A new blog post argues attribution should follow the geometry used to train the model
aryaman2020 · x · 2026-07-23
What the post says
This post links to a blog post called Attribution in the Optimizer's Geometry. The author argues that parameter attribution should use the natural metric induced by the optimizer, and says there is evidence that attribution works best when done in the geometry in which the model was trained.
In short, the piece is about matching the attribution method to the training geometry rather than treating all parameter directions as equivalent.
More from Research
- Kimi K3 finds 32 of 86 vulnerabilities for just $19.37 in benchmark run — zeeg · 2026-07-23
- A chart comparing AI detectors says most are barely better than a coin toss — burkov · 2026-07-23
- CoLT teaches multimodal models to think in latent steps and cuts inference 10.1x — jiqizhixin · 2026-07-23
- AI agent finds a refutation to a 30-year-old graph theory conjecture in Slack — veggie_eric · 2026-07-23
- Duke study says better memory, not a bigger model, lifted ARC-AGI-3 scores by 18 points — imjustnewatai · 2026-07-23
- Trace builds a reproducible RLVR environment for 11 visual reasoning domains — Md Tanvirul Alam · 2026-07-23