Startup's new mechanistic interpretability tool lets you debug LLMs

(technologyreview.com)

1 points | by Brajeshwar 6 hours ago ago