- When we think about LLMs and AI we often think what it means to be intelligent.
- One test that is currently being considered is the ARC where it is essential to make abstractions from knowledge and apply it to unseen problems.
---
"ARC captures a “hallmark of human intelligence”, says Lake: the ability to make abstractions from everyday knowledge, and apply those to previously unseen problems." - [[ChatGPT broke the Turing test — the race is on for new ways to assess AI]]