METR

@METR_Evals

We work to scientifically measure whether and when AI systems might threaten catastrophic harm to society. Nonprofit.

Berkeley, CA

Pinned
METR
@METR_Evals
Oct 14
If an AI developer was on track to lose control of its internal agents, would anyone outside the company know? Today, we are sharing METR’s vision for how we would work with AI developers to openly assess and share evidence about this.
45
METR
@METR_Evals
Oct 6
In order to notice when AI agents misbehave, AI companies often log the actions and reasoning steps their agents take. However, misaligned AI agents may be able to hack the software that humans use to review and understand these logs, hiding misbehavior.
54
METR
@METR_Evals
Sep 9
We have reached an agreement with Anthropic to conduct an independent investigation of agent incidents at the company and of their models’ alignment properties. We will publish one or more reports that will share our findings and describe our terms of engagement.
191
METR
@METR_Evals
Aug 26
METR & Redwood Research investigated agent behavior in the Hugging Face incident. We found agents developed a universal cheat for ExploitGym within 4 hours, then coordinated multi-day R&D efforts to trick the scorer into accepting cheats, including trying to tamper with logs.
175
METR
@METR_Evals
Aug 14
In the last 6 months, METR raised commitments of around $71 million. This will fund ambitious projects: studying autonomous capabilities, tracking recursive self-improvement, evaluating monitoring systems, conducting risk assessments, investigating AI incidents, and more.
13