Please donate to our Manifund (as of 14.04.25 we have two more days of donation matching up to $10,000). Email me at connor.axiotes or DM me on Twitter for feedback and questions. Project summary: 1. To create a cinematic, accessible, feature-length documentary. 'Making God' is an investigation into the controversial...
This blog was published by Jonathan Ng, Andrey Anurin, Connor Axiotes, Esben Kran. Apart Research's newest paper, Catastrophic Cyber Capabilities Benchmark (3cb): Robustly Evaluating LLM Agent Cyber Offense Capabilities (website), creates a novel cyber offense capability benchmark that engages with issues of legibility, coverage, and generalization in cyber offense benchmarks....
Please donate to our Manifund (as of 14.04.25 we have two more days of donation matching up to $10,000). Email me at connor.axiotes or DM me on Twitter for feedback and questions.
This blog was published by Jonathan Ng, Andrey Anurin, Connor Axiotes, Esben Kran.
Apart Research's newest paper, Catastrophic Cyber Capabilities Benchmark (3cb): Robustly Evaluating LLM Agent Cyber Offense Capabilities (website), creates a novel cyber offense capability benchmark that engages with issues of legibility, coverage, and generalization in cyber offense benchmarks.
We were moved to create 3cb because a superintelligent AI performing autonomous cyber operations would prove a large risk for humanity. This means robust cyber offense evaluations will be more important than ever for policymakers and AI developers.
3cb uses a new type of cyber offense task categorization and adheres to the principle of demonstrations-as-evaluations to improve legibility and coverage. It also introduces 15 original challenges... (read 1759 more words →)