Must be just bogging down the context window, especially if you've uploaded reference docs, etc.
LLMS
-
Weights Sharding Strategy: Full Replication Except Last Dimension
By
–
Also, tip on weights sharding: as a generic rule of thumb you can use full replication for all variable dimensions except the last one, which should be sharded across the "model" axis
-
Can LLMs Learn Visual Understanding Without Images?
By
–
I don’t think you can learn to understand the visual world from text alone. but congenitally blind people learn a ton about the world without visual input, with much richer representations than I would ascribe to LLMs. Landau & Gleitman had a terrific book about this in the
-
Humans trained AI on tiny fraction of internet
By
–
I am sure there is. Brilliant humans have trained on a tiny fraction of the internet.
-
New AI Tools Canvas and NotebookLM Transform Workflows
By
–
I hope no one is getting numb to the continual innovation in our AI tools. There is so much to learn and the rewards in doing so are far more than just the fun of learning. New OpenAI ChatGPT Canvas and new Google NotebookLM have changed @ooda workflows already. Very exciting.
-
ChatGPT Shows Better Bedside Manner Than Physicians
By
–
In medicine, #ChatGPT and other LLMs have surprisingly and consistently been shown to promote empathy better than physicians
"You might find it disturbing that A.I. can have a better bedside manner than humans."—
@jonreismanMD @nytopinion Gift link -
Strubell Study Flawed: Training Cost Estimates Off by 88X
By
–
Did you get this number from Strubell et. al? Because that number was a very flawed estimate that turned out to be off by 88X, and the paper also portrayed it as an every time cost when in fact it was a one time cost. The actual cost of training a model of the size they were
-
Augmenting Statistical Models with Natural Language Parameters
By
–
Augmenting Statistical Models with Natural Language Parameters https://
bit.ly/3XS5DmX
#AI #MachineLearning #DeepLearning #LLMs #DataScience -
Roundup of recent AI tool and model updates
By
–
New MS Copilot UI / voice on iOS On a smaller note: – perplexity new financial cards
– kling ai lips sync – flux 1.1 – updated Notion chat
– Grok image history Most of the upcoming big features are still cooking