Skip to content
onsuggest
Log in
Article en

Gemini 4 safety reports don’t demonstrate low loss-of-control risk

Source: LessWrong ·

Google DeepMind’s safety reports for new Gemini 4 models, Argon and Carbon , conclude that the models are “highly unlikely” to cause severe harm by evading oversight (p. 79). This conclusion “rests primarily on our ability to detect and intervene on harmful trajectories via CoT and action monitoring” (p. 89) – that is, they’re arguing for safety based on their control techniques, rather than alignment. We think that the arguments in these reports have serious weaknesses, and the reports overall don’t convince us that the risks of deploying these models have been mitigated reliably. The biggest

What is Gemini 4 safety reports don’t demonstrate low loss-of-control risk?

Gemini 4 safety reports don’t demonstrate low loss-of-control risk is a article discovery curated from LessWrong. Google DeepMind’s safety reports for new Gemini 4 models, Argon and Carbon , conclude that the models are “highly unlikely” to cause severe harm by evading oversight (p. 79)…

How can I access or experience this?

You can experience this content directly on its original host (LessWrong) by clicking the "Open Original" button.

Where does this content come from?

Curated from LessWrong by the OnSuggest team to surface the most interesting and serendipitous content on the web.

Article · Culture & ideas

The Great London Smog of 1952

Podcast Transcript In December 1952, a deadly catastrophe descended upon one of the world’s largest cities. Within hours, visibility dropped to almost nothing. Transportation ground to a halt, hospitals were overwhelmed, and thousands of people began dying. Yet there was no earthquake, no flood, and no invading army. The disaster came from something the city’s ... Read more

Everything Everywhere View Details →

Keyboard shortcuts

Previous discovery
←P
Next discovery
→N
Like
L
Not for me
X
Save
S
Show this help
?
Close
Esc

Shortcuts work on the discovery page.