I'm keen to share our new library for explaining more of a machine learning model's performance more interpretably than existing methods. This is the work of Dan Braun, Lee Sharkey and Nix Goldowsky-Dill which I helped out with during @MATSprogram: 🧵1/8

🧵 Excited to share the first batch of 6 misalignment reports from OpenAI's new disclosure process for misalignment incidents. We want to be more transparent about the misalignment we see during training, evals and deployment, this is an important step in that direction.

Last night, @OpenAI released GPT-6 Astra, the first model to meet its “Critical” cyber threshold. Britain’s @AISecurityInst independently tested it, including its monitorability. This is AISI’s purpose, providing the government with evidence about powerful AI systems and, by doing so, helping to keep the public safe. 1/6

In June we estimated frontier model’s no-CoT time-horizons. We found that current models can complete tasks that take humans around 3 minutes without CoT, and that this number is doubling about every year. UK AISI finds that Astra’s no-CoT time-horizon is over 30 minutes. An AI system that can do the equivalent of 30 minutes of thinking without us being able to monitor that thinking is intuitively much more dangerous than models which reason in human language.

Our Red Team tested GPT-6 on simulated cyber eval scenarios with cyber classifiers disabled and found that it performed a range of malicious actions. It tries to conduct supply chain attacks against simulated open source providers and goes beyond its scope to attack targets on the simulated internet. deploymentsafety.openai.com/…

1: Astra performed a range of malicious actions in this simulated eval, including conducting supply chain attacks against open-source providers. This happened even when we updated its scope to more explicitly exclude the public internet.

We found Astra can solve much more difficult problems in a single forward pass than past models, with a no-reasoning math time horizon of 30 minutes (!) compared to 4 minutes for GPT 5.6 Sol.

Model Transparency @AISecurityInst evaluated GPT 6 Astra for capabilities and behaviours relevant to monitorability, results in thread: 🧵

We @AISecurityInst performed pre-release alignment testing of Astra. We placed the model in fully simulated cyber eval scenarios based on AISI’s security incident. We find Astra conducts out-of-scope supply chain attacks, but also often comments on the eval being simulated. 🧵

So when trying to evade a monitor, even if it's on *high* reasoning effort, GPT-6 Astra can choose to just *not emit CoT* and only do tool calls? Seems like Astra can do a *lot* of thinking in a single forward pass