Models and meaning

Garden theme 02

How do models turn observations into useful representations, and where do those representations stop?

Published

Can padding change audio-classification metrics?

A completed DistilHuBERT study comparing ordinary and masked positional BatchNorm under minimal and additional padding.

Complete study

Note · 2026

Published

A five-seed DistilHuBERT experiment asks whether a model can learn around BatchNorm statistics contaminated by padding.

Article · 2026

Published

Evaluate a local agent by the useful result after preparation, review, correction, and unsuccessful attempts.

Note · 2026

Published

A psychology-inspired experiment inside Gemma: strong category responses, a few useful surprises, and a hypothesis that did not survive its controls.

Article · 2026

Published

Explore the recorded Gemma 3 1B architecture and 6,912 MLP units across 26 layers.

Project · 2026

Published

A bounded local worker for repository exploration, first-pass reviews, diagnostics, and candidate patches.

Project · 2026

Published

A small hands-on path into the mechanics behind a model.

Article · 2025

Published

Text data visualization and analysis of Netflix titles.

Article · 2023

Published

An experiment in NLP and text classification using football news.

Article · 2023

Published

An experiment with Elastic-net regression and decision-tree boosting on FIFA 23 player ratings.

Article · 2023