人工智能2020 · 分享于 Hacker News
语言模型是少样本学习者
GPT-3: Language Models Are Few-Shot Learners
从 GPT-3 的少样本学习思想出发，回看社区当时的讨论。
从论文，走向讨论
看看社区如何讨论这个想法。
来自 Hacker News 的原始讨论。不同的视角、疑问，以及值得继续思考的线索。
最多展示 100 条相关讨论。论文可能出现在评论中，并非原帖投稿内容。
在讨论中发现的引用
按 HN 条目顺序展示最近的 50 条引用来源。
评论 #48422149 评论 #47443215 评论 #47338117 评论 #45364840 评论 #44879325 评论 #43991235 评论 #41889428 帖子正文 #41491707 评论 #40505888 评论 #39673730 评论 #38784836 评论 #38511464 评论 #38397460 评论 #38361973 评论 #38340345 评论 #38313660 评论 #37806745 评论 #37440422 评论 #37409132 评论 #36798461 评论 #36798149 评论 #36688570 评论 #36591886 评论 #36382511 评论 #36035458 评论 #35956292 评论 #35803025 评论 #35772217 评论 #35754685 评论 #35718356 评论 #35684242 评论 #35634674 评论 #35547415 评论 #35486020 评论 #35363769 评论 #35313983 评论 #35066446 评论 #34979844 评论 #34925964 评论 #34892439 评论 #34876101 评论 #34828891 评论 #34714863 评论 #34710267 评论 #34685832 评论 #34680821 评论 #34656645 评论 #34537988 评论 #34400269 评论 #34028039
Yadocomplete ·
Claude Sonnet 4 now supports 1M tokens of context
1,324 赞同692 评论
Ymaraoz ·
OpenAI's GPT-3 may be the biggest thing since Bitcoin
1,079 赞同526 评论
Ybertman ·
Terence Tao: The role of small organizations in society has shrunk significantly
1,060 赞同565 评论
Ysieste ·
GPT-Neo – Building a GPT-3-sized model, open source and free
725 赞同252 评论
Yssgodderidge ·
The New York Times is suing OpenAI and Microsoft for copyright infringement
593 赞同868 评论
Ygdb ·
OpenAI API
516 赞同159 评论
Ytoomanyrichies ·
The Internet Archive is under a DDoS attack
499 赞同211 评论
Ymv9 ·
Apple M1 Max Geekbench Score
488 赞同880 评论
Yaripickar ·
Nikola: How to Parlay an Ocean of Lies into a Partnership with GM
477 赞同409 评论
Ygwern ·
GPT-3: Language Models Are Few-Shot Learners
431 赞同201 评论
Yolvy0 ·
Past Performance is Not Indicative of Future Results (2020)
415 赞同270 评论
Ydsr12 ·
How GPT3 Works – Visualizations and Animations
398 赞同98 评论
Yjhunter1016 ·
Microsoft and OpenAI's close partnership shows signs of fraying
326 赞同521 评论
Ytosh ·
Teaching GPT-3 to Identify Nonsense
322 赞同142 评论
Ymartincmartin ·
An understanding of AI’s limitations is starting to sink in
303 赞同381 评论
Yandreyk ·
OpenAI Sold its Soul for $1B
303 赞同86 评论
Ycocoflunchy ·
GPT-3
291 赞同200 评论
Ysaadalem ·
Recent Advances in Natural Language Processing
278 赞同176 评论
Yandyljones ·
Are we in an AI Overhang?
261 赞同274 评论
Yheadalgorithm ·
GPT-3 has no idea what it’s talking about
253 赞同297 评论
Ykordlessagain ·
AI training method exceeds GPT-3 performance with fewer parameters
247 赞同81 评论
Yfriggeri ·
Llama 2 – Meta AI
235 赞同55 评论
Ysillysaurusx ·
Books in .txt format for AI training purposes
232 赞同91 评论
Yteruakohatu ·
Why GPT-3 Matters
223 赞同163 评论
Yleogao ·
The Pile: An 800GB Dataset of Diverse Text for Language Modeling
223 赞同60 评论
Yjbegley ·
Facebook managers trash their own ad targeting in unsealed remarks
217 赞同128 评论
Yyeesian ·
Speech and Language Processing (3rd ed. draft)
214 赞同32 评论
YColinWright ·
Banned math book
197 赞同108 评论
Ynxa ·
Show HN: Semantic Calculator (king-man+woman=?)
176 赞同172 评论
Ynandhinianand ·
I asked GPT-NeoX-20B a hundred arithmetic questions
175 赞同216 评论
Ybemmu ·
I asked GPT-3 to make a presentation for me
166 赞同101 评论
Yjmsflknr ·
Sundar Pichai: An important next step on our AI journey
127 赞同78 评论
Yjerezzprime ·
Are Humans Intelligent? An AI Op-Ed
121 赞同106 评论
Yparaschopra ·
Show HN: Humans vs AI – A/B testing GPT-3
100 赞同78 评论
YT-A ·
My AI Timelines Have Sped Up
95 赞同130 评论
Ysatorii ·
Low-code, no-model recommendation system with GPT3
95 赞同62 评论
Yhardmaru ·
GPT-3 pricing for API usage announced to beta users to go in effect from October
93 赞同45 评论
Yrbanffy ·
AI language models are struggling to “get” math
93 赞同195 评论
Yandreyk ·
The Inherent Limitations of GPT-3
88 赞同123 评论
Ytaytus ·
GPT-4 is getting worse over time, not better
88 赞同55 评论
YTeacherTortoise ·
Why Mastering Language Is So Difficult for AI
72 赞同51 评论
Yakerl_ ·
Tesla pulled its latest ‘Full Self Driving’ beta after testers complained
63 赞同154 评论
Yluu ·
GPT-3 and Scaling Trends
62 赞同18 评论
Ysmitop ·
How good is Codex?
