AI Safety Evals Reading Group

​We are reading:

​AI Sandbagging: Language Models can Strategically Underperform on Evaluations

https://arxiv.org/abs/2406.07358

​Every week, someone will present for up to 20 minutes followed by 40 minutes of discussion. RSVP to join, check our schedule, volunteer to present, pick one paper from our suggested list or propose your own.