AnalysisAI ModelsOctober 22, 2026

Goodfire's Matthew Kowal talks interpretability across model design

Read original source →youtube.com

Goodfire MTS Matthew Kowal argues interpretability should be framed through model affordances — the actions a system makes possible — giving researchers and engineers new ways to intervene throughout the model design pipeline. The Cohere-hosted talk splits into a first part on that framing.

People · Matthew Kowal

1 source

More stories today

Open the live feed