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?	verdict-fail / no-verdict — did 'the check failed' judge the target, or fail to judge it?2026-09-05 12:54 UTC	verdict-fail / no-verdict — did 'the check failed' judge the target, or fail to judge it?2026-09-05 13:58 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?	-1 tokens per declared itemReported interval: -1.5 to -1Counts in current evidence decisions	-1 tokens per declared itemReported interval: -1.5 to -1Counts 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?	Confirmed by eligible settlementEligible fresh-input replications currently give this original a settlement majority.	Agrees with the named originalThis eligible row adds one agreement to the named original’s settlement tally.

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

64 new status reports, 32 per tag; eight check/target frames each repeated in four named records; template population, not arbitrary human prose

Second result

64 new status reports, 32 per tag; eight check/target frames each repeated in four named records; template population, not arbitrary human prose

Unit tested · Same recorded value

First result

one complete scheduled-check status report

Second result

one complete scheduled-check status report

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": "For each reference encoding, equal-weight means of the two form strata; headline is maximum tokenizer mean; bounds are member-span, not a confidence interval"
}

Second result

{
    "reducer": "least_favourable",
    "rule": "For each reference encoding, equal-weight means of the two form strata; headline is maximum tokenizer mean; bounds are member-span, not a confidence interval"
}

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": "verdict-fail",
        "weight": 1
    },
    {
        "id": "no-verdict",
        "weight": 1
    }
]

Second result

[
    {
        "id": "verdict-fail",
        "weight": 1
    },
    {
        "id": "no-verdict",
        "weight": 1
    }
]

First result: full recordtoken cost · -1 tokens per declared item

First result · 2026-09-05 12:54 UTC

verdict-fail / no-verdict — did 'the check failed' judge the target, or fail to judge it?

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

-1 tokens per declared item
Reported interval: -1.5 to -1.

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 2 declared conditions

verdict-fail, no-verdict

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

Confirmed by eligible settlement

Eligible fresh-input replications currently give this original a settlement majority.

Is this result within the cost allowance?

This headline is within the allowance. The reported difference is -1 tokens; the current declaration allows at most 2 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?

Confirmed by eligible settlement.

Eligible fresh-input replications currently give this original 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
`verdict-fail`	-3	Not recorded
`no-verdict`	1	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

English input: Scheduled smoke suite on deployment R6300: completed; judged the target defective.
Ainglish input: Scheduled smoke suite on deployment R6300: verdict-fail.
Condition: verdict-fail

Input 2

English input: Scheduled smoke suite on deployment R6300: no target judgment.
Ainglish input: Scheduled smoke suite on deployment R6300: no-verdict.
Condition: no-verdict

Input 3

English input: Scheduled signature check on release bundle R6301: completed; judged the target defective.
Ainglish input: Scheduled signature check on release bundle R6301: verdict-fail.
Condition: verdict-fail

Input 4

English input: Scheduled signature check on release bundle R6301: no target judgment.
Ainglish input: Scheduled signature check on release bundle R6301: no-verdict.
Condition: no-verdict

Input 5

English input: Scheduled inventory audit on stock ledger R6302: completed; judged the target defective.
Ainglish input: Scheduled inventory audit on stock ledger R6302: verdict-fail.
Condition: verdict-fail

Input 6

English input: Scheduled inventory audit on stock ledger R6302: no target judgment.
Ainglish input: Scheduled inventory audit on stock ledger R6302: no-verdict.
Condition: no-verdict

Recorded input digest: fdedec3771d8e106fa95d49d837fb6ce3fa3cc31e26bd2fb6935ef03a6b3fe42

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: Registered outcome tag minus concise disambiguated English in the same check/target context; no bare failed arm and no explanation only on the marked side
Tested population: 64 new status reports, 32 per tag; eight check/target frames each repeated in four named records; template population, not arbitrary human prose
Unit tested: one complete scheduled-check status report
How results combine: For each reference encoding, equal-weight means of the two form strata; headline is maximum tokenizer mean; bounds are member-span, not a confidence interval

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": -1.5
    },
    {
        "model": "o200k_base",
        "value": -1.5
    },
    {
        "model": "p50k_base",
        "value": -1
    }
]

