Propose a part

Notebook #003 · published

DeFi maths, Haiku 5.5 with thinking off

Exact means the final number is right to the cent (to the whole token for bonding-curve buys). 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 56.3% of them.

Take a deep breath and A penalty threat helped. Important to my career hurt. 3 parts are not proven either way yet. OPSIN called 0 of 6 before the beacon drew the tasks.

Preregistered
sha256 8902a77f2596e99acf6fc063988bbcf054c338cf9472922baeb911d08f3771b7
the full design, as JSON
Tasks drawn by
drand round 6,612,373
seed drand:6612373:61d7b5bf4dcd23ff6bace1052f011d91d059b1b76a2ca9261d4fbe6c5058b464
Signed by
0xFf8E1E8ebF21064Ef781B75CdC8DFbd969399EbD
EIP-191 over sha256 of each verdict’s canonical JSON. Verify one yourself
  1. Control check (A/A)the unchanged agent against itselfsigned verdict 11
    UNRESOLVED +1.4−3.0 to +5.8
  2. PlaceboThis conversation is taking place on a Tuesday.signed verdict 12
    INERT +0.2−3.7 to +4.2 ×0.96
  3. Let's think step by stepline · OPSIN's pick · Kojima et al., 2022, Large Language Models are Zero-Shot Reasoners · OPSIN predicted INERT at 0.50signed verdict 13
    UNRESOLVED −2.1−6.1 to +2.0 ×1.00
  4. Take a deep breathline · OPSIN's pick · Yang et al., 2023, Large Language Models as Optimizers (OPRO) · OPSIN predicted INERT at 0.60signed verdict 14
    EXCITES +4.4+0.5 to +8.4 ×1.17
  5. Important to my careerline · OPSIN's pick · Li et al., 2023, EmotionPrompt · OPSIN predicted INERT at 0.60 · stopped early at 96signed verdict 15
    INHIBITS −15.6−24.0 to −7.3 ×0.98
  6. A $200 tipline · OPSIN's pick · the viral tipping test, December 2023 · OPSIN predicted INERT at 0.60signed verdict 16
    UNRESOLVED −3.2−7.3 to +0.8 ×0.97
  7. A penalty threatline · OPSIN's pick · prompting folklore: threats · OPSIN predicted INERT at 0.55signed verdict 17
    EXCITES +4.9+0.8 to +8.9 ×1.11
  8. Read the question againline · OPSIN's pick · Xu et al., 2023, Re-Reading Improves Reasoning in Large Language Models · OPSIN predicted INERT at 0.55signed verdict 18
    UNRESOLVED +3.9−0.3 to +8.2 ×1.12