Redwood Research
1 open role
description
Redwood Research is an AI safety research organisation. Their first big project attempted to make sure language models (like GPT-3) produce output following certain rules with very high probability, in order to address failure modes too rare to show up in standard training.
Learn more
80,000 Hours links
External content
Some of Redwood's latest research
Research
An 'Ask Me Anything' with Redwood Research staff members (from late 2021)
FAQ