AI Safety Evals - Paper Reading Club

​We are reading: General Scales Unlock AI Evaluation with Explanatory and Predictive Power

​https://arxiv.org/pdf/2503.06378

​​We’re kicking off a bi-weekly Reading Group on Evals in Safety Context this Spring 2025, and you’re invited!

This is your chance to dive into the cutting-edge ideas shaping the field.

• ​​When: Every week, 12-1pm EST

• ​​Format:

  • ​​Give a 20-minute paper presentation during the paper club

  • ​​40-minute discussion— share your thoughts and questions

• ​​Spring 2025 Theme: Foundations and Critiques—let’s explore the building blocks of evals and question the status quo.

​​Ready to dive in?
Hit us up to RSVP to join or volunteer to present (Pick one paper from our suggested list or bring your own, as long as it fits the theme). See here for the calendar!

​​See you there!