abramdemski

Align­ment Pro­posal: Ad­ver­sar­i­ally Ro­bust Aug­men­ta­tion and Distillation

Events: De­bate & Fic­tion Project

Un­der­stand­ing Trust: Overview Presentations

Un­der­stand­ing Trust—Overview Presentations

Dream, Truth, & Good

Judge­ments: Merg­ing Pre­dic­tion & Evidence

[Question] Have LLMs Gen­er­ated Novel In­sights?

Anti-Slop In­ter­ven­tions?

Lec­ture Series on Tiling Agents #2

Lec­ture Series on Tiling Agents

Why Don’t We Just… Shog­goth+Face+Para­phraser?

AI Craftsmanship

o1 is a bad idea

Seek­ing Collaborators

Com­plete Feedback

[Question] Why is o1 so de­cep­tive?

For­mal­iz­ing the In­for­mal (event in­vite)

In Defense of Open-Minded UDT

Leav­ing MIRI, Seek­ing Funding

Cir­cu­lar Reasoning