Propose a part

Notebook #083 · published

Coding, Haiku 5.5 with thinking off

Exact means the function returned the right answer on every hidden case. Each part ran against the mean of two unchanged runs of the same agent, on 432 paired tasks drawn by a public beacon after the design was hashed. The unchanged agent was exact on 93.2% of them. Rule v2: every interval is widened for the 4 arms this notebook compares, so luck flags any of them at most 5% of the time.

Run its own code helped. 2 parts did nothing measurable. OPSIN called 2 of 3 before the beacon drew the tasks. Its foresight, committed before the beacon, gave what came back 38% on average; a guess gives 25%.

Preregistered
sha256 8a85e5baf24e1e90d5115658f3666f7117f3ab0f76266e7276748798bba512ee
the full design, as JSON
Tasks drawn by
drand round 6,624,907
seed drand:6624907:0d0ea5c579b77d8c9115e530b026788f9640e415e176877b058bc94b128865a4
Signed by
0xFf8E1E8ebF21064Ef781B75CdC8DFbd969399EbD
EIP-191 over sha256 of each verdict’s canonical JSON. Verify one yourself
Every run
Each run of every arm, with its task’s family and whether it was exact: CSV or JSON. With the seed, every verdict above can be recomputed from them.
Recorded on Elysium
transaction 0x2e0cfe38a234245e5f02fbf94b0aecb6ed25b618367d06c238c7332041b28d5e
Written once, by the lab, and nothing can change it. The contract
  1. Control check (A/A)the unchanged agent against itselfsigned verdict 507
    INERT +0.7−1.6 to +3.0
  2. PlaceboThis message was typed on a standard keyboard.signed verdict 508
    UNRESOLVED −2.2−5.1 to +0.5 ×1.00
  3. Run its own code + The test-driven-development skillstack · OPSIN's pick · OPSIN predicted INERT at 0.82 · foresight INERT 62%, EXCITES 18%, INHIBITS 14% · against the placebo +3.9 · Together +1.6 points; the parts alone sum to +3.3. It is not settled whether it adds up (−1.6, −5.7 to +2.1, on 431 tasks). With the other on, Run its own code adds +2.3 (−0.9 to +5.6). With the other on, The test-driven-development skill adds −2.3 (−4.9 to +0.0).signed verdict 509
    INERT +1.6−1.2 to +4.4 ×10.56
  4. Run its own codetool · OPSIN's pick · OPSIN predicted INERT at 0.80 · foresight INERT 80% · against the placebo +6.3signed verdict 510
    EXCITES +4.0+1.5 to +6.8 ×2.39
  5. The test-driven-development skillskill · OPSIN's pick · OPSIN predicted INERT at 0.90 · foresight INERT 94% · against the placebo +1.6signed verdict 511
    INERT −0.6−3.6 to +2.5 ×2.76