AI CEOs are not saying it’s dangerous just to hype their companies
(Inspired by Katja Grace’s post on the same subject. Maybe this is all obvious to this audience,…
Category: Opinion
Preventing extinction from ASI on a $50M yearly budget
ControlAI's mission is to avert the extinction risks posed by superintelligent AI. We believe that…
I used this repo to partially replicate Anthropic’s Emotion Concepts paper in a day
tl;dr This post introduces the traitinterp repo I used to partially replicate Anthropic's Emotion Concepts…
Introducing LinuxArena
We are releasing LinuxArena, a new control setting comprised of 20 software engineering environments. Each environment…
Takes on Automating Alignment
Recently, AI Models have become good at long horizon tasks. In addition, they seem especially…
Finetuning Borges
My newest hobby is fine-tuning a Chinese open-source LLM to generate Pierre Menard, Author of…
How do LLMs generalize when we do training that is intuitively compatible with two off-distribution behaviors?
Authors: Dylan Xu, Alek Westover, Vivek Hebbar, Sebastian Prasanna, Nathan Sheffield, Buck Shlegeris, Julian StastnyThanks to…
A lesson in courage from science camp
The summer before my freshman year of high school, I attended a science-themed summer camp…
Opus 4.7 Part 1: The Model Card
Less than a week after completing coverage of Claude Mythos, here we are again as…