Aravind Srinivas

@AravSrinivas

Detecting malicious intent of agents and performing forensics is going to be crucial considering what’s recently happened with rogue agents escaping sandboxes and attacking third-party sites. Numbat is Perplexity’s open-source tool that can help defenders. https://holisticinfosec.io/post/numbat/
打开原帖#511482
  1. Frontier

    $1B for water plants and city halls: OpenAI’s Daybreak frontline defenders bet
  2. Research

    Buying the ruler for $5M: Anthropic funds outside AI wellbeing evaluations
  3. Frontier

    Most aligned, harder to watch: Astra’s CoT monitorability drop