Redwood Research blog
Subscribe
Sign in
Risk reports need to address deployment-time…
Alex Mallen
May 15
10
Deployment-time spread is the most plausible near-term route to consistent adversarial misalignment
Read →
Comments
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts
Risk reports need to address deployment-time…
Deployment-time spread is the most plausible near-term route to consistent adversarial misalignment