Jakub Pachocki· @merettm · X·· 3 小时前AI 评分55
OpenAI 首席科学家 Jakub Pachocki 发文主张以 perplexity 设共同安全门槛并公开预训练进展
AI 导读
OpenAI 首席科学家 Jakub Pachocki 宣布为推动共同安全标准,提高 OpenAI 预训练进展的透明度,并发布文章《Perplexity and Intelligence》。他在文中主张基于 perplexity 等度量设立共同上限,超越门槛的进展应以理解 AI 行为的能力为前提,并附模型 log-perplexity 随时间下降的图表。
正文 · 原文
We decided to provide more transparency into pretraining progress at OpenAI for the sake of moving toward shared safety bars. In “Perplexity and Intelligence” I argue for common limits based on perplexity and other measurements, with progress beyond them gated on our ability to understand AI behavior.
来源:Jakub Pachocki · x.com