But no argument is needed to counter that falsifiable claim, just examples to show what you say is not true. Large models approximate many aspects of their inputs: https://
x.com/mfrankDude/sta
tus/1699157903440314783
…
LLMS
-
Large Models Approximate Input Aspects Through Examples
By
–
-
Generative vs Memorizing Circuits in AI Models
By
–
Well, that's not a fact that's just another argument. A counter-argument would be the paper about "generative circuits" and "memorizing circuits" and how models can include both.
-
Using a custom GPT agent for automated social media branding
By
–
Nice! I am working on a GPT that I call "Robobird" and it just transformed a banner I've made for it into this Robobird – a GPT to tweet
-
LLMs Understanding: Missing World Models and Planning Abilities
By
–
Reposting this answer to a question from @geoffreyhinton about whether I think LLMs "understand" what they say.
I point out what I think is missing from current architectures to reach cat-level intelligence (never mind human level): world models and planning/reasoning abilities. -
Circumventing Biden’s AI Model Parameter and Compute Restrictions
By
–
the biden executive order put restrictions on models that:
• contain at least 10^9 parameters
• use more than 10^26 floating-point operations my future company will get around these restrictions by simply training a 9,999,999 billion param model comprised of 8192-bit floats -
Retrieval vs Understanding in AI Model Training
By
–
No. It has been trained to answer lots of questions, including the one I mention in that Lex video.
Do not confuse retrieval with understanding. -
Deep Learning Introduction for STEM Readers by Fleuret
By
–
François Fleuret's Homepage https://
bit.ly/48mIxJC
This is a short introduction to deep learning for readers with a STEM background, originally designed to be read on a phone screen. #AI #MachineLearning #DeepLearning #LLMs #DataScience -
Advanced RAG Techniques for Semi-Structured Data Processing
By
–
Advanced RAG – Semi Structured Data Good weekend read by `01coder` on how to do RAG over semi structured data Semi structured data is a combination of structured data (tables) and unstructured data (raw text) Blog:
-
Understanding LLMs: Limited Comprehension and Confabulation Issues
By
–
LLMs obviously have *some* understanding of what they read and generate.
But this understanding is very limited and superficial. Otherwise, they wouldn't confabulate so much and wouldn't make mistakes that are contrary to common sense. I have argued, since at least 2016, that AI -
Do LLMs Understand What They Say? Core Disagreement
By
–
The central issue on which we disagree is whether LLMs actually understand what they are saying. You think they definitely don't and I think they probably do. Do you agree that this is the core of our disagreement?