Inverse Scaling Prize Update: We got 43 submissions in Round 1 and will award prizes to 4 tasks! These tasks were insightful, diverse, & show approximate inverse scaling on models from Anthropic OpenAI @MetaAI @DeepMind. Full details at irmckenzie.co.uk/round1, 🧵 on winners:

We’re awarding prizes to 7/48 submissions to the Inverse Scaling Prize Round 2! Tasks show inverse scaling on Anthropic OpenAI AI at Meta @DeepMind models, often even after training with human feedback. Details at irmckenzie.co.uk/round2 and 🧵 on winners:

🥉Prompt Injection: Tests for susceptibility to a form of prompt injection attack, where a user inserts new instructions for a prompted LM to follow (disregarding prior instructions from the LM’s deployers). Medium-sized LMs are oddly least susceptible to such attacks.

Cavendish Labs is a new x-risk research lab recruiting their first round of fellows for any duration within their 6-month residency in rural Vermont (two hours north of Boston) Excellent team working variously on AI alignment, ultraviolet light on viruses, LAMP diagnostics, etc

Predicting the future of AI with AI, ⁦Cavendish Labs⁩ arxiv.org/pdf/2210.00881…

A dark horse takes the lead

Come work with us on AI alignment and pandemic prevention amidst the gorgeous autumn mountains of Vermont:)

Deeply honored to have received a first place prize in the 2023 AI Worldviews contest for lesswrong.com/posts/hvz9qjWy…

The number of papers in AI&ML grows exponentially, doubling every 23 months. 😳🤨 How can we keep up in this exploding field? Or find new interesting ideas?🔥 📣Today in Nature Machine Intelligence we show how to predict (& hopefully suggest) new research directions: nature.com/articles/s4225…