IIT Delhi x IBM course on ‘Responsible LLMs’ begins
after their popular in-depth course on LLMs, this one covers watermarking, uncertainty, jailbreaks, interpretability, bias, etc from basics to depth
i’m attending these lectures live from an IIT D.
YT:
youtu.be/p8T48sqIGao
"Open source is dangerous"
This is a story Frontier Labs has been selling for a while.
but recently, a pre-released OpenAI model (closed-source) launched an autonomous cyberattack
and an open model (GLM 5.2) came to the rescue.