WORKS / STUDYFIVE DIRECTIONS

A room for
work in motion.

Five ways to hold the same growing collection: as a bench, a notebook, an archive, a decision map, or a changelog.

04 / DECISION MAP

Make the change visible.

urge → test → readout → revision

Works is where a curiosity becomes a method, then leaves behind a trace of what changed.

  1. 01STARTED WITH

    A pattern in five years of notes.

    One long sequence of Chinese diary entries made a small question feel worth measuring.

    Read the field note
  2. 02PUT TO TEST

    Build a benchmark around evidence.

    DiaryBench asks whether a model can track nearby emotional states without over-reading the record.

    Open the experiment
  3. 03READ THE SIGNAL

    Keep the score, hide the source.

    Aggregate Jaccard scores make comparison possible while the diary and scoring key stay offline.

    See the public readout
  4. 04RELATED PRACTICE

    Carry the method into another domain.

    The benchmark is one case; trading reviews keep the same habit of stating a thesis, recording a result, and revising the next move.

    Read a related practice
03 / PUBLIC READOUT

DiaryBench

2026 / 09-28 · set 50 · aggregate only

TOP 3 / JACCARD
Diagram of DiaryBench's private diary input and model readout
Evidence without exposure.
  1. GPT-6 Astra64.83
  2. GPT-6 Sol EH63.06
  3. Grok 4.655.10
Open the full board
OTHER SIGNALS

The same question, in other materials.