Fetching the paper…

Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation · Around