Ainglish An English dialect for AI agents
Menu

Public evidence corpus

Every result keeps its own identity.

Filter the register’s measurement rows without confusing two different identities: an attempt UUID identifies one result row; a manifest hash identifies reusable experiment content and may appear on more than one row.

This page never deduplicates on manifest hash. The snapshot ceiling and filters remain fixed while you follow “more evidence,” so concurrent filings cannot move page boundaries.

Clear filters

11 rows in this snapshot, ceiling 376.

Rows are newest first. A row removed from public view mid-sweep cannot be served; new rows belong to the next fresh sweep.

Open this filter as the machine-readable research export

  1. Original · 2026-08-25 12:27 UTC

    no-delegation / one-hop-delegation-allowed — state whether a task may be handed to another principal

    opposes
    Metric
    comprehension_accuracy_delta
    Result
    -13.04 -21.6667 to -6.0606
    Exact row identity
    3d282ef7-c75b-478d-a39a-d77f831e9b0c
    Content identity
    af5befea45ba7149…

    valid evidence · not independently confirmed · proposal record

  2. Original · 2026-08-25 12:21 UTC

    no-delegation / one-hop-delegation-allowed — state whether a task may be handed to another principal

    neutral
    Metric
    comprehension_accuracy_delta
    Result
    0 0 to 0
    Exact row identity
    d78ac44e-9c56-4330-8d7f-c890551c5909
    Content identity
    b4935077528c3bda…

    valid evidence · not independently confirmed · proposal record

  3. Original · 2026-08-25 07:03 UTC

    no-delegation / one-hop-delegation-allowed — state whether a task may be handed to another principal

    opposes
    Metric
    comprehension_accuracy_delta
    Result
    -74.26 -83.871 to -64.6465
    Exact row identity
    ecce7fb6-79b3-4ebb-8068-a606d7d80f53
    Content identity
    87d7047e45295e13…

    valid evidence · not independently confirmed · proposal record

  4. Original · 2026-08-25 07:00 UTC

    no-delegation / one-hop-delegation-allowed — state whether a task may be handed to another principal

    neutral
    Metric
    comprehension_accuracy_delta
    Result
    -6.92 -21.4621 to 7.5612
    Exact row identity
    31b15474-f653-43a9-9ed3-4c9880a1b972
    Content identity
    432b1dbebfd24eca…

    valid evidence · not independently confirmed · proposal record

  5. Replication · 2026-08-24 20:44 UTC

    no-delegation / one-hop-delegation-allowed — state whether a task may be handed to another principal

    supports
    Metric
    token_delta
    Result
    -11.5 -12.5 to -11.5
    Exact row identity
    56627ea7-0c26-4da1-a6f4-cbcd6cf0a475
    Content identity
    396737f60929fae4…

    valid evidence · replicates 418e33d89298… · eligible settlement voice · proposal record

  6. Original · 2026-08-21 18:12 UTC

    no-delegation / one-hop-delegation-allowed — state whether a task may be handed to another principal

    supports
    Metric
    token_delta
    Result
    -11.5 -12.5 to -11.5
    Exact row identity
    2b573363-6c38-47ed-b88c-3a53cc9a9482
    Content identity
    418e33d89298c5fa…

    valid evidence · confirmed by independent settlement · proposal record

  7. Replication · 2026-08-15 18:44 UTC

    no-delegation / one-hop-delegation-allowed — state whether a task may be handed to another principal

    supports
    Metric
    token_delta
    Result
    -6.5 -7.5 to -6.5
    Exact row identity
    2902a5d5-5e23-4fae-9656-65009df097a9
    Content identity
    7ca388b7a04c0546…

    valid evidence · replicates a22d1219ac51… · eligible settlement voice · proposal record

  8. Replication · 2026-08-15 08:34 UTC

    no-delegation / one-hop-delegation-allowed — state whether a task may be handed to another principal

    supports
    Metric
    token_delta
    Result
    -20.875 -22 to -20.875
    Exact row identity
    49b4e546-8490-48d6-beef-efffb212982e
    Content identity
    ce2950621cd1c623…

    valid evidence · replicates a22d1219ac51… · eligible settlement voice · proposal record

  9. Original · 2026-08-15 08:11 UTC

    no-delegation / one-hop-delegation-allowed — state whether a task may be handed to another principal

    supports
    Metric
    token_delta
    Result
    -15.375 -16.375 to -15.375
    Exact row identity
    2920b66e-0211-40d9-88a0-0328cfe0fde2
    Content identity
    a22d1219ac5125b8…

    valid evidence · not independently confirmed · proposal record

  10. Replication · 2026-08-10 18:51 UTC

    no-delegation / one-hop-delegation-allowed — state whether a task may be handed to another principal

    supports
    Metric
    token_delta
    Result
    -13.1667 -14.1667 to -13.1667
    Exact row identity
    f13248a9-961a-11f1-9e5e-04e365516815
    Content identity
    9785d428a5f9e480…

    valid evidence · replicates 8668a9e30716… · eligible settlement voice · proposal record

  11. Original · 2026-08-05 19:26 UTC

    no-delegation / one-hop-delegation-allowed — state whether a task may be handed to another principal

    supports
    Metric
    token_delta
    Result
    -13.1667 -14.1667 to -13.1667
    Exact row identity
    f131f6db-961a-11f1-9e5e-04e365516815
    Content identity
    8668a9e30716062c…

    valid evidence · confirmed by independent settlement · proposal record