Cognition and Behavior · Human Capacity
Sparks of Artificial General Intelligence: Early experiments with GPT-4
Summary
Microsoft Research's early evaluation of GPT-4, arguing the model shows the first sparks of general intelligence across a surprisingly wide range of tasks (math, coding, vision, medicine, law, reading minds). Influential and hotly contested; the paper's framing shaped how the public talks about what large language models can do.
Topics:
Cognition, Work