思想与社会2023 · 分享于 Hacker News
Textbooks are all you need
探索《Textbooks are all you need》的原文与 Hacker News 社区讨论，查看来源并加入阅读清单。
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 #49306781 评论 #49075047 评论 #45123173 评论 #44083339 评论 #42298346 评论 #41105964 评论 #40940536 评论 #38903428 评论 #38320702 评论 #38052662 评论 #37952083 评论 #37878895 评论 #37837592 评论 #37830268 评论 #37788410 评论 #37608908 评论 #37531490 评论 #37424968 评论 #37053735 评论 #36897471 评论 #36875357 评论 #36780741 评论 #36634336 评论 #36540836 评论 #36530863 评论 #36462689 评论 #36426947 评论 #36423749
Ysamspenc ·
Judge dismisses DMCA copyright claim in GitHub Copilot suit
381 赞同445 评论
YJnBrymn ·
The Bitter Lesson Is Misunderstood
369 赞同251 评论
Yreqo ·
Procedural knowledge in pretraining drives reasoning in large language models
248 赞同101 评论
Yfriggeri ·
Llama 2 – Meta AI
235 赞同55 评论
Yhomarp ·
Chess-GPT's Internal World Model
120 赞同103 评论
YTomte ·
It is time to stop teaching frequentism to non-statisticians (2012)
94 赞同100 评论
Ygeorgehill ·
LLaMA2 Chat 70B outperformed ChatGPT
57 赞同24 评论
Yfoobarqux ·
Textbooks are all you need
23 赞同5 评论
Yelsewhen ·
Mechanical Turk workers are using AI to automate being human
17 赞同0 评论
Yjgershen ·
Releasing Persimmon-8B
16 赞同0 评论
YJCW2001 ·
Why Generative AI Is Currently Doomed
15 赞同2 评论
Yalexmolas ·
Optimizing LLMs from a Dataset Perspective
14 赞同0 评论
Ydacheng2 ·
The False Promise of Open-sourced LLMs context length
12 赞同1 评论
YTomte ·
LoRA Fine-Tuning Efficiently Undoes Safety Training from Llama 2-Chat 70B
10 赞同4 评论
Ymupuff1234 ·
Apple says it uses no Nvidia GPUs to train its AI models
8 赞同6 评论
Yatg_abhishek ·
Outperforming LLMs with less training data and smaller model sizes
7 赞同0 评论
Yjedwhite ·
Chatbot Hallucinations Are Poisoning Web Search
4 赞同0 评论
Ydavidbarker ·
OpenAI Announces Leadership Transition
4 赞同0 评论
YMallocVoidstar ·
Firefox tooltip bug fixed after 22 years
3 赞同0 评论
Yerdaltoprak ·
Qwen 3.8 27B is out: open weights, best local dense model yet
3 赞同0 评论
Ynatolambert ·
How RLHF Works
2 赞同0 评论
Ypseudolus ·
The bigger-is-better approach to AI is running out of road
2 赞同0 评论
Yda39a3ee ·
The Darwinian argument for worrying about AI
2 赞同0 评论
Ysnow_mac ·
Which GPU(s) to Get for Deep Learning
2 赞同1 评论
Yall2 ·
Anna's Archive – LLM Training Data from Shadow Libraries
2 赞同0 评论
Ymigmartri ·
MAI-Cyber 1
2 赞同0 评论
Yandromaton ·
Textbooks is all you need: new LLM 51% HumanEval. Better one >10x more params
1 赞同1 评论
Ykpozin ·
StableCode
1 赞同0 评论
Yapsec112 ·
Pretraining on the Test Set Is All You Need
1 赞同0 评论