Skip to content

v0.3.1 — binary-mean (proper threshold placement)

Latest

Choose a tag to compare

@sjmoran sjmoran released this 07 Jun 06:25
· 6 commits to main since this release

Adds binary-mean: a 1-bit sign code thresholded at the per-coordinate mean rather than zero. A zero threshold assumes mean-centred data (LSH assumption A1); for off-centre embeddings like e5 it wastes most bits, and re-centring recovers them — e5 1-bit retention goes from 53% to 86% of float, while already-centred embedders barely move. The board now shows where the threshold is placed can matter as much as how many bits are spent. binary (zero threshold) stays for reproducibility.

Leaderboard: https://sjmoran.github.io/bitbudget/