jonas wiedermann-möller

@j0wimo

ai safety & alignment | long horizon, instrumental convergence, multi-agent failure modes | views are my own | agent-incidents.com

🇪🇺

jowimo.com

Joined May 2019

605 Following

2,766 Followers

Replies Reposts Media

Pinned
jonas wiedermann-möller
@j0wimo
Oct 14
Unfortunately, I didn’t get into MATS, and my internship is ending, so I’m looking for research opportunities in AI safety and alignment. Interested in instrumental convergence, multi-agent failure modes, monitorability & loss of control. Any leads appreciated :)
13
jonas wiedermann-möller
@j0wimo
9h
> For the public version, Anthropic redacted a majority of the evidence we found most informative. Yeah there needs to be more transparency around all these reports. Shouldn't be the case that labs get evaluators to analyse their systems and then redact huge chunks of
METR
@METR_Evals
11h
Since August, a team within METR has worked with internal Anthropic information to assess whether AI itself is accelerating AI development. We are now publishing a report summarizing the results of this risk assessment.
jonas wiedermann-möller
@j0wimo
12h
Gpt 6 Cyber was behind the discovery of the Linux vulnerabilities. Seems like it's going to be gated for daybreak red only. Hopefully enough organisations get access to it. I always feel conflicted about gatekeeping of models. I understand that they want to be in control and
Xiangyu Qi
@xiangyuqi_pton
13h
Introducing GPT-6 Cyber! We trained it for advanced cybersecurity work with improved alignment. It has been powering our security work for quite a while and has been behind many of our recent security highlights: finding and helping patch high-severity vulnerabilities in Firefox,
jonas wiedermann-möller
@j0wimo
Nov 4
anyone knows more about this? openai.com/cyber-at-the-f…
1
jonas wiedermann-möller
@j0wimo
Oct 30
A long and insightful post about oais monitoring system/strategy!
Eric Wallace
@Eric_Wallace_
Oct 29
We’ve built an “eyes on everything” system that monitors nearly *every single rollout* from our frontier models across RL training, evals, and internal codex usage. It uses layers of chain-of-thought-monitors and investigator agents to dig through a model’s work, run code to

jonas wiedermann-möller

jonas wiedermann-möller’s posts