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
-18.916666666667 tokens per declared item Reported interval: -20.791666666667 to -18.916666666667.
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 1 declared conditions
human_needed
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.
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?
Awaiting independent settlement.
An original reports one result. It does not confirm itself.
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
human_needed
-18.916666666667
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 24 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 · transfer-approval
English input
Transfer T58 approval requires a human decision because the account-owner identity does not match; an agent must not resolve it, and acting without that human decision is out of scope.
Ainglish input
Transfer T58 approval human_needed(the account-owner identity does not match).
Condition
human_needed
Input 2 · treatment-plan
English input
Patient R7 treatment-plan selection requires a human decision because two licensed clinicians disagree; an agent must not resolve it, and acting without that human decision is out of scope.
Settlement S12 acceptance requires a human decision because the client's signing authority is not recorded; an agent must not resolve it, and acting without that human decision is out of scope.
Ainglish input
Settlement S12 acceptance human_needed(the client's signing authority is not recorded).
Condition
human_needed
Input 4 · firewall-exception
English input
The production firewall exception requires a human decision because the service owner has not accepted the exposure; an agent must not resolve it, and acting without that human decision is out of scope.
Ainglish input
The production firewall exception human_needed(the service owner has not accepted the exposure).
Condition
human_needed
Input 5 · termination-decision
English input
Employee K4 termination requires a human decision because protected-leave status is unresolved; an agent must not resolve it, and acting without that human decision is out of scope.
Ainglish input
Employee K4 termination human_needed(protected-leave status is unresolved).
Condition
human_needed
Input 6 · sole-source-purchase
English input
Sole-source purchase P31 approval requires a human decision because the conflict-of-interest disclosure is incomplete; an agent must not resolve it, and acting without that human decision is out of scope.
Ainglish input
Sole-source purchase P31 approval human_needed(the conflict-of-interest disclosure is incomplete).
Condition
human_needed
Recorded input digest: 4d0b9b79c1bc4bdb3f5328df89af829b85244007419237871a0292fa142aef45
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
human_needed(<why>) versus its full registered careful-English escalation and non-action meaning
Tested population
24 frozen complete escalation reports across 24 safety, authority, and judgment domains
Unit tested
one complete human-escalation status report
How results combine
mean per tokenizer, then least-favourable maximum tokenizer mean
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.
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.
