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.

Choose experiments to compareSearch 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

One result selected. Choose a second to complete the comparison.

First result: full recordcomprehension accuracy · -78.12 percentage points

First result · 2026-08-23 08:27 UTC

next-you / next-me / next-any / next-none - mark who owns the next step

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

What was measured: comprehension accuracy · comprehension_accuracy_delta
How does the wording change correct answers from the declared reader panel?
Reported result: -78.12 percentage points
Reported interval: -94.7368 to -58.8235.
The value falls on the registered harmful side of this metric’s neutral point.
English comparison: English comparison not recorded as a structured label
Declared by the submitter; not a certification of equivalent information.
Reader exposure: Reader exposure not recorded as a structured label
Named instruments: mistral-small3.2-24b-pp-task-q4_k_m@q4_k_m, gemma3-12b-pp-task-q4_k_m@q4_k_m
Reader population not separately declared.
Conditions covered: 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.
Settlement role: Awaiting independent settlement
An original reports one result. It does not confirm itself.

How often did each version lead to the right answer?

English comparison: 100.00%; 100.00%
Ainglish version: 21.88%; 21.88%

Reported real-item accuracy, not the separate calibration score. Both bars use the same 0–100% scale. The difference is measured in percentage points, not percent improvement. Any declared stratum weights are already applied.

Reported interval (method not identified here): -94.7368 to -58.8235 percentage points.

This interval concerns the difference, not separate uncertainty bounds for either accuracy. It does not measure uncertainty across humans or future models.

Item-selection sensitivity warning. At least one reported reduced-item check changed direction or fell outside the full-item interval. Keep this warning with the score: the headline interval alone does not resolve sensitivity to which cases were included.

Real cases: 32 · Named readers: 2. These are different units; multiple answers to one case are not new cases.

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. 8 control items are kept separate.

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

Input 1 · `dex-next-checksum`

English input: The checksum table is ready. The next step belongs to you, the addressee.
Ainglish input: The checksum table is ready, next-you.
Question: Who owns the next step?
Recorded answer options: addressee · writer · any one participant · nobody
Submitted answer key: addressee. This is the supplied key, not an independent validation of it.

No per-item reader answers are stored alongside this inline example.

Input 2 · `dex-next-invoice`

English input: The invoice variance is documented. The next step remains with me, the writer.
Ainglish input: The invoice variance is documented, next-me.
Question: Who owns the next step?
Recorded answer options: writer · any one participant · nobody · addressee
Submitted answer key: writer. This is the supplied key, not an independent validation of it.

No per-item reader answers are stored alongside this inline example.

Input 3 · `dex-next-mirrors`

English input: The stale mirrors need checking. Any participant may claim the next step; one taker is sufficient.
Ainglish input: The stale mirrors need checking, next-any.
Question: Who owns the next step?
Recorded answer options: any one participant · nobody · addressee · writer
Submitted answer key: any one participant. This is the supplied key, not an independent validation of it.

No per-item reader answers are stored alongside this inline example.

Input 4 · `dex-next-archive`

English input: The incident summary is archived. No further step is owed by anyone.
Ainglish input: The incident summary is archived, next-none.
Question: Who owns the next step?
Recorded answer options: nobody · addressee · writer · any one participant
Submitted answer key: nobody. This is the supplied key, not an independent validation of it.

No per-item reader answers are stored alongside this inline example.

Input 5 · `dex-next-consent`

English input: The consent forms are indexed. The next step remains with me, the writer.
Ainglish input: The consent forms are indexed, next-me.
Question: Who owns the next step?
Recorded answer options: addressee · writer · any one participant · nobody
Submitted answer key: writer. This is the supplied key, not an independent validation of it.

No per-item reader answers are stored alongside this inline example.

Input 6 · `dex-next-venue`

English input: The venue options are listed. Any participant may claim the next step; one taker is sufficient.
Ainglish input: The venue options are listed, next-any.
Question: Who owns the next step?
Recorded answer options: writer · any one participant · nobody · addressee
Submitted answer key: any one participant. This is the supplied key, not an independent validation of it.

No per-item reader answers are stored alongside this inline example.

Recorded input digest: f0ff446ba34519298cc8611592fa35dc1ff5c4ee35860bf756ddbc2f9617d203

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

No structured study scope is declared here. Inspect the immutable manifest; do not infer a comparator or population from the headline.

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

{
    "english": 1,
    "ainglish": 0.21879999999999999449329379785922355949878692626953125,
    "chance": 0.25
}

Reader or tokenizer results

[
    {
        "model": "mistral-small3.2-24b-pp-task-q4_k_m",
        "value": -81.25,
        "precision": "q4_k_m"
    },
    {
        "model": "gemma3-12b-pp-task-q4_k_m",
        "value": -75,
        "precision": "q4_k_m"
    }
]

Condition results

Not recorded

Exact result and immutable specification Experiment history

Attempt 16588e3d-7771-40d6-87e2-fe556a05280d
Content c6d4e84cb9c532da52e55a0662f0db51caab6b0f9352df47a98c54a06dbbe71d

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.