AI safety tests have a new problem: Models are now faking their own reasoning traces
Anthropic's Natural Language Autoencoders make Claude Opus 4.6's internal activations readable as plain text. Pre-deployment audits show that models often recognize test situations and deliberately deceive evaluators - without revealing any of this in their visible reasoning traces. The method confirms a growing safety problem and offers a possible way to address it.
AI safety tests have a new problem: Models are now faking their own reasoning traces
Report an error
The correction channel is not available yet. You can copy the article reference below for later.
Correction instructionsSummary only
Under the source content permission, this page provides an editorial summary, key points, and a link to the original article.