Euodia Dodd

Fellow @MATSProgram working on AI Control
PhD student: @ImperialCollege, AI Security & Privacy
Prev: AI @GoldmanSachs

[Euodia Dodd @DoddEuodia]

📢 We dropped another paper! And this time we're backdooring voice agents to send $5000 every time Judi Dench speaks 🗣

[Nataša Krčo @povrcka]

🎙️ New preprint: You Are the Trigger: Speaker-Activated Backdoors
We introduce a backdoor attack on speech models where the trigger is a target speaker’s voice. The model behaves normally for everyone else, but switches to attacker-chosen behavior when it hears the target. 1/

[Euodia Dodd @DoddEuodia]

New paper alert 🚨

We build a model organism for finding security vulnerabilities in code which selectively withholds hard vulnerabilities, and is more resistant to standard elicitation methods like pirate training and weak-to-strong generalization than prior MOs!

[Lennart Finke @LennartFinke]

New paper out of @MaryPhuong10's @MATSprogram Stream, with @hrdkbhatnagar and @DoddEuodia! We made a testbed for eliciting latent knowledge in a code vulnerability scanning task. See below how we made the sandbagging model organism robust to both prompting and training! ( 1/14 )

[Lennart Finke @LennartFinke]

New paper out of @MaryPhuong10's @MATSprogram Stream, with @hrdkbhatnagar and @DoddEuodia! We made a testbed for eliciting latent knowledge in a code vulnerability scanning task. See below how we made the sandbagging model organism robust to both prompting and training! ( 1/14 )

[Igor Shilov @_igorshilov]

New Anthropic research!

We study how to train models so that high-risk capabilities live in a small, separate set of parameters, allowing clean capability removal when needed – for example in CBRN or cybersecurity domains.

[Igor Shilov @_igorshilov]

Check out our website for more info: https://t.co/vzpvDwSIsR

arxiv: https://t.co/9orYFSkO5f

See you in Seattle!

And thanks to my amazing co-authors: @josephbpollock, @DoddEuodia and @yvesalexandre

[Yves-A. de Montjoye @yvesalexandre]

Join us at Imperial College for an exciting event on the future of privacy in machine learning! 🔒🤖 Keynotes and lightning talks 🗓️ Date: Nov 7 @ 6pm📍 Imperial College London