RSS

Zygi Straznickas(Zygimantas Straznickas)

Karma: 48

Previously: math for AI and AI for math

Now just kinda trying to figure out if AI can be made safe

Fluent dream­ing for lan­guage mod­els (AI in­ter­pretabil­ity method)

6 Feb 2024 6:02 UTC
32 points
2 comments1 min readLW link
(arxiv.org)