The AI Fix cover image

Claude Code, the rise of AI agents, and the AI that called its human

The AI Fix

00:00

Disentangling safety with subspace orthogonalization

Discussion of the researchers' fix using sparse autoencoders and subspace orthogonalization to partially protect safety features.

Play episode from 36:35
Transcript

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app