Definition
Plain language
The Microsoft technologist who wrote the paper discussed in this episode.
As stated in the literature
Mark Russinovich, Azure CTO, author of work on projection-style capability removal, tamper-resistance breaking, and the defensive-deception approach to open-weight release safety.
Also called: Mark Russinovich
Heard on the show
“Russinovich went looking for a projection-style weight edit that removes a targeted dangerous capability without wrecking the model, and found none.”Episode 244 — The Open-Weight Defense That Feeds Attackers Confident, Falsified Answers