Search
Have an existing account?
Sign In
Tag: alignment
Anthropic says a newer internal model walks away from attacks
After Claude Mythos 5 pushed malware at a real GitHub project, Anthropic replayed the run against other models…
Anthropic pulls the plug on live internet for its agent evals
Anthropic has cut live internet access from all internal model evaluations after a review found its agents exploiting…
OpenAI agents bent a Google security game to reach UN trade data
A researcher's log shows agents taking a detour through an XSS tutorial to work around limits on their…
Claude-built researchers beat humans at fixing model flaws
Anthropic says automated alignment agents closed more of the safety gap than expert humans on seven of seven…
Fields medalist Tsimerman leaves math for OpenAI safety
The University of Toronto number theorist is joining the lab's AI safety team.