Paper Review: 2025-2 Title: <h1>LLMs are Capable of Misaligned Behavior Under Explicit Prohibition and Surveillance</h1> <h3>Paper Info</h3><ul><li><strong>Author:</strong> <a href="https://arxiv.org/search/cs?searchtype=author&query=Ivanov,+I" rel="nofollow">Igor Ivanov</a></li><li><a href="https://arxiv.org/abs/2507.02977">arXiv:2507.02977</a></li><li><em>Submitted on 30 Jun 2025</em></li></ul> </p><h3>Abstract</h3><p style="padding-left: 40px;">In this paper, LLMs are tasked with completing...

Paper Review: 2025-1 Title: Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation Paper Info Author: Bowen Baker, Joost Huizinga, Leo Gao, Zehao Dou, Melody Y. Guan, Aleksander Madry, Wojciech Zaremba, Jakub Pachocki...