Skip to content
onsuggest
Log in
Article en

Towards embedded evaluations for AI monitoring

Source: LessWrong ·

Executive summary This year we observed multiple rogue agent incidents . AI monitors, often built from the developers' own internal models, are now the main oversight of internally deployed agents. Embedded evaluations address this gap: independent evaluators work inside the developer and test its safety claims. Our first pilot towards an embedded evaluation was announced with OpenAI . This is the third post in a series. Principles for Embedded Evaluations describes our claim-based approach to embedded evaluations. Towards embedded evaluations for scheming propensities details the claims acros

What is Towards embedded evaluations for AI monitoring?

Towards embedded evaluations for AI monitoring is a article discovery curated from LessWrong. Executive summary This year we observed multiple rogue agent incidents . AI monitors, often built from the developers' own internal models, are now the main oversight of…

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

OVERDOSE

In the wake of an affair and a workplace fentanyl overdose in his intern year, anesthesiologist Matt Robbins wants only to advance professionally and provide for his wife and daughter. When proposed legislation threatens the financial interests of Central Penn Anesthesia, his money-hungry employers become entangled in the corrupt reelection campaign of Pennsylvania senator Brad McNally. As stolen pharmaceuticals and envelopes of illicit cash change hands, Matt finds himself both investigator and scapegoat in a succession of shady deals. A web of convincing character motivations unfolds across

Article · Culture & ideas

The Case for A.I. Optimism

The tech reporter Kevin Roose discusses the rise of Anthropic, the fall of Jensen Huang, and why not to freak out—too much—about the new age enveloping us all.

The New Yorker 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.