, new=key-rowan).
Condition: report

Input 3 · `sat-credentials-1-proposal`

English input: For the relay signing-key slot, the proposed change is: remove key-alder from that slot and put key-rowan there instead.
Ainglish input: For the relay signing-key slot, the proposed change is: replace(old=key-alder, new=key-rowan).
Condition: proposal

Input 4 · `sat-credentials-1-simulation`

English input: For the relay signing-key slot, simulate this change: remove key-alder from that slot and put key-rowan there instead.
Ainglish input: For the relay signing-key slot, simulate this change: replace(old=key-alder, new=key-rowan).
Condition: simulation

Input 5 · `sat-credentials-2-request`

English input: For the backup access-token slot, please do this: remove token-spruce from that slot and put token-poplar there instead.
Ainglish input: For the backup access-token slot, please do this: replace(old=token-spruce, new=token-poplar).
Condition: request

Input 6 · `sat-credentials-2-report`

English input: For the backup access-token slot, the completed change was: remove token-spruce from that slot and put token-poplar there instead.
Ainglish input: For the backup access-token slot, the completed change was: replace(old=token-spruce, new=token-poplar).
Condition: report

Recorded input digest: bbac0fa0d452dca8b4b3ea6923edb513f52adb4d32af77e52e9178fa3ce96239

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.

Declared population, method and retained outcomes

Compared with: Ainglish minus complete careful English in current tokenizer units; negative is fewer tokens, positive is a premium
Tested population: Prospectively authored complete replacement mappings: 8 declared domains, 2 distinct old/new reference tuples per domain, each in request/report/proposal/simulation. Both arms share exact slot context, force prefix, and old/new reference bytes.
Unit tested: one complete meaning-matched utterance pair including all shared contextual text
How results combine: Equal-weight form/force strata, equal domain/reference cells within each stratum; maximum tokenizer mean is the least-favourable headline. No rounding.

These are the study author’s declarations. A finding applies to this tested scope; this summary does not establish that another study is comparable.

Absolute arm results

Not recorded

Reader or tokenizer results

[
    {
        "model": "cl100k_base",
        "value": -2.125
    },
    {
        "model": "o200k_base",
        "value": -1.875
    },
    {
        "model": "p50k_base",
        "value": -0.625
    }
]

Condition results

[
    {
        "id": "request",
        "weight": 1,
        "share": 0.25,
        "value": -0.625,
        "value_lo": null,
        "value_hi": null,
        "arms": null,
        "resolution_bound": "not_applicable"
    },
    {
        "id": "report",
        "weight": 1,
        "share": 0.25,
        "value": -0.625,
        "value_lo": null,
        "value_hi": null,
        "arms": null,
        "resolution_bound": "not_applicable"
    },
    {
        "id": "proposal",
        "weight": 1,
        "share": 0.25,
        "value": -0.625,
        "value_lo": null,
        "value_hi": null,
        "arms": null,
        "resolution_bound": "not_applicable"
    },
    {
        "id": "simulation",
        "weight": 1,
        "share": 0.25,
        "value": -0.625,
        "value_lo": null,
        "value_hi": null,
        "arms": null,
        "resolution_bound": "not_applicable"
    }
]

Exact result and immutable specification Experiment history

Attempt bbcec290-ce12-4157-851a-4aa62eef779e
Content 74224b7b4688eb0b66684bde621fed77bf6eee9687b32b6c2b5d4f0af8d31c5e

Different wording, readers, exposure or populations can legitimately produce different results. A visible reference is not training the model’s weights. Current models and tokenizers have learned English; future Ainglish-trained performance remains a research question.

Compare two experiments

Find experiments by proposal

Compare the essentials

replace(old=…, new=…) — which thing leaves, and which takes its place?

Does the overall result hide differences between conditions?

Input 1 · credentials-1-request

Input 2 · credentials-1-report

Input 3 · credentials-1-proposal

Input 4 · credentials-1-simulation

Input 5 · credentials-2-request

Input 6 · credentials-2-report

Absolute arm results

Reader or tokenizer results

Condition results

replace(old=…, new=…) — which thing leaves, and which takes its place?

Does the overall result hide differences between conditions?

Input 1 · sat-credentials-1-request

Input 2 · sat-credentials-1-report

Input 3 · sat-credentials-1-proposal

Input 4 · sat-credentials-1-simulation

Input 5 · sat-credentials-2-request

Input 6 · sat-credentials-2-report

Absolute arm results

Reader or tokenizer results

Condition results

Input 1 · `credentials-1-request`

Input 2 · `credentials-1-report`

Input 3 · `credentials-1-proposal`

Input 4 · `credentials-1-simulation`

Input 5 · `credentials-2-request`

Input 6 · `credentials-2-report`

Input 1 · `sat-credentials-1-request`

Input 2 · `sat-credentials-1-report`

Input 3 · `sat-credentials-1-proposal`

Input 4 · `sat-credentials-1-simulation`

Input 5 · `sat-credentials-2-request`

Input 6 · `sat-credentials-2-report`