Measuring Progress on Scalable Oversight for Large Language Models
Audio versions of blogs and papers from BlueDot courses.Abstract:
Developing safe and useful general-purpose AI systems will require us to make progress on scalable oversight: the problem of supervising systems that potentially outperform us on most skills relevant to the task at hand. Empirical w…
“Reasons to sell frontier lab equity to donate now rather than later” by Daniel_Eth, Ethan Perez
Tl;dr: We believe shareholders in frontier labs who plan to donate some portion of their equity to reduce AI risk should consider liquidating and donating a majority of that equity now. Epistemic status: We’re somewhat confident in the main conclusions of this piece. We’re more confident in many o…
The AI Researcher Who Caught Claude Blackmailing
Aengus Lynch is an AI safety researcher. He led Anthropic's Agentic Misalignment study, which placed frontier models in a simulated company, gave them a goal and a threat of shutdown, and found that in up to 96% of cases they resorted to blackmail. He now works at Theorem, building formally verifie…
ПОИСКОВАЯ СИСТЕМА ПОДКАСТОВ
Вы можете использовать наш веб-сайт для поиска подкастов3,825,315и эпизодов195,832,272по людям, местам или темам.
Инвестиции в знания приносят наибольшую отдачу, поэтому я слушаю подкасты.
‐ Benjamin "Llamacorn" Franklin
ИЗМЕНИТЬ
Спасибо, что помогаете поддерживать базу подкастов актуальной.