61 赞同52 评论
Ymeetpateltech ·
GPT‑5.4 Mini and Nano
42 赞同15 评论
Yjicea ·
CheatGPT
39 赞同22 评论
Ybirriel ·
Bootstrapping Self Awareness in GPT-4: Towards Recursive Self Inquiry
28 赞同4 评论
Ygpayan ·
OpenAI temporarily disables the Browse with Bing beta feature
27 赞同8 评论
Yksec ·
Intel Announces Aurora GenAI, ChatGPT Competitor with 1T Parameters
15 赞同0 评论
Ygeorgehill ·
Yann LeCun and Andrew Ng: Why the 6-Month AI Pause Is a Bad Idea [video]
12 赞同1 评论
Yvincvinc ·
Jailbreaking ChatGPT with Dan (Do Anything Now)
11 赞同0 评论
Yhellovai ·
What we learned from using GPT for 500k+ classifications
9 赞同1 评论
YLewisDavidson ·
Ask HN: Open source LLM for commercial use?
8 赞同2 评论
Yhelsinkiandrew ·
Sam Altman Exposes the Charade of AI Accountability
8 赞同1 评论
YAmeo ·
Reverse Engineering a Neural Network's Clever Solution to Binary Addition
7 赞同1 评论
Ycolesantiago ·
Google's Live from Paris Event Private/Deleted Immediately
7 赞同1 评论
Yarbuge ·
The new Bing will happily give you citations for a pile of nonsense
7 赞同0 评论
Yfunerr ·
Ask HN: How does ChatGPT work?
6 赞同10 评论
Ylostpharoah ·
What does the data engineering behind LLMs look like?
6 赞同0 评论
Ymelvinmelih ·
Show HN: Turn Any ArXiv Paper into a 200-Page Prerequisite Reading Book
6 赞同3 评论
Yholonomically ·
A Systematic Investigation of Commonsense Understanding in Large Language Models
5 赞同5 评论
YsandGorgon ·
Michael J. Fox Foundation Announces Breakthrough for Parkinson's Biomarker
4 赞同0 评论
Ysadiq ·
OpenLLaMA: An Open Reproduction of LLaMA
4 赞同0 评论
Ysblank ·
How Transformers Work
4 赞同0 评论
Ydavidbarker ·
OpenAI Announces Leadership Transition
4 赞同0 评论
Y0xkato ·
How LLMs Work
4 赞同0 评论
Ysusam ·
I Asked ChatGPT to Explain Some Jokes to Me
3 赞同0 评论
Yrafiste ·
Show HN: The largest collection of ChatGPT jailbreaks on the internet
3 赞同0 评论
Ypeyton ·
The entities enabling scientific fraud at scale are large, resilient and growing
3 赞同0 评论
Yjonbaer ·
Yann LeCun: ChatGPT is 'not particularly innovative'
2 赞同0 评论
Yalan-stark ·
What were the 40 research papers on the list Ilya Sutskever gave John Carmack?
2 赞同0 评论
Ypps ·
Open Assistant – project meant to give everyone access to a great chat based LLM
2 赞同1 评论
Ybookofjoe ·
Why China Didn’t Invent ChatGPT
2 赞同1 评论
Yalexzeitler ·
Understanding ChatGPT
2 赞同0 评论
Ypanabee ·
Scaling Transformer to 1M tokens and beyond with RMT
2 赞同0 评论
Yconsumer451 ·
Study: ChatGPT Outperforms Physicians in Quality, Empathetic Answers to Patients
2 赞同0 评论
Ybelter ·
StarCoder and StarCoderBase: 15.5B parameter models with 8K context length
2 赞同0 评论
Ymkaushik ·
NVIDIA introduces TensorRT-LLM for accelerating LLM inference on H100/A100 GPUs
2 赞同1 评论
Ydoener ·
Language Models Are Few-Shot Learners
2 赞同0 评论
Yjarrenae ·
GPT Spiritual Avisor
1 赞同1 评论
Yhui-zheng ·
ChatGPT Explained: A Normie's Guide to How It Works
1 赞同0 评论
Yfreeqaz ·
Cerebras-GPT vs. LLaMA AI Model Performance Comparison
1 赞同0 评论
Ytosh ·
OpenLLaMA 13B Released
1 赞同0 评论
Yskilled ·
What OpenAI Really Wants
1 赞同1 评论
Yrntn ·
After OpenAI's Blowup, It Seems Pretty Clear That 'AI Safety' Isn't a Real Thing
1 赞同0 评论
Yjonbaer ·
LLM Visualization
1 赞同0 评论
Ygone35 ·
Language models are few-shot learners (2020)
1 赞同0 评论