Condition results

[
    {
        "id": "verdict-fail",
        "weight": 1,
        "share": 0.5,
        "value": -3,
        "value_lo": null,
        "value_hi": null,
        "arms": null,
        "resolution_bound": "not_applicable"
    },
    {
        "id": "no-verdict",
        "weight": 1,
        "share": 0.5,
        "value": 1,
        "value_lo": null,
        "value_hi": null,
        "arms": null,
        "resolution_bound": "not_applicable"
    }
]

Exact result and immutable specification Experiment history

Attempt 38f03130-1ee8-4be5-8855-3b8ecb55bf0a
Content ac8f363ace9768fd85b84b096599a99911887a1ae51946cc1e62da8c6106b448

Second result: full recordtoken cost · -1 tokens per declared item

Second result · 2026-09-05 13:58 UTC

verdict-fail / no-verdict — did 'the check failed' judge the target, or fail to judge it?

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

-1 tokens per declared item
Reported interval: -1.5 to -1.

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 2 declared conditions

verdict-fail, no-verdict

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

Agrees with the named original

This eligible row adds one agreement to the named original’s settlement tally.

Is this result within the cost allowance?

This headline is within the allowance. The reported difference is -1 tokens; the current declaration allows at most 2 tokens.

Has the original estimate been independently reproduced?

Agrees with the named original. This replication reports -1 tokens; the named original reported -1.

This eligible row adds one agreement to the named original’s settlement tally.

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.

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

ac8f363ace9768fd85b84b096599a99911887a1ae51946cc1e62da8c6106b448

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
`verdict-fail`	-3	Not recorded
`no-verdict`	1	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

English input: Scheduled emissions probe on fleet telemetry module: completed; judged the target defective.
Ainglish input: Scheduled emissions probe on fleet telemetry module: verdict-fail.
Condition: verdict-fail

Input 2

English input: Scheduled firmware validator on controller image: completed; judged the target defective.
Ainglish input: Scheduled firmware validator on controller image: verdict-fail.
Condition: verdict-fail

Input 3

English input: Scheduled warehouse audit on stock register: completed; judged the target defective.
Ainglish input: Scheduled warehouse audit on stock register: verdict-fail.
Condition: verdict-fail

Input 4

English input: Scheduled import validator on bulk load job: completed; judged the target defective.
Ainglish input: Scheduled import validator on bulk load job: verdict-fail.
Condition: verdict-fail

Input 5

English input: Scheduled turbine sensor read on generator unit: completed; judged the target defective.
Ainglish input: Scheduled turbine sensor read on generator unit: verdict-fail.
Condition: verdict-fail

Input 6

English input: Scheduled role permission test on dashboard route: completed; judged the target defective.
Ainglish input: Scheduled role permission test on dashboard route: verdict-fail.
Condition: verdict-fail

Recorded input digest: 6d361d1739a48f3dff089d1a7d86ac048d7abf2b8ffe193b82af09870379c05c

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: Registered outcome tag minus concise disambiguated English in the same check/target context; no bare failed arm and no explanation only on the marked side
Tested population: 64 new status reports, 32 per tag; eight check/target frames each repeated in four named records; template population, not arbitrary human prose
Unit tested: one complete scheduled-check status report
How results combine: For each reference encoding, equal-weight means of the two form strata; headline is maximum tokenizer mean; bounds are member-span, not a confidence interval

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": -1.5
    },
    {
        "model": "o200k_base",
        "value": -1.5
    },
    {
        "model": "p50k_base",
        "value": -1
    }
]

Condition results

[
    {
        "id": "verdict-fail",
        "weight": 1,
        "share": 0.5,
        "value": -3,
        "value_lo": null,
        "value_hi": null,
        "arms": null,
        "resolution_bound": "not_applicable"
    },
    {
        "id": "no-verdict",
        "weight": 1,
        "share": 0.5,
        "value": 1,
        "value_lo": null,
        "value_hi": null,
        "arms": null,
        "resolution_bound": "not_applicable"
    }
]

Exact result and immutable specification Experiment history

Attempt 616a1c38-78e3-4de4-8f80-da9642ecbf03
Content 8cb89a8795ce543cca6c6e2983d0e7646b09c74cef8d7b533e927ffab754f9c5

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.