编辑“︁OpenAI Codex”︁

{{Rough translation|time=2026-03-23T22:31:19+00:00}}
'''OpenAI Codex'''是由[[OpenAI]]开发的[[人工智能]]模型。它能解析[[自然语言]]并生成相应的[[代码]]。该模型驱动了[[GitHub Copilot]]，一个为选定的[[集成开发环境]]（如[[Visual Studio Code]]和[[Neovim]]）提供的编程[[自动补全]]工具<ref name="OAI">{{cite web|last=Zaremba|first=Wojciech|author-link=Wojciech Zaremba|date=August 10, 2021|title=OpenAI Codex|url=https://openai.com/blog/openai-codex/|url-status=no|access-date=2021-09-03|website=[[OpenAI]]|archive-date=2023-02-03|archive-url=https://web.archive.org/web/20230203201912/https://openai.com/blog/openai-codex/}}</ref>。Codex是OpenAI的[[GPT-3]]模型的后代，经过[[微調 (深度學習)|微调]]以用于[[编程]]。

OpenAI为Codex发布了一个[[封闭测试]]版的[[API]]。2023年3月，OpenAI关闭了对Codex的访问<ref>{{Cite web |last=Kemper |first=Jonathan |date=2023-03-22 |title=OpenAI kills its Codex code model, recommends GPT3.5 instead |url=https://the-decoder.com/openai-kills-code-model-codex/ |access-date=2023-03-29 |website=THE DECODER |language=en-US |archive-date=2023-06-01 |archive-url=https://web.archive.org/web/20230601195835/https://the-decoder.com/openai-kills-code-model-codex/ |dead-url=no }}</ref>。然而，由于来自研究者的公开呼吁，OpenAI改变了这一决定<ref>{{Cite tweet |user=OfficialLoganK |author=Logan Kilpatrick |number=1638336152800206858 |title=Hey Carolyn, we will continue to support Codex access via our Researcher Access Program. Sorry for any confusion and hopefully the research is going well! |access-date=2023-04-08}}</ref>。现在，Codex模型仍可供OpenAI研究访问计划的研究者使用<ref>{{Cite web |title=Researcher Access Program application |url=https://openai.com/form/researcher-access-program |access-date=2023-04-08 |website=openai.com |language=en-US |archive-date=2023-10-10 |archive-url=https://web.archive.org/web/20231010073704/https://openai.com/form/researcher-access-program |dead-url=no }}</ref>。

== 功能 ==
基于 GPT-3，一种在文本上训练的[[人工神经网络]]，Codex 还接受了来自 5400 万个[[GitHub]]仓库的 159 GB [[Python]]代码的训练<ref name="VB-bias">{{Cite news|last=Wiggers|first=Kyle|date=July 8, 2021|title=OpenAI warns AI behind GitHub's Copilot may be susceptible to bias|work=[[VentureBeat]]|url=https://venturebeat.com/2021/07/08/openai-warns-ai-behind-githubs-copilot-may-be-susceptible-to-bias/|access-date=2021-09-03|archive-date=2023-02-03|archive-url=https://web.archive.org/web/20230203201912/https://venturebeat.com/business/openai-warns-ai-behind-githubs-copilot-may-be-susceptible-to-bias/|dead-url=no}}</ref><ref name="IQ">{{Cite news|last=Alford|first=Anthony|date=August 31, 2021|title=OpenAI Announces 12 Billion Parameter Code-Generation AI Codex|work=InfoQ|url=https://www.infoq.com/news/2021/08/openai-codex/|access-date=2021-09-03|archive-date=2022-07-09|archive-url=https://web.archive.org/web/20220709221205/https://www.infoq.com/news/2021/08/openai-codex/|dead-url=no}}</ref>。 Codex 的一个典型用例是用户键入评论，例如“<code>//计算给定窗口大小的数组的移动平均值</code>”(<code>//compute the moving average of an array for a given window size</code>)，然后使用 AI 来建议满足该评论提示的代码块<ref name="RegTA">{{Cite news|last1=Anderson|first1=Tim|last2=Quach|first2=Katyanna|date=July 6, 2021|title=GitHub Copilot auto-coder snags emerge, from seemingly spilled secrets to bad code, but some love it|work=[[The Register]]|url=https://www.theregister.com/2021/07/06/github_copilot_autocoder_caught_spilling/|access-date=2021-09-04|archive-date=2023-06-02|archive-url=https://web.archive.org/web/20230602214528/https://www.theregister.com/2021/07/06/github_copilot_autocoder_caught_spilling/|dead-url=no}}</ref>。 OpenAI 表示，Codex 可以完成大约 37% 的请求，旨在让人类编程更快，而不是取代它。根据 OpenAI 的博客，Codex 最擅长“将……简单问题映射到现有代码”，他们将其描述为“可能是编程中最不有趣的部分”<ref name="SH">{{Cite news|last=Dorrier|first=Jason|date=August 15, 2021|title=OpenAI's Codex Translates Everyday Language Into Computer Code|work=[[SingularityHub]]|url=https://singularityhub.com/2021/08/15/openais-codex-translates-everyday-language-into-computer-code/|access-date=2021-09-03|archive-date=2023-05-26|archive-url=https://web.archive.org/web/20230526045651/https://singularityhub.com/2021/08/15/openais-codex-translates-everyday-language-into-computer-code/|dead-url=no}}</ref><ref name="VB">{{Cite news|last=Dickson|first=Ben|date=August 16, 2021|title=What to expect from OpenAI's Codex API|work=[[VentureBeat]]|url=https://venturebeat.com/2021/08/16/what-to-expect-from-openais-codex-api/|access-date=2021-09-03|archive-date=2023-02-03|archive-url=https://web.archive.org/web/20230203201913/https://venturebeat.com/ai/what-to-expect-from-openais-codex-api/|dead-url=no}}</ref>。

== 问题 ==
OpenAI 演示展示了代码示例中的低效代码和一次性怪癖等缺陷<ref name="VB" />。  在接受 [[The Verge]] 采访时，OpenAI [[首席技术官]] Greg Brockman 表示，“有时 [Codex] 不太清楚你在问什么”，这可能需要反复试验。”

== 参见 == 
{{div col|4}}
* [[OpenAI]]
* [[ChatGPT]]
* [[GitHub Copilot]]
* [[Transformer模型]]
* [[GPT-3]]
* [[對話程式語言模型|LaMDA(對話語言模型)]]
* [[深度学习]]
{{div col end}}

== 外部链接 ==
* [https://platform.openai.com/playground  Playground for GPT3.5 and Codex] {{Wayback|url=https://platform.openai.com/playground |date=20230610092836 }}  

== 参考文献 ==
{{reflist}}

{{OpenAI}}
{{生成式人工智能}}
{{Differentiable computing}}

[[Category:OpenAI]]
[[Category:交互式叙事]]
[[Category:人工智能應用]]
[[Category:語言模型]]
[[Category:2022年面世的產品]]
[[Category:人工智慧]]
[[Category:機器學習]]