Ideas & society2023 · shared on Hacker News
Transformers struggle with generalizing tasks beyond pre-training data
Explore Transformers struggle with generalizing tasks beyond pre-training data, its original source and Hacker News discussions. Discover community perspectives and save it to your reading list.
GO BEYOND THE ABSTRACT
The conversation around the idea.
Original Hacker News threads. Different perspectives, questions, and things worth thinking about.
Showing up to 100 related threads. A paper may appear in a comment rather than the original submission.
Links found in the discussion
Up to 50 recent source links, in HN item order.
Ymfiguiere ·
OpenAI and Elon Musk
1,037 points977 comments
Ytodsacerdoti ·
Markov chains are funnier than LLMs
525 points233 comments
Yspwestwood ·
Seven replies to the viral Apple reasoning paper and why they fall short
343 points312 comments
Ysaliagato ·
Will scaling work?
242 points283 comments
Yhomarp ·
Chess-GPT's Internal World Model
120 points103 comments
Yryandvm ·
New theory suggests LLMs can understand text
79 points106 comments
Ydmotz ·
Gemini – Google DeepMind
59 points11 comments
Yhislaziness ·
Transformers struggle with generalizing tasks beyond pre-training data
36 points36 comments
Yagomez314 ·
Evidence that transformers cannot generalize beyond their training data
3 points0 comments
Ygardenfelder ·
Google researchers deal a major blow to the theory AI about to outsmart humans
2 points1 comments
Yistotex ·
AI is about to change how you use computers
2 points1 comments
YYeGoblynQueenne ·
Pretraining Data Mixtures Enable Narrow Model Selection Capabil. In Transformers
1 points0 comments
Yvagabund ·
Weak-to-Strong Generalization
1 points0 comments