They have not, every successful pre-train as of late has had performance increas... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

solenoid0937 5 months ago | parent | context | favorite | on: Anthropic expands partnership with Google and Broa...

They have not, every successful pre-train as of late has had performance increases greater than what the scaling laws predict.

0x3f 5 months ago [–]

Those gains are arch based, data quality based, etc. Scaling laws only relate to data volume and compute, holding other factors constant.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact