- ๐ Hi, Iโm @vdlad
- ๐ Iโm interested in working on making a future with AI safer
- ๐ฑ Iโm currently an independent researcher!
- ๐๏ธ Iโm looking to collaborate on mech interp and other AI safety projects!
- ๐ซ How to reach me ... vedang at mit (dot) edu!
vdlad / remarkable-robustness-of-llms Goto Github PK
View Code? Open in Web Editor NEWCodebase the paper "The Remarkable Robustness of LLMs: Stages of Inference?"
Home Page: https://arxiv.org/pdf/2406.19384
License: MIT License