back
Get SIGNAL/NOISE in your inbox daily

MIT and MIT-IBM Watson AI Lab researchers have developed a universal guide for estimating how large language models (LLMs) will perform based on smaller models in the same family.