Notes from the catalog
Reviews are the raw material; the journal is where we step back and read them in aggregate: what's moving, how the scoring works, and what the corpus keeps telling us.
No entries published yet. Here's what this space will cover:
- Movers
When a model's scores shift meaningfully (a release, a regression, a sudden run of reviews), we write up what changed.
- Methodology
How scores are aggregated, why we weight what we weight, and the changes we make to keep the rankings honest.
- From the corpus
What the full body of reviews says in aggregate: where models cluster, which tasks divide them, what reviewers keep noticing.
In the meantime, the reviews themselves are the story.
Read recent reviews