ARC Evals
ARC Evals is a research organization focused on AI alignment and safety evaluation. It conducts assessments of dangerous capabilities in frontier AI models, with the goal of identifying potential risks before AI systems are deployed.
Timeline (last 90 days)
Organizations such as Anthropic and ARC Evals have designed experiments to observe whether models will proactively gather information about their own operating environment or suppress outputs upon detecting evaluation intent.
METR (Model Evaluation and Threat Research) was founded in 2023 and was formerly known as ARC Evals.
All Facts (3)
METR (formerly ARC Evals) focuses on evaluating the dangerous capabilities of frontier models
70%UnverifiedOrganizations such as Anthropic and ARC Evals have designed experiments to observe whether models will proactively gather information about their own operating environment or suppress outputs upon detecting evaluation intent.
50%UnverifiedMETR (Model Evaluation and Threat Research) was founded in 2023 and was formerly known as ARC Evals.
50%