Daily Productive Sharing 715 - Shepherding ChatGPT

Daily Productive Sharing 715 - Shepherding ChatGPT
Photo by Jonathan Kemper / Unsplash

One helpful tip per day:)

In a previous share, there was a profile of OpenAI CTO Mira Murati. Today's share is an interview with her by the Associated Press. This article mainly introduces OpenAI's view on the current GPT-like technologies:

  1. All large language models should be regulated, and government agencies should intervene in a timely manner;
  2. OpenAI has made a lot of efforts in the safety of large language models, such as erasing sensitive data during the training phase, adjusting the gender ratio of image data, etc. At the same time, six months were spent on developing various safety measures before releasing GPT-4;
  3. Before the launch of ChatGPT, OpenAI went through numerous product iterations, such as adding chat interface, increasing data, so the success of ChatGPT was not achieved overnight.

Daily Productive Sharing 680

If you enjoy today's sharing, why not subscribe

Need a superb CV, please try our CV Consultation

之前的分享中有一篇是 OpenAI CTO – Mira Murati 的人物侧写,今天的分享则是美联社对她的专访。这篇主要介绍了 OpenAI 对于当下 GPT 这类技术的看法:

  1. 所有的大语言模型都应该被监管,政府机构应该及时介入;
  2. OpenAI 在大语言模型的安全性中做出了大量努力,比如在训练阶段,就抹除了敏感数据,调整图片数据的男女比例等等。同时在发布 GPT-4 之前,有六个月的时间专门开发各种安全措施;
  3. 在 chatGPT 发布之前,OpenAI 做了大量产品迭代,比如增加了聊天页面,增加数据,所以 chatGPT 的成功也不是一蹴而就的。

Daily Productive Sharing 680

如果你喜欢我们的内容,不如支持我们 :)

需要更棒的简历,不妨试试我们的 CV Consultation