korrents

A korrentour readingWhat is a korrent?

Anthropic now uses mechanistic interpretability techniques to audit new models for deception and scheming before releasing them.

Drawn from what Dario Amodei said

In the currentAnthropic
Interesting
Keep this without saying you believe it. Private, and never counted.

What Dario Amodei actually said

Their words, word for word, each linked to where it was said. They did not write this page.

  1. Dario Amodei Even more recently, we’ve begun to use mechanistic interpretability techniques to improve our safeguards and to conduct “audits” of new models before we release them, looking for evidence of deception, scheming, power-seeking, or a propensity to behave differently when being evaluated.

    Co-founder and CEO of Anthropic

    Their wordsEven more recently, we’ve begun to use mechanistic interpretability techniques to improve our safeguards and to conduct “audits” of new models before we release them, looking for evidence of deception, scheming, power-seeking, or a propensity to behave differently when being evaluated.

    ↗The Adolescence of Technology (January 2026)darioamodei.com

First recorded here 1 Jan 2026 · How quotes work · Something wrong? Tell us

Do you hold this korrent?

Sign in to record that you hold this, with a confidence number of your own.

Related korrents

Our reading — they may agree, disagree or merely touch the same thing.

See this on the map →