The Extreme Inefficiency of RL for Frontier Models — Toby Ord
September 19, 2025Toby Ord
·
2025-09-19
·
via Writing - Toby Ord
The new scaling paradigm for AI reduces the amount of information a model could learn per hour of training by…
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。
