RSS

Joseph Bejjani

Karma: 0

CS student at Harvard. AI alignment at the Kempner Institute.

[CS2881r][Week 8] When Agents Pre­fer Hack­ing To Failure: Eval­u­at­ing Misal­ign­ment Un­der Pressure

7 Nov 2025 5:45 UTC
2 points
0 comments23 min readLW link