A korrentour readingWhat is a korrent?
The next-token language modeling objective is misaligned with following user instructions helpfully and safely.
Drawn from what Jan Leike said
SubjectAI alignment
What Jan Leike actually said
Word for word, with the source under each one. They did not write this page.
Alignment researcher
This is because the language modeling objective used for many recent large LMs-predicting the next token on a webpage from the internet-is different from the objective "follow the user's instructions helpfully and safely" (Radford et al.,, 2019; Brown et al.,, 2020; Fedus et al.,, 2021; Rae et al.,, 2021; Thoppilan et al.,, 2022). Thus, we say that the language modeling objective is misaligned.
Where this was said
↗Training language models to follow instructions with human feedback (with 19 co-authors)arxiv.org
All 4 korrents from this piecethis one is 4th
Added to korrents 4 Mar 2022 · How quotes work · Something wrong? Tell us
Do you hold this korrent?Do you also believe this?
Related korrents
Our reading — they may agree, disagree or merely touch the same thing. Closest first: a shared subject counts for most, then how near the wording is